CIRCUITS - americanradiohistory.com · 1984 (p. 28) issue of Elektor Electronics. In the design stages, stability problems were en-countered when opamps with JFET inputs (TL074;

303CIRCUITS

I Elektor Electronics

CIRCUITS

A WOLTERS-KLUWER COMPANY

Elektor Electronics1 Harlequin Avenue Great West Road

BRENTFORD Middlesex TW8 9EW ENGLANDTelephone: 01-847 2618 Telex: 917490

All rights reserved. No part of this book may be reproduced or transmitted in anyform, or by any means, including photocopying and recording, without the priorwritten permission of the publishers. Such written permission must also be ob-tained before any part of this book is stored in a retrieval system of any nature.

The circuits and other diagrams contained in this book are for domestic use only.Patent protection may exist with respect to circuits, devices, components, etc.described in this book. The publishers do not accept responsibility for failing toidentify such patent or other protection.

This book is sold subject to the Standard Conditions of Sale of Net Books andmay not be re -sold in the United Kingdom below the net price given by thepublishers in their current price list.

First printed 1988

Copyright © 1988 Elektuur BV

Printed in the NetherlandsISBN 0 905705 26 2

WARNING!ELECTRICITY CAN BE DANGEROUS!The electronic projects in this book are, tothe best of the author's knowledge andbelief, both accurately described and safe.None the less, great care must always betaken when assembling electronic circuitswhich carry mains voltage, and neither thepublishers nor the author can acceptresponsibility for any accidents which mayoccur.

Because electricity is dangerous, its use,application and transmission are subject torules, regulations and guidance. These arelaid down in numerous laws, ElectricityGenerating Board regulations, British Stan-dards, and IEE recommendations. Some ofthese may be obtained from your localelectricity showroom, but most, if not all,should be available for reference in yourlocal library.

contentsCIRCUITS FOR:

BRAKE LIGHTS MONITORCAR BURGLAR ALARMCAR FUSE MONITOR

555657

1. Audio & hi-fi CAR LIGHTS MONITOR 572. Cars & bicycles CAR RADIO ALARM I 583. Computers & microprocessors CAR RADIO ALARM II 604. Design ideas COURTESEY LIGHT DELAY 605. Electrophonics FLASHING REAR LIGHT 616. HF & VHF GARAGE STOP LIGHT 627. Hobbies & pastimes HALOGEN LAMP PROTECTOR 638. Home & garden LED REVOLUTION COUNTER 639. Power supplies MOTOR -CYCLE GEAR INDICATOR 6710. Test & measurement11. TV & video

3. COMPUTERS AND MICROPROCESSORS4 -WAY DAC EXTENSION 68

1. AUDIO AND HI-FI 8 -BIT ADC 69AUDIO -CONTROLLED MAINS SWITCH 9 8 -BIT DAC 70AUDIO LINE DRIVER 10 16 -KEY INPUT FOR MSX MICROS 71AUDIO TRANSFER EQUALIZER 10 32 KBYTE PSEUDO ROM 72COMPRESSOR 11 40 -TRACK ADAPTOR 75CURRENT CORRECTED AF AMPLIFIER 13 2708 ALTERNATIVES 75DIGITAL AUDIO SELECTOR 14 6502 TRACER 76DIGITAL VOLUME CONTROL I 15 A -D CONVERTER FOR JOYSTICKS 78DIGITAL VOLUME CONTROL II 17 BI-DIRECTIONAL PARALLEL INTERFACE FORDISCOMIXER 18 C64 79HEADPHONE AMPLIFIER 22 BI-DIRECTIONAL SERIAL -PARALLEL CON -HI -Fl HEADPHONE AMPLIFIER 25 VERTER 80HIGH DYNAMIC RANGE MIXER 26 BUS DIRECTION ADD-ON FOR MSX EXTEN-INTEGRATED STEREO AMPLIFIER 27 SIONS 82LOUDSPEAKER PROTECTION I 27 COMMUNICATION PROGRAM FOR C64 83LOUDSPEAKER PROTECTION II 29 CPU GEAR/BOX 86LOUDSPEAKER PROTECTION III 30 CURRENT LOOP FOR MODEM 87LOW -NOISE RIAA PREAMPLIFIER 32 DIRECT READING DIGITIZER 88MICROPHONE PRE -AMPLIFIER WITH DISCRETE DAC 90MUTE SWTICH 36 DRIVE SELECTOR 91

MICROPHONE -SIGNAL PROCESSOR ... 39 FILTERED CONNECTOR 92MINI -AMPLIFIER 40 FLOPPY CENTRING UNIT 92MINI STEREO AMPLIFIER 40 FLOPPY DISC DRIVE 93MOSFET POWER AMPLIFIER 41 HEXADECIMAL KEYBOARD 94NOISE GATE 42 IMPROVED SOUND FOR THE BBCSIMPLE PRE -AMPLIFIER 43 MICRO 95SINGLE -CHIP 40 W AMPLIFIER 44 JOYSTICK ADAPTOR 96SMD HEADPHONE AMPLIFIER 45 LEVEL ADAPTOR FOR ANALOGUESPEECH PROCESSOR WITH JOYSTICKS 97BACKGROUND SUPPRESSION 46 LISTEN -IN KEY FOR DATA RECORDERS . 98

STEREO INDICATOR 48 MAINS INTERFACE 98STEREO PRE -AMPLIFIER WITH MANDELBROT GRAPHICS 99TONE CONTROL 48 MORSE TRAINING WITH THE JUNIOR

SUBWOOFER FILTER 50 COMPUTER 100TRUE CLASS/B AMPLIFIER 51 PIA FOR ELECTRON 102TUNING AF POWER STAGES 52 QL RAM EXTENSION 103

RAM EXTENSION FOR QL 1042. CARS AND BICYCLES RS232 INTERFACE 105AUTOMATIC CAR ALARM 54 SAMPLE & HOLD FOR ANALOGUEBICYCLE LIGHTS AND ALARM 54 SIGNALS 107

5

SERIAL DATA CONVERTER 109SERIAL LINE DRIVER AND RECEIVER 111SIDEWAY RAM FOR BBC AND ELECTRONPLUS ONE 112

SIMPLE D -A CONVERTER 113SIMPLE VIDEO INVERTER FOR ZX81 115SYNC INVERTER FOR THE QL 116SYNCHRONIZATION SEPARATOR 116TWIN KEYBOARD FOR APPLE II 117TWO -FREQUENCY CLOCK 119

4. DESIGN IDEAS6 -WAY CHANNEL SELECTOR 120ANALOGUE AND DIGITAL 120BAND -GAP VOLTAGE REFERENCE 121BUZZER DP,!VER 123COMBINING DIGITAL CIRCUITS 123CURRENT DRIVE FOR STEPPERMOTORS 124

DC -OPERATED 50 HZ TIMEBASE 126DECOUPLING IN LOGIC CIRCUITS 126DEGLITCHER 128DESIGNING A LOW NOISE AMPLIFIER 128DISPLAY INTENSITY CONTROL 129DUTY FACTORY ANALYSER 130ELECTRONIC ROTARY SWITCH 132FAST OPTO-COUPLER 133FAST 0 PTO -ISO LATOR 133HC -BASED OSCILLATORS 134HCMOS VCO 134HCU/HCT BASED OSCILLATOR 135HEARTBEAT MONITOR 136HEAT SINK MONITOR 138LOGIC FAMILIES 139LOW VOLTAGE DROP REGULATORS .. 139MAINS ZERO -CROSSING DETECTOR .. 141OPAMP-BASED CURRENT SOURCE 141PIERCE OSCILLATOR 142POWER SUPPLY SEQUENCING FOROPAMPS 143

PRECISION CRYSTAL OSCILLATOR 143SMART LED SELECTOR 144SPEED CONTROL FOR DC MOTORS 145STEPPER MOTOR CONTROL 146SYMMETRICAL CASCODE OSCILLATOR 148THRIFTY LED INDICATOR 149TIME STRETCHER 150TRACKING WINDOW COMPARATOR . 151TRANSMISSION LINES FOR TTL

CIRCUITS 152TUNING AF POWER STAGES 152TWO -FREQUENCY OSCILLATOR 153TWO -GATE BISTABLE 154UP/DOWN CLOCK GENERATOR 154UP/DOWN COUNTER CONTROL 155

VERSATILE TIMER 157VOLTAGE -TO -CURRENT CONVERTER 158

5. ELECTROPHONICSBLOW THAT SYNTHESIZER! 159DISCO SOUND LIMITER 160GUITAR FUZZ UNIT 161LIMITER FOR GUITARS 162MELODIC SAWTOOTH 163METAL PERCUSSION GENERATOR 164PATCH CATCHER 164QUARZ-CONTROLLED TUNING FORK 166SOUND -LEVEL INDICATOR 166SWELL PEDAL 170WAH-WAH BOX FOR GUITARS 172

6. HF AND VHFELECTRONIC VHF/UHF AERIAL SWITCH 173FOUR-WAY AERIAL SWITCH 173FRONT END FOR FM RECEIVER 175FRONT END FOR SW RECEIVER 176HIGH LEVEL PASSIVE DBM 177HIGH LEVEL WIDEBAND RF

PREAMPLIFIER 177LOW NOISE AERIAL BOOSTER 178MORSE FILTERS 180MULTI -MODE ,uP-CONTROLLED IF

MODULE 181NOISE BLANKER 184NARROW -BAND IF FILTER 185NAVTEX RECEIVER 185RTTY CALIBRATION INDICATOR 186RTTY/CW FILTER 187S METER 187SEND/RECEIVE IDENT 188SIMPLE FIELD STRENGTH INDICATOR 189SPOT FREQUENCY RECEIVER 189SWITCHABLE BANDSELECTOR 190SYNTHESIZER FOR SW RECEIVER .. 192TUNEABLE ACTIVE AERIAL FOR SW 194TUNEABLE FM BOOSTER 195VLF CONVERTER 197WEATHER SATELLITE INTERFACE 198

7. HOBBIES AND PASTIMESAUTO FOCUS FOR SLIDE PROJECTOR 200DIGITAL JOYSTICK INTERFACE 201ELECTRONIC TOSS-UP 202FLASHING LIGHTS 203HALOGEN -LAMP DIMMER 204MODEL AIRCRAFT MONITOR 206MODEL RAILWAY MONITOR PANEL 207"ON THE AIR" INDICATOR 209PWM DRIVER FOR DC MOTORS 210

6

SECTION INDICATOR FOR MODEL STAIRCASE LIGHT CONTROLLER 260RAILWAY 211 SUPER DIMMER 261

SERVO ROBOT DRIVER 212 TELEPHONE -BELL SIMULATOR 263SOLID-STATE DARK -ROOM LIGHT 212 TEMPERATURE REGULATOR WITH ZEROSPEED CONTROL FOR R/C MODELS .. . 213 CROSSING SWITCH 263STARTING PISTOL SIMULATOR 215 TEMPERATURE SENSOR 264TIME-LAPSE UNIT 216 THERMOMETER 264TIMER FOR FIXING BATH 217 THERMOSTAT -CONTROLLED SOIL

HEATING 265

8. HOME AND GARDEN7 -DIGIT CODE LOCKABSORPTION -TYPE METAL DETECTORALTERNATING FLASHERAUTOMATIC SLIDING DOORBURGLAR DETERRENT

218219220221222

TWIN BELL -PUSHTWIN DIMMERTWO-TONE CHIMEVENTILATOR CONTROLWATCHDOGWATER DIVINER

266266268269270271

CALL COUNTER 223CENTRAL HEATING CONTROL 224CH -BOILER CONTROL 226 9. POWER SUPPLIES

COLOUR WHEEL 227 12 V NICD BATTERY CHARGER 273CURRENT MONITOR AND ALARM 228 ACTIVE RECTIFIER WITHOUT DIODES 273

DECEPTIVE LOCK 229 BATTERY CHARGE/DISCHARGE

DESICCATOR 230 INDICATOR 274ELECTRONIC BELL -PULL 231 BATTERY CHARGING INDICATOR 275

ELECTRONIC DOG 231 BATTERY FITNESS CENTRE 275

FLASHING LIGHT WITH TWILIGHT BATTERY GUARD 276

SWITCH 232 CURRENT INDICATOR 277

FOUR POSITION TOUCH DIMMER 233 CURRENT INDICATOR FOR 723. 278HOTEL SWITCH 234 DC/DC CONVERTER 279INDUSTRIAL CLOCK CONTROLLER 235 DIGITAL VOLTAGE/CURRENT DISPLAY . 280INFRA -RED LIGHT BARRIER 236 DIRECT -CURRENT MONITOR 282

JUMBO DIMMER 238 DIRECT -VOLTAGE DOUBLER 283

JUMBO DISPLAYS 238 ECONOMICAL POWER SUPPLY 284LED DIRECTION INDICATOR 243 LEAD -ACID BATTERY CHARGER 285

LIGHT-SENSITIVE SWITCH 243 LOSS -FREE SUPPLY PROTECTOR 286LIGHT SENSITIVE TRIGGER 244 LOW -DROP VOLTAGE REGULATOR 287

LONG -INTERVAL TIMER 245 MAINS POWER SUPPLY WITH PRIMARY

MAINS -BASED REMOTE CONTROLLER . 245 REGULATION 287MAINS FAILURE ALARM 246 MAINS ZERO -CROSSING DETECTOR . . 288MAINS VOLTAGE MONITOR 247 NEGATIVE SUPPLY CONVERTER 289MAINS WIRING LOCATOR 248 NICD BATTERY CHARGERS 290METAL DETECTOR 248 ONE -CHIP DC CONVERTER 291

METAL -PIPE DETECTOR 249 PRECISION RECTIFIER 291

MINIATURE RUNNING LIGHTS 250 SIMPLE NICD CHARGER 292MUSICAL GREETING CARDS 251 SIMPLE ZERO CROSSING DETECTOR .. 293RANDOM LIGHTS CONTROLLER 252 SUPPLY PROTECTION 293REMOTE CONTROL FOR LIGHT SWITCH -MODE POWER SUPPLY 294SWITCHES - 1 253 VARIABLE 3A POWER SUPPLY 295REMOTE CONTROL FOR LIGHT VISIBLE POWER -ON DELAY 295SWITCHES - 2 254 VOLTAGE INVERTER 296RODENTS DETERRENT 255SET POINTER 256 10. TEST AND MEASUREMENTSIREN 257 8 -CHANNEL VOLTAGE DISPLAY 298SMD DIE 258 AUDIO TESTER 298SMOKE AND GAS DETECTOR 259 AUTOMATIC SWITCH OFF 300

7

CALIBRATION GENERATOR 301CRYSTAL TESTER 301DIVIDER CASCADE 302FAST VOLTAGE -CONTROLLED PULSEGENERATOR 303FAULTFINDING PROBE FOR µPs 303FUNCTION GENERATOR 304GHZ PRESCALER 305INSTRUMENTATION AMPLIFIER 306LINE BAR GENERATOR 307LOGARITHMIC SWEEP GENERATOR . 307LOW -CURRENT AMMETER 308MEASURING WITH THE BBC MICRO . . 309METER AMPLIFIER 310METERING SELECTOR 311NOISE GENERATOR 312OPAMP TESTER 313POCKET FREQUENCY METER 314PROGRAMMABLE BAUD -RATEGENERATOR 315

RECTANGULAR PULSE GENERATOR 316RMS-TO-DC CONVERTER 317SERVO MOTOR TESTER 318SIMPLE SWEEP GENERATOR 319SIMPLIFIED WORD COMPARATOR 321TWO-TONE RF TEST OSCILLATOR 322VARIABLE WIEN BRIDGE OSCILLATOR 323WIEN BRIDGE OSCILLATOR 324

11. TV AND VIDEORF MODULE FOR IDU 325RGB-TO-MONOCHROME CONVERTER . 326SCART SWITCH 327SYNCHRONIZATION SEPARATOR I 329SYNCHRONIZATION SEPARATOR II 329VIDEO AMPLIFIER FOR B/W TV SETS 330VIDEO BUFFER/REPEATER 331VIDEO DISTRIBUTION AMPLIFIER 331VIDEO SELECTOR 332

8

CHAPTER. I AUDIO 8/

001 AUDIO -CONTROLLED MAINS SWITCH

It is often useful for audio or video equipment to beswitched off automatically after there has been noinput signal for a while.The function of the on -off switch in such equip-ment is then taken over by switch Sz in the accom-panying diagram. It remains, however, possible toswitch off manually by means of Si. Automaticswitch -off occurs after there has been no inputsignal for about 2 minutes: this delay makes it poss-ible for a new record or cassette to be placed in therelevant machine.The audio input to the proposed circuit may betaken from the output of the relevant TV set, ampli-fier, or whatever. The input earth is held at +6 Vwith respect to the circuit earth by potential dividerRi-R2-R3-R4. The two 741s function as com-parators: the output of ICI goes high when the in-put signal is greater than + 50 my, whereas the out-put of IC2 goes high when the input signalbecomes more negative than -50 mV. ResistorsR6, R7, and Rs form an OR gate that drives tran-sistor Ti. If the output of either ICI or IC2 is logic1, Ti conducts.

100mA

12V500mA

The 555 operates as a retriggerable monostable,whose period is determined by Rio and Ci. Thedevice is triggered when its pin 2 is earthed by theclosing of S2. Its output, pin 3, then remains highfor 1 to 2 minutes, depending on the leakage cur-rent of the 555. The monostable resets itself as soonas the potential across Ci exceeds a certain value.As long as there is an input signal to the circuit, Ticonducts and CI remains uncharged. As soon asthe audio signal ceases, Ti switches off, and Cicharges until the potential across it is sufficient toreset the 555. The monostable may also be reset byclosing Si, which connects pin 6 of the 555 to+12 V.When IC3 is reset, Ci is discharged via its pin 7.Resistor Rn serves as protection, because withoutit Ti could short-circuit the supply lines.When the output of IC3 goes high, Tz conducts,the relay is energized, and the relay contacts switchon the mains voltage as appropriate. To counter theinduced potential when the relay contacts close,which could damage T2, diode Di has been con-nected in parallel with the relay coil.

12V

86408.1

9

002 AUDIO LINE DRIVER

Integrated operational amplifiers are not alwayssuitable for applications where a high signal level(1J0:510 V.) is required for driving a relativelylow impedance (Z=50-600 Q). The amplifier de-scribed here is eminently suitable as a high dynamicrange line driver or power buffer in public addresssystems and AF distribution amplifiers.The input amplifier of the line driver is formed bya low noise opamp Type OP -37 from PMI. This en-sures the following technical specification of theline driver: U0=70 Vpp max.; 10=400 mApp max.;Dtot =0.01% at U0=10 Vrms, ZL = 50 Q andSIN 90 dB.Regulators T, -T2 bring the supply voltage for theOP -37 down to ± 15 V. The complementary poweroutput stage is formed by T3 -T4 The amplifier hasa standard negative feedback circuit R, -R2, whichresults in a voltage gain A, = -(R2/121). A localfeedback R3 -R4 has been included to keep the out-put voltage of the opamp within safe limits, whilecapacitors Cl-C2 serve to improve the stability. Itshould be noted that the value of Cl and C2depends on the construction of the line driver:typical values are 680 pF for Cl and 22 pF for C2.In a prototype of the circuit, neither capacitor wasrequired for the frequency response to remain flat( ± 1 dB) up to 100 kHz.Resistors Re should drop just enough voltage forT3 and T4 to start conducting (class A -B operation).The quiescent current of IC, is about 3 mA, so that150 Q can be taken as a suitable starting value forRe. The quiescent current in the power output

R2Av 711

112W

* = see text

36V

87459

stage should be between 20 and 50 mA. Highervalues of Re cause the quiescent current, andhence the power dissipation, to increase, resulting inless distortion. The power output stage is not pro-tected against thermal overloading, so that due dareshould be taken in adjusting the quiescent current.

003 AUDIO TRANSFER EQUALIZER

Limiting the bandwidth of an audio system to20 kHz affects the behaviour of the system in thepass band. The steeper the filter characteristic, thegreater the phase shift in the pass band. That phaseshift stands in non-linear relation to the frequency,and this causes a frequency -dependent delay of thesignals (increasing with frequency from about4 .. .6 kHz). This effect is audible.The CD (compact disc) player is an example of asystem in which the bandwidth has been so limited.Particularly the Sony CD player and its clones suf-fer from a frequency -dependent transfer time. The

Philips (and Philips -derived) system does not sufferfrom this effect.The effect can be negated by introducing a delay inthe transfer time of the frequencies below4 . . .6 kHz, which equalizes the delay over vir-tually the entire audio range. In other words,transfer of all audio frequencies is carried out at thesame speed as it should.Such a delay is realized by phase shifter A2 (left-hand channel) and A4 (right-hand channel) in theaccompanying figure. The maximum delay for thelowest frequencies is

10

A1,A2 = IC1 = NE5532NA3,A4 = IC2 = NE5532N

05462

2R5C5 = 2R6C6 = 36 f4s.The circuit is connected between the output of theCD player and the AUX or CD input of the mainamplifier.

15 V20 ,,A

ci

=20 mA

IC1

004 COMPRESSOR

This versatile circuit serves to raise the average out-put power of an AF amplifier. Its simplicity makesit suitable for applications in intercom systems,public address and discotheque equipment, and alsoin various types of transmitter.Compression of music and speech essentially entailsreducing to some extent the dynamic range of theAF input spectrum in order to drive an AF poweramplifier with a fairly steady signal level just belowthe overload margin, thus increasing the averageoutput power of the system. However, some distor-tion is inevitably incurred in the process of amplify-ing the relatively quiet input sounds and at-tenuating the louder sounds. It is evident, therefore,that the control of the amplifier/attenuator func-tion in the compressor determines to a large extentjust how much distortion is introduced by the cir-cuit.Before inserting any type of compressor in an AFsignal path, due consideration should be given tothe attack time i.e., the time it takes the circuit todetect and counteract a sudden increase in the am-plitude of the incoming signal. Allowing for per-

sonal preference and the character of the inputsignal (speech, popular music, etc.), the attack timeof a compressor generally lies in the range from 0.5to 5 ms. The release time of the compressor is thetime it takes the circuit to return to the settings thatexisted before the rise in amplitude occurred. Con-trary to the attack time, the release time is usuallyof the order of seconds. If it is made too short, thecompressor's attenuating action may cause inter-ference with the lowest components in the fre-quency spectrum. On the other hand, too long a re-lease time (10-15 s) is also undesirable as this willgive rise to an unrealistic and unpleasant effectcaused by the output sound remaining completelymuted long after the increase in input ampli-signalamplitude. In practice, the release time of a com-pressor will need to be adapted to meet the demandof the particular input signal; speech generally re-quires a longer release time than music. Some com-pressors have a provision for the setting of the re-lease time, but the one proposed here is an auto:ranging type, that is, it arranges for the release timeto change automatically with the instantaneous

11

amplitude of the input signal.Figure 1 shows the circuit diagram of this com-pressor. Despite its simplicity, the design respondsadequately to a good number of contradicting re-quirements. As to its dynamic characteristics, an in-put signal change from 25 mVpp to 20 Vpp( r ---s 58 dB) is compressed into an output signalchange from 1.5 Vpp to 3.4 Vpp (.4s--7.1 dB). For aless extreme signal change, e.g., from 25 mVpp to2.5 Vpp (A40 dB), the compressed output signalchanges from 1.5 Cpp to 2.25 Vpp (=3.5 dB). Thecircuit has an extended frequency response fromabout 7 Hz to 67 kHz nominally, thanks to the useof a fast opamp, the Type LF357 (IC,), which is setup here to provide an amplification of about 471[(R6 + R5)/R5]. Capacitor C3 blocks the directvoltage at the inverting input of ICi, and with R5sets the low -frequency roll -off of the opamp alone atabout 16 kHz.Resistors R3 and R4 bias the non -inverting input ofthe opamp-and hence its output-at half thesupply voltage, ensuring optimum linearity. Capaci-tor C2 feeds the input signal to the opamp whileblocking the bias voltage at pin 3. Its value is notcritical, but it has some effect on the low -frequencyresponse of the compressor. The attenuator sectionin this circuit is essentially composed of RzandThe collector of this transistor is held at 0 V withthe aid of R1 and R2. In this way, Ti is always oper-ated in its saturation region, and its collector -emitter junction acts as a variable resistance con -

2a

87411-23

3V

trolled with the current fed to the base. The higherthis current, the lower the c -e resistance, and thehigher the instantaneous attenuation of the signalfed to IC,. The controlling rectifier is composed ofDl -05-R7. Transistor T2 functions to provide thecharge current for C7 so as to avoid distortion other -

12

wise incurred by too heavily loading the IC, output.The rectified voltage across C5 is a direct measureof the output signal amplitude, and forward -biasesthe base of T,, which regulates the attenuation asdiscussed. The use of a diode with a low internal re-sistance, Di, and a buffer, T2, ensures fast chargingand slow discharging of C5, and thus a short attacktime and a long release time, respectively. As C5 isdischarged via R7 and the base resistance of T1, therelease time of the compressor is the product of thevalue of these three components. When the basebias is reduced, the base resistance of Ti increases,lenghtening the release time. This is a mostwelcome feature, especially with speech signals.The output of the opamp is fed to Ca -Pi -Rio, whichprovide DC insulation and level adjustment.Two compressors are readily combined to make astereo version by feeding them from a common bat-tery and connecting points X and points Y (neverX to Y!). In this case, Ti and Di in both com-pressors must be matched types to ensure proper op-eration. Figure 2 shows two simple test circuits forselecting transistors and diodes with matching DCcharacteristics. The basic method is to start withnoting the voltmeter reading for a particular device,and then find a matching type from an available lotby inserting devices until one is found that gives thepreciously noted test voltage. In the diode test cir-cuit, the LED lights to indicate the absence or re-verse connection of a diode under test.

A = 0 VB = +4.5 VC = 6 mAD = 3.9 V

All values are typical andwithin 10%.

All voltages measuredwith respect to groundwith a DMM (Zin = 1 MO).

Provision has been made to use the circuit as anoise suppressor. Referring to Fig. 1, closing Si con-nects C8 across the regulator transistor to form alow-pass filter in conjunction with R1 and R2. Thecut-off frequency of this LPF is a function of thecurrent sent into the base of Ti. The overall effectthus obtained is an effective elimination of noisefrom quiet passages in the programme. For louderpassages, the suppression of noise is not so import-ant, as it is then virtually inaudible.Finally, when using this compressor, make sure thatyour amplifier has ample cooling provision, becauseit may well be continuously operated at the top ofits power rating. For the same reason, checkwhether the loudspeakers can handle the availablepower.

005 CURRENT CORRECTED AF AMPLIFIER

The majority of modern AF power amplifiers drivethe loudspeaker(s) with a voltage that is simply afixed factor greater than the input voltage. It is

fairly evident, therefore, that the power delivered bysuch amplifiers is inversely proportional to the loud-speaker impedance, since the cone displacement ofa loudspeaker is mainly a function of the currentsent through the voice coil, whose impedance mayvary considerably over the relevant frequencyrange. In multiway loudspeaker systems, this diffi-culty is overcome by appropriate dimensioning ofthe crossover filter, but a different approach is calledfor when there is but one loudspeaker.This amplifier is based on current feedback to en-sure that the current sent through the voice coil re-mains in accordance with the input signal. The cur-rent through the voice coil and R7 develops avoltage across the resistor. A negative feedback loop

0

_7=1W =MKT

A=OVB=OVC=17V40=-17V4E=OV

07410

13

is created by feeding this reference voltage to the in-verting input of ICI. The overall amplification ofthe circuit depends on the ratio of the loudspeaker'simpedance, ZL, to the value of R7. In the presentcase the amplification is 16 times (ZLJR7 = 8/0.5 =16).

The connection of the opamp's output to ground isslightly unusual, but enables the base current foroutput transistors Ti -T2 to be drawn from thesupply rails, rather than from the opamp. CapacitorC6 functions to set the roll -off frequency at about90 kHz. The quiescent current of the amplifier is ofthe order of 50 to 100 mA for class A operation,

and is determined by R3 -R4 and Its -R6. The com-plementary power transistors should be closelymatched types to avoid fairly large offset currents(and voltages) arising. Some redimensioning ofeither R3 or R4 may be required to achieve the cor-rect balance for the power output stage. The emittercurrent of Ti and T2 is about 500 mA when theamplifier is fully driven.The harmonic distortion of this amplifier is lessthan 0.01% at P0= 6.25 W and Ub = ±18 V.

Source: Texas InstrumentsLinear Applications.

006 DIGITAL AUDIO SELECTOR

Switching audio signals digitally could be done withthe aid of CMOS analogue switches or multiplexers.Simple as this may seem, there is, however, an in-evitable loss in the quality of the sound due to thenoisy nature of CMOS switches. Furthermore, thehigh on -resistance of these devices together with thelarge parasitic capacitances generally present in

CONTROLINPUT

{815

D2

13

312

4

A

B

C

IC1

0001

02

03

04

05

06

R13

2 _-3

6

07 - -08 10 --09

11

C111220n

C101 1220n

11

R22 VD11 R12

CMOS circuits causes a high susceptibility tocrosstalk. The circuit given here is a novel way ofselecting one out of ten audio signals digitally with-out any of the foregoing drawbacks.As shown in the circuit diagram, the ten inputsignals numbered 1-10 are applied to the bases oftransistors T, -Tio via capacitors Ci -C, 0 respect -

A= OV9B= 1V6C = 0V2I)=14V6E = 14V6F = 13V60= 14V2H= 2V1J= 2V1

RI RIO

1N41481D12

R11

T11BC

560 C

11,800pA

Th

16

IC 378 L 05

IC2 IC1C15 C12

0 100n 1076V

C14min70n

ZuD1D10

(D ,

10x % 110BC 560 C

10x1 N 4148

15 V13 mA

C13

101 25V

6C11

2121

IC 2LF 356

NOM

IC = 74 LS 45, 74 LS 145,74 HCT 45, 74 HCT 145

87443

015V

7 mA

e

14

ively. The bias voltages for the transistors are ob-tained with the aid of RI -Rio. Depending on thebinary state applied to IC), one of its outputs Qo-Q9 goes low. For example, if the input code is 0010,Q2 goes low, pulling the base of T3 to 0 V, while thebases of all other transistors are raised to nearly+15 V. Therefore, T3 works as an emitter followerwhile the other transistors are effectively reversebiased. The output rail of the transistor array is con-nected to voltage follower IC2, which provides theoutput signal of the digital audio selector.Voltage regulator IC3 is required only if a + 5 V railis not available. If the number of channels requiredfor a particular application is less than 10, the rel-evant components can be omitted. If a mute facilityis required, simply short one input to ground tosilence the output on selection of the correspondingchannel.

This circuit can handle input signals up to 4 Vrms.The total distortion does not exceed 0.01% for fre-quencies up to 20 kHz. The crosstalk incurred inthis circuit is less than -80 dB. This value can beattained by paying due attention to the layout ofthe practical circuit, the decoupling of the supply.lines (fit C14 and C15 direct to the relevant pins ofthe opamp), and the use of good quality compo-nents.The measuring values indicated in the circuitdiagram were obtained in a prototype. All voltagesare measured with respect to ground with the aid ofa DMM (Zn = IMO). The channel selected wasnumber 1.

007 DIGITAL VOLUME CONTROL I

This digital potentiometer circuit is a hybridanalogue and digital design offering push-buttoncontrolled programmable attenuation as well ashigh to low impedance conversion by means of asingle active device. Digital noise is eliminated as ef-fectively as possible through galvanic isolation ofdigital and analogue parts in the input attenuator.At the heart of the digital control section is a Type2716 EPROM, which can be programmed either asshown in Table 1 or to individual requirements, aswill be detailed below. At power -on, debouncerbistables N1 -N2 and N3 -N4 force logic low levelsonto EPROM address lines As and A6 respectively,selecting a programmed address range that suppliesthe digitally coded, initial volume settng. R -C net-work R16 -C2 causes gates N7 and Na to generate aclock pulse for IC2, which latches the 8 -bit wordfrom ICI, passes this information to driver ICs, andthus determines which relay(s) is/are energized,thereby fixing the attenuation before the AF signalis applied to opamp IC6. Depression of S (up) or S2(down) causes the corresponding address line As orA6 to go low, selecting a certain address range inthe EPROM. The exact address location is deter-mined by the value last latched into IC2 aftereither key has been released. It is readily seen thatthe five available databits at the Q1 Qs outputsof 1C2 allow 32 (25) simulated potentiometer set-tings.

The digital control section has been designed to of -

Table 1

0000 00 01 02 03 04 05 06 07 08 09 OR OB OC OD OE OF0010 10 II 12 13 14 15 16 17 18 19 IA 113 1C ID IE IF

0020 00 00 01 02 03 04 05 06 07 08 09 0A OB OC OD OE0030 OF 10 II 12 13 14 IS 16 17 18 19 IA 18 IC 10 IE0040 01 02 03 04 05 06 07 08 09 OA OB OC OD 0E OF 100050 11 12 13 14 IS 16 17 18 19 IA 18 IC ID IE IF IF

0060 OE OE OE OE OE OE OE OE OE OE OE OE OE OE OE OE0070 OE OE OE OE OE OE OE OE OE OE OE OE OE OE OE OE0080 00 01 02 03 04 05 06 07 08 09 OR OB OC OD OE OF0090 10 11 12 13 14 15 16 17 18 19 IA 113 IC ID IE IF

00A0 00 00 00 01 02 03 04 05 06 07 08 09 0A OB OC 000080 OE OF 10 11 12 13 14 15 16 17 18 19 IA IB IC ID0000 02 03 04 05 06 07 08 09 0A 08 OC 00 OE OF 10 110000 12 13 14 15 16 17 18 19 IA 1B IC ID IE IF IF IF

00E0 clE OE OE OE OE OE OE OE OE OE OE OE OE OE OE OE00E0 OE OE OE OE OE OE OE OE OE OE OE OE OE OE OE OE

fer an auto -repeat function when either one of thestep control keys is kept depressed; oscillator gateN6 then provides a clock pulse train to N7 -No, andso causes successive addresses in ICI to be scannedautomatically, until either the lowest or highestpossible volume setting is reached, at which mo-ment the circuit forces itself to a hold state, whichcan also be selected at any time by simultaneouslydepressing the up and down key.S3 enables the user to select a further addressblock, programmed with another set of volumesteps; the circuit as shown, along with the data fromTable 1 arranges for 3 dB steps.

15

-48 dB -24dB -12 dB +3 dB -6 dB

U

DOWN

S2

RI R3 R5

N1...N4 = IC3 =74LSOON5...N8 = IC4 =74LS132IC6 = 5534

The analogue section of the circuit is basically afour -section, relay -controlled attenuator composedof resistor networks to achieve a signal attenuationin 3 dB increments, as defined with the relevant bitpattern at the Qi . .Q5 outputs of IC2. Rea (Q5), ifdeactivated, enables IC6 to amplify its input signalby 3 dB. The inset resistor and preset combinationmay be used take over the function of Cio, sincethe latter should be a high stability foil type, whichmay be a rather difficult to obtain part. Both circuitalternatives function as click suppressors when step-ping through the available range of volume settings.The preset, if used, should be set for zero offsetvoltage at pin 6 of the opamp; replace Cm with awire link.It is suggested to use miniature DIL relay types inthe Rei Res positions, while all resistors in theattenuator are preferably close toler-ance (1%), high stability types. Also observe thatthe supply voltages to analogue and digital sectionare kept well apart and decoupled so as to precludeintroducing switching pulses and digital inter-ference in the sensitive attenuator sections as wellas the opamp output stage.Finally, Table 1 offers a suggestion for programming

Vcc

Gem

OE

CLK

R9

1,911. 115

09 7 6 5 4 3 2

74LS374

D8 7 6 5 4 3 2

24 V" D 7 6 5 4 3 2 1 0

21 VvaIC1

21 GN°R17 2716

'76 a A10 9 8 7 6 5 4 3 2 1 0

...."7"."r19 27123 1 2 3 4 5

®189 y9 m A

CLL 11

C)leV

IC6T74(;)

R19

-120mA*

° 5V

0.

CaL C.51T C.:31. a ciIC3 IC410°7 T7n

4V- 77n0 T6V

000k

T

86423-1

* without relay

the EPROM with data to achieve circuit operationas set out above.

16

008 DIGITAL VOLUME CONTROL II

Many of today's hi-fi amplifiers feature a "clicking"volume control, but this is only rarely a real steppedattenuator based on a wafer switch. In nearly allcases, this expensive system is based on a normal

9RR14

potentiometer, whose spindle is fitted with a mech-anical construction to simulate the stepping move-ment. A normal rotary switch is not suitable for ad-justing the volume of an amplifier because it briefly

5V

R1

100n

Up

2

10

ES1 1

1

ES 1 ...ES 4 = IC 1 = 4066ES 5 ... ES 8 = IC 2 = 4066

N 1 N 4 = 2/3 1C 3 =4049

N5, N6 = V2 1C 4 =4011

15

U/D

CLK

01

1C 5 02Q3

04

E. 2 2 a¢ T. in

4516

ES3 8

9

ES4 :1

12

10ES5 1

13

R3

R4

R5

R6

R7

C2

6

11

4 112 113 13 19

14

2

0

2

ES6 4

O

3

ES 7 8

E 8

12

9

11

R9

1110

R11

R12

100n00

(SL,;LL105 1C4 1C3 1C2 1C1

T TT TTR2 R8

5V

O87401 17

disconnects the input from the signal source whenoperated, and so readily gives rise to clicks and con-tact noise.Different problems crop up when designing an elec-tronic volume control. Of these, distortion is prob-ably the hardest to master, but reasonable resultsare still obtainable, as will be shown here.Basically, there are two methods for making an elec-tronic potentiometer. One is to create a tapped re-sistor ladder (which is not much different from anormal potentiometer), the other is to change theresistance of the two "track sections" such that thetotal resistance remains constant. The circuit pro-posed here is based on the second method, andfeatures 16 steps in its basic form. The number ofsteps can be increased to, say, 64 by adding fourswitches and resistors.The electronic potentiometer is composed of twoequal sections, which have a total resistance of15 k52 each. The electronic switches in each sectionare controlled by binary counter 105. Since the

switches in section ESI-ES4 and those in ES5-ES8are controlled in complementary fashion, the totalresistance of the potentiometer remains constant.Resistors Ri-R2 and R7 -Rs serve to keep the poten-tial at the input and output at 0 V so as to precludeclicks when the step switch, S2, is operated. SwitchSi is the up/down selector. Gates N5 -N6 form abistable to ensure that the counter is clocked withdebounced step pulses.The number of steps can be increased by adding acounter and the required number of electronicswitches, divided over the two "track sections".These switches are then connected in parallel withresistors whose values correspond to binary order 1-2-4-8, etc., as shown in the circuit diagram. For-tunately, precise binary ratios are not required here,since adequate results are obtainable with approx-imations of the theoretical resistance values, and aslong as the actual resistors are kept equal in bothsections.

009 DISCOMIXER

This mixer is a typical example of the way moderncomponents can, and do, simplify the realization ofgood quality audio circuits. In the given configura-tion it is eminently suitable for use as a discomixer,but the number of input channels can easily beenlarged.As can be seen in figure 1, in its basic form the mix-er has four input channels. These could, for in-stance, serve as inputs for a microphone, stereo pick-up, and cassette player or tape recorder.The power supply has been kept as simple as poss-ible; if it proves difficult to obtain the XR4195 regu-lator IC, it may be replaced by a combination of a78L15 and 79L15. The transformer is preferably ofthe PCB type to keep the mixer as compact as poss-ible.The values of Ci and Ri are dependent on the typeof microphone used. If this is a high -impedancetype, the values should be 470 nF and 22 k52 re-spectively, whereas with low -impedance types,10 µF and 680 Q are required.Unfortunately, miniature bipolar electrolyticcapacitors (CI, CI', C9, and CO are not yetavailable everywhere, although they are almostindispensable in applications such as described here.Standard electrolytics may be used with maximumreverse voltages of 1 V, but their use introduces dis-

tortion and premature ageing (because of the re-verse polarity).Provision has been made on the printed circuitboard for up to four channels. Two or more PCBsmay be connected together; the output and supplysections may then be cut off as required.Current consumption is about 10 mA per channel.

Parts list

Resistors:

R.1*...Rs*,Rv*...R5'* see tableRs*,Rs*,Ra,Ra' = 47 kR7,RT = 22 kR9*,Rw* = 100 kPia*,Pib* = 22 k stereo slide potentiometer,log, 58 mm long

Capacitors:

= see tableCs*,Cs' = 470 nCs",C7",C1o,C11,C1a,C1s = 100 nCs,Ca' = 10 pCs,Cs' = 10 /1/25 VC12,C13,C14,C15 = 22 nC1e,C17 = 470 p/25 VC2o,C21 = 10 p/16 VC22,C23 = 100 p

18

Semiconductors:Di . . .D4 = 1N4001D5,D6 = 1N4148ICi* = NE5532 or LM833IC2 = TL072IC3 = XR4195

Miscellaneous:

Tri = mains transformer, secondary2 x 15V/100 mA

Fi = fuse, 50 mA, delayed actionSi = DPST on/off switchSingle -hole fixing chassis phono socket - 2per channel

PCB 85463

*One of each required per channel.

Table 1.

1:15. go table

Al ,A2 = IC1 = NE 5532, LM 833

0 C2 Ca

DRO 'I±H±F

C1

C1'C2

C2'C3C3'

C4C4'

R1

R1`

R2

R2'

R3

R3'

R4

134'

R5

R5'

pick-up 220 n 1n5 1n5 3n3 47 k 2k2 2k2 100 k 1 M

tape/cassette *** *** *** *** *** *** see Note 1

microphone(high im-pedance) 470 n *** *** 10 p 22 k 1 k *** 0 -0 100 k see Note 2

microphone 10 ,,,i/(low impedance) 25 V *** *** 10 p 680

Q 1 k *** 0-0 100 k see Note 2

Note 1. Wire links A-B and A'-B' required; ICi, Cs, and C7 not required.

Note 2. With mono microphones, use input R; do not connect Pi b; wire link C-C' required; allaccented components not required.

o-o = wire link*** = not required

15 V

A3,A4 = IC2 = TL 072

19

o-®c%*1/4"712o_o_o

fa; jiao-o-Ake4;

oWA

A 2°

04,12 4

\DA

.., 0 -:1A3'

II-Wreri. e-0_0;\eIsebi

Vo\o oo

20

IJ

iidiai 0 w

0 \311,0 so

;'

.4o

0 930;

ci.to,

-4,- x....31,,

nE

l

Z' -

..$

°1,i4114)

2t=t1,0

.211

El+

)04*C

1=1141';

1;)1 .49otl,th0

0711-00PFPO

-11-#

40 0-1P-0 2.911-0F44

91s Fb-latt

alm

0?--a

a

ba

Fg) caL

d0-1101`010-11-0

9114 29-11-0

FO_2_ W

Wso

33-

0-0.00

4OV

93

.<1

010 HEADPHONE AMPLIFIER

OP -50 Power operational amplifierFeatures: Open -loop gain: Input offset voltage: Input bias current: Offset voltage drift: Common mode rejection ratio: Power supply rejection ratio: Noise level:

V/V min.25 IN max.5 nA max.0.3 NV/°C max.126 dB min.126 dB min.5.5 nV//Hz (f=10 kHz)4,5 nV/V-Hz (f =1 kHz)

Output current: ± 50 mA Drives capacitive loads up to 10 nF. On -chip thermal shutdown circuit.

Data taken from manufacturer's data sheet.

There is little doubt that the headphone amplifierdescribed here belongs in the so-called high endclass of audio equipment, and is, therefore, perfectfor incorporation in, or adding to, the Top -of -the -range Preamplifier described in (ii, although it isalso suitable as an autonomous, high quality, unit.The circuit diagram of the headphone amplifier ap-pears in Fig. 1. The unit is based on Type OP -50

1 15V

15V

15V

power operational amplifiers, whose technicalfeatures are summarized in Table 1. Clearly,everything feasible has been done by the manufac-turers, Precision Monolithics Inc., to ensure opti-mum operation of the device, and it is with this inmind that the remainder of the amplifier was de-signed.Both supply rails to the amplifier ICs are adequatelydecoupled and filtered with a small series resistor,(R4 -R5) and a combination of an electrolytic and asolid capacitor (C4 -C2 and C5 -C3). With reference tothe upper of the two identical channels, preset P2

enables compensating the (small) offset voltage atthe output of the OP -50, while Cl-R3 forms a com-pensation circuit to minimize overshoot for a givenclosed -loop voltage amplification, AVCL. In thepresent application, AVCL is about 6, since

Rt -=R2/(Avu-1)

When it is intended to alter the amplification, R2

25241 1N4001

87512

22

should be left at 20 kg. Also observe that the indi-cated values for R3 and C, are valid when AVCL isbetween 5 and 20, while R3 = 3.3 kg and C, =1 nFwhen AVCL is between 20 and 50. No R -Ccompensation is required when Am. is greaterthan 50.The +15 V supply for the headphone amplifier isa relatively extensive circuit based ona precision regulator Type LM325, which featuresexcellent noise suppression whilst ensuring smoothand simultaneously rising output voltages at power -on. Mains -borne interference and clicks from S, aresuppressed in varistor R9 and high -voltage capacitorC19. The four diodes in rectifier bridge B, arebypassed with rattle suppression capacitors to en-sure minimum noise on the supply rails to theopamps.

The headphone amplifier can function optimallyonly if great care is taken both in the choice of thecomponents and in the construction on PCB Type87512, details of which are shown in Fig. 2. Asalready stated, the headphone amplifier is suitablefor building into the Top -of -the -Range Preamplifier.This makes it possible to feed the ± 15 V regulatorfrom the raw voltage across C9 (+) and C, 0 I-) ofthe existing ± 18.5 V supply, while the inputs of thevolume control of the headphone amplifier aredriven direct from the outputs of IC4 (R) and IC4'(L).Opamps IC, and IC2 should be soldered direct ontothe PCB, and are preferably fitted with a DIL-typeheatsink. Provision has been made to screen theamplifiers and the supply on the board by means oftwo sheets of brass or tin plate, which are mountedvertically onto the dotted lines, and secured withthree soldering pins each. Series regulators T, andT2can do without a heat -sink. When the board is com-plete, its underside should be thoroughly cleanedwith a brush dipped into white spirit or alcohol toremove any residual resin. Next, the track side issealed with a suitable plastic spray.

When possible, use insulated sockets for the stereoinput and output of the amplifier. At the input side,few problems are expected to arise when using gold-plated phono sockets mounted onto a separate ABSor epoxy plate. When a good quality, insulated,6.3 mm, stereo headphone socket proves unob-tainable, the nearest alternative is a non -insulatedtype, whose common tag is connected direct to theground point on the PCB, between C17 and C18 toeffect central earthing. Mains transformer Tr, is

preferably a toroidal type fitted behind a metalscreen to ensure minimum hum and other inter-

ference picked up by the amplifier inputs. Presets P2and P2' are trimmed for minimum offset voltage atthe respective amplifier output-this is likely to re-quire a very sensitive DMM. The headphone ampli-fier can be terminated in 100 Q to 1 kg, and istherefore perfect for use as a high -quality line driveralso. The outputs are short-circuit resistant.Finally, a brief summary of the amplifier's expectedperformance at Vo = 6 VMS and AvcL 6:Total harmonic distortion: 0.0025% (100 Hz);0.003% (1 kHz); 0.011% (10 kHz).Signal-to-noise ratio: ..80 dB.Response flatness: ± 0.4 dB from 10 Hz to 20 kHz.

Literature references:(1) Top -of -the -Range Preamplifier. Elektor Elec-tronics, November and December 1986, January1987.(2) Linear and Conversion Applications Handbook(1986). Precision Monolithics Incorpor-ated.

Parts list

Resistors (± 5%):Fi1;Ri'=4K02F*R2; R2' = 20KOFRs; Rs' = 560RR4;R4';R5;R5';R6;R7=2R2R8 = 820R; 0.5 WRe= SIOV S10 K250 varistor (Siemens;ElectroValue (0784) 33603).

Rio=2M2Pi = 25K logarithmic stereo potentiometer.P2;P2' = 100K multiturn preset.

Capacitors:

Ci;Ci'=4n7C2;C2';C3;C3' = 220nC4;C4';C5;C5'=4701.,z; 16 V; radialC6;C7;C8;Co=22nCio;Ci2=100014; 25 V; radialC1i;C13;C15;C16=100nC14=11.4; 16 V; tantalumC19 = 22n; 250 VAC

Semiconductors:... D6 incl.= 1 N4001

D7 = LED redICI;IC2=0P-50 (Precision Monolithics Inc.)IC3=LM325Ti;T2=BD241

23

Miscellaneous:

Fi = 250 mA delayed action fuse plus panel -mount holder.

Tri = 2 x 15 V; 15 VA 2 x 0.50 A) toroidalmains transformer, e.g. ILP Type 03013.

DIL-14 heat -sink for IC, and IC2.Mains entrance socket.PCB Type 87512

= SPST miniature mains switch.Stereo 6.3 mm headphone socket, preferably

insulated.2 off phone input sockets.Suitable metal enclosure.

* See text

ceor]l ri

j't° %10UI Iji

24

011 HI-FI HEADPHONE AMPLIFIER

This 1 -watt amplifier lends itself par excellence foruse as driver for a low impedance headphone or asoutput stage in a hi-fi preamplifier driving an activeloudspeaker. Many preamplifiers do not permitlong, unscreened leads to be connected to them, butthe present amplifier accepts these happily.The circuit - figure 1 - consists of an opamp typeLF 356 and a push-pull transistor output stage.Low-pass filter R,/C2 at the input limits the slewrate of the input signal. In conjuction with the rela-tively fast LF 356, this results in very low delay dis-tortion. The fixed quiescent current of 30 mAdrawn by the output transistors, and set by diodes131 . .D4 in conjunction with emitter resistors R7and R8, ensures very low crossover distortion.Feedback resistors R3 and R4 fix the gain at about15 dB. The consequent overall distortion with a -3 dB bandwidth from 10 Hz to 30 kHz is only 0.1per cent.The amplifier delivers a maximum power of 1 watt

2

cc

C

C1

R2

a

T1

into 8 Q for an input signal of about 500 mVrms.High -impedance headphones and 4 Q loudspeakersmay also be connected without detriment.The amplifier is best built on the printed circuitboard shown in figure 2. To enable it surviving ashort circuit at the output, the two transistorsshould be mounted on heat sinks - do not forgetthe insulating washers and the heat conductingpaste!The power supply need not be more than a simpleaffair, consisting of a mains transformer with acentre -tapped, 6 . . .8 V, 0.5 A secondary, a suitablebridge rectifier, and two 1000 plF/16 V electrolyticcapacitors in a conventional arrangement.To drive high -impedance headphones at highvolume, you need a ± 15 V regulated power supply:in some cases, this may be derived from the pre-amplifier supply. In this arrangement, care must betaken not to short-circuit the output terminals.

Parts list

Resistors:

11.1 = 10 kR2,R4 = 100 kR3 = 22 kR5,R6 = 1 kR7,R8 = 22 Q

Capacitors:

Ci = 22 nC2 = 330 pC3 = 1 p

C4,C8 = 100 n...D4 = 1N4148

Ti = BD 135 or BD 139T2 = BD 136 or BD 140IC, = LF 356

Miscellaneous:

PCB 85431Heat sinks for Ti and T2

4;j6{kz..\.31

Ak _9JVilrYN15):

25

Naaml oos-6.indd 1 28-08-2008 10:01:01Naamloos-6.indd 1 28-08-2008 10:01:01

012 HIGH DYNAMIC RANGE MIXER

A mixer is expected to have low -noise and highdynamic performance. Most standard mixers use in-verting operational amplifiers. Unfortunately, thenoise figure of many opamps is poor, and opampswith a good noise figure are normally not suitablefor operating with large signals.The noise factor of standard circuits is often madeeven poorer because the source and amplifier arenot properly matched.The characteristics of a mixer can be greatly im-proved, therefore, by the use of buffers at the inputstages, and the constructing of operationalamplifiers from good -quality transistors. This hasbeen done in the accompanying circuit. The inputis buffered by T1 and T2. The input impedance ofT1 can be ignored, so that the source merely needsto be matched with PtThe opamp is formed by transistors T3 to T8 incl.Good -quality RF transistors have been used in dif-ferential amplifier T3 -T4 -T5. These transistors have

15V

®

15V

0O

2k2

C801 -100p25V

R11*

2k2R12

R13

P2

BF256C47klog

BF256B

C1

2p

a better noise figure at a greater bandwidth thanAF types.The proposed circuit has a frequency range (-3 dBpoints) of 10 Hz to 80 kHz; third harmonic distor-tion of not more than 0.05 per cent at 10 kHz andan output voltage of 9 Vpp; and a signal-to-noiseratio of 100 dB.The signal-to-noise ratio applies to an output signalof 9 Vpp with open -circuit input, and a bandwidthof 20 kHz. The maximum value of the outputsignal is about 12 Vpp, measured across a load im-pedance of 560 ohms. If the mixer is terminated bya higher impedance, the output voltage will begreater.A further advantage of the circuit is that thepopular valve sound may be realized in a simplemanner. To this end it is necessary that Ti and T2commence limiting at a slightly lower level, i.e.

12 Vpp input, than the composite opamp. Thesupply voltage of Ti and T2 must then lie between

T3

BF494

R1

82p

BF494

R5

74 C411

82p

D1BF494

1N4148

1N4148

1N4148

BC550C75 T7

red

C6

Cl 10pit 25V

BC337

R10

CIE

BC327

10P/25V

®

86516

26

± 6 V and ± 9 V. Since T2 is connected as a currentsource, the exact supply voltage can be set with the2k2 preset at the wanted clipping level.If desired, the output off -set may be zeroed by inser-ting a 50 kilo -ohm preset in the base circuit of T4.This base should also be decoupled by a 1µF, 63 Vcapacitor.

The current consumption of the opamp is about35 mA and that of the buffer stages not more than10 mA. If, therefore, ten buffer stages are used, thepower supply should be capable of providing150 mA at ±15 V.

013 INTEGRATED STEREO AMPLIFIER

The Type TDA1521 from Valvo/Mullard is an inte-grated HiFi stereo power amplifier designed formains fed applications such as stereo TV. Thedevice works optimally when fed from a ± 16 Vsupply, and delivers a maximum output power of2 x 12 W into 8 Q. The gain of the amplifiers isfixed internally at 30 dB with a spread of 0.2 dB toensure optimum gain balance between the chan-nels.A special feature of the chip is its built-in mute cir-cuit, which disconnects the non -inverting inputswhen the supply voltage is less than ± 6 V, a levelat which the amplifiers are still correctly biased.This arrangement ensures the absence of unwantedclicks and other noise when the amplifier is

switched on or off. The TDA1521 is protectedagainst output short circuits and thermaloverloading. The SIL9 package should be bolted

P1

100klog. Cl

onto a heatsink with a thermal resistance of nomore than 3.3 K/W (RL= 852; Vs= ± 16 V;Pa =14.6 W; To = 65 °C). Note that the metal tabon the chip package is internally connected to pin 5.The accompanying photograph shows that this highquality stereo amplifier has a very low componentcount, and is readily constructed on a piece ofVeroboard.The following technical datathe datasheets for theVs = ± 16 V):Distortion at Po =12 W:Quiescent current:Gain balance:Supply ripple rejection:Channel separation:Output offset voltage:3 dB power bandwidth:

are stated as typical inTDA1521 (RL= 8 Q;

B1

880C5000/3300

0.5%40 mA0.2 dB60 dB70 dB20 mV20-20,000 Hz

87490

014 LOUDSPEAKER PROTECTION I

There are many ways of protecting loudspeakersagainst the switch -on 'plop': many of these rely ona clamp circuit across the power amplifier input to

hold this at 0 V for a few seconds after switch -on.Others, like the one suggested here, depend on arelay to switch off the loudspeaker(s).

27

la

b

85404 -la

85404 -lb

C OA

ON/OFF

85404-1c

e

f

85404-1e

B9 V 85404-1f

500 mA

2885404-1d

Terminals A and B of the circuit in figure 2 are con-nected to one of the sensing circuits in figuresla . . .1f, of which the pros and cons will be dis-cussed shortly. Whichever of these circuits is used,A is shorted to B immediately the power is switchedon. This cuts off transistor TI instantly, whichcauses capacitor Cl to charge. After a few seconds,the voltage across Cl causes zener D2 to breakdown. Transistor T2 and T3 then conduct; the relayis energized, and the loudspeakers are connected incircuit.When the power is switched off, T1 conducts andthis causes Cl to discharge very rapidly. The voltageacross Cl quickly drops below the breakdown levelof D2; transistors T2 and T3 are cut off, and therelay returns to its quiescent state, which discon-nects the loudspeakers.Input circuit la relies on a light -dependent resistor(LDR) fitted close to the mains on indicator lamp.When the lamp lights, the resistance of the LDRdrops sharply, so that terminal A is virtually shortedto B.The input in lb relies on a reed relay connected tothe secondary winding of the mains transformer. Assoon as the mains is switched on, the relay contactsclose.The third possibility, shown in Ic, is that the mainson/off switch has a third contact that connects A toB when the mains is switched on.A further option is illustrated in ld, where a transis-tor is connected to the secondary of the mainstransformer via a diode and resistor. The transistorconducts when the mains is switched on.The inputs in le and if also provide power for theprotection circuit. That in le has a bridge rectifierconnected across the secondary winding of themains transformer. When the mains is switched on,the BC 547 conducts and shorts A to B.Finally, the circuit in If is connected direct to themains. Here again, as soon as the mains is switched

2

A

B

* see text

T1,T2 = BC 5466T3 = BD 139

85404-2

u

on, the BC 547 conducts and terminal A is shortedto B.Whichever of the input circuits is used depends oncircumstances and/or individual preferences. If oneof circuits la . . . Id is used, a separate power supplyis required for the protection circuit. As suggested,the output voltage, Uv, of this should be40...60 V d.c. For lower values of Uv, the ratingof D2 must be reduced accordingly.Resistance R1 depends on the relay used, and iscalculated from

= [(Uv -Ur - 2.5)//,]where U, and Jr are the operating voltage (in volts)and current (in amperes) of the relay used respect-ively.The relay contacts must be able to carry a large cur-rent: 10 A is not unusual in many amplifiers.The rating of Rv is [U,/,] W.If the 'plop' is still heard, increase the value of R3as required - in reasonably small steps.

015 LOUDSPEAKER PROTECTION II

This is an all -transistor design for incorporation inAF amplifiers that produce nasty clicks in the loud-speakers when turned on or off, jeopardizing thevoice coils by passing a large current surge.Assuming that AF amplifier and protection circuitare off, Ci and C2 are empty of charge and Re isdeactivated. At power -on, Di rapidly charges Cl.Provided both the negative and the positive supplyvoltage are present and at the correct level, T2 and

T3 conduct, while Ti is off, enabling C2 to be slowlycharged via R4. If the voltage across C2 is suf-ficiently high for T4 to conduct, T5 will draw basecurrent and energize Re, which connects theloudspeakers to the amplifier outputs. Zener diodeD4 fixes the voltage across the coil of Re, so that dif-ferently rated relays may also be used in the circuit,provided D4 is changed accordingly. However therelay coil current should not exceed about 50 mA,

5W

29

1N4002 = 01

C1O 1006mom 830

2

- Ub

+ Ub

0

BC5566

BC546B

while the changeover contacts should be rated inaccordance with the amplifier output power andimpedance; for a 2 x 100 W at 8 Q type, for in-stance, the relay contacts should be rated at least8 A.Should either one or both supply voltages (-Ub;+ Us) disappear for some reason or other (amplifiermalfunction, short-circuited smoothing capacitor,etc.), the relevant transistor T2 or T3 will be disabl-ed, causing Ti to receive base current via R1; C2will be discharged forthwith and Re is deactivatedin consequence since T4 and T5 are turned off. Theamplifier channels can now produce clicks they like;

BC5568RB

*D42401W

D3 LED I

Imaz = 5061A

BD139

see text

86471.1

LS

the output is safely applied to two resistors match-ing the output impedance.The protection circuit is fed off the voltage acrossCl, which is purposely rated at only 10014F toenable Re to be deactivated almost immediatelyafter the amplifier has been switched off. Power -offclicks, if produced, will therefore end up in the dum-my resistors rather than the expensive loudspeakervoice coils.The protection unit is most readily fitted on a pieceof veroboard, while Re should be mounted close tothe loudspeaker output terminals to keep contactlosses as low as possible.

016 LOUDSPEAKER PROTECTION III

Many modern AF power output stages are capableof delivering considerable power levels in the super-sonic frequency range. When the loudspeaker cannot handle that power, the voice coil is rapidlyoverheated, and causes a short-circuit. If the poweroutput stage is not properly protected, it breaksdown and supplies a direct current that effectivelydestroys the loudspeaker.The present loudspeaker protector is composed ofthree sections: a measuring amplifier, a detector,and a relay driver. Four channels are shown here asan example. Potential divider R1 -R2 determines the

sensitivity of the protection circuit, while D1 -D2protect the input of Al. Opamp A5 is set up as a lowpass filter with a cut-off frequency of 0.5 Hz, sothat it can function as a DC detector. The secondsection of the circuit is composed of four detectorsAs - Al2. As compares any negative direct voltagesto a reference set with Rs -Rs, while C3 -R7 deter-mine the delay time. Opamp A 10 has a similar func-tion for positive direct voltages. The circuit is ac-tuated when

VisR2 0.65 > 15R2It + R2 128+119

30

Comparators All and Al2 function as the powerlimiter. Positive and negative peak voltages are recti-fied in D3 -D4 and averaged with the aid of R -Ccombinations R36 -C33 and R26 -C23 . The relativelylong periods of these networks precludes erroneoustriggering of the circuit on peaks in the input signal.The power limiter is actuated when

VinR2f2 15R280.65 >Ri + R2 R28 + R29

This equation is also valid for the positive detectorset up around Al2. The stated component valuesresult in Pmax"='30 W in 8Q.When the input signals are all right, the open col-lector outputs of As - Al2 are in their high im-pedance state, so that the output voltage is + 15 Vvia R40. When a fault condition exists at one ormore of the inputs, junction Rao -R41 is pulled downto -15 V.The central part in the relay driver is bistableGate Ni is a resettable power -up delay circuit whichclocks FF, . The logic high level at the D (data) in-put is only transferred to output Q when the R(reset) input is logic high. It is seen that a reset pulse

can originate either from the mains detector N3 -N4,or from the fault detectors As-Al2.The loudspeaker protector is conveniently fed fromthe amplifier's symmetrical supply, but care shouldbe taken to dimension D48 and such that the in-dicated voltage across C44 and C45 is not exceeded.If the amplifier supply delivers less than 28 V, IC6may be omitted, and the loudspeaker relay, Re, re-placed with a 12 V type fed from the + 15 V rail.Voltage divider R43 -R44 should then be redimen-sioned such that the input of N4 is held at about+13 V when R43+ k52.

IN

-45

©C)

IN

AS.. e ICE AC.. An. IC3.1.1.339

IC4 HU 013

31

017 LOW NOISE RIAA PREAMPLIFIER

This high quality phono preamplifier is based onthe Type HA12017 integrated circuit from Hitachi.The principal technical data of this chip aresummarized in Table 1. The circuit diagram, Fig. 1,shows that output off -set correction is provided byintegrator IC2. The output signal of IC is firstpassed through low-pass filter R7 -C9, then inte-grated in IC2-C8. The error signal is fed to the in-verting input of amplifier ICI via 47 162 resistorR6. The amplitude of this signal is always suchthat the off -set voltage at the output of ICI is vir-tually nought. The off -set correction used hereenables the preamplifier to drive a power amplifierdirect.The correct capacitive termination of the pick-upcartridge can be selected with the aid of Si. The in -

1

R1

0 MIMI10pMKT

20R3 0TTTT

Si/ 11 I 1

CitaLl2L

* R10

* see text

1 F STYROFLEX

-1=1- = 1%, E960

elt100p

put impedance is 50 k52, but can be altered byredimensioning R2 -R3. The output impedance ofthe preamplifier is 510 Q, i.e., low enough for driv-ing a relatively long cable.The RIAA equalization filter in the negative feed-back circuit of ICI is fairly complex, which wasnecessary to meet the required IEC specification(note the use of high stability capacitors andresistors).The regulated power supply for the phono preampli-fier is shown in Fig. 2. This is once again a rela-tively extensive circuit which, in combination withthe low-pass filters on the ± 24 V lines to ICI andIC2, gives excellent suppression of RF signals,hum, rectification noise, and mains borne inter-ference. Presets Pi and P2 serve to adjust the out -

C10 C11

220nMKT

IC I

DRIVER

C4a C5a

'°'-I 2 n713-.1tel0 C4b m1n00

--Alli, IF-C4c 33n0

.-A 1150p1111e R11b R12

EU CEZE1

0 0 0

R4

C6 C7

100p 470R9

R7

C12

R15 I x OmA

1000P25V

VDcx 1 rnV

220nMKT

C13

R16

1000p25V

R13

C9 C151:=1100,.

470n20n 25VMKT KT

C16J C77

220n00p

25V

R14

I= 4mA

24V

24V

0 12V

121-2mA

I .1...2mA

,15 (:)12V

87429 - 1

32

Table 1. HA12017 Low noise preamplifier

Features:IN Low noise: Vn[in] = 0.185 µV typ. (measured in IHF-A network,

Rg = 43 kQ, IEC RIAA). V0=-95 dB relative to Vo=1 Vrms.MI Wide dynamic range: V:=235 mVrms max. (Vco= ± 24 V, f =1

kHz, THD=0.1%, Ao----100A40 dB). Low distortion: THD =0.002% typ. If =20-20 000 Hz, Vo=10

Vrms, RIAA equalization).MI Supply ripple rejection: SVRI +Vco)=56 dB; SVRI-Vcc) = 45 dB

(typical values at f = 100 Hz and Rg = 43 K2). Maximum operating voltage: ± 26 V. Maximum power dissipation: 500 mW at Ta =75°C.

Note: Rg =114. in this design.

2S2

22n630V

R18SIOVS10K250

250mA

D1...D8 = 8x 1N4001R19...R22 = 4x 1n8C19...C22 = 4x 22n

2 x 24V100mA

put voltage on the ± 24 V rails.The printed -circuit boards for the preamplifier andthe power supply are shown in Figs. 3 and 4 respect-ively. The correct values for R 1, C4 and C5 areachieved by parallel and series connection. All fourvoltage regulators can be fitted onto a commonheatsink if electrical insulation is provided. The ac-companying photograph shows a suggested con-struction of the preamplifier.It is strongly recommended to use good quality

LM317

R27

1057812

1C4

LM337-,

R28

C29 C31O =nom imm

1000it 2200p25V 40V

4_0

C30 C32=I I=Nom Mom

1000p 2200p25V 40V

87429-2

24V

12V

OV

12V

24V

components for the volume adjustment and inputsource selection --consult the references given below.

References:

1. Top of the range preamplifier, Elektor Elec-tronics, January 1987.2. Valve preamplifier, Elektor Electronics, March1987.3. Electronics potentiometers, Elektor Electronics,April 1987.

33

Parts list

Resistors (±5%):Ri =240RR2;R3=100KR4=43KR5=51ORRs=47KR7;R8 = 1 MORs=1K6Rio= 165RFRi 1 = 3K1 6F + 4K 64FRi2= 95K3FR13. . .R16 inCi.;R27...R30 incl.=1ORR17=2M2R18= varistor SIOV S10K250 (Siemens)R . . R22 incl. = 1R8R23=22ORR24=3K9R25= 120RR26=2K0Pi;P2=250R presetCapacitors:Ci =10t2; MKTC2= dimension to suit capacitive termination

of cartridge.

R

toU

C4

AF

Et*

C3= 100p; polystyreneC4=2n7F//6n8F//150p0F; styroflexC5=1n0F//33n0F; styroflexCs= 100p styroflexC7 = 470p styroflexCs;Cs=470n MKTCio;C12.;C14;C1s=220n MKTCii;C13;C2s;C3o=100014; 25 V; radialC15;C17=100µ; 25 V; radialCis=22n; 630 VC19 ...C22 incl.=22nC23;C24=47001.4; 40 V; radialC25;C26= 1µ; 40 V; radialC27;C28= 100nC31;C32=2200µ; 40 V; radial

Semiconductors:Di ...Da incl.=1N4001ICi = HA12017 (Hitachi) +IC2=0P-77 (Precision Monolithics Inc.)IC3=LM317IC4=LM337IC5=7812IC6= 7912

E

Cl

CB

ClCl

V

V

V

V

34

Miscellaneous:

Si = 4 -way DIP switch block.S2 = SPST mains switch.Fl = 250 mA delayed action fuse with PCBmount holder.

TRi = 2x24 V;100 mA mains transformer forPCB mounting.

Heat -sink for IC3. . ICs incl.Insulating washers for IC3. ..IC6 incl.PCB Type 87429-1 and 87429-2

Available from ElectroValue Ltd MI 28 StJudes Road Englefield Green Egham Surrey TW20 OHB. Telephone: (0784) 33603IIII Telex: 264475.

9CD

011

04

U.

Ato +

0

ti.

U

+ Available from Cirkit PLC III Park Lane Broxbourne Hertfordshire EN10 7NQ. Tele-phone: (09921 444111 III Telex: 22478.Stock number: 61-170-12017.

General note: many special AF components forthis project are available from Audiokits Pre-cision Components 6 Mill Close Borrowash Derby DE7 3GU. Telephone:103321 674929.

cerocimq1072

35

018 MICROPHONE AMPLIFIER WITH MUTESWITCH

Microphones, unfortunately, produce only a smallsignal and they, therefore, require a special pre-amplifier to boost their output. Because smallsignals are involved, the signal-to-noise ratio of thepre -amplifier is a very important parameter.In this article, we present two circuits for a pre-amplifier suitable for virtually all occasions: a sym-metrical and an asymmetrical version. We have in-corporated a mute switch, which speakers can usewhen they want to clear their throat. As there is annumber of low -noise operational amplifiersavailable nowadays, the cost of these pre -amplifiersis relatively low.The asymmetrical version is shown in figure 1.Switching between high and low impedance match-ing is possible with switch S2. Opamp Al is ar-ranged as an AC amplifier with a gain of around27 dB. This stage may also be used as a DC ampli-fier: R3 and Ci are then omitted, and the value ofR2 is lowered to 22 k. Capacitor C2 limits the band-width of the amplifier to ensure stable operation.

1

HI/LOIMP

*see text

Sl

C1

R1

Av= 27.2 dB

22pR4---

Al... R3 = metal film

15 V

15 V

S1

MUTE

Irrespective of whether Al functions as a DC or anAC amplifier, the DC component in its output isblocked by C3. The amplified AC signal is appliedto muting stage T1. This field-effect transistor (FET)normally conducts and the output of Al is thenfurther amplified in A2 by about 5. Finally, thesignal is taken to the output terminal via high-passfilter R13 -C6. The load must be greater than 10 k52.When mute switch Si is pressed, the FET receivesa negative voltage at its gate and is switched off. Ca-pacitor C5 determines the speed with which mutingoccurs within certain limits. Capacitors CI, C3, andC6 may be electrolytic types: measure the DC levelat both terminals to determine which way theyshould be connected!The symmetrical version of the pre -amplifier is

shown in figure 2. The only difference between thisand that in figure 1 is that the input stage now con -

Figure 1. Circuit of the pre -amplifier with asymmetri-cal input.

RS BF 256C

R11

Av= 13.2 dB

lopR7

* RL> 10kcs

IF":5110

R12 R13

BC 5476

®15V

Al, A2 = IC1 = LM 833; NE 5532; TL 072

BF 256C

GII II IID

85450-1

36

2

IMP

14.11 -01Q-

IC see text

92

15 V

15 V

0

0

R4

C4

R6

Av = 20 dB Av = 13.3 dB

R5

R7

R1 ... R4 = metal film

220n

LO.) an I IC1 IC2

C5 100n

15 V R14

0 SZE

Av = 6,7 dB

Cj1122p

R13

Di C3 R15MIM

AA 119T47°

31

MUTET2

A4

E

BC 547B

015 V

* RL> 10 kc2

1-6-4ips

BF 256C

GiVVD

A1, A2 = IC1 = LM 833; NE 5532; TL 072A3, A4 = IC2 = LM 833; NE 5532; TL 072 85450-2

Figure 2. Circuit of the pre -amplifier with symmetri- sists of Al, Az, and A3 to obtain symmetry. OpampsAl and A2 provide a total gain of about 20 dB.Opamp A3 functions as a differential amplifier to

Figure 3. Printed circuit board for the asymmetrical ensure that common -mode noise and interference iseffectively suppressed.

cal input.

pre -amplifier.

3

37

Naamloos-6.indd 2 28-08-2008 10:11:52Naamloos-6.indd 2 28-08-2008 10:11:52

4

R 4 21

C

<24.1...nivmeriwoorn1

Figure 4. Printed circuit board for the symmetricalpre -amplifier.

Parts list (figure 3)

Resistors:

= 680 52 metal filmR2,R3 = 47 k metal filmR4,137,1312 = 22 kR5 = 1 k

R6 = 4k7= 1 M

Re = 150 kRio,Rii = 120 kR13 = 270 k

Parts list (figure 4)

Resistors:

111,134 = 330 52 metal filmR2,R3 = 22 k metal film115,R7 = 6k8R6 = 1k5R8,1311) = 1k2R9,1111,R12 = 5k6R13 = 12 kR14 = 1MRis = 150 kR16,Riz = 120 kR18 = 10 kR19 = 270 k

Capacitors: Capacitors:Ci = 1 p/16 V MKT (see text) C1 = 22 pC2 = 22 p C2 = 1p5 MKT (see text)C3 = 2p2 MKT (see text) C3 = 47 nC4 = 10 p C4,C5 = 220 nC5 = 47 n C6 = 100 nC6 = 1p5 MKT (see text)C7,C8 = 220 n Semiconductors:

Semiconductors: = AA119 = BF256CT2 = BC547BIC1 = LM883; NE5532; TL072

Miscellaneous:

Si = spring -loaded push to make switchSz = miniature SPST switchPCB 85450-2

38

= AA119 = BF256CT2 = BC5476IC1,1C2 = LM833; NE5532; TL072

Miscellaneous:

= spring -loaded push to make switchS2 = miniature SPST switchPCB 85450-1PCB 85450-2

MKT = metal -plated plastic polythereftalate foil


019 MICROPHONE -SIGNAL PROCESSOR

In broadcasting systems, intercoms, and mobileradio telephones it is necessary to amplify themicrophone signal over a restricted range only. Thismay be achieved with the aid of a compressor or aclipper. The former provides low distortion, but itsdesign is rather complex, whereas a clipper is ofsimple construction, but suffers from appreciableharmonic and intermodulation distortion. Of thesetwo, intermodulation distortion is far and away themost troublesome; in fact, the acceptability of aclipper in an audio signal processor would be fargreater if clipping would not cause such severe inter -modulation distortion.

1

2 tr6390p

I 6WFB

Ctl

Tp 6V

C5

220p

3130

6-.12V < 5mA

0

11*10013

c e *PIM

R7

1M

cl220n

-4C)

C *=NM

70p06479-2.

100k

600mVpp

In the accompanying diagram, intermodulation dis-tortion is reduced by signal -control of the cross -overpoint. The principle of operation is shown in Fig. 1.The amplifier has a very high impedance input(value of R O. When the signal level is so low thatthe diodes do not conduct, the cross -over point isdetermined by R1 -C. As soon as the diodes con-duct, the input impedance of the amplifier is re -

duced, which causes the cross -over point to shift up-wards. The lower amplification of the frequencies isthen smaller, and this enhances intelligibility. Infact, intelligibility of a signal processed in this man-ner is much better than a conventionally clippedsignal.

3 c.L

" 3

The diagram in Fig. 2 shows the detailed realizationof the principle. Transistor Ti is a microphone low -noise microphone preamplifier. The clipping circuitis based on A t: the limiting level is set by Pi. Thevalues of certain components depend on the appli-cation: guide lines are given in the table.

Application C4 C6 C8

Hi-fiCommunicationsor intercom

-100-220 pF

47 nF

0-4.7 nF

470 pF

4.7 nF

For input signals above about 100 mV, the micro-phone preamplifier may be omitted. The inputsignal is then applied to the junction C4-05 via aresistor (R in Fig. 3). The value of R should be suchthat the sum of it and the microphone used is about10 kilo -ohms.

39

020 MINI AMPLIFIER

This little amplifier, operating from 3 . . . 9 V, andproviding 1 W output into a 4 Q loudspeaker, is oneof those circuits of which you never have enough.The amplifier is based on one 8 -pin DIL IC typeLM1895N. Electrolytic capacitors C2 and C6 de -couple the supply lines; C7 prevents d.c. reachingthe loudspeaker; and C3 and C5 provide a low -impedance path to earth for audio frequencies.The input signal is applied to pin 4 of theLM1895N via 131 and C4. Resistor R4 and capacitorCa suppress any tendency to oscillation, i.e., im-prove the stability.The amplification is determined by R., and R3: it isof the order of 50. Capacitor G, in parallel with Ri,ensures that the amplification drops off for fre-quencies above about 20 kHz. If the amplifier is in-tended for use with a small AM receiver, it is

desirable that the amplification starts falling off ata lower frequency. This is brought about by enlarg-ing Cl; for instance, if its value is doubled, the am-plification starts dropping at 20/2 =10 kHz.On the printed circuit board shown in figure 2(which is not available ready made), 131 may be re-placed by a wire link; the volume control is thencarried out by an external logarithmic poten-tiometer connected to the PCB via a short length ofscreened audio cable.Current consumption is 2.5 mA at 3 V or 7.5 mAat 9 V under no -signal conditions, and 80 mA at3 V or 270 mA at 9 V under fully driven con-ditions: in the latter condition, the output power is

C4

R2

+3...9 V

C1470p

4

100n

P1

50k

Elm Cf

5

R3

Si

C2

.-1]220P

2 10V

8

IC1

LM 1895 N

3

C5 C3I=1

1073V 1 17001/5

R4

Ca

C6O14705

10y

C7

0

DI -47010V 0

L S

4

al1W

70n 00 0854/5

100 mW or 1 W respectively into 4 ohms.The output power for different supply voltages andloudspeaker impedances can be estimated by deduc-ting 1 V from the supply voltage, and raising theresult to the power 2. Divide the number obtainedby 8 and then again by the loudspeaker impedance.The sensitivity of the amplifier is about 50 mV. Thiscan be reduced by lowering the value of Rl.National Semiconductor Application.

021 MINI STEREO AMPLIFIER

This mini amplifier is based on the Thomson TypeTEA2025. In this 16 -pin DIL device hides a stereoamplifier that with a supply voltage of 9 V will pro-vide 1 watt output per channel into a 4 -ohm loud-speaker. At full output, the input sensitivity is

about 25 mV77. If this is too sensitive, a resistor Rmay be connected between pin 6 and C7 and be-tween pin 11 and C2. The sensitivity then becomes(25 + 1/2R) mV, provided R >1 k52. Furthermore,the supply voltage may lie between 3 V and 12 V.The operation of the IC cannot be discussed here,but for those interested its internal circuit diagramis reproduced in Fig. 1. One useful feature of the

TEA2025 is that it has a soft -start circuit on board,thus obviating annoying plops in the loudspeaker atswitch -on.Construction of the amplifier is fairly simple, buthas its peculiarities. First, there is the earth, whichin this case should not be of wire, but rather consistof a metal earth plane (if you design your own PCB,this would be of copper). If at all possible, pins 4and 5 as well as pins 12 and 13, should be connectedto a (copper) area of not less than 5 cm'. The twoareas should be connected in a suitable manner, andin such a way that a heat sink is formed under theIC as shown in Fig. 2. This ensures both good heat

40

1

1:3ED

P1 .10k log STEREO

Pla

470nP1b

10

TEA2025

C12=I C7

100171-11P-ioop16V

3V

Fig. 1. Circuit diagram of the mini amplifier.

conduction and a good earth. Moreover, all otherconnections should, of course, be kept as short aspossible. This is particularly important in the caseof the supply lines, which should be decoupled by

as close as possible to the IC. The negative ter-minal of this capacitor should be soldered directonto the earth plane; the positive terminal issoldered in the normal manner to pin 16.Finally, the distortion for a power output of about0.25 W is roughly 0.3 per cent.

100416V

2SC .StartcircuitDC.Decoupling

CM

150n

C11

100p16V

C9

=I=150n

C5

470p16V

C10

6--0470p

16V

86412-1

40

Fig. 2. This construction of earth plane cum heat sinkis both practical and saves space.

022 MOSFET POWER AMPLIFIER

The output power of an operational amplifier isoften increased by a complementary emitterfollower. It can also be done with a MOSFET, butit is not a good idea to connect such a device as acomplementary source follower because the maxi-mum output voltage of the opamp is then reducedappreciably by the gate -source control voltage ofthe MOSFET, which can be a couple of volts.Another approach is to connect two MOSFETs asa complementary drain follower. The (alternating)output current provided by the MOSFETs is limitedby the level of the supply voltages and the satura-tion voltages of T3 and T4 Resistor R8, together

with Rs, provides feedback for both the opamp andthe MOSFETs. The open -loop amplification of theopamp is, therefore, increased by (1 +Rs/Rs). Theclosed -loop amplification of the complete amplifieris (1 +Rs/R2), i.e., 11.

The current source formed by Ti and T2 is requiredfor arranging the quiescent current of T3 and T4 at50 mA. The values of resistors R4 and R5 are suchthat, without the current source, the voltage dropacross the resistors resulting from the direct currentthrough the opamp is not sufficient to switch on T3and T4 With the current source, and depending onthe setting of 13-; the voltages across R4 and R5 rise,

41

which increases the quiescent current through T3and T4 In view of the temperature dependence ofthe quiescent current, T2 must be mounted on thecommon heat sink (c. 5 K/W) of the MOSFETs.The output power is not less than 20 W into 8 Q,at which level the harmonic distortion amounts to

R1

R3

0.075 per cent at 100 Hz and to 0.135 per cent at10 kHz.

Source: Voice coil drives using complementarypower MOSFETSby M Alexander in Motor -Conproceedings, April 1984

R4

'Cl

(120 SL +

120521

(1k2+100 RI

100k

T2

R8

20 V

13 SI/ IG

1

N.. I

IRF 9520 D

150 mA

14 D

R11 /IG

1

1

\./ \

1RF 520 s

020 V

IFR 520I FR 9520

,/ I's.3 D

85500

LS

20 W

023 NOISE GATE

Noise on an audio signal becomes moretroublesome as the signal itself becomes smaller.When a mixer is connected to a number of signalsources, it becomes particularly disturbing whenone or more of these sources produce only noise. Inthese situations, a noise gate is a real help. Such agate continuously monitors the level of the audiosignal and switches it off, after a predeterminedperiod, if the level drops below a preset value.

The circuit consists of two parts: a control sectionand a regulator section. The control section, basedon opamps At to A4 incl., derives a voltage fromthe audio signal that is used to drive the regulator.The regulator is a voltage -controlled amplifier, forwhich one of the two operational transconductanceamplifiers contained in a Type LM13600 orLM13700 is used. For a stereo system, one controlsection and two regulator sections are required. For

42

Drive

CI

D1...D6=1N4148 A1...A4 = IC2 = TL074

a double mono version, two control sections andtwo regulators are needed. One LM13600 orLM13700 will thus suffice for all these re-

quirements.Opamps At and Az form a full -wave rectifying cir-cuit. Opamp A3 compares the peak value of thesignal with the direct voltage set by P2. If the peakvalue is larger, capacitor C7 is charged via Ti: theattack time is set by P3. The time lapse after whichthe audio signal is switched off is determined withP4. The control of the voltage -controlled amplifier(VCA) and the LED indicating whether there is asignal present is effected by A4. Diode D4 ensuresthat the amplification of the VCA is really zerowhen the output of A4 is low (i.e. less than-15 V).The input of the regulator section has an im-pedance of about 10 1(52 and is designed for audiosignals of 1 Vrms. However, even for a 12 dB higherinput signal, the distortion is still not greater than1 per cent. Where higher input voltages are thenorm, the value of Ri should be altered according -

I = 27mA

IC2 .

C9

,5voi2 = -26:7 (i) 7C8....C7 = 100n

M

86419

Olsv

12 :1F_k OUTty

ly. Where lower inputs are the norm, a preamplifiershould be used.It is, therefore, seen that the noise gate shouldpreferably be connected between the preamplifierand power amplifier.The output level is set with R5, while Pi enablesthe circuit to be adjusted for minimum switchingnoises. To this end, the drive input is switched onand off by Si, while the audio input remains open -circuit.It is best to use a 3.5 mm chassis socket with breakcontact for the drive input: the break contact thenreplaces Si. As soon as the jack is inserted into thesocket, the connection between the audio input andthe regulator is broken.This type of drive input affords a number of specialeffects, such as the switching in of, say, an echo unitat the command (sufficiently high signal level) of agiven instrument (e.g. a snare drum). The commandinstrument is plugged into the drive input for thispurpose, while the regulator is connected into theeffects unit.

024 SIMPLE PREAMPLIFIER

This design answers the need for an inexpensive, yetgood quality, preamplifier equipped with a tonecontrol section.

Fig. 1 shows the circuit diagram. The amplificationof the input stage set up around opamp Al is adjust-able between 10 and 20 with preset Pi The 0 dB

43

1

R3

A1,A2=1C1 =TLC272

P3R5 100k R8

level at the input is 50 mV, while the input im-pedance and capacitance are 47 K52 and 47 pF, re-spectively to enable ready connection of most rec-ord players and cassette decks. The tone control sec-tion is a standard Baxandall type with P3 and P4 asthe respective bass and treble controls. The gain vsfrequency curves for various settings of the tonecontrols appear in Fig. 2. Here the 0 dB level cor-responds to I V.The current consumption of this preamplifier ismodest at about 5 mA. When the circuit is cor-rectly balanced, the indicated measuring pointsshould all be very nearly at ground potential. The

1491 0

10

20

IC 1

87425-1

-20,/k BOOST

FULL BOOST -

FULL CUT

CUT

.1'80 200 '560 Yk 2k 5k 20k

f

0 1491

circuit shown here must, of course, be duplicated toobtain a stereo preamplifier.

025 SINGLE -CHIP 40 W AMPLIFIER

To answer the need for a compact amplifier that iscapable of satisfactory operation when driven froma compact disc player, Philips have developed theType TDA1514 AF amplifier chip, which is

remarkable for its excellent specifications, rug-gedness and output power. The device is housed ina 9 -pin SIL POWER enclosure which has a thermalresistance of less than 1.5 K/W, so that the heatsinkrequired must have a thermal resistance of no morethan 3.8 K/W if the chip is operated at its maxi-mum dissipation of 19 W (Ub = ± 27.5 V,Ta = 50 °C).The circuit diagram shows that very few compo-

nents are needed to make this high-performanceamplifier. The power supply to feed the chip mustbe capable of delivering a current of at least 3 A;the quiescent current demand of the amplifier asshown is about 60 mA. The supply voltage shouldnot exceed ± 27.5 V.Although this project is not supported by a ready-made printed circuit board, you should not ex-perience too much difficulty in constructing theamplifier if it is built on a piece of Veroboard. Makesure, however, that the tracks and connections tothe supply and output terminals are as short aspossible, and use double tracks where this is

44

necessary. In this context, it is advisable to fitdecoupling capacitors C3 and C8 as close as possibleto the chip supply pins. Resistors R2 and R3 deter-mine the amplifier's closed loop voltage gain, whichhas a range of 20 to 46 dB.Finally, some measurement data obtained with aprototype of the amplifier:

P. at Dtot = -60 dB;Ub = ± 27.5 V; RL = 8Q: 40 W

S/N at P0=50 mW: 82 dBSupply ripple rejectionat f=100 Hz: 72 dB

Harmonic distortion atP0=32 W: -85 dB

Intermodulation distortionat P0=32 W: -80 dB

3 dB bandwidth at Dtot=-60 dB: 20-25 000 Hz

Slew rate: 15 \T/µs

The gain vs frequency curve and the harmonic dis-tortion table show that this amplifier provides verygood sound reproduction at a considerable outputpower level.

siti,, c1.11.1.AV SCALE I ARMOR

uA

500

I

1.

Po= 10 WrmsTotal harmonic distortion level

Order no. 1 2 3 4 5 6

100 Hz -79 -84 -84 - - -1 kHz -69 -82 -78 -86 -82 -10 kHz -55 -76 -65 x x x

-: below analyser's noise floor (-87 dB)x: analyser unsuitable for measurement.

WED

s.

i=j ran o04

E.

FIEVEile TM% 160

I I s It I 1. I I

026 SMD HEADPHONE AMPLIFIER

Although the use of SMDs (surface mount devices)is not yet widespread among electronics hobbyists,and the availability of these parts is still problematicin certain areas, there appears to be no way of stop-ping the ever increasing miniaturization of chipsand circuits. A good instance of this happening atan accelerating pace is the Type TDA7050 head-phone amplifier, which used to be available in astandard DIL enclosure, but is currently onlymanufactured in SMA technology.

The Type TDA7050 is a complete stereo amplifierwith a gain of 26 dB and an output power of2 x 75 mW. As seen in the circuit diagram, two elec-trolytic capacitors are required to block the offsetvoltage at the amplifier outputs. It is also possibleto set up the amplifier in a bridge configuration toobtain an output power of 150 mW: simply omit thecapacitors, and connect pins 2 and 4 to ground.Pins 1 and 3 are connected to form the amplifier'sinput, while the loudspeaker is connected between

45


R >320

R >320

'ClCr3j=1

TOP6V

0 1V6...6V

OA1, A2 = 1C1 IDA 7050

87408

pins 6 and 7.The current consumption of the chip at maximumoutput power is of the order of 100 to 150 mA,while the quiescent current amounts to a mere5 mA. The amplifiers should be terminated in 3252,a common value for modern headphones. Thesupply voltage is normally 4.5 V, and pins 6 and 7are at half the supply potential during quiescent op-eration.

027 SPEECH PROCESSOR WITH BACKGROUNDSUPPRESSION

A speech processor is commonly used in public-address installations and in utility transmitters. Itaugments the average value of the speech signal, sothat in spite of a high level of background noise or,in the case of a radio transmission, a lot of inter-ference, speech recognition remains possible. Inmany cases it is, however, undesirable that thisbackground noise or interference is enhancedtogether with the wanted signal. A possible remedy,as outlined here, is to provide an adjustablethreshold at which the speech processor becomesactive.With reference to the diagram, the signal from themicrophone is amplified in T (a low -noise ampli-fier) and in AI. Limiting (or clipping) of the signaltakes place in A3.The signal (taken from the output of AI) is alsoamplified in Az. When the output of this opampreaches a certain level, electronic switch ES1 is ac-tuated. Consequently, the monostable formed by

ES2 changes state, and this closes ES3, whereuponES4 is opened, which in its turn increases the am-plification of A3. When ES4 is closed, the amplifi-cation of A3 is determined by the ratio PI:Rs;when the switch is open, by the ratio (P2+ R8):Rs.The mono -time, determined by the time -constantRzo-C19, has been chosen such that speech is notclipped. The low-pass filter between A3 and A4 en-sures that frequencies above 3 kHz are severely at-tenuated. The required output level is set by P3.Calibration is somewhat unorthodox: a signalsource with a continuous output of speech by train-ed speakers is used. The microphone is positioned infront of the loudspeaker at normal speakingdistance and the sound level adjusted to roughly thelevel of the user. Next, connect a pair of head-phones to the output of the processor and makesure that only the output of these phones can beheard. Adjust P4 for maximum resistance, and thenset the clipping level with P2 (which is a matter of

46

R11

10

toa

8645

a a

personal taste). At maximum clipping level, in-telligibility of the speech will remain good in thepresence of interference, but it will have asomewhat harsh, metallic character. Then, adjustP: for maximum resistance, and P4 till allbackground noise disappears. Finally, set the ratio

88494

signal: background noise with Pi; this is best doneby making a recording of the user's speech via themicrophone and the processor. When the processoris active, i.e. clips, D4 lights.Li to L4 incl. are 6 turns 36 SWG CuL through3 mm ferrite beads.

C15

P3

08

C

01

3 P

I

1a

A1.424 IC I=TL082/13,44=1C2=T1_082DI 03 = 1N4148ESI 8844 IC3.406613

+1 V

C13 T1

C1

a a a

a m5a

47


028 STEREO INDICATOR

On most FM tuners, the stereo indicator lightsupon detection of the 19 kHz pilot tone. However,this need not mean that the programme is actuallystereophonic, since the pilot tone is often trans-mitted with mono programmes also. A similar situ-ation exists on stereo amplifiers, where the stereoLED is simply controlled from the mono -stereoswitch.The LED -based stereo indicator described herelights only when a true stereo signal is fed to the in-puts. Differential amplifier Ai raises the differencebetween the L and R input signals. When these areequal, the output of AI remains at the same poten-tial as the output of Az, which forms a virtualground rail at half the supply voltage. When Aldetects a difference between the L and R inputsignals, it supplies a positive or negative voltagewith respect to the virtual ground rail, and so causesC3 to be charged via 1), , or C4 via D2. The resistorsconnected in parallel to these capacitors ensure slow

= IC1 = LM 324D1...D6 .1N4148

R3

C1

282/160

00C2

282/16V

discharging to bridge brief silent periods in the pro-gramme. Comparator A3 -A4 switches on the LEDdriver via OR circuit D3 -D4.When building the circuit into an amplifier, careshould be taken to select the right point from whichthe input signals are obtained. In general, thisshould be before the volume and balance controls,but behind the mono/stereo selector. The signallevel should not be less than 100 mV to compensatefor the drop across Di or D2. Also observe that theimpedance at the selected "tap" location is rela-tively low. Should the stereo light come on when amono programme is being received, the inputsignals are different, and the sensitivity of one ofthe amplifier channels should be altered. If this isimpossible or undesirable, R3 may be replaced by aseries connected preset and a resistor. The sensi-tivity of the stereo indicator is adjustable with PtThe current consumption is less than 7 mA whenthe LED is off, and about 20 mA when it is on.

8...30V

0

T1

BC547

R13

087420

029 STEREO PREAMPLIFIER WITH TONECONTROL

This simple, one -chip, stereo preamplifier is ideal forbuilding into an existing AF power amplifier. It isbased on a recently introduced integrated circuit,the Type TCA5500 or TCA5550 from Motorola.

This double AF amplifier chip with inputs forbalance, volume, and bass and treble controls formsa sound basis for a good quality preamplifier witha minimum of components. The onset points for

48

C15R

)24-L r) )I[1._ 12

4p763V

C3

220n

C5

100p/14-C-11V

C6

220n

15

13 17 11

IC1TCA 5500

(TCA 5550)

18 14 10

2

8

C1 C9infm moo- mum

100n 100n

CIL CIL C141C1114p7 4p777T 40V

C17 C11.=I

7400vp7On

IC27815

the bass and treble controls are defined with C3 andC4 respectively. All (mono) potentiometers are bestfitted direct onto the circuit board to make forsimple mounting into a cabinet, and also to preventhum and noise being picked up in the wiring thatwould otherwise be required.The preamplifier has a current consumption of35 mA, of which 5 mA is drawn by voltage regu-lator IC2. Zenerdiode Di and power resistor Rsshould be added if the positive supply voltageavailable in the power amplifier is more than about30 V.

Specifications of the preamplifier:

Distortion: .0.1% at nominal output level.Channel separation: a 45 dB.Supply voltage: 8.8-18 V.Tone control range: 14 dB.Volume control range: >75 dB.Maximum input voltage: 100 mV.Amplification: 10.Low output impedance.

R5*

, [1:13C18

'A4alic*12.20n1W

Parts list

lop63V

10p63V

* see text

87405

Resistors (± 5%):Ri R4 incl.=100K/35 = see textPi... P4 incl.=100K linear potentiometer

Capacitors:

Cl; Cs; C16 = 100nC2;C8 = 10µ; 63 V; radialC3;C4;Ce;C7;C18=220nC5;C17=100µ; 40 V; radialC1o;C15=4µ7; 63 V; radialCii;Cla =4117; 40 V; radialCi2;C13=47n

Semiconductors:

Di = zenerdiode 27 V; 1 W (see text)IC, =TCA5500 or TCA5550 (Motorola)1C2 = 7815

Miscellaneous:

PCB Type 87405

49

e EPS 87405

030 SUBWOOFER FILTER

The filter described here is intended primarily for human ear cannot sense direction in a standingexperimenting with a (central) subwoofer (see Ac- wave, directional sensitivity is generally poor at lowLive Subwoofer, EE March 1986, p. 28). As the frequencies, so that it would seem superfluous to

86417 .1

A1,A2 = 101A3,A4 =102A5,A6 = IC3A7,A8 =104A9,A10 = IC5

1C1...IC5 =LM833NE5532TL072LF353

50

DS

TR1

50

°'01D2

01...06 = 154001

00,39

75Y.r

3 1

D321-:

25V 25V

1007Ftz

;.7IC6 LM325

cno

VI n Vo

50

;::07 (Y21 (I c'l (1 ,; c'- (I)

C22 C21/Rim De

De

88,

use a stereo set-up below about 200 Hz. Therefore,the low frequencies can be concentrated on onegood bass enclosure, which, of course, keeps thecost of the overall system down. The satelliteloudspeakers (see EE, April 1986, p. 22) will thenhave to cope with the higher frequencies only.The requisite cross -over network described here isbased on 24 dB/octave Bessel filters: the cross -overfrequency lies around 200 Hz. With reference to thecircuit diagram, Ai and A2, buffer the left-handand right-hand signals respectively. The high-passfilters for the two channels are formed by As -A4and A9 -A lo respectively.At the same time, the two channels are combinedin As, and the resulting signal is passed throughlow-pass filter A6 -A7. The amplification of A8 canbe varied with Pi, so that the level of the low-

IC2 C2I IC3 C2IIC4 IC5I

? T ? ? 3T?50mA

C23...C30 = 100n07 , D8 = 164148

86417.1

frequency signal can be matched to that of the high -

frequency signals. Note that the component valuesgiven in parentheses are the calculated values, withperfectionists may try to approach.The power supply is a symmetrical design withshort-circuit protection, which also prevents annoy-ing "plops" at on and off switching.If a different cross -over frequency is required, referto Active Cross -over Network in the September1984 (p. 28) issue of Elektor Electronics.In the design stages, stability problems were en-countered when opamps with JFET inputs (TL074;LF353, for instance) were used, whereas types withbipolar inputs, such as the NE5534 and theLM833, worked perfectly. The reason for the in-stability in the JFET types is not known.

031 TRUE CLASS B AMPLIFIER

The quiescent current in this amplifier is alwaysnought, so there is no need for zero setting or for acircuit to prevent thermal run -away. Complexity isfurther reduced by the use of a single supplyvoltage,Voltage divider RI -R2 -R3 fixes the voltage level atthe base of Ti at just above half that of the supplyvoltage. Since a current source, consisting of T3,R7, Di, and D2, has been included in the collectorcircuit of T2, this stage provides a very high voltageamplification. The return line of the current sourceis connected to the output, so that the voltagenecessary to stabilize the source does not limit thedynamic push-pull characteristic. The currentsource has, therefore, a high -impedance character.The complementary power amplifiers, T4 and Ts,are darlington transistors, which, of course, enablethe collector current in the driver stage to be keptrelatively low.The feedback to the emitter of Ti via Rs and R6

0

C3

10p 25V

R3

2202V

6D6791

D1...D4= 1N4148 86514

12...24V

O

LS

4...160

0

51

determines the overall voltage gain, here 20 dB, andirons out any non-linear components.Class B operation is normally obtained by direct in-terconnection of the bases of the power transistors.In practice, this gave an overall distortion of notmore than 0.16 per cent (at a drive power of 0.25 Wat 1 kHz). The simple addition of diodes Di and D2improved the distortion to not more than 0.1 percent. Note that these diodes do not alter the oper-

ation, because the darlingtons have a relatively highbase -emitter potential.With a supply voltage of 12 V, the amplifier deliverssome 2 W into 4 ohms (input sensitivity 200 mV),or rather more than 1 W into 8 ohms. A highersupply voltage will increase the output power (to amaximum of 10 W into 4 ohms at 24 V), but thepower transistors then need cooling.

032 TUNING AF POWER STAGES

Simple, economically priced audio output stages,such as, for instance, those using the hybrid ICs inthe STK series, may be improved in a simple man-ner as regards distortion, noise, and off -set voltage.To this end, the output amplifier is included in thefeedback loop of an op -amp. Fig. 1 shows the set-upfor inverting output amplifiers, and Fig. 2 that fornon -inverting ones (the normal situation).In the calculations to arrive at the new gain of theoutput amplifier, determined by R, and R2, it is as-sumed that the LF356 provides an undistortedsignal of 5 Vrms; note also that this type of op -ampmust work into a load of not less than 5 kilo -ohmsto prevent distortion.For an output power of 50 W into 4 ohms, the out-put stage must provide a voltage,U = PR =14.2 Vrms. If the amplification of thestage is 3, the op -amp should deliver 4.73 V. For theset-up in Fig. 1, the value of R2 is then R2= 3R1,while for that in Fig. 2, R2 = 2Ri. Note that in bothversions only the value of Ri should be altered. Thetotal amplification may be calculated from the ratioof RA and Rs as follows: A = (RA + Rs)/Rs.Furthermore, because of the load impedance of theop -amp, R, >10 k (Fig. 1); R2>10 k (Fig. 2);

RA >10 52; and Rc >10 Q (Fig. 1 and 2).To compensate for the off -set voltage of the outputamplifier, the input capacitor should be replaced bya wire link. The capacitor in series with R, in Fig.2 should also be short-circuited. The lower fre-quency limit of the complete circuit is then deter-mined by Cs =1/2rrflimRs. The off -set voltage isthen smaller than 3 mV, provided both RA and Rcare equal to, or greater than, 100 k52. Where greateraccuracy is required, 131 can be used to set the off-set to exactly 0 V.To ensure that there is no direct voltage at the newinput of the amplifier, capacitor Cc should have avalue of Cc = l/fiimRc.Since the amplification of the output stage has been

01

reduced to 3, its feedback factor has gone up, andthe distortion has gone down. The ad-ditional feedback of the LF356 reduces the distor-tion even further. An overall reduction in the distor-tion from 1 per cent to 0.1 per cent is fairly typical.The altered feedback unfortunately results in a

52

change in stability. If there is a tendency to oscillate, Cx must be used: their value lies between 100 pFthe first thing to do is to bring the upper frequency and I nF. Our prototype (using STK ICs) workedlimit back to its previous value with the aid of satisfactorily without either Cx or Cy.Cy =1/2nfiimRA. If the tendency persists, capacitors

53

033 AUTOMATIC CAR ALARM

Even the best car alarm is useless if you forget to setit upon leaving your car, whence this circuit.The relay has a make and a break contact: theformer is necessary to delay the switching in of thealarm after you have got out of your car, and thelatter serves to switch on the car alarm proper.Immediately on re-entering your car, you must pressthe hidden switch, Si. This causes silicon -controlledrectifier Thi to conduct so that the relay is ener-gized. At the same time, the green LED lights to in-dicate that the alarm is switched off.

12V

O

32.

R1[1 R3

R2

T2

T1 .. T3 = BC 547Th1 = TIC 106, TAG 103

As soon as the ignition is switched off, T, is off, T2is on, and the buzzer sounds. At the same time,monostable ICi is triggered, which causes T3 toconduct and the red LED to light. The silicon-

controlled rectifier is then off, and D4 is reversebiased, but the relay remains energized via its makecontact for a short time, preset by Pt As soon as thistime has lapsed, the relay returns to its quiescentstate, and the alarm is set via the break contact. Thedelay time can be set to a maximum of about 1 mi-nute.

re lb

1N4148

Ref

L=1-12 V-30052

SI

131710

R99

85512

034 BICYCLE LIGHTS AND ALARM

A bicycle or tricycle should, as everyone knows, befitted with front and rear lights. The noteworthyaspect of the lights circuit described here is that italso provides a visible alarm, which is primarily in-tended for invalid road users. When such handi-capped people are in need of assistance during theday, this is quickly spotted by passers-by. At night,this is, unfortunately, not so, whence the present cir-cuit.

The usual dynamo or battery is replaced by a 6 Vrechargeable lead -acid battery, which ensures thatthe bicycle lights are operational even when thebicycle is not moving. When the rider is in need ofassistance, the alarm can be switched on: in ad-dition to the normal lights, a small display with theword "HELP" will then flash. Such a signal for helpis not easily overlooked!The circuit is based on an astable multivibrator,

54

Fl

which does not operate when alarm switch S2 is

open. Provided Si is closed, the front and rear lightsare on, however.When the alarm switch, S2, is closed, themultivibrator operates, which causes the normallights and the HELP lights to flash alternately.The circuit is powered by a 6 V 1.8 Ah lead -acidbattery which, when properly charged, is sufficientto keep the lights on for about three hours.The circuit can be fitted in a small, preferablywater -proof, case. Lamps Lao . . Las light the letters"HELP" that have been cut out in the lid. The

Lai = front lightLa2,La3 = rear lightsLao ... Lag = emergency lightsS1 = main switchS2 = alarm switch

F2771 22 /600 mA

=1380 C1500

2k5

1W

85471 -lb

+ 6V9650 61/1

0

BC141 should be fitted onto a small heat sink.Because of the need of regularly charging the bat-tery, the case should be fitted to the vehicle in amanner which allows easy removal and attachment.A circuit for a suitable charger is given in figure lb.This provides a constant charging voltage of 6.9 V(preset with Pi), while the charging current is

limited to about 650 mA. This enables the batteryto be fully charged in around 3 hours. The chargingvoltage should be set carefully, otherwise the bat-tery will not be charged correctly.

035 BRAKE LIGHT'S MONITOR

The circuit described below monitors your car'sbrake lights, and indicates by a light-emiting diodewhether they both function correctly. In that sense,it can save you money by preventing your being fin-ed for driving with defective brake lights, and it alsoleads to increasing road safety.The monitor depends inevitably on the voltage dropacross the supply lines to the two lamps. For the cir-cuit to work correctly, that drop needs to be greaterthan 0.6 V. If this is not so, the drop must be in-creased by adding a 5 V diode in series with eachlamp. Transistor Ti and T2 in figure 1 form aSchmitt trigger, which reacts to the voltage dropacross the supply lines to the two brake lights. Thisreaction manifests itself in Di lighting via T3. Ifone of the brake lights is faulty, the switch -on cur-rent drawn by the other lamp will cause Di to lightbriefly when the brake pedal is pressed. If bothbrake lights are defective, Di will not light at all.All three possible states of the brake lights are thusindicated.The hysteresis of the trigger, and, therefore, the sen-

SI

sitivity of the circuit, can be adjusted within narrowlimits with Pi. The preset is best adjusted with onelamp out of action in a manner which makes Dilight briefly as described above.If you find it disturbing that Di lights every timeyou brake, the operation can be reversed by replac-

55

ing the BC557B in the T3 position by a BC547B(n -p -n). The collector of T3 is then connected to thepositive supply line, and the emitter to R6. On theprinted circuit board this means that the flat edgeof T3 must be turned the other way. A second baseconnection has also been provided on the PCB.Note, however, that this configuration no longer

makes it possible to ascertain whether one or bothbrake lights are faulty, i.e., when the LED lights,one or both lamps need replacing.The printed circuit board is not available readymade.In figure 1, Si is the brake pedal switch, and Laiand Lae are the brake lights.

036 CAR BURGLAR ALARM

This versatile and yet easy to build circuit may beused as an effective deterrent against criminals at-tempting to steal what you are bound to consider ahighly valued and indispensable piece of property:your car.Extremely simple to control, the circuit leaves thecar owner 15 seconds to get out of the vehicle afterhe has set the alarm. Upon return, he deactivates itagain by pressing a hidden reset switch within 7seconds after having opened the car door(s).Criminals who (hopefully) have not been able tolocate the reset switch within the 7 second delaywill regale themselves and their accomplices, if any,with a 100 seconds long, intermittend car horn con-cert which, ideally, should stop them from pursuingtheir nefarious activities and, in short, scare themoff. Also, the lawful owner of the vehicle is alertedby the horn sound that something is amiss, requir-ing appropriate action.

SET ALARM

mom100 I+

0 V

100 n

D9

1 1141141 S2

RESET ALARMN1...N4=IC1=4093N4...N8=IC2=4093

131...D11. 1N4148

The present circuit offers the possibility to connectseveral types of alarm activating devices, such as avibration and/or ultrasonic detector, a windowbreakage sensor, etc., provided these supply an ac-tive low output level when an alarm condition ex-ists. However, it is also possible to use the courtesylight switches for this purpose, since these usuallyconnect to the car body when a door is opened. Tounderstand the operation of the alarm, refer to thecircuit diagram and assume that the circuit is in thenon -activated mode. On leaving the car, the userpresses the 'set alarm' button, which leaves himsome 15 seconds to actually get out and lock thedoor(s); the 15 second interval is determined by net-work R2-Ci; the N2 -N3 bistable will toggle afterthis delay and activate the alarm proper (watchfunction). Note that this condition may be signalledby a suitable LED driver circuit instead of RE2 asshown in the circuit Only when on of the alarm in -

78L08

C. 0 100925 V

.5mA xis 011 500mA

BC 5476

BD 139

86407.1

4.129

56

puts goes low (i.e. active) will monostable N6 -N7toggle and start a 100 second interval, as deter-mined by network Rs -Ca. However, the horn willnot sound immediately, since network R9-05 pro-vides a 7 second delay to reset (de-activate) the alarm before oscillator gate Ns inter-mittently switches the horn relay transistors TiandT2. Note that the horn will stop sounding after 100seconds, but the alarm will remain in its activatedstate, i.e. any alarm condition signalled by thesensor devices or the door contacts will set it offanew and cause another round of horn sounding.As already stated, T3, Ta, and Ri3 may be connec-ted to the N2 -N3 bistable to provide a LED indi-cation of the activated state of the alarm. Instead ofthe LED, a relay may be connected to break the ig-nition coil primary connection. It should be noted,

however, that this relay can not be used in cars withelectronic ignition; in this case, another means fordisabling the car ignition system should be arrangedwith the alarm in its activated state. The relaysemployed in this circuit are standard types asavailable from motorists' shops. The contacts ofREi are simply connected in parallel with the exist-ing horn relay contacts.Finally, note that it is of utmost importance tomount the entire circuit and the relay wiring in anout of the way position; the reset switch may be acoded or key operated type and must be fitted wellhidden. Current consumption of the circuit in thenon -alarm condition is so low as to hardly load thecar battery. A voltage regulator section has beenadded to prevent the alarm from being triggered inerror when the car is started.

037 CAR FUSE MONITOR

This extremely simple to construct contrivance of-fers motorists a visible indication as to the nature ofmalfunctions occurring in the car electric system,which, as we all know or come to find out sooneror later, is protected by means of fuses which havea tendency to melting at times and places most in-convenient to driver and his passengers, if any.This circuit, if constructed with a little mechanicalskill, may be plugged across all fuses in the fusecompartment to quickly locate the defective onewithout having to remove the whole set for visualinspection.Given the very low cost of the undertaking, it maybe worthwile to fit all fuseholders with indicators ofthe type described; in case a malfunction occurs,

86501 1 51f

you are immediately notified which fuse had best bereplaced (after the necessary repairs have beenmade, of course).

038 CAR LIGHTS MONITOR

Many traffic accidents are caused by failing carlights. Often, the driver is not aware of such a mal-function, because the warning lights provided onthe dashboard do not, strictly speaking, monitor therelevant lights, but rather the switch position sincethey are almost invariably connected in parallelwith the relevant car lights.The proposed circuit is intended to indicate thefailure of one light in a pair: sidelights; headlights(up to 55 W); rear lights; brake lights; or fog lights.The two lamps must have the same rating.

Counter -wound coils, Li and L2, carry the samecurrent when both lamps are working correctly, sothat the magnetic fields created by these currentscancel one another. When one of the lamps fails,the magnetic field caused by the current throughthe other induces a voltage in L3. This pulse causesthe TIC106D to switch on, and this in turn makesD5 light. If both lamps fail simultaneously (the pro-bability of which is, however, minute), the circuitdoes, of course, not function.Because in practice the two lamps do not come on

57

1.1= 'I Burns

Lal Lag= 5....55W

L2= Mums L3 = 20turns

or go out simultaneously, R1 -C2 -R2 provide a delayto enable the magnetic field to stabilize. Note, how-ever, that Cl must be matched to the particularlamps being monitored: increasing its value makesthe circuit less sensitive (longer delay).The coil is easily made from an old (or new) core ofa choke or dimmer switch. First, wind two times 11turns SWG22 enamelled copper wire around thecore as shown in the drawing. Inductor L3 consistsof twenty turns SWG40 enamelled copper wire (thiscoil does not carry a large current). Note that theblack spots in the drawing are the same as those in

`see IBYt

the circuit diagram. If the circuit does not work, italmost certainly means that the connections ofeither Ll or L2 have to be interchanged.To monitor all the lights of car, the circuit will haveto be built as many times as there are pairs of lamps.The indicator diodes are best fitted in thedashboard. It is, however, possible to use only oneLED for a number of circuits: when this lights, it isthen, of course, necessary to walk around the car tosee which lamp has failed. Once the LED lights, itremains on until either the thyristor or the ignitionhas been switched off.

039 CAR RADIO ALARM I

It is an unfortunate as well as a generallyacknowledged fact that the car radio (plus cassetterecorder) ranges among the most desirable andoften surprisingly easy to steal objects on many aburglar's "shopping list".This circuit may help to prematurely end thecriminal practice by sounding the horn if it is at-tempted to remove the radio set; cutting or unplugg-ing an additional ground wire, which has been hid-den in the cable for connection to the battery and

loudspeaker(s), causes the alarm to be set off, sincethe connection to the car chassis (ground) is inter-rupted.The circuit for the car radio alarm is composed ofa single timer, the well-known Type 555, surround-ed by a few additional odds and ends to make anastable multivibrator, whose on -time is determinedwith C!. Horn relay Re should have a coil resist-ance to enable the timer chip to energize it direct bymeans of the voltage at output pin 3.

58

12V

V

1,10rnA111

U

3

164148 a3

2

Re

L__17:1 1164001

62 R1R1 Ta

IS

'Cl555

RST

164001

C

It is seen that the multivibrator is in the reset stateas long as point M is connected to earth, i.e. whenthe set is in the place where it should be. Removingthe car radio inevitably causes the voltage at M torise to nearly 12 V, ending the reset state of ICi,which responds with activating Re, i.e. the car horn,since this is energized via the relay contacts in paral-lel with the horn switch in the steering wheel.Note that Re is any small changeover relay havinga 12 V coil, provided the 555 is capable of handlingthe coil current; many motorists' and car repairshops can, no doubt, supply you with a suitablerelay for the alarm circuit.

6

OUT

TRG

IRS

04AL

164001

C2mIm

C=31

7057500e

66406

The sense wire to point M should be hidden in themulti -wire cable to the radio set, while the circuititself must be fitted in an out of the way position,somewhere behind the dashboard.In order that not even an attempt is made to breakinto your car, it is, as will be readily understood,prudent to stick adhesives to the car side windows,warning of the presence of the radio alarm.

59

040 CAR RADIO ALARM II

The purpose of this one -chip circuit is to give anaudible alarm in case a thief attempts to steal thecar radio, which is generally considered an item ofprime importance to the motorist's well-being dur-ing any trip with his vehicle.Since removing the car radio necessarily involvescutting or unplugging the supply cables, the presentcircuit detects disconnection of an extra earth lead,which has been fitted to the rear side of the carradio (metal) housing. In the circuit diagram, thispoint is marked as M. If M is at earth potential, T;is off (high collector voltage); if the earth connec-tion is cut or unplugged, the voltage at M rises toa positive level, Ti conducts, and a negative -goingpulse triggers timer ICI, which has been arrangedto provide a 30 -second timing interval as defined

with R6 -C3. The second timer contained in IC;functions as a 0.5 Hz (R7 -R8 -C4) oscillator sectionwith an output duty factor of 50% (D3). Note thatthe Type 556 dual timer chip directly energizes a12 V, low -power relay, whose contacts are connec-ted in parallel with the horn switch in the car'ssteering wheel.If it is attempted to steal the car radio, the alarm in-termittently sounds the horn for 30 seconds. It is,of course, imperative that constructors of this carradio alarm locate the additional earth connectionon the radio set in such a way as to necessitatedisconnection at an early stage of attempted theft,otherwise the alarm would come on too late, en-abling the thief to get off at his leisure.

I=.30mA (stand-by)

041 COURTESY LIGHT DELAY

Ever been groping about for the safety belt, ignitionslot, choke control or a map while in utter darknessand happy to have closed the car door(s) because ofthe cold, or foul weather? Wouldn't it be convenientto have the courtesy light on for a few more instantsin order to get the vehicle started and ready to moveoff?Figure 1 shows a courtesy light delay circuit foreasy incorporation in almost any type of car. Thecourtesy light is switched by power MOSFET T2,which is a Type BUZ72A ensuring a low voltagedrop (0.2 V typ.) across drain and source andtherefore the lowest possible power loss. The doorcontact, connected to terminals B and C, is nor-mally a push to break type. Ti is therefore off and

C-; discharged when the door is closed; MOSFETT2 does not conduct, sothat the courtesy light re-mains quenched. Opening the door, however, causesTi to charge Cl, and the courtesy bulb willtherefore light in a gradual manner. Althoughclosing the door turns Ti off again, C-; continues tosupply gate drive to T2 for a few more seconds; thecourtesy light will be dimmed slowly. The suggestedMOSFET type should not switch more than about10 W, which is the usual power rating for thecourtesy light.Figure 2 shows how the circuit may be modified toenable the courtesy light to go out immediatelyafter the ignition key is turned. The terminalnumbers refer to the wiring code convention as rel-

60

2 30

31b

15(50)

31

evant to most types of European car:15 = + Vbatt - ignition on.30 = + Vbatt - unswitched.31 = ground.31b = door contact (connects to ground).50 = + Vbatt - starter motor on.

Figure 3 clearly shows the circuit connections in ac-cordance with the foregoing convention.In case the suggested MOSFET Type BUZ72A(Siemens) is a difficult to obtain item, any equiva-lent n -channel power MOSFET to the followingspecifications will also do adequately: Vds -100 V;Id >_9 A; Pd L 40 W; Rds(on) 5 0.25 Q.

Source: Siemens Components XX (1985) No. 6.

86474.3

86474-2

042 FLASHING REAR LIGHT

This rear light for bicycles is fed from a batterycharged with current from the dynamo, and startsto flash when the cyclist halts. To preserve batterypower, the unit automatically switches off 4minutes after halting.The circuit is essentially composed of a batterycharger and a logic switching section. The NiCdbattery is charged from a voltage doubler C1 -C2 -Di -D2 -C3 to ensure a charge current of about20 mA when riding at a reasonable speed. Thismakes it possible for a charge of 3 mAh to beavailable after a 10 minute ride, i.e., enough for the

light to flash for about 4 minutes after the bicycleis halted. A relay is used to switch between oper-ation while riding and while standing still. Whenthe bicycle is in motion, the voltage from thedynamo, G, ensures that N4 is enabled, so that Tiactuates Re, and the small 6 V bulb is illuminated.Since C3 is only slowly discharged via R1, N4 re-mains enabled for about 4 minutes after halting.Push-button Si enables immediately switching offthe rear light, because R2 then discharges C3 in afew seconds.Gate Ni monitors the dynamo voltage, which is rec-

61

N3

D1...05 = 1N4148Ni...N4 = 1C1 = 4093

tified by D4 -C4 -R3. When the direct voltage dropsbelow approximately 2 V, N, switches onmultivibrator N2 -N3 -N4 which causes the relay totoggle at a rate determined by R4-05. The 5 V DILrelay requires only 11 mA, while the current con-sumption of the 4093 is virtually negligible at aboutI µA.

N2N1 D4

87446

It should be possible to fit the circuit and the bat-tery in a somewhat larger than normal bicycleheadlight, equipped with terminals for connectingthe dynamo and the rear light. Of course, due caremust be taken to avoid the battery contactstouching the metal inside of the light.

043 GARAGE STOP LIGHT

A novel use of solar cells makes positioning yourcar in the garage rather easier than old tyres, a mir-ror, or a chalk mark.The six solar cells in figure 1 serve as power supplyand as proximity sensor. They are commerciallyavailable at relative low cost. The voltage developed

across potentiometer Pl is mainly dependent on theintensity of the light falling onto the cells. The cir-cuit is only actuated when the main beam of one ofthe car's headlights shines direct onto the cells froma distance of about 200 mm (8 inches). The distancecan be varied somewhat with Pi

62

Under those conditions, the voltage developedacross Ci is about 3 V, which is sufficient to triggerrelaxation oscillator N,. The BC547B is thenswitched on via buffer N2 so that D3 begins toflash. Diodes Di and D2 provide an additional in-crease in the threshold of the circuit. The totalvoltage drop of 1.2 V across them ensures that thepotential at pin 1 of the 4093 is always 1.2 V belowthe voltage developed by the solar cells. As the triplevel of Ni lies at about 50 per cent of the supplyvoltage, the oscillator will only start when thesupply voltage is higher than 2.4 V.The circuit, including the solar cells, is best con-

structed on a small veroboard as shown in figure 3,and then fitted in a translucent or transparent man-made fibre case. The case is fitted onto the garagewall in a position where one of the car's headlightsshines direct onto it. The LED is fitted onto thesame wall, but a little higher so that it is in easyview of the driver of the car. When you drive intothe garage, you must, of course, remember to switchon the main beam of your headlights!A descriptive article on the operation and use ofsolar cells appeared in the June 1985 issue of Elec-tronics: solar battery - p. 6-65.

044 HALOGEN LAMP PROTECTOR

Halogen lamps are, unfortunately, rather prone toburn out when they are switched on, and this ismainly owing to the high current consumption ofthese devices during the initial stage of heating upto the normal operating temperature of the filamentin haloid gas.A typical value for the cold resistance of a 6 V -4 W halogen lamp is about 0.3 ohm, demanding aturn -on current of 20 A. In view of the relativelylow internal resistance of car and motor -cycle bat-teries, such a current surge is not at all to be dismiss-ed as purely theoretical, and it is easily seen that theensuing rapid heating inside the lamp is a primecause for the thin filament to melt at the suddentemperature effect. What is required, therefore, is aseries regulator system to limit the current duringthe heat -up phase; in other words, a soft turn -on fa-cility.The circuit diagram shows that Ci is charged to thebattery voltage by means of It and R2, causingFET T, to become slowly conductive after Si hasbeen closed. The Type BUZ1O(A) power FET isused in view of its low drain -source resistance in thefully conductive state; a typical value for Rdsoio is0.19 ohm, which ensures a low voltage drop acrossthe FET, and, therefore, a sufficiently highoperating voltage for the halogen lamp. Parts Di

BUZ10(A)

GDS

86468-1

and R3 discharge Ci after opening Si, so that thepower -on delay functions correctly any time thelamp is turned on.Lamp voltages other then 6 V require R2 to bemodified according to R2 = 200,000/(Vbatt - 2) [52].In case the BUZIO(A) proves hard to obtain, othertypes of n -channel power MOSFET may be used inthe circuit. The minimum requirements are: drain -source voltage Inds = 50 V, drain current Id =19 A, and drain -source on resistance Rdsoi) =0.2 Q.

045 LED REVOLUTION COUNTER

A close look at the dashboards of a number of cars counter: first, the still most commonly found needlemay reveal the use of three basic types of rev and round scale, analogue combination; second, a

63

R2

CIrats

170n

D61

`see text

D1 -1:160 = LED.

1000

8C54713

2000

481RCV CcC

OUT

ICI 555

CV TRSGND

116

IonIloop160

R11

P1

100k

4

3000

U1096 B

set of digital displays (often LCDs); and third, apseudo -analogue meter in the form of multi-coloured LED bar, looking much the same as aLED -based VU meter on modern recording equip-ment.The circuit presented here belongs to the thirdcategory. However, contrary to the straight LEDbar indication, this design features a round scalewith a coloured LED needle imitation, just as thegood old mechanical rev counter.The circuit is based on the Telefunken Type U1096Banalogue input LED driver which can light one of30 LEDs on the rpm scale, whose lower and upperindication limits may be set to individual re-

quirements; e.g. the 30 LEDs may merely indicatea limited rpm range to attain a higher resolution.The circuit diagram shows ICI to receive the con-tact breaker pulses and to reshape them for conver-sion to an analogue voltage in an R -C filter, whichpasses the signal to the input of the LED driver.The detailed operation of the circuit is as follows.Zener diode Doi and parallel capacitor CIsafeguard the base of inverter transistor Ti against

6000

lC2L4810

R13

062

C9

1 2750Vp0

12V

¢O

1.30mA

1N4001

green

red

yellow

D59

g 060

5000 6000 7000 864611

64

R9

+12

aCl

R

E

B R5

.+ C9 .00000--

--"46.0*.

receiving high voltage pulses induced in the ignitioncoil secondary winding. The Type NE555 timer hasbeen configured to function as a monostable withan output pulse period time of 3 ms, during whichtime R3 causes Ti to conduct so as to prevent er-roneous triggering of the monostable. The analoguevoltage, proportional to the engine rpm rate, is es -

Parts list

Resistors:

131...R4;R6:Rii =47 kRs;Rio= 100 kR7= 270 kRs= 1 k

Rs= 10 kR12= 220 kR13=4.7 QPi;P2= 100 k preset

Capacitors:

Ci;C4=10 nC2= 4n7Ca= 100 nCs= 10014;16 VCs=10 11;16 VC7= 111;16 VCs=4711;16 VC9=47014;25 V

Semiconductors:IC1=555IC2 = L4810IC3=U1096B (Telefunken)Di ...Dso=LED

(see text)Dsi = zenerdiode 8V2; 400 mWD62 = 1N4001Ti =BC547B

Miscellaneous:

PCB Type 86461

tablished by means of smoothing network R8-05,R9 -C6 and Rio -C7. The indication range for theLEDs may be set with Pi and P2, the presets forthe lower and upper limit, corresponding to LEDsDI -D2 and D59 -D60 respectively. Note the relativesimplicity of the LED array connection to IC3;only nine IC output lines suffice to drive any one

65


of 30 pairs of LEDs, whose colour may be chosento individual taste, while it is also possible to useseries -connected LEDs to achieve a brightly as wellas functionally lit rpm scale.The circuit diagram shows two rows of LEDs; theupper one is the normal rpm indication scale, forwhich the following coloured subdivision may beused: 0 to 5000 rmp are green LEDs; from 5000 to6000 rpm yellow or orange types; 6000 rpm and upare bright red types. This range and subdivision may,of course, be adapted for the specific type of engine.

The lower row of LEDs may be used to indicate anumber of fixed rpm rates on the scale, for instanceat 1000 rpm intervals.The PCB track layout and component overlay withthis design should enable anyone to readily con-struct the LED scale revolution counter, but notethat the LEDs are mounted at the PCB track sideto get the correct indication in clockwise directionwith increasing the rpm rate. Also note the use ofthe low voltage -drop regulator IC2 which suppliesICI and IC3 with a stable, noisefree 10 V rail.

046 MOTOR -CYCLE GEAR INDICATOR

This circuit provides motor -cycle riders with a gearindication to the foot -operated lever at one side ofthe engine block. The proposed indication unit willbe appreciated by those riders in the habit of forget-ting which gear they have selected when attemptingto drive off at traffic lights or crossroads and findingthat the engine stalls because it had been switchedto second gear.The circuit as shown is based on the use of twogear-lever operated, plunger or roller type

0

S2

MIC7

7808

IC8

78L08

C5-Ci1C1 - IC4 IC5, IC6

720pt16

7100n(2.

microswitches, along with the neutral gear indi-cation lamp, which is a standard item on most typesof modern motor -cycle.Bistables NI -N2 and N3 -N3 serve as debouncer cir-cuits for micro -switches Si (lever down) and S2(lever up). If either one switch is actuated, Nia orNi5 will cause bistable N12 -N13 to be set or reset;counter IC5 counts up (U/D = 1) or down (U/D =0) as a result of actuating S2 or Si respectively. Onrelease of the relevant microswitch, AND simulator

C2

T2p74,0 J.

V.[.(44 2kN1

N2

N1...N4 = IC1 = 4093N5...N8 = IC2 = 4081N10...N13 = 1c3 = 4011N14...N16 = 1C4 = 4075

8V

O

R8-014

5

E3Ill

131 12T1114 d e f

IC6 4511

A

74565 n

miE8 10 11

2

N7 1 1 N8

81 9 121 13

B

C

11

6

2

00 01 02 00

IC5 4029

0/.13 C R P1 P2 P3 gE

R6

4C3

EOM

49716V

86464.1

66

Di -D2 -R6 supplies ICs with a clock pulse, in-crementing or decrementing the gear readout com-posed of IC6 and the indication -panel mounted7 -segment LED display.Input pin 5 of gate N6 may be wired to point A, B,or C to suit 4-, 6-, or 5 -gear types of motor -cycle re-spectively. N6 inhibits OR gate Nis from supplyingfurther clock pulses if S2 is operated when drivingin top gear. N16 and Nit have the same function forthe bottom gear, preventing the counter fromdecrementing the display reading at gearing upfrom neutral to 1.If the neutral switch -S - is closed, 1C8 suppliesthe A and B inputs of 106 with logic low levels; the

level at the C input need not be forced low, since theneutral gear is in between first and second, both ofwhich positions cause the most significant bit -C- to be low anyhow.Parts Rs-C2-Nio-Ns have been included to preventan erroneous display reading at gearing down from2 to neutral and up again; for two seconds, Nis isdisabled from clocking 105, so that the lever -uppulse is not detected.At power -on, R7 and C3 preset counter 105 to state1.

In conclusion, it goes without saying that Si andSz should be good quality microswitches, sealedagainst moisture and dirt.

67

047 4 -WAY DAC EXTENSION

This extension circuit makes it possible to use asingle DAC (digital -analogue converter) forgenerating four analogue voltages. Evidently, thecost of the extension described here is only a frac-tion of that of four DAC chips.The operation of the 4 -way DAC is fairly simple.Assuming that inputs A, B and E of multi-plexer/demultiplexer IC, are driven low, the outputof Al is fed to the + input of A2, while the outputof this opamp is connected to the - input of Ai viathe demultiplexer and 121. Capacitor C2 functionsas a storage device. The output voltage available at

UDAC

° Al5

0

6O

0

G4

MUXDMUX

P125k

Cl

100p

R1

R2

0...3

IC 14052

0

1

2

12

terminal 1 equals UDAC because Ai is dimensionedfor unity gain. When the E input is driven high, orwhen a new code is applied to inputs A -B, the inputvoltage for A2 is derived from C2 , so that the programmedvoltage remains available at the output. The function ofthe other output buffers and capacitors is, of course,similar to that of A2 -C2 .

For optimum performance, C2-05 should be low leakagecapacitors, e.g. multilayer MKT, and the input current toA2 -A5 should remain low. The latter condition is satisfiedby using opamps with FET inputs (typical bias current:

15

11

0...3

an

CL*

A2

MOn

A3

160-®

C5 *

A4

60-()

zOnA5

(50ir-C)

2

5

2

1C2 IC3 IC4

* see text

Al = IC2 = TLC 271A2, A3 = IC 3 = TLC 272A4, A5 = IC 4 = TLC 272

C7

70n

IC1

60

5V

0

68

1 pA). Only Al requires an offset compensation since feed-back is provided via the lower multiplexer in IC,. The E(enable) input serves to disable IC, during switchover toanother channel. R2 then gives A, unity gain to preventthe - input being left open.When a Type HCT4052 is used in the IC, position, stan-dard TTL levels can be used to drive inputs A, B and E.

A "normal" CMOS 4052 requires 5K6 pull-up resistors tobe fitted on these inputs, but only if TTL signals are usedto drive the extension. The current consumption of thecircuit is less than 10 mA. UDAC should be between- 3.5 V and + 3.5 V.

048 8 -BIT ADC

Before any analogue voltage can be measured andsubsequently processed by a computer, a converterdevice with the necessary precision is required toprovide the computer with the digital n -bit equiva-lent of the voltage as applied to the DAC circuit.Obviously, the higher n, the more steps involved inthe conversion process, but also the higher the accu-racy that can be obtained.This 8 -bit ADC circuit works with very few parts;

SOC

Rd

2107V

L3

yet it is versatile, fast, and sufficiently accurate formost purposes. The maximum input voltage to thecircuit is arranged at 5V, as determined by the re-sistor network connected to the Ain terminal of theType ZN427 ADC chip. Given this upper limit forVin, the conversion accuracy equals 5V/(28-1) = 19.6mV/step. Other input voltage levels may beaccommodated by appropriate redimensioning ofthe input voltage divider.

5V (30 mA)

0

4 10

WR

Ri

Ro

R1C1

ZN 427

Ain

busyGND

DO

D2

D3

D4

Da

D

CLK

11

12

13

14

15

16

CPUdata bus

17

18

5 9

N1

R4 R5

R63300

N2

* busy

1n900 kHz

C2 N1...N2 = 1/2 4093=1C2Nom

2n2

86404O

69

Since the proposed ADC chip features an analogue -to -digital conversion time of only 10µs (typicalvalue), alternating voltages may be measured(digitalized) and processed under machine languagecontrol; just as with the above DAC circuit, BASICis usually not very suitable for this purpose, and itsuse is restricted to applications where timing re-quirements are less stringent. It will be understoodthat fast and therefore smooth computer responseto, say, joystick movement is only feasible if theADC reading subroutine is written in machinecode.

A low SOC (start of conversion) pulse at the WR in-put of the chip triggers the internal voltage conver-sion process and the BUSY output is activated (i.e.pulled low); this, in turn, enables Schmitt triggergate Ni to generate the ADC clock frequency ofabout 900kHz. On completion of the clock -

controlled conversion, BUSY goes high, and theCPU may read the 8 -bit value contained in theADC latch by activating the read line. Note that theSOC and read signals must be decoded with suitablecircuitry as required by the type of computer orCPU. Provision has been made in the ADC circuitto select either the BUSY or BUSY signal in orderto flag the conversion condition to the host com-puter CPU.Calibration of the present circuit is straightforward,since this merely involves setting two presets. First,a simple test loop may be written in machinelanguage; next, adjust Pi (offset) for a computerreading of 0 with no input voltage applied to the cir,cuit; P2 is set to give a reading of 255 (FFhex) withthe maximum input voltage at Vin, i.e. 5V. Finally,test the ADC linearity by applying 2.5V from a suf-ficiently accurate source; the computer should read

,13ex,128 (80 hex).

049 8 -BIT DAC

This simple circuit enables computer users togenerate analogue voltages under software control,which, no doubt, offers interesting possibilities for

R4

intelligent control of, for example, volume adjust-ment of audio equipment, light dimmer circuits, etc.It is also possible to write machine language

R11

N-le -IC

MIL

NN IN

N

3

14

2

13

12

11

10

9

*see text

DO

D2

D3ZN426

D4

D5

D6

D7

minr

O

R

RO

C4

20mV.5V®

C2

p

100 p

3130

10k

C3

56p

O70 86403

® -5V

algorithms for the generation of several different,complex periodic output voltages, in short, to con-struct a computer -controlled function generatorusing a minimum amount of hardware.The circuit is based on the Type ZN426 digital -to -analogue converter (DAC), which is an 8 -bit resol-ution (255 -step), high conversion speed (1 pts) devicefor direct microprocessor interfacing. The circuitmay be connected to an 8 -bit output port whichprovides TTL or CMOS compatible digital levels;most computers currently on the market have sucha port, or the manufacturer has made provision toadd one or more of these in the form of an expan-sion. The conversion time of the DAC chip allowsthe use of machine code for high frequency outputvoltages; BASIC is usually too slow for this purpose.The DAC output voltage is buffered with an BIFETopamp, which can be adjusted for a step response of

15mV/step, which means that the maximum outputvoltage of the present circuit is 3.825V, since 8 bitsrepresent 255 steps (28-1).Adjustment of the circuit is straightforward: con-nect a DVM to the output and adjust PI for an in-dication of 0.00V with nought (0) written to theDAC; next, write 255 (FFhex) and adjust P2 for themaximum voltage indication of 3.825V.The circuit is also very suitable as an D -to -A con-verter driven by 8 -bit I/O port (EE, December 1985)as part of the universal I/O bus. It should be noted,however, that writing FFhex to this port gives ananalogue output voltage of OV, since the ULN2003buffer IC in the 8 -bit output port is an invertingdevice: moreover, the eight data lines to the DACchip should be fitted with pull-up resistors as shownin the circuit diagram.

050 16 -KEY INPUT FOR MSX MICROS

This simple circuit is an unusual, but interesting,application of the joystick port available on anMSX microcomputer. With some modifications, it

1

1.131 I fro 1

V E.,

should also work with other types of computerequipped with a similar "game" input. The use ofthe joystick port for reading 16 switches is advan-

Vi

O 71

tageous because very little additional hardware is re-quired, and programmers can avail themselves ofstandard BASIC instructions relating to thejoystick.On MSX computers, the position of the joystickhandle is read with the aid of instruction ST I C K (n),where n is 1 or 2, i.e., the number of the relevantjoystick. The instruction returns an integer between1 and 8, from which the handle position is deducedas shown in Fig. 1. Instruction ST RIGIn) enablesdetermining the state of the trigger (fire) button onjoystick n, and returns -1 when this is actuated.A diode matrix is used here to enable connectingeight pushbuttons Si -S8 to the four direction inputson the joystick port. When actuated, either one ofthese buttons forces a logic low level upon one ortwo of the input lines, enabling the computer toidentify the key number. Eight additional diodes,D21 -D28, make it possible to double the number ofkeys (S9 -S16). These can be kept distinct from theformer 8 by connecting them to the trig. A input.The 16 keys are identified in BASIC with the aid ofinstructionsX = STICK (1) (or X = STICK (2)) andY = STRIG (1) (or Y = STRIG (2))so that the key number is simplyZ = X-(Y*8) + 1.This goes to show how a versatile extension canmake good use of existing hardware whilst beingcontrollable with BASIC commands. Finally, Fig. 3shows the pin assignment on the 9 -way sub D con-nector used for connecting the present circuit to theMSX joystick port.

2

3

4

051 32 KBYTE PSEUDO -ROM

This versatile, exchangeable, memory moduleshould appeal to programmers developing softwarefor computers other than the one being used forwriting, testing and debugging the program. Thebattery back-up function of the module ensuresthat data is retained, and so makes it possible to use"portable", software that is ROM -based and yet canbe altered readily without having to program anderase an EPROM a number of times.The memory module is based on the use of a Type43256 32 Kbyte static CMOS RAM from NEC-see Fig. 1. Other 32 K types, such as the 62256,should also work here. A battery (2 button cells, ora 2.4 V NiCd cell when Di is bypassed with a re-sistor to enable charging) enables the chip to retainits contents when the computer is off. When the

+ 5 V supply from the computer is on, Ti drivespin 1 of Ns high, so that this gate can enable theRAM via the CE input. The supply set up aroundT3 -T2 then feeds all the chips on the board withabout 4.8 V. The drop across the C -E junction of T2is less than 0.2 V here since the transistor is driveninto saturation. When the computer is switched off,the circuit is fed from the battery via germanium di-ode Di. Voltage divider R6 -R6 causes Ti to beturned off when the supply level drops below some4.5 V.Input I of N5 is grounded via R7, so that CE on theRAM is held high, causing the chip to switch to thepower -down (standby) mode. A prototype of theplug-in RAM consumed only 1.5 µA in the dataretention mode, after briefly taking about 3 mA

72

ROMsocket

A14

NWDS

02

S1

CE2

R3

OE

a

2

2

0E2

R4

R5

R6

R2

CE2

R1

121 13

11

12BC8576

N1...N4 = IC 1 = 74HC00N5...N8 = IC 2 = 74HCOO

when the input voltage dropped from 1.5 to I V.

This effect is normal, however, and is due to the in-puts of the HC gates briefly being in an undefinedstate. The ICs fitted were Types 74HC00 (SMD)and a 43256C -12L (120 ns).The module is configured as a 32 Kbyte RAMblock by fitting wire jumper A -C, while jumper B -Cselects 2 x 16 Kbyte. The latter configuration is re-quired when the socket that receives the module isintended for a maximum memory capacity of 16

CfY

NWDS,Cr

27

* see text

Clnim

100n

R10

20

AA119

®s +2V42V4

r - -

[Ti-0

8 8IC 1 IC 2

87500

WE

Al 4

IC3

RAM6225643256

lx 32K: AC2x 16K: BC

62641x 8K

Kbyte (ROM or RAM), as on the BBC sideway ex-tension board. A Type 6264 RAM can be used inthe IC3 position when only 8 Kbytes are required.Neither jumper need then be fitted.Successfully constructing the RAM module re-quires great care in soldering the SMA parts ontothe board shown in Fig. 2. It is absolutely necessaryto first fit all the SMA parts at both sides of theboard, then the three wire links and jumper B -C orA -C as required. Do not forget to solder the ter -

73

capper- sideB

minals of Di (not an SMA part), and the batteryconnections, at both sides of the board. Also,through -contacting with short lengths of compo-nent wire should be effected at four locations. Pushall the pins of two 14 -way IC terminal stripsthrough the straight rows of holes on the compo-nent side of the board, i.e. the side that holds thetransistors, then solder the pins to the islands on thecopper side, i.e., the side that holds the 74HCOOs.The pins should protrude at least 4 mm. The use ofa centrally cut wire -wrapping socket is not rec-ommended here in view of the thickness of the pins.Locate the pin that protrudes from the hole marked1, and cut it off. Mount a turned IC pin holder nextto pin 28, 27, 22 and 20 of the right-hand side ter-minal strip, and solder these at both sides of thePCB. These pins should not protrude at the copperside, and their tops should be 1.5 to 2 mm abovethose in the terminal strip. When it is intended touse the RAM in its 2 x 16 Kbyte configuration,wires are connected to points 0E2 and CE2 at thecopper side, and guided between pins 5-6 and 9-10respectively. Remove pin 1 of a standard 28 -way ICsocket, before carefully push -fitting this onto the 27protruding pins at the copper side. Connect the bat-tery supply wires and the wire to Si (NWDS) to therespective points at the component side. Use a pairof precision pliers to carefully bend pins 28, 27, 22and 20 of the 43256 or 6264 slightly to the right ofthe other pins in the row. This enables pushingthese four IC pins in the previously mentioned, sep-arate, socket pins, while the 24 others are insertedin the usual manner. The battery is convenientlymounted at some distance from the module. Whena miniature battery is available, this can be fittedunderneath the RAM chip. For BBC users: wires0E2 and CE2 are conveniently connected to pins22 and 20 respectively of a 28 -way IC socket forplugging into the adjacent ROM/RAM socket onthe BBC's sideway extension board; the NWDSsignal is available at pin 8 of IC77. Switch Si ismounted at a convenient location on the com-

a

(18z

comp.-eudecopperE

Parts list

Note: all parts SurfaceMount Assembly types unless marked

Resistors:

Ri ... R4 incl.;Ri;139;Rio =47KRs= 180RR6;R8= 1K0

Capacitor:

Ci =100n or 47n

Semiconductors:Di=AA119Ti;T2=BC857B or similar pnp SMA type.T3 = BC847B or similar npn SMA type.1C1;1C2=74HC00 (Do not use HCT types).IC3=43256C-10/12/15L (NEC) or 62256

LP10/12 32Kbyte CMOS static RAM '.

Miscellaneous *:PCB Type 875002 off 14 -way terminal strips with 7 mm pins.4 off turned pins for IC leads.Suitable battery (see text, V b 2.4 V)

puter's rear panel, and when opened inhibitswriting into the RAM. It is recommended to openSi after turning the computer off to prevent the bat-tery having to supply some 50 µA for prolongedperiods: this current flows into the NWDS drivervia Rio. Non -BBC or Electron Plus -1 users shouldnote that the NWDS signal is the same as WRITE,not READ/WRITE.The MOVE command in the ADT ROM availablefor the BBC computer enables exchanging data be-tween resident and sideway memory. Programmersshould have little difficulty, however, in writing ashort routine that selects the relevant sidewaysocket(s) via the socket latch at FE3Ffi, and copy-ing one or two 16 Kbyte blocks.

74

052 40 -TRACK ADAPTOR

Over the past few years, the cost of 51/4 inch floppydisk drives has gone down to the extent thatmodern, 80 -track, double -sided drives now cost lessthan a simple, 40 -track, single -sided type some threeyears ago. It is, therefore, not surprising to see manycomputer owners upgrade their systems with a setof 80 -track, slim -line drives to boost the massstorage capacity of their micro.However, 40 -track stored programs are not readilyretrievable in the new system, because the distancebetween tracks in the 40 -track drive is twice that inthe 80 -track model.This circuit offers a solution to the problem, in thatit doubles the step distance for the R/W head in the

80 -track disk drive, so as to make it "look like" a 40 -

track type to the computer which should, of course,be programmed with a 40 -track disk operatingsystem (DOS).It is seen from the circuit diagram that Gate Nireceives the FDC controller STEP pulse, which isused in the circuit as a timing reference for the auto-matic generation of another STEP pulse to followthe first after 3 ms.It should be noted that, when incorporating the cir-cuit in an 80 -track drive, the track -to -track accesstime in the 40 -track mode is double that as given inthe drive specifications, which refer to 80 -track use.

N1...N2=IC2=74LS33MMVI,MMV2 = IC1 = 74LS221

053 2708 ALTERNATIVES

Thanks to the development of an ever expandingrange of capacious EPROMs in the 27xxx and25xxx series, the Type 2708 has become completelyobsolete. Not only is this forerunner in EPROMtechnology relatively hard to program, it is also ex-pensive in view of its modest 1 Kbyte holding ca-pacity.It stands to reason that replacement of the 2708with either the 2716 (2 Kbytes) or the 2732 (4Kbytes) is most readily accomplished if the dif-ferences in pin functions are first taken into con-sideration.

The pinning overview and associated table go toshow quite conclusively that the replacement is nodaunting task, since the former positive andnegative supply pins to the 2708, 19 and 21 respect-ively, may be hard wired as suggested for either the2716 or 2732.It should be noted that pin 18 (CE for the 2716 aswell as the 2732) is tied to ground, while pin 20(OE) is driven by the computer CS signal. This newarrangement is of no consequence for neitherEPROM nor computer, since OE may function asCE if it is realized that the EPROM can not be

75

A7

A6

A5

A4

A3

Al

AO

DO

D1

D2

GND

+5V

A8

A92708 2716 2732

ED VBB - VPP + SV All

C) CS/WE CS OE CS OENPP CS

(I) VDD - A10 * A10 *

0 PRG. GND CE GND CE GND

D7

D6

DS

D4

D3

switched to its low power standby state anymore.However, this minor drawback merely causes an in-crease in current consumption, whilst at the sametime offering a faster EPROM access time, as onlythe three -state bus drivers are enabled internally,rather than the entire chip logic.As the Type 2716 and 2732 EPROMs offer doubleand four times the capacity of a 2708, respectively,

5V 86489

a manual address block selection may be added tothe circuit; this set-up, composed of a switch and re-sistor (to be constructed double for the 2732) ismarked with an asterisk in the accompanyingdiagram. Wire Ai() (and All, if applicable) toground if you intend to stick to the 1 KbyteEPROM contents, located in the first 1024 bytesblock.

054 6502 TRACER

A program that has been written into an assemblerwill rarely run error free on the first run. It often ex-hibits blurbs and other ramblings: in bad cases,there is a complete hang up and it is then necessaryto start the computer afresh with a RESET.To find such faults in a relatively easy manner, thetracer described here will be found very useful.The circuit layout of the tracer is shown in figure 1.Gate N, is an address decoder, whose output in theaddress range $F000. . . $FFFF is logic 0. NANDgate N2 is fed with the SYNC signal from the com-puter and the 0 signal; it is disabled by either the ad-dress decoder, Ni, or bistable FF2. The addressdecoder disables N2 when the EPROM is addressedfrom the CPU. This prevents the SYNC line of the6502 processor generating an MI (maskable inter-rupt). If the processor passes through a machineprogram somewhere in the RAM, N2 generates an

interrupt as soon as the processor reads an opcode,which makes the SYNC line logic 1. This non-maskable interrupt directs the processor to an inter-rupt program in the monitor program. All CPUregisters are safeguarded by this interrupt programand subsequently displayed on the monitor screen.At the same time, the processor disassembles thenext command.The programmer can, therefore, see beforehandunder what conditions the processor starts with theexecution of the next opcode. Since the statusregister and all its flags are also displayed on thescreen, the programmer can easily ascertainwhether a flag in the status register has been set in-correctly.Bistable FF, serves as a debounce stage; FF2toggles on receipt of a leading edge from FF1: thatis, every time Si is pressed. When the tracer is

76

1 Parts listNMI

415 03

A14 (;)

A13054

2

a

A1206 5

SYNC 01

4,2

5V

12

2 10

5V.5 V 0

13

N2

5V

2

5 V

Si Lf

00

13(5

11 C 9CLK

FF1lxD

D.-

-7/1"3

3PR

CLK 0

FF2

5

60

D1

R4

10k

C1CI10p1135

N1,N2 = 1C1 = 74LS22FF1,FF2 =1C2 = 74LS74

2

R3

85466

Resistors:

Ri = 1 k

Rz...R4 = 10kR5 = 220 Q

Capacitors:

= 10µ/16 VC2 = 100 n

Semiconductors:= LED (red)

IC, = 74LS22IC2 = 74LS74

Miscellaneous:Si = miniature spring -loaded press -to -

make switchS2 = miniature spring -loaded press -to -

make switch (see text)PCB 85466

S 2

ISOa0g*XESZ2 (12

S I

switched on, D, lights. Resistor R4 and capacitor Clform a power -on reset network that automaticallyswitches the tracer off when the computer is

switched on.The printed circuit board for the tracer is shown infigure 2. If you want to build the tracer into thecomputer case, the PCB can be cut along the dash-ed line, so that the section containing Si and S2 maybe fitted in the most convenient position. Switch Simust be connected to the tracer via a suitable cable,but S2 may be connected to the manual RESET ofthe system.

77

055 A -D CONVERTER FOR JOYSTICKS

Although joysticks come in an astounding varietyof versions, their internal organization is virtuallyalways a standard concept, based on either a set ofrelatively fragile, springy, membrane contacts, ortwo potentiometers. Many computer en-thusiasts will agree that the latter, analogue, type of-fers better reliability and quality. Unfortunately,however, these can not be used in conjunction witha popular home micro such as the CommodoreC64, and that is where the present circuit comes in.The four comparators in IC, function as switches to

Al ... A4= IC 1 = LM 339N1... N6 =1C2 =74LSO4N7... N12 =IC 3 = 74LS 05

translate the handle movement into digital signals.The outputs of the comparators are buffered in 1C2to enable interfacing to the computer's joystickport. The two remaining inverters in 1C2, N5 andN6, along with two inverters in 1C3, function asdrivers for the LEDs that indicate the handle pos-ition. Gates N9 -N12 are set up as a wired NORfunction to enable LED D5 to light when thejoystick handle is in the centre position. Finally, thecurrent consumption of the converter is about25 mA.

R84,1472,2Rollo,R10

J) 14

id, 1C2 1C3

5V

O

78 87417

056 BIDIRECTIONAL PARALLEL INTERFACE FORC64

The so-called User Port on the Commodore C64home micro is intended for connecting peripheralssuch as a modems, RS232 interfaces, and controlcircuits. In some applications, it is also used for om-munication with other C64s. This circuit makes it

User PortC 64

PBS 0PB1 0PB2

PB3

PB4

PBS

PB6

PB7

PA2

I z 25mA

z

11113124 21141

possible to use port lines PBO-PB7 as inputs andoutputs. Software enables the computer to select be-tween input and output by means of the PA2 line(terminal M). Examples:

5

9

16

8/ 2022

IC 18212

6

10

15

17

19

21

2

O1,2441

6

8

10

15

17

19

21

IC 28212

b

S1

R31 R4 R5 R6 R7 R8 R9 RIO

O

00OOO

0

OO

16

18

2022

O

OO

121 2

D2 +4 2

N1, N2 = 2/6 IC 3 = 7404; 7414

red

green

User Port0 > ,%!

W5 a, 5

CYuzHid 4

>

4

> 0(2

2 3 4 5BCDE 6F

7H

8J

9K

10 11 12LMN98., 4 E

e.

E E 2

a = active lowb = active high

87519

79

Data input:10 POKE 56579,0:REM user port is input.

20 POKE 56576,255:REM interface is input.

30 A = PEEK(56577):REM read variable A.

Data output:10 POKE 56579,255:REM user port is output.

20 POKE 56576,251:REM interface is output.

30 INPUT B:REM read dataword.

40 POKE 56577,B:REM and send to interface.

The circuit is essentially composed of 2 three -stateoctal bus drivers Type 8212. Via the logic level onPA2, each driver can be enabled individually so asto select between the input or output function ofthe interface, whose current state is indicated by apair of LEDs. Switch Si selects between pull-up (a)or pull -down (b) termination of the input lines.Finally, an example for interactive data processing:

10 POKE 56567,255:REM interface is input.

20 POKE 56579,0:REM user port is input.

30 A = 255-PEEK(56577):REM read variable A.

100

:REM example of logic control:

110 IF A =1 THEN B= 64111 IF A = 2 THEN B =128112 IF A = 4 THEN B=192113 IF A =1 THEN B= 32

300 POKE 56577,B:REM load data register

310 POKE 56579,255:REM user port is output

320 POKE 56576,251:REM interface is output

330 GO -10 10

057 BIDIRECTIONAL SERIAL- PARALLELCONVERTER

This interface circuit enables doing rather morethan normally possible with the computer's serial(RS232) port. Serial output data from the computeris converted into parallel format, and parallel dataapplied to the interface is converted into a serial bitstream for reception by the computer.The interface is based on the industry standardUART (universal asynchronous receiver/transmit-ter) Type AY -5-1013, or the CMOS version of it, theCDP1854 from RCA. Serial data from the com-puter is received at input RXD, and inverted in Tifor driving the RSI input on the UART, which con-verts the received word into 8 -bit parallel format(RDo-RD7). The shifting in of serial bits is clockedby the 19,200 Hz signal applied to the RCP andTCP input. This fixes the baud rate of the interfaceat 1200 (19,200/16). The baud rate generator is aconventional design based on a binary counter/div-ider with built-in clock oscillator, which is crystalcontrolled here and operates at 2.4576 MHz. Theparallel output of the UART is buffered with the aid

of IC2 to enable controlling 8 relay drivers 131-138.The parallel word applied to the UART at its TDo-TD7 inputs is converted into serial format and out-put via the TSO terminal, where the signal is in-verted and fed to the TXD output.The serial data format can be defined with the aidof wire links B -F: Table 1 lists the function of eachof these. Inverter T4 automatically resets the re-ceiver in the UART by driving RDAR (receiveddata available reset) low when RDA (received dataavailable) goes high to signal that a complete wordhas been shifted into the receiver hold register.When wire link A is installed, RDA can also controlthe TDS (transmitter data strobe) input, so that anew parallel word (TDo-TD7) is loaded into thetransmitter holding register. Thus, jumper A makesit possible to use the CTS (clear to send) hand-shaking signal. The TEOC (transmitter end ofcharacter) pulse is used here to generate the RTShandshaking signal, and also to control the TDS in-put, together with CTS. This handshaking input,

80

when active, prompts the UART to output a newserial word. Set -reset bistable N1 -N2 precludes con-flicts arising between the signals in question. Power-

on network Cl-R1 ensures that the UART is prop-erly reset and initiated. TSO and TEOC then gohigh, while RDA is forced low. When link A is notfitted, the presence of the inverted TEOC pulse atinput TDS causes the transmission process to com-mence.The author has developed this circuit mainly toenable two IBM PCs to communicate with the aidof the Turbo Pascal program listed in Table 2. Beforethis can be run, the status of serial port COM1:(AUX:) should be defined by typing DOS commandMODE COM1:1200,n,8,2 <CR>(1200 baud, no parity, 8 data bits and 2 stop bits).Pins 6 (DSR) and 20 (DTR) on the 25 -way D socketshould be interconnected, and the same goes for

CTS

RTS

TXD

RXD

5V

5V

O

=C9

rvp

10160

121

1N4148

T1...T4 = 8C5478N1...N4 = IC5 = 74HC132

OC6r IC 4mom

1oon1 12 CEO ,74'

.I. 4060

8

pins 4 (RTS) and 5 (CTS) when no handshaking isbeing used. When it is intended to use the hand-shaking facility on the bidirectional interface, link Ashould be removed, and socket pins 4 and 5 connec-ted to interface terminals CTS and RTS respectively.

35, 36 3738139

(SIPB13

POE

8C RDAR

RDA

IC 1

TDS

TEOC

ROO

RD I

902

R03

RD4

RD5

1306

RD7

TDO

Tel

2

Table 1

link fitted not fitted

A no RTS & CTS RTS & CTS

B no parity bit parity bit

C 2 stop bits 1 stop bit

D/E see below

F even parity odd parity

D E data word

0 0 5 bits

0 1 6 bits

1 0 7 bits1 1 8 bits

C7

11-110n11-0-0 5V

2

1 20

11

10

7 13

6 15

5 17

26 18

IC 2

74HC241

El

(") 18>Me2

/4 83

B4

85

3 87

27 16

28 14

COP 1854A2502 T23

8502AY -5-1013704

TD5

30

31

TD6 32

TD7

ISO

; tt401 17 21 41J Q16

RSI

6

3

3

19,1 kHz

12

1 1991

El 2

4 810

6 B11

IC 31311

9 74 " 813HC

7 241 13 614

5 5 Bis

37' 1316

r6 010 20

C8

100n

815

1 4576MHz

C4

39p

O

O

81...888x

B9...616

87499

3 - 50V

81

058 BUS DIRECTION ADD-ON FOR MSXEXTENSIONS

The majority of MSX computers do not require aBUSDIR (bus direction) signal from add-on circuitsplugged into slots. A problem arises, however, if theextension circuits published in Elektor Electronicsare used in conjunction with, for example, a SanyoMSX machine, which has a few peculiarities in itsexternal I/O concept. In general, the more slots onan MSX computer, the higher the probability thateither one of, or both, these circuits are required tobe able to use the home-made extensions.Two solutions are offered to provide for theBUSDIR signal. One is usable for the Universal I/OBus and the I/O & Timer Cartridge, the other forthe Cartridge Busboard. Each of these circuits con-sists of one IC only.Circuit A is used with the two I/O extensions, andis readily incorporated in the computer, at a suitablelocation near the slot that receives the extension. Ifnecessary, all slots on the computer are fitted withthis circuit, but this makes it impossible to utilizecartridges that do supply a BUSDIR pulse, unlessSi is included to disconnect the output of N4 fromslot pin 10. Note, however, that this switch must notbe operated when the computer is on.As I/O range 40h-FFh is reserved for thecomputer -resident hardware, address lines A6 andA7 must be low for the selection of external I/O cir-cuitry. Moreover, IOREQ and RD must be low toensure that BUSDIR is only active when the CPUreads data from an I/O device. Interrupts from anexternal device can only be processed correctlywhen BUSDIR is low in response to M1 andIOREQ being low also. This requires an OR func-tion for logic low levels:BUSDIR =MI IOREQ + IOREQ RD A7 A6If you are hesitant about opening the computer toinstall circuit A, you may consider the use of a partof the EPROM cartridge board to hold the74HCT32 as shown in the accompanying photo-graph. Note that the 50 -way track connector plugsstraight into a computer slot, and that a slot con-nector is fitted at the other side of the "adaptor -PCB" to receive cartridges.Circuit B is intended for use on the CartridgeBusboard. Its function is to pass BUSDIR pulsesfrom cartridges to the computer. To this end, it is

necessary to first break the interconnecting tracksbetween slot pins 10 so as to make all cartridge

N 1...N4 = 1C1 = 74HCT32Di, D2 =1144148

b)5V

0

10 - K1

10 -K2

10-K3

10 -K4

10 -05

10 -K6

10-K7

10-K8

BUSDIR10 -K9

0 BC 547

R1... R88 4k7

5

11

12

N5 =IC 2 = 7430

87430

BUSDIR outputs separately available for wiring to8 -input NAND gate N5. Inverter T1 turns thissimple add-on unit into an 8 -input OR gate for logiclow levels. The collector of this transistor is wired topin 10 of Ks on the busboard.

82

It may well happen that both circuit A and B are re-quired for a specific I/O arrangement. In that case,it is suggested to fit circuit A on one slot of the Car-tridge Busboard, and consequently use only thatslot for external I/O. Pin 8 of N4 is then connecteddirect to the relevant input of Ns.

Note: articles in the series MSX Extensions werepublished in the following issues of Elektor Elec-tronics:January 1986, February 1986,March 1986, January 1987,March 1987, April 1987.

059 COMMUNICATION PROGRAM FOR C64

This program enables users of the popular Com-modore C64 home computer to exchange messagesbetween two machines.No hardware whatsoever is needed to accomplish: communication over several tens of metres using

a three -wire connection-see Fig. 1. Longerdistances, or communication over the telephone, ofcourse require the use of a modem. split screen operation: the upper half of the

screen displays the operator's input (LOCAL),the lower half displays the received messages(REMOTE). full duplex communication, i.e. transmission and

reception are quasi -simultaneous processes.

The flowcharts in Fig. 2 illustrate the structure ofthe proposed program. TX is short for transmitter,R X for receiver. Note that screen pointer updatingroutines are not apparent from these diagrams.Unfortunately, since the C64 BASIC interpreterdoes not allow structured programming to be car -

I (C 64) :

1

ried out, the constructs shown in the flowcharts arenot readily detected in the practical BASIC programlisted in Fig. 3.Keyed -in text is transmitted to the far computerafter pressing the RETURN key. The BORDERcolour changes to warn the user when the screen isfull. Typing errors can be corrected in the usual waywith the aid of the INST/DEL key. A short beep issounded to signal the receipt of a message from theREMOTE computer.Testing the program is straightforward, and does notrequire two computers. Figure 4 shows the connec-tions that can be made temporarily on the com-puter's user port. This creates a zero modem, andcauses LOCAL text to be echoed on the REMOTEscreen.For those computer enthusiasts interested in analys-ing the BASIC program, and for those who intendto rewrite it for other types of computer, the func-tion of the major lines can be summarized asfollows:

II:

O 0000.00000001 2 34567891011 12ACDEFHJKLMNBO 090 0000000

O 0 00000000001 234567891011 12ABCDEFHJK LMNO 00000000000

83

2 * MAIN LOOP

INITIALISATION

BLINK TRANSMIT CURSOR

READ TX CHARACTER

THEN

TX CHAR. = DELTHEN

ERASE PREVIOUSCHARACTER

FROM TX SCREEN

TX CHAR. = RET

THEN

(TRANSMIT)

THEN

TOGGLEBORDERCOLOUR

TX CHAR < > EM

TX SCREENFULL

ELSE

ELSE

ELSE

TX CHAR.TO TX SCREEN

PTY

1

ELSE

(RECEIVE-)

REPEAT FOREVER

* (TRANSMIT) SUBROUTINE

TRANSMIT MARKER ">SPACE"

DO FOR ALL CHARACTERS:

TRANSMIT CHARACTER

ERASE CHARACTER

RECEIVE

TRANSMIT MARKER "RET" (END OF MESSAGE)

* (RECEIVE) SUBROUTINE

BLINK RECEIVE CURSOR

READ RX CHARACTERYmm.M.

THEN

BEEP

LF FOR RX CURSOR

CLEAR CURRENT LINE

RX CHAR.<>

RECEIVED CHARACTERTO

RECEIVE SCREEN

EMPTY

1

ELSE

87461-2

100-125: initialize the screen and the sound gener- 140: T is the base address of the transmit screen,ator. and T0 is the associated index. R and R0 are similar130: open the serial port with parameters 300 baud, variables for the receive screen, while R1 in addition8 data bits, 1 stop bit, no parity, no handshaking, gives the maximum number of character per line.full duplex. 160: blink the cursor and read the keyboard buffer.

84

180-200: test for DELETE, and erase the previouscharacter.210-230: test for RETURN and transmit message.240-260: toggle the BORDER colour when thescreen is full.270: go to the receive subroutine.280: repeat the above loop.710: transmit the "begin of message" marker.720-750: transmit and erase all characters. Moni-tor the receive channel for messages, after trans-mission of every character; reception has thehighest priority.760: transmit the "end of message" marker.810: blink the cursor and read the receive buffer.820: buffer empty?830: end of message.840: have the sound generator produce a beep.850-870: advance the cursor to the next line.880: clear the new line.900: display received character on REMOTEscreen.910-920: advance cursor to next position.

C64:

0 0 0 0 0 0 0 0 0 0 0 01 2 3 4 5 6 7 8 9 10 11 12

ABCDEFHJKLMN00000000000

Flag 2 Sin

Isoul

3100 POKE53281,12:PRINT"":FOKE53280,9:POKE53281.0:PRINT CHRS(152):POKE53272,23110 61=54272:POKE 24+51,15:POKE 61,207:POKE 1+SI.34:POKE 5+61,10128 FOR H=1033 TO 1044: READ A: POKE H,A: NEXT H125 FOR H=1273 TO 1283: READ A: POKE H,A: NEXT H130 OPEN 2,2,0,CMRS(6)+ONR$(0)140 T=1104: TO=0: R=1344: R0=0: R1=0150 REM MAIN160 POKE T+T0,60: POKE T+T0,32: GET TS170 IF TS."" THEN GOTO 270180 IF TS<>CHRS(20) THEN GOTO 210190 IF T0>0 THEN TO=T0-1200 POKE T+T0,32: GOTO 270210 IF TS<>CHRS(13) THEN GOTO 240220 GOSUB 700230 GOTO 270240 IF T+TO>=R-80 THEN GOTO 260250 POKE T+113,ASC(TS): TO=T0+1: GOTO 270260 POKE 53280,1: FOR H=0 TO 15: NEXT H: POKE 53280,9270 GOSUB 800288 GOTO 150700 REM TRANSMIT710 PRINTS2,CHRS(62):: PRINT42,CHRS(32);720 FOR K=T TO T+TO-1730 PRINT02,CHRS(PEEK(K));: POKE K.32740 GOSUB 800750 NEXT K760 PRINTS2,CHRS<13/;: TO=O770 RETURN880 REM RECEIVE810 POKE R+R0,60: POKE R+R0,32: GETO2,RS820 IF RS="" THEN GOTO 930830 IF RS<>CHRS(13) THEN GOTO 900840 POKE 54276,0: POKE 54276,33850 IF R1=40 OR R1=0 THEN GOTO 870860 POKE R+R8,32: R1=R1+1: RO=R0+1: GOTO 850870 R1.0: IF R+R0=2024 THEN R0=0880 FOR H=R+190 TO R+R0+39: POKE H.32: NEXT H890 GOTO 930900 POKE R+RO,ASC(RS): R0=R0+1: RI=RI+1918 IF R1=40 THEN R1=0920 IF R+R0=2024 THEN R0=0930 RETURN950 DATA 42,32,84,82,65.78,83,77,73,84,32,42960 DATA 42,32,82,69,67,69,73,86,69,32,42970 END

READY.

85

060 CPU GEAR -BOX

While many computer enthusiasts are keen on get-ting their system to run at the highest'possible clockspeed, there are often quite awkward constraintsposed by relatively slow, bus -connected supportchips, and the ensuing frustration after failing to getreliable system operation at, say, double the 'old'clock speed may readily lead to abandoning thespeed-up project altogether, for lack of precise infor-mation regarding the necessary clock -basedsynchronization between CPU and peripheralchip(s).A noteworthy example of this happening in practiceis the go at incorporation of the Type 9367 CRTcontroller in a 6502 -based computer system runn-ing at 2 MHz; the specific application concerns thehigh -resolution graphics card published in ElektorElectronics, November 1985 ff.This circuit ensures a correctly timed, synchronizedslow -down of the system clock speed, when appro-priate for CPU access to a memory -mapped (E150 -El 5F) device. Following the reception of a highlevel on the relevant I/O line, the proposed circuitarranges for the clock signal frequency to be dividedby two, while a low I/O causes division by four.It is important to point out why the commonly usedmethod of using 4)2 to enable the address decoderchip is to no avail when it comes to synchronous

1

I/0

4MHz

and glitch -free clock speed switching under soft-ware control; the following paragraphs thereforeaim at offering an insight into the basic operationof the gear -box circuit and its incorporation in a6502 -plus -graphics card system.Figure 1 shows the hardware to the gear -box. Alogic level at the I/O input is passed to the D(data)input of bistable FF3, as well as to the R(reset) input of FF4. FF3 toggles and activates its Qoutput; this causes the 4 MHz clock signal, dividedby two in FF4, to be output as 2 MHz towards theCPU 43 terminal. Division by four (1 MHz clockoutput) should take place in a synchronous timingarrangement as soon as I/O goes low; just prior tothis pulse transition, 4).n has already gone low, sothat the level change at the FF4 reset input is of noconsequence to the CPU operation at that time,however the bistable can not change state anymore.Thus, FF3 will have to supply the output clocksignal; the D input follows the I/O signal tran-sitions, since Q of FF2 was forced to go low in con-sequence of S (set) being activated. The firstleading edge coming from the FF, Q output willcause Q of FF3 to go logic high, ending the set con-dition of FF2. Given an input clock frequency of4 MHz, the outlined timing sequence results in Qof FF2 going high after 250 ns, followed by a low

F F1,F F2 = IC1 = 74LS74

F F3,F F4 = IC2 = 74LS74

N1,N2 ='/2 1C3 = 74LS32

86496-1

86

2 000©

O

©

2 MHz 1 MHz 2 MHz

level at Q of FF3 after another 250 ns. The timingdiagram shown in Fig. 2 clarifies this, admittedlyrather complicated, timing arrangement in the gear-box circuit. It is noted that a complete 1 MHzperiod has lapsed, provided FF1 is properly syn-chronized during the CPU initialisation cycle.

Theoretical research into this matter, however, hasshown that this is not always the case; the result isan asymmetrical output clock period with a logiclow and high level duration of 250 and 500 ns re-spectively. The remedy for this undesirable effect issimple, since it merely involves interchanging theclock signal connections to FF2 and FF3.It is seen that (ID2-based I/O decoding is lessdesirable, since it involves too long a delay; what re-mains is to indicate the method of obtaining I/O

from the graphics card system (EE, November1985, p. 71).XX5X is dismissed for now obvious reasons, butP = Q at pin 19 of IC, can be used for our purpose,while the possible objections to the resultant, rathercoarse address decoding are readily rendered devoidof relevance by the incorporation of a single 3 -to -8decoder Type 74LS138, mounted piggy -back ontoIC2 and connected direct to pins 1. . .5, 16 and 8.The remaining pins of the additional IC are eithercut off or bent to preclude wrong contacts from be-ing made in the circuit. However leave pins 6 and10 in function, since the former should be tied per-manently to + 5V (small wire to pin 16), while thelatter can now be used to supply the correct I/Opulse for the CPU gear -box.

061 CURRENT LOOP FOR MODEM

A modem, such as the direct -coupled modemfeatured in the October 1984 issue of Elektor Elec-tronics, opens a whole new world to the computeruser by making possible communication betweentwo computers anywhere in the world (provided, ofcourse, they can be coupled to a telephone line).Ironically, although the distance between the com-puters may be very large, that between computerand modem is strictly limited. This is because theRS 232 input is voltage driven and is, therefore,very susceptible to noise. This is not a new problem:it existed many years ago when, for instance, two

telex machines had to be interconnected. Thesolution then found, and still in use today, is thecurrent loop. Such a current loop can also be usedwhen the distance between the modem and thecomputer is relatively large: up to 1 km.A current loop so used converts RS 232 compatiblevoltages into RS 232 compatible currents. The stan-dard in the RS 232 protocol is a current loop of20 m A.In view of the arrangement of the circuit it is poss-ible for the current loop to be used as a voltagedriven input and output. In the receiver, the opto-

87

isolator converts the input current into an outputvoltage via T,. The output voltage is ± 12 V. As thecurrent loop is closed via V + and V-, mind thepolarity. If you want to use an input voltage insteadof a current, apply the input between V- andearth.The input voltage to the transmitter may vary fromTTL level to ± 12 V. Its output signal is available aseither a voltage or a current: the former betweenV+ and earth and the latter between V+ and V.Current consumption in the quiescent state is zero;with full load, it amounts to 20 mA.The maximum bit rate at which the circuit operatesreliably is 1200 baud, but this can be increased bythe use of a faster opto-isolator.

062 DIRECT READING DIGITIZER

The computer to which this digitizer is coupledreads a 3 -digit number that is a direct representationof the measured voltage in millivolts.The analogue -to -digital converter is an RCA typeCA3162, which was designed for use in a 3 -digitdigital voltmeter. The input range of the IC stret-ches from -99 mV to 999 mV: the resolving poweris, therefore, 1098 units. In other words, this con-verter offers a resolving power that is better thanthat of a standard 10 -bit device for the price of an8 -bit device.The 3 -digit information at the output of the 3162 ismultiplexed. The data can, for instance, be writteninto the micro via seven PIA (peripheral interfaceadapter) input lines. That means, however, thatsome machine language is required to be loadedinto the RAM every time the converter is to beused. The present circuit uses hardware to obviatethis difficulty.The 3 -digit information, which is emitted every20 ms, is automatically loaded into three 4 -bit buf-fers, IC8, ICs, and 1/2IC, 5, whose outputs are con-nected direct to the data bus. Each of these buffershas its own address. Writing the converted valueinto the computer has become simply a matter ofreading the three memory locations, which can becarried out by PEEKs in BASIC.The address decoder consists of IC3. . . IC5 and IC7.The present circuit occupies a block of eight ad-

dresses of which only the first four are used. Whenthe first address is read, monostable IC, is started,which causes IC, to commence the conversion pro-cess. When the monostable returns to its stablestate, IC, goes to the HOLD mode, and themeasured voltage can be read.An interval of not less than 50 ms is required be-tween the start of the conversion process and thereading of the buffers.The eight successive memory locations required forthe digitizer may be placed anywhere in thememory range by means of the open inputs of gatesN5.. .1\116. If any of these inputs is connected to+5 V, the relevant address line becomes logic 1; ifthe input is linked to 0 V, the address line goes logiclow.

Assuming that the decoding has been set to address$E300, the first address is read with a PEEK, whichstarts the conversion.Wait for 50 ms.Write the data from address $E301, which is theleast significant bit (LSB), i.e., the extreme right-hand digit of the 3 -digit number. Then write $E302and finally $E303. At each of these transfers, anAND action must by carried out with 00001111(binary) or 15 (decimal), because only the fourlowest data bits are of import.If the converted voltage during the further process-ing of the three written digits is negative, this is indi-

88

N1 . N4 = IC2 = 74LS132N5 ... N16 = 1C3,1C4,105 = 3 x 74LS266N17 ... N20 = IC6 = 74LS08

5V

C31 C41 CLMKT

7 0n 107100p 270n10V

5V*0P1,P2 = multiturn

-99 my

P1

50k

MIN

9

5V*0DIma 2xAlN4148

11

+999 n1V

O

5VO

R10

13

C9

680n 1

4-

MKT

121 14

NT

ZEROadjust

HI

IC1

0

CA 3162

2

2

2

LSD

R3 R4 R5.L

2

5V

R6

5

5

R7 RS R9

1 ,2.)s

N2011

NOD 3

LO MSD

GAIN .1. HL013

P2

7

10k

14 41 31 2(1)5(1)

0 2 m IQ CLRR/C

IC1174LS122,.

MI=

*see text

5V

r- 5 V

6

4

C5 C6 C7SIM 4S149

731373

912

K3N 0135V

6

5V0

tooto

id6 Tin ICS 47n

CBis

14

13

12

IC8

11

M1D 10

2D 203D 304D IC8 40

74LS173CLK

CLR ra2r5ro y9c?

..L

14

(51C10

4

13

12

11

7

N210 11

eN3

AO

)CoDo

M10 10

20 20

3D 304D IC9 40

74LS173CLKN CLR 11; laTry 9?

11

6

3

4

7

8

14

10

OS

ID 1020 iCi0 20

430D 7430

5DLS 12

6D 374 80 157D 70

BD 80

4

5

6

2

5

6

9

5VO 15014ci3O2n11606 96 76

16 I> I> r>.-IC7

74LS138 GIA D

0 0 0 0 0R 55 AO Al A2 A3

N4

Z40 al

ZZ2ZZ2Z( Z I Z

6

19

R13 5 V

0 0 0 0 0 0 0 0 0 0 0 0A4 A5 AG Al AB A9 AIO All Al2 Ala A14 A 5

85498

0 DO

0 010 D2OD3OD4OD5OD6007

01)2

5V

cated by the data at address $E303, which is 10.

Overflow is also easily recognized: if the value readfrom address $E301 is 11, the voltage is greater than999 mV; if the value is 10, there is a negativeoverflow.

The small BASIC program given here is an exampleof a possible conversion routine for the Junior com-puter.The circuit as shown can be used with a 6502 µ13;if it is required to be used with a Z80, RD must be

89

connected to the R/W line via an inverter, andIORQ or MREQ is put onto the 02 line via an in-verter. The choice between IORQ and MREQdepends on whether the digitizer is located in theI/O or the memory range respectively.Take care during the construction to keep the con-nections marked with an asterisk (0 V and +5 V toIC)) as short as possible. These lines go together toC3 and C4, from where the connection is made tothe 0 V and + 5 V lines of the digital part of the cir-cuit. Keeping these lines short prevents possible in-teraction between the analogue and digital parts ofthe circuit.Four inputs of ICio are not used in the present cir-cuit, and they can, therefore, serve as four ad-ditional digital inputs. During the reading of ad-dress $E303 (in the example), the highest four databits indicate the state of these four inputs.

10 A=1616A3+3816"2; REM ADDRESS 8E300

20 IWEDUA/1 REM START CONVERSION

30 FOR T=1 TO 15: NEXT: REM DELAY

40 WEER(01) AND 15

50 Y=1,913((002) AND 15

60 b0M9(i4+8) AfiD 15

70 S=I

80 IF 2=10 THEN Z=0: S= -S: REM SIGN IS NEGATIV IF 2=10

90 AD=SX(10012+10XY+X)

100 IF X=11 THEN PRINT ' POS.OVERFLON '; CHR$(13);: GOTO 130

110 IF X=10 THEN PRINT NE6.CIVERFLCI4 '; CHR$(13);: 0010 130

120 PRINT ' U=';AD;' mV '; C/82$(13);

130 GOTO 10

063 DISCRETE DAC

A digital -to -analogue converter (DAC) that is easyto build from a handful of readily available parts.The 8 -bit digital input for the circuit is applied toresistors R17 -R24 incl., each of which drives an as-sociated current source composed of two series -connected diodes, a transistor and a current defin-

01

D2

R9

ing resistor fed from the positive supply rail. A logichigh level at the input causes the relevant currentsource to be switched on, a logic low level switchesit off. The sum of currents from Ti -T8 incl. is ar-ranged to pass through preset P1, which thus dropsa voltage Uo in accordance with the magnitude of

0

C.T9

R17

R1 D3 R2 051 R3 D7 R4 D9 R5 011 R6 013 R7 D15 R8

04

U

I.T1 T2

R10

0

C.T10

D6

R11

0

C.TI

C408

I

CO13 14

R12

0

112

DIO

R13

0

113

C.012

C.TS 16

R14

0

114

014

R15

T15

D16

R16

0

116

C.18

ClOmom

100p 40V

P1

Ub

0 0

R18 R19 R20 R21 R22 R23 R24

*see text

0 0

LSB DO 01 D2 D3 D4 D5 06 D7 MSB.1.

1.1..:r8=13C557B

T9...T16=BC547BD1...D16=1144148

87435

90

the 8 -bit word written to the circuit.The current supplied by each current source isabout 700/R. [mA], where R. is the value of theassociated resistor between the emitter and the + Vrail. In order to ensure satisfactory linearity of theanalogue output voltage, resistors 16 -Rs incl. mustbe dimensioned to obtain a current ratio of 1:2 be-tween any two adjacent sources. In practice, it iswise to first apply a logic high voltage to the MSB(most significant bit) input of the circuit, leaving theremaining inputs low, and measure Uo with the aidof a good -quality voltmeter. Next, drive D6 highand all other inputs low, and make sure that Uodrops to half the previously obtained level by dimen-sioning R7 as required. The other current determin-ing resistors are similarly established; the value ofR1 -R8 incl. that gives the correct level of Uo is ob-tained by making suitable combinations of series

andlor parallel connected high stability resistors.Alternatively, it is possible to use multi -turn presets.As all resistors R, -R7 incl. must be dimensionedstarting from a particular value of R8, this resistormust first be calculated considering that the outputlinearity of the circuit is affected unless

1.4131/R8 < lUb1-2

In practice, the maximum feasible level of U0 isabout 1/2Ub-1 EV] with only MSB high, and thislevel should be observed in the dimensioning of R8and the setting of P1Although this 8 -bit DAC should be sufficiently ac-curate for most practical applications, it is of coursepossible to opt for a greater or smaller number ofcurrent sources with a corresponding increase ordecrease in the available resolution of Uo.

064 DRIVE SELECTOR

This circuit makes it possible to use double -sideddisk drives with a computer that supports onlysingle -sided units. Many of the older generation ofcomputers were designed to operate in conjunctionwith Shugart -compatible, single -sided disk drives.These have rapidly been superseded, however, bythe more economical double -sided drive, which hasa greater storage capacity.The Shugart standard supports the use of four diskdrives, which are selected with drive select linesDSO-DS3. Two further lines, HS0 and HS1, controlthe head selection on each of these drives. Whenthis circuit is installed between the computer's diskcontroller output and two double -sided drives, thedisk operating system (DOS) can recognize fourlogical drives. When the computer selects drive Aor B, the situation is similar to before the conver-sion. Selection of drive C or D, however, causes thesecond head in the relevant drive A or B to be ac -

Table 1

N1...N4 = IC1 = 74LS08, 74HCT08N5...N8 = % 1C2 = 7407

logicaldrive DS3 DS2 DS1 DSO DSO DS1 HSO HS1

physicaldrive

A (1) = 1 1 1 0 -, 0 1 0 1 = A side 08 (2) = 1 1 0 1 -' 1 0 0 1 = B side 0C (3) = 1 0 1 1 --0 0 1 1 0 = A side 1D (4) = 0 1 1 1 -. 1 0 1 0 = B side 1

5V

87445

91

tivated. In this way, the total storage capacity of thedouble -sided drives is available even under"primitive" circumstances.Note that the use of drive denotations A -B -C -D or

0-1-2-3 is particular to the type of computer, or theDOS version. Finally, Table 1 provides informationabout the combination of the criginal four DS linesinto two HS and two DS lines.

065 FILTERED CONNECTOR

Computers and computer -driven peripherals arenotorious sources of RF interference, and receiverjamming may occur at frequencies well above100 MHz, even though the computer is said to runat a mere 16 MHz or so. The cause of this problemlies in the very fast pulse rise time of the switchingand timing signals internal and/or external to thecomputer system and its peripherals, which areoften located well away from one another (printer,modem, mass storage).Much of the interference originating from longperipheral wiring systems may be suppressed quiteeffectively by inserting simple low-pass filters in thesignal lines for data and handshaking. The proposedL -C filters are composed of small (3 mm) ferritebeads with 10 turns of 0.2 mm (36 SWG) enamelledcopper wire, plus a ceramic 1 nF capacitor; the coilinductance is about 80 µ}1, which gives a cut-offfrequency of about 60 kHz (120 Kbaud).The filters are mounted on a small piece ofveroboard which may be cut and filed to fit into astandard D -connector housing. Other cut-off fre-quencies may be defined by modifying the smallcoils; inductance is proportional to the square of thenumber of turns, while constructors boasting ofgood (near) eyesight and lots of patience mayendeavour to use thin (0.05 mm) copper wire to runthrough the beads. However, the L -C ratio as givenshould not be modified.

RxD

SIGN.GND

TxD

RTS

CTS

PROT.GND

In conclusion, it should be noted that a filtered con-nector dimensioned for, says 10 kHz, should not beconnected to a high frequency (20 MHz) computeroutput, since the excessively high capacitive loadmay cause damage to the line driver IC.

066 FLOPPY CENTRING UNIT

In modern disk drive mechanisms, as, for instance,the TEAC FD55x, the motor starts automaticallywhen a disk is inserted into the drive. When the lidis closed, the motor stops again. This arrangementensures better centring of the disk. Better centringmeans less wear on the centre fixing hole, the life ofthe disk is extended, and read/write errors owing toeccentricity off the disk are prevented.Owners of older drive mechanisms, such as the

BASF 6106, can incorporate that facility with thecircuit proposed here. The signal from the write pro-tect phototransistor is used to determine when adisk is being inserted (this signal is normally gatedwhen the drive is closed), and to start the motor forthe total period of monostable MMV1. TheSPEED signal is not absolutely necessary: it stopsthe motor direct when the lid is closed. If it is notused, pin 3 must be connected to the + 5 V line.

92

The motor will then run for the duration of theperiod of MMV1, i.e., about 10 s. The monostableperiod can be reduced by lowering the value of thecapacitor.The points where to connect the circuit in the 6106are easy to find. Looking at the pcb from the front,you will see a cut-out in the front centre of theboard. Immediately to the left of this are three ICs(see photograph). The one at the front is a 7474, theone in the middle a 7432, and the one at the backa 7404. The signal SPEED is taken from pin 6 ofthe 7474, and the signal DI from pin 2 of the 7404.The signal MOTOR ON is applied to pin 3 of the7404. As all existing connections remain, the con-necting wires of the auxiliary circuit can be soldereddirect to the relevant IC sockets. In the same way,it is possible to derive the supply voltage for theauxiliary circuit: for instance, +5 V from pin 14 ofthe 7404, and 0 V from pin 7 of this IC.It is important to note that there are two types ofpcb used in 6106 drives: the ICs and the IC functionare the same in both versions, but the constructionmay look different .

IC 1E(4217404)

DI

SPEEDIC 3E

(06/7474)

5V

5V

Rt-TR

MMV1 0

R

IC1

MMV1 = 1/SIC1 = 4098

Cl

6V3 Tant.MOTOR ON

IC 1E(03/7404)

BC 547

85428-1

067 FLOPPY DISK DRIVE

This is a much simplified version of the circuit pub-lished in the April 1984 issue of Elektor Electronics,but it is, unfortunately, not usable with all diskdrive motors.First, a recap of the operation. The drive motors areswitched on when one of the drives is accessed bya DISK SELECT signal. There is a delay of a fewindex pulses before access proper to give the motorspeed time to stabilize. A few seconds after all thedrives have been deselected, the motor is switchedoff. This arrangement reduces operation of thedrive mechanisms, the heads, and the disks to aminimum, which ensures a longer life of thesedevices.In contrast to the earlier published article, theREADY output of the drive mechanism is used,wherein lies the reason that the older circuit cannotbe as compact and simple as the present one: it hasto take into consideration that not all drivemechanisms have this output. However, as far as wecan find out, most drive mechanisms do have it, butthere must be some, of course, that do not.Figure 1, which is part of the circuit of the floppycontroller board (Elektor Electronics, November

2 K2

nMOTOR ON

85427-2

93

1982), shows the new wiring of port A7. The x atplug PL2 represents pin 3 of the type FD -55x drivemechanism, and pin 6 of the BASF 6106. As thislatter input corresponds to Disk Select 4, not morethan three BASF 6106 drives can be connected tothe present circuit.It is a wise precaution to break the connection be-tween pin 10 of gate N25 and pin 6 of PL2, but it isnot strictly necessary. As long as you do not selectdrive 4 (with the Ohio DOS, drive D), nothing cango wrong.

One connection that must be broken is that be-tween pin 16 of PL2 and earth. Instead, pin 16 mustbe connected to pin 8 of IC2 as shown in figure 2.If you are really a dab hand at soldering, you maybe able to make the changes, with the appropriatelengths of wire, on the relevant printed circuitboard. Most of you will, however, find it mucheasier to use a 15x20 mm piece of veroboard, whichafter completion can be glued or screwed on shortspacers underneath C16 on the floppy controllerboard.

068 HEXADECIMAL KEYBOARD

There are various ways of producing a hexadecimalkeyboard. Normally, it is based on a number of keycontacts in a matrix, but here a rather simplermethod is used: 16 key contacts (0 . F) that arecommoned to the positive supply line. Suchkeyboards are commercially available.Code conversion is carried out by two priority en-coders, IC3 and IC4. If one of the inputs Io . I7 ofthese ICs is connected to the positive supply line via

1J)

3 ... 18 V

14 18

IC1 IC2 IC3

T 4)

()IC4

N1... N4= IC1 = 4093N5 ... N8 = IC2 = 4071

617

0

C

3

5

N2

6

1000

one of the contacts Sl . .S16, i.e., made logic high,the relevant binary code appears at the associatedoutput, Qo... Q2, of which Qo is the least signifi-cant bit (LSB). As the encoders are cascaded, thereis a total of 16 inputs.Corresponding outputs of the encoders are com-bined in OR gates N6 . . N8 to form the lowestthree output bits Do . D2. the fourth data bit istaken from the GS (group select) output of IC4. This

_FL0 "'°'"

-;rte

N6

51

7

10

13

is 0.,

0

oo 01 02

IC34532

13 14 15 16 I

8x475 4-1

11 12 13 1 2 4

14

5 15Eou,

10

0

S 00

IC44532

2 13 14 5 I

12 13 1 2 3

02

4

51 818 285468-1

5 e3 ...18 V

94

output is logic high when one of key contactsSs . . .S16 (8 . . . F) is closed.As the GS outputs of the two ICs are combined inOR gate N5, D3 is active high when a key is pressed.The signal at pin 9 of N3 is delayed by R18 -C2. Atthe same time, the signal at pin 15 of IC3 triggersmonostable N1 -N2. During the pulse period ofabout 10 ms, pin 8 of N3 is logic low so that, in-dependent of the delayed signal at pin 9, the outputof N3 remains logic high. If pin 9 of N3 is still highwhen the pulse begins to decay, the output of N3goes low and remains so until pin 9 becomes logic 0again. During this time, pin 6 of N2 remains low, sothat the monostable cannot be triggered erroneous-ly. The timing diagram in figure 2 further clarifiesthe operation, which results in a debounced strobeor strobe pulse.If more than one key is pressed, the highest is selec-

2 .14,

IC3, 015

N3

N3, 09

ted, as is to be expected from a priority encoder!The circuit requires a power supply of 3 . . .18 V:current ccnsumption is not greater than 10 mA.

069 IMPROVED SOUND FOR THE BBC MICRO

Despite the many laudable qualities of the BBC with the sound quality of the standard version asmicrocomputer as to speed and ease of peripheral manufactured by the Acorn company. An investiga-interfacing, many users are slightly disappointed tion into this matter has revealed that Acorn have

externalsound output

86402-195

disregarded the optional connection of an externalaudio amplifier to the computer; this is the moresurprising since special holes have been provided tothis purpose on the main PCB. The result of thisomission manifests itself in a very poor sound qual-ity, caused by the small loudspeaker in the cabinet,the high noise level of the improperly driven audioamplifier chip, and the rather coarse volume setting.However, a minor modification to the BBC com-puter is sufficient to boost its sound production bymeans of an external, more powerful audio ampli-fier which may be connected to a sound outputsocket on the computer. Proceed as follows:1. Open up the computer, remove the keyboard and

the main PCB.2. Locate the PCB holes for plug 16, to the left of

IC7, the Type LM386 audio amplifier chip.3. Use desoldering braid to open up the holes for

plug 16, if these are filled with solder.4. Cut off the centre pin of a three -pin, 0.1 inch

pitch single row PCB header, and solder it in the

holes provided for plug 16.5. Mount a 3.5 mm jack -type audio socket with a

breakcontact at the rear side of the computer,and wire P16, P15, and the internal loudspeaker asshown in Fig. 1.6. Reassemble the computer and test the new audio

output by connecting an external amplifier set tothe jack socket. Insertion of the jack plug shouldsilence the internal loudspeaker.Now that we are on the subject of the BBC com-puter, it is just as well to give a few hints concerningreduction of the total power consumption of thecomputer. The Type 6522 VIA chips may be re-placed with their new CMOS equivalents 65C22 toreduce the total current consumption by some240mA. The 6850 chip may also be replaced witha 6350, but this is a riskier matter since the formerchip is soldered direct onto the PCB.

070 JOYSTICK ADAPTOR

Some popular computer games require the joystickto be turned 45° in order to get the correct cursormovement on the screen. Obviously, this presentsproblems if the joystick is desk mounted or of thetype that is ergonomically styled and hand-held.The electronic solution to this inconvenience startsfrom a redefinition of the joystick axes, as shown inFig. 2. Direction A is defined as in between thepositive X and Y axes; direction D as in betweenthe negative X and positive Y axes. Directions Cand B are opposite to A and B respectively. Table 1summarizes the old and new direction assignmentsand associated activated outputs.The circuit diagram of the adaptor circuit - Fig. 2- shows that the output levels to the computer areactive low rather than high as in the unmodifiedjoystick connection; this necessitates the use of in-verter gates between adaptor and computer input.A Type 74LSO4 hex inverter may be used to thisend, and the trigger (fire) function(s) can also be in-verted at the same time, since this IC contains sixinverters.The double trigger function enables the turnedjoystick to be connected to MSX types of computeras well. Table 2 lists the relevant connections forboth the C64 and the MSX computer type.The adaptor input and output signals may be

visualized with red and green LEDs, clearly in-dicating the electronic signal turn over 45°. Whenthe joystick is moved into direction A, for instance,input LED + Y lights, as well as output LEDs + Yand +X. Current consumption of the adaptor cir-cuit is about 75 mA.

Table 1

Direction Contact

A +X and +YB +X and -YC -X and -YD -X and +Y

Direction Contact

A and B --> +XB and C -YC and D -XD and A +Y

Table 2

CBM64 MSX

(11 +Y (1) +Y(2) -Y (2) -Y(3) -X (3) -X(4) +X (4) +X(5) - (5) +5V(6) trigger (6) trigger 1(7) +5V (7) trigger 2(8) ground (8) output(9) - (9) ground

96

5V 2

A

13

C

E

O

01 DO

Vo.V4'O

green

O

O

0

O

1)44°4(1

0 0 0 0as oe

red

86445.1

Web.6o To 8. 90

N1...N4 = 74LS02165...N10 = 74LSO4N11...N14 = 74LS38N15...81l8 = 74LS05

071 LEVEL ADAPTOR FOR ANALOGUE JOYSTICK

An analogue joystick usually contains two poten-tiometers, whose wipers are controlled from thecentral handle on the unit. Unfortunately, the anglecovered by the handle is generally only about 90°,whereas the potentiometer's spindle and wiper canbe rotated over 270°. The voltage range provided bya potentiometer in a joystick is, therefore, relativelysmall. Two of the circuits described here make itpossible to enlarge the output voltage range of bothpotentiometers in the joystick. The circuit is readilydoubled, thanks to the use of dual CMOS oper-ational amplifier Type TLC272.Each of the two wiper voltages from the joystick isprocessed separately, which enables interesting ef-fects to be achieved. The amplification of the circuitis determined by P3. This preset enables the enlarg-ing of the potentiometer's "range" to individual re-quirements. Preset P2 serves to shift the operativerange of the potentiometer within the limits of thesupply voltage, which may lie between 3 and 16 V.Setting up this circuit is straightforward. Com-mence with setting P3 for minimal resistance, i.e.,Al should give unity gain. Set the joystick handleto its centre position, so that the wiper of Pi is atmid -travel. Adjust P2 to make the output voltage ofthe circuit equal to V2Vdd. Move the joystick

Al = 1/2 IC1 = TLC 272

(TOP VIEW)

OUT

IN-W+GND

VDD

OUT

IN-IN+

handle to the outer positions in the relevant plane,and note the corresponding output voltages fromthe circuit. Adjust P3 such that the circuit outputsthe required voltage span. The adjustment of P2enables changing the toggle point of the circuit,

97

that is, the voltage it outputs when the joystickhandle is set to its centre (rest) position.The current consumption of the circuit depends onthe supply voltage level, and also on the value of PiWhen Vdd = 5 V, and P1= 4K7, the current drain is

less than 10 mA. The Type TLC272 was chosenbecause it works fine from a single supply voltage,and also because it has an extensive input voltagerange, 0 to Vdd-1.5 V.

072 LIS'T'EN -IN KEY FOR DATA RECORDERS

The pros and cons of using data (cassette) recordersfor mass memory storage in a computer system arelikely to be so well-known that any further dis-cussion as to the relative cost efficiency of thecassette tape would seem to be superfluous.There is, however, one distinct disadvantage to thedata recorder that is relatively easy to get rid of, viz.the trouble many users experience in positioningthe tape to the leader note of the desired programor file to load into the computer. Manydatarecorders, while offering the highest possiblesave and load speed, fail to produce the sound ontape when the computer audio cable is plugged intothe earphone socket, forcing the user to plug andunplug this cable in a desperate search for theprogram.The solution to this sorry plight consists of a simplecombination of resistor and push to make button,which are to be built into the cassette recorder. Thecircuit diagram shows the method of connectingthese parts; pressing the button with the earphoneplug inserted in the socket will enable the user tolisten to the recorded data as the tape is played. The

value of the resistor may have to be adapted to suitthe specific output power of the data recorder, giventhe optimum playback level for the computer.Now that you have opened the recorder for theoutlined modification, it is just as well to mount asecond button enabling tapes to be wound andplayed while the remote control plug rests insertedin the associated socket; this simple modificationmay also be of appreciable interest for the improvedefficiency in locating files on tape.

073 MAINS INTERFACE

This circuit is of use, for instance, when a computeris required to monitor a mains -operated equipment.Opto-isolator TIL111 ensures complete isolation be-tween the mains and the computer.With the mains on, during every positive half -wavea current of about 1 mA flows through the LED inthe opto-isolator. The associated transistor thenconducts and its collector current of about 100 µAis sufficient to drive T1. Remember, however, thatthis is a pulsating current: capacitor C1 ensures thatTi conducts continuously as long as the mains ison. If a 50 Hz square wave is required at the collec-tor of T1, C, should, of course, be omitted.The two 100 k resistors in series with the LED

should not be replaced by one 220 k resistor,because the maximum permissible voltage dropacross a standard '/4 W resistor should not exceed150 Vrms.

98

074 MANDELBROT GRAPHICS

:,,...,,

r 4.--4,,41

,....t.'

as

rs'

864551

The computer -based implementation of certainiterative types of calculation may offer highly at-tractive graphics screen representations, as we got toknow when keying in a program to crunch a fewnumbers in the Mandelbrot series, and found thatdoing so with the support of the computer'sgraphics facilities took us through a regular graphicsadventure. On further investigation, it was foundthat the degree of complexity of the resultantgraphics image is in direct proportion with thenumber of iterative steps the control program is ar-ranged to perform. However, since the necessarycalculations to obtain a Mandelbrot series becomethe more complex, and therefore time consuming,as the computer crunches through its approxi-mations and evaluations, it should not strike theprogrammer as odd that obtaining a nicely detailedgraphics image may take as long as 12 to 24 hours,even with the fastest types of personal or semi-professional types of computer, such as the BBCequipped with a second processor.The Mandelbrot series of numbers is basically ob-tained with the use of complex numbers, in a calcu-lation that converges rather than diverges the in-termediary results according to the equationZ=Z2+C, where C is the complex number con-.stant having a real part between -2 and 1, whilethe imaginary part ranges between -1.5i and 1.51;Z is the result of the preceding calculation.Stepping through a section of the series is possible

by assigning start values and/or differently dimen-sioned step rates to either the real or the imaginarypart of C. It goes without saying that calculationtime and image resolution increase with thenumber of iterations used for obtaining results inaccordance with the set requirements; the calcula-tions may be stopped when the result is larger than2. The colour assigned to any pixel on the screendepends on the number of iterative steps required tosatisfy the Mandelbrot equation; if this is not thecase, the iteration loop is consequently aborted.The program shown in Listing 1 has been writtenfor the Electron or BBC computer, and arranges for15 iterative steps; the screendump of Figure 1 showsthe result. Figure 2 illustrates how a section of thegraphics image is enlarged by means of relevantredefinition of the equation variables, as outlinedabove. Obviously, the suggested program allows agood deal of further patching and experimenting toarrive at even more attractively styled graphicsdesigns, but it should be pointed out that producingFig. 2 took our BBC no less than . . . 2 days!

99


Listing 1.

>L.102030405060708090

100110120130140150160170180190200210220230240250260270280290300310320330340350360370

REM MANDELBROTMODE 1REM MAXIMUM X AND Y PICTURE COORDINATESMAX%=200:REM MAX%<700VDU23,1,0;0;0;0;;REM CURSOR OFFVDU19,2,2,0,0,0: REM GREEN FOR YELLOWREM DEFINE DISPLAY WINDOW AT CENTRE OF SCREENVDU24,640-MAX%/2;512-MAX%/2;640.MAX%/2;512+MAX%/2;VDU29,640-MAX%/2;512-MAX%/2;REM DEFINE TEXT DISPLAY AT BOTTOM OF SCREENVDU28,0,31,39,28REM DEFINE ANGLE AT BOTTOM LEFT. ANGLE=AngleR+AngleIiAngleR=-2: Angle1=-1.25REM LENGTH OF SIDE IN COMPLEX SURFACESide=2.5REM DISTANCE BETWEEN TWO POINTS IN COMPLEX SURFACEDistance=Side/MAX%T=TIMEREM CALCULATIONFOR Y%=0 TO MAX% STEP 4

FOR X%=0 TO MAX% STEP 4REM C=CR+CliCR=X%wDistance+AngleR: CI=Y%wDistance+AnglelREM Z=ZR+Z11. Start value for Z equals CZR=CR: ZI=CIIteration%=0REM Z=Z-2+C where Z-2=ZR-2-Z1-2,(2wZR.Z1)1REPEAT

A=ZR-2: B=ZI-2: Length=SOR(AwB): ZI=2*ZR*ZI+CI: ZR=A-B+CRIteration%=Iteration%wl

UNTIL Length>2 OR Iteration%>15GCOLO, Iteration%MOD4PLOT69,X%,7%

NEXTCLSPRINT"TIME"(TIME-T)/100" S"

NEXT

075 MORSE TRAINING WITH THE JUNIORCOMPUTER

Here is yet another small program to be added tothe large amount of software already available forthe Junior Computer. It is intended to teach pro-spective short wave listeners to read morse code.The program can be used even with the basic ver-sion of the JC. The only additional hardware is theamplifier stage shown in the accompanying figure.The input to this is taken from port line PB5.The number and speed of the morse characters canbe predetermined. After the program has started,the JC will generate 1 to 6 morse characters, whichthe trainee should decode and write down. The let-ters corresponding to the generated characters ap-pear on the display after a short delay, so that thetrainee can check his decoding with the actual text.During this phase, the computer is on stand by untilan arbitrary key, other than ST and RST, is pressed.The hex dump given is sufficient to write theprogram into the JC. Once that has been done, youcan prepare the start, but the program needs the fol-lowing information before it can run.

in address 0010 write data 00 . . . 05; in address 0011 write data 01. . .55 (max); in address 0014 write data from table 1 for the

first character to be generated minus 1; in address 0015 write data from table 1 for the

last character to be generated.

100

Table 1. Table 2.

alphanumericcharacter

hexadecimalcode

alphanumericcharacter

hexadecimalcode

A 01 S 13B 02 T 14

C 03 U 15D 04 V 16

E 05 W 17F 06 X 18

G 07 Y 19H 08 Z 1A

I 09 1 1B

J OA 2 1C

K 0B 3 1D

L 0C 4 1E

M OD 5 1F

N OE 6 20

0 OF 7 21

P 10 8 22

Q 11 9 23

R 12 0 24

address function026F alphanumeric display routine028D tone generation routine02A8 random number routine02CB display code table02EF morse code table0000 to0005 display buffer0010 number of letters0011 length of dots and dashes

(speed)0014 lower limit of block of

characters to be generated0015 upper limit of block of

characters to be generated

0 1 2 3 4 5 6 7 8 9 ABCD E F

0200: A9 FF 8D 83 1A 8D 81 1A 85 01 85 02 85 03 85 040210: 85 05 18 A5 11 65 11 65 11 85 12 A5 10 AA 20 A80220: 02 A8 B9 CB 02 95 00 B9 EF 02 85 21 29 07 85 200230: 06 21 BO 07 A5 11 85 13 4C 3F 02 A5 12 85 13 200240: 8D 02 C6 20 DO EA A5 12 85 13 C6 40 DO FC C6 130250: DO F8 CA 10 C9 20 6F 02 20 AC 1D F0 F8 20 6F 020260: 20 AC 1D FO F0 20 6F 02 C6 40 DO F9 4C 00 02 8A0270: 48 A9 FF 8D 81 1A 8D 83 1A A2 08 A5 04 20 E3 1D0280: CE 7C 02 10 F6 A9 05 8D 7C 02 68 AA 60 A9 FF 8D0290: 83 1A EE 82 1A DO FB C6 13 DO F7 A5 11 85 13 C602A0: 40 DO FC C6 13 DO F8 60 8A 48 38 A5 E9 65 EC 6502B0: ED 85 E8 A2 04 B5 E8 95 E9 CA 10 F9 C5 15 BO EA02C0: C5 14 30 E6 85 30 68 AA A5 30 60 08 03 27 21 0602D0: 0E 42 09 7A 72 OA 47 48 2B 23 0C 18 2F 52 07 6302E0: 41 01 36 11 64 79 24 30 19 12 02 78 00 10 40 4202F0: 84 A4 83 01 24 C3 04 02 74 A3 44 C2 82 E3 64 D40300: 43 03 81 23 14 63 94 B4 C4 7D 3D 1D OD 05 85 C50310: E5 F5 FD

Now, the program can be run; it starts in address0020 when key GO is pressed. Programmingexample: the JC is to generate morse characters forthe letters B to G. Before the start, the folowingdata should be written: in address 0010 - data 05 in address 0011 - data 55

in address 0014 - data 02 in address 0015 - data 07As soon as these data have been written, theprogram starts when key GO is pressed.The hex data for the letters of the alphabet andnumbers 0...9 are given in table 1. The most im-portant addresses are given in table 2.

101

076 PIA FOR ELECTRON

Despite its neat design and relatively low cost, theAcorn Electron computer suffers from an unfortu-nate lack of I/O support, which is remarkable, con-sidering the fact that it is a relatively simple matterto add, say, two I/O ports to enable the computer todrive a printer, plotter, modem, or other peripheralsby means of the proposed PIA (peripheral interfaceadapter).

5V

A3

202 Vcc

PO

4- P1 016

P 2 02

8

A4 11

A5 13

P3 in Q3

P4 N 04

P5 05

A6 15P6 06

A7_17,_ 07P =0 GND

A8

A9

A10

All 8

412

Al3

414

Al

019

IC1 = 74LS688IC2 = 74LS688IC3 = 6821 PIAN1,N2 = 1/2 IC4 = 74LS00

The circuit diagram of the PIA -based extensionshows that address decoding over the full 64 Kbytesis by means of two 8 -bit magnitude comparatorsType 74LS688. Address selection is manual withswitches S, . .S14, which provide a logic low levelwhen closed; observe this when writing out theones and zeros to arrive at the desired address in theI/O map. The PIA chip is enabled when the preset

BC547B

10

20

DO 33 Vcc 2DO PAO

D 1 32 3D1 PA1

31D 2-4"D2

30D3

D 4 29D4

D

2 7 D6D6

D 26D7

RW 21

02 25E

D3

RES--o034 RES

38I RQA

37

AO 3 6

Al 3

22

4PA2

5P A3 4

6PA4'

7PA5

8PA 6 4-'1'

9PA7

10R/W P BO 4

11

P B 1

112PB213

P B3 "4---1"14

B PB415

PB 5 4P'16

RSO P B6

P B7 117---.RS140

CSO CA1

24 39*----CS1 CA2 4

CB1

CS2 CB2Vss

18

1

0

00

96491 1

102

address matches that on the computer's addressbus; writing simple I/O drivers is there-fore mainly a matter of assigning the relevant ad-dress block to control words and PIA I/O data.Ti has been included to enable the PIA circuit togenerate and forward interrupt request pulses by

means of the wired -OR arrangement for this con-trol line.In case it is desirable to switch heavier loads thanis normally permissable with the PIA outputs, it issuggested to employ power drivers/inve, iers such asthose in the ULN2000 series.

077 QL RAM EXTENSION

1.

0 V

C

5V

RESET L

37 36 9F.40

ASL

A00225Al 024bA20 2"A3 0 2"A4 0 2"A5022'AS 02"A70 2"A8 01"A901"

A10 0171All 01"Al2 01"A1301"A14 0 1"Al5 0 "bAIR 01

CLKCPU O 11a

7.5 MN.

DSL 0 66

34

DSMCL 0 27.

FS0 FS1

185ALE

ACR11

16

17

2326

2930

RAO

RA1

RA2

RA3RA4

RA5RA6

35 RA712 CAO15 CAI16 CA222 CA3

25 CA428 CA531 CA6

CA7

3 REN 1 S; ti CS

C

2 38120

IC2 IC3

FTCVN 0ITAM D6

REFRECI 016

MAO 13MA1 14

MA2 11IC1

24

MA5 ILTMS MA6 32

4500 A MA7 33

CAS

IC)

D

FF1

o

51

5V

60

IC4 A10

OTACKL Bb C "1" 74

11 LS156

12 OAT

3- 5V

D-

10

7

R2

2N2905

5V

2

A3

A4 4164A5

13 A6

IC6CAS

FF1 = %IC2 = 74LS74N1 %IC3 = 74LS32

65

14

I 1C13

5V

O EN0

1 105 A

*see text

416

26 GROWL

85483

All

A18

103

Sinclair's QL has as standard a 128 K RAM, whichsounds like a lot in comparison with most 64 Kmachines. Unfortunately, the software writers, inthe knowledge that there is more than enoughmemory, have been rather wasteful in their work, sothat at the end of the day, there is not all that muchmore in the QL than in the 64 K machines. So, youneed more memory . .

The accompanying circuit is an application of theTMS4500A as RAM extension for the 68008. Thischip can drive a maximum of 128 K dynamic RAMand provides virtually everything: multiplexing ofthe address lines, RAS, CAS, and REFRESH.The memory ICs are 64 K x 1 (128 or 256 refreshare both permitted) and have a speed of better than150 ns. Since the QL uses a clock frequency of7.5 MHz rather than the normal 8 MHz, such aRAM can run without wait cycles. An 8 MHz CPUthat regularly has to carry out a wait cycle is ap-preciably slower than a 7.5 MHz type!The 68000 family is provided with a dataacknowledge input. As with other processors, theCPU places addresses and data onto the bus and in-dicates the validity with an address strobe and datastrobe respectively. It continues to do so until thememory sends a DTACK signal. The present exten-sion generates this signal with the aid of the LS156.

Normally, this acknowledgment is given almost im-mediately, but it may happen that the 4500 is in themiddle of a refresh. In that case, the CPU has towait, which is arranged via the ready output (pin 2).To prevent the QL waiting forever when an addressis read that has no memory, the DTACK isgenerated internally: this must, however, be disabl-ed for addresses where the RAM extension is

located, and fortunately this can be done easily viaDSMC. By making this logic high as quickly aspossible, the internal DTACK is cancelled.If you cannot get the 2N2905 transistor, you mayuse a BS250, in which case resistor R can be omit-ted and R2 should be replaced by a wire link.The circuit as shown is for the 128 K version. It isalso possible to omit the eight RAMs connected toRAS1 and make a 64 K extension. Input A of theLS138 must then be connected to A16 and pin 11 in-stead of pin 13 must be used as CS.There is no 5 V supply available on the connector,but there is a 9 V line. This can be reduced to 5 Vby a standard 7805. The current drawn depends onthe types of RAM and will be 200 . . . 300 mA. It isimportant to decouple the supply lines properly:each RAM IC and the 4500 require a 100 n capaci-tor!

078 RAM EXTENSION FOR QUANTUM LEAP

The Sinclair Quantum Leap (QL) computer iseminently suitable for a low-cost introduction intoworking with Motorola's 68000 true 16 -bit micro-processor. Many computer enthusiasts did not failto note the spectacular price cuts for the QL whenits production was discontinued. An excellent sup-port program, TOOLKIT II, became available andis still considered indispensable by many for gettingto grips with the QL. The present 512 Kbyte RAMextension should be very welcome for running aRAM disk, and/or programs such as ICE andQIMP.The circuit is based around the Type THCT4502RAM controller from Texas Instruments. Thisdedicated controller takes care of all the DRAMcontrolling, including the refresh timing, and theaddressline multiplexing. The address decoder ismade with a single XOR gate, N7. The DSMCLline is made high within 30 ns with the aid of three -state buffer N5. Bistable FFi delays the ASL signalsomewhat, so that DTACL is only activated whenthe RDY output of IC: is stable. The databus is

buffered by bidirectional octal transceiver IC23.The extension memory is divided in two banks of256 Kbyte. Note that CAS, unlike WKS-, is commonto both banks. It is possible on the QL to omit thesecond bank without altering the address decoding.This is thanks to QDOS, which searches for cor-rectly operating continuous, and unique, i.e., non -mirrored, memory. It is interesting to note thatmachine code in the extension memory runs atalmost double the normal speed.The RAM chips used should have an access time of150 ns or less. Current consumption of the exten-sion is low at 50 mA or 150 mA in the non -activeand active mode respectively. Non -used inputs ongates should be tied to ground.Finally, note that the Type THCT4502 controllermay not be available everywhere yet.

Distributor for TI Semiconductors in the UK is DCDistribution Freepost Hitchin Road Arles-ly Bedfordshire SG156BR. Telephone: (0462)834444 or (0454) 273333.

104

BANK II

RESET L 0 120

ASL 05b

220

Al 0240

A2 0 259A3 0 244

A4023'A5 02"A6 021'A7029'A90 16a

Ala 017aAll 01"Al2 015aA13 0 "aA14 O 134A15 0110A16 0194

A8 0A17 0A180CLKCPU 0

7.5 MHz

N1...N3 = 3/41C2 = HCT 32FF1 = 1/21C3 = HCT 74N4...N6 = 3/41C4 = HCT 125N7...N8 = 1/2105 = HCT 86

5V0

29 28 1F*35FS0 FS1

tftALE

3 RAO8 RAl9 RA2

RA3RA4

RA5RA6

RA7

CAO

CAICA2

CA3

CA4

CA5CA6

CA7

REN 1RA8

4 ANROCLK c I- mm

18

21

2227

4

7

10

14

17

20

2326

9a8a

IC1THCT4502

ACIN 0RASO

REFREQ

MAO

MA1

MA2MA3

MA4 18MA5 19

24

5,4,7 25

MA8 345.50CASI

RAS2

RAS3

8451

CS

346

11

13

ASL 05b

OSL6b

41 421 30 12 V3 6CLK SET

3839

40

7

1.4

NI ,NO

44

15V

D F FIRES 0-0

0

OTACKL 0 8b

91)°---N3 810

8

818

Al9 2

DSMCL 0

RWDL

27a

N6 9

IC14IC RAS

5 AO7 Al

11

10

1!

p15 CAN

15o

7

10

13/ 49-

O

AO

AlA2

A3A4 41256A5

A6A7A8

A-479

5V

IC21

BANK I

IC13

IC23245

DO

O 1

D2

D3

Dd

135

O 067

SIR

DO

DI

02

D3

04

DS

D6

07EN

9

6

CS

R/W

5

32a

Cl

11665

la

0 V 14

IC227805

C3(3 14 14 8 ()

C2 C2014141 1C2 1C3 1C4 IC5 1C23

?MT

87514

O

5V

079 RS232 INTERFACE

This circuit is intended as an interface between theElektor modem (Elektor Electronics, October 1984)and a computer. The software for each individualcomputer must, of course, be written separately.Since the writing of a terminal program can only becarried out in machine language, the interface canbe kept quite simple.

Signals at TTL level are sufficient to operate themodem and LSO5 buffers are therefore used. Com-plete address decoding of the 6551 is ensured by 1C2

and IC3 so that only four locations in the memoryare required, and these should be available on virtu-ally any computer. The fourteen common addressbits are selected with Si .S14: a closed switch

105

5 V 0

TIM0201600AO 0Al 0

711115T 0

DO 001 002003 0.005 0DO 0D7 0

0 V 0

5V

1.8432 MHz

15 O

28 05028 R,27

36C r8513 ASO14 Ral

2155

18

21

22 D4

23 DO

IC1

6551

J- .13U

TX0

9

2

RS232

TXD 8

CT5

P ,

OW1 6 --0,151 -4-

65C51 650 6 4ROD t2 tt<4/.111'7D

Rxc 5

5 V

WIWI]P=0015 0 P7 07014 0 75 P6 1C2 06 =-40130 13 14

0120 11 12

A11 0 p3 74 53 9010 0 6 P2 LS 02A9 0 4 P1 688 01 5AB 0

5V

:52IC3 IC5

Cl .C3=3.100

A7 0A6 0A5 0A4 o03 0A2 0

15

13

11

6

SO

190P=0

1807wi IC3 06Ps 05 76P4 04 73F.3 74 03 9P2 LS 02 7P1 653 Gt 5po 00-3+

IN1 N3 = h IC4 = 74LS05N4 N7 = IC5 = MC 1489

TT T5V

8.10k

59 514

0.105

65408.1

sa

respresents an address bit or 0. The input buffers arestandard RS232 line receivers so that they can copewith any voltage levels that may be present on anRS232.The interface is also suitable for connecting a serialprinter to a computer, provided it can operate fromTTL levels, which normally is the case.The accompanying tables show some of thepossibilities of the 6551 and are intended as an aidin the writing of the terminal program.

Register Select Coding

RS1 RSO Write Read

0 0 Transmit Data Receiver DataRegister Register

0 1 Programmed Status RegisterReset (Data is"Don't Care)

1 0 Command Register

1 1 Control Register

Note that only the Command and Control Registers can beaccessed during both Read and Write operations. ProgrammedReset operation does not cause data transfer, but is used toclear (reset) all G65SC51 internal registers. Programmed Resetis used in a slightly different way as compared to the hardwareReset (RES). These differences are described under each indi-vidual register description.

STOP BITS

CONTROL REGISTER

7 6 4 3 2

0 - 1 Stop Bit1 2 Slop Bits

1 Stop Bit if Word Length8 BITS and Parity', Stop Bits if Word Length5 Bits and No Parity

WORD LENGTH

BIT DATA WORD6 5 LENGTH

0 0 8

0 1 7

1 0 6

1 1 5

RECEIVER CLOCK SOURCE

0 External Receiver Clock1 = Baud Rate Generator

'This allows for 9 -bit transmission(8 data bits plus panty)

HARDWARE RESET

PROGRAM RESET

6 S 4

BAUD RATEGENERATOR

0 0 0 0 16x EXTERNAL CLOCK

0 0 0 50 BAUD

0 0 1 0 75

0 0 1 1 10892

0 1 0 0 134.58

0 1 0 1 150

0 1 1 0 300

0 1 1 1 600

1 0 0 0 1200

1 0 0 1 1800

1 0 1 0 2400

1 0 1 1 3600

1 1 0 0 4800

1 0 1 7200

1 1 1 0 9600

1 1 1 1 19,200

3 2

0 0 0 0 0 0

Control Register Format

106

COMMAND REGISTER

PARITY CHECK CONTROLS NumBIT

OPERATION7 6 5

0 Parity Disabled-No Parity BitGenerated-No Parity Bit Received

0 0 1 Odd Parity Receiver and Transmitter

0 1 I Even Parity Receiver andTransmitter

1 0 1 Mark Parity Bit Transmitted.Parity Check Disabled

1 1 Space Parity Bit Transmitted,Parity Check Disabled

NORMAL/ECHO MODEFOR RECEIVER

0 Normal1 Echo (Bits 2 and 3

must be 0-1

HARDWARE RESET

PROGRAM RESET

7 6 5 4 3

DATA TERMINAL READY

0 = Disable Receiver and AllInterrupts (DTR high)

1 = Enable Receiver and AllInterrupts (15-TTI low)

RECEIVER INTERRUPT ENABLE

0 = IRO Interrupt Enabled trom Bit 3of Status Register

1 = Interrupt Disabled

TRANSMITTER CONTROLS

BIT TRANSMIT RTS TRANSMITTER3 2 INTERRUPT LEVEL

0 0 Disabled High Off13 1 Enabled low On

1 0 Disabled Low On

1 1 Disabled Low Transmit BRK'

2 0

0 0 0 0 0 0 0 0

- - - 0 0 0 0 0

Command Register Format

7 4 3

LSET BY CLEARED BY

Parity Error' 7 : Errorvrom Self Clearing"

= ttoErorFraming Error' O Self Clearing"

0 = No ErrorOverrun'1 = Error Self -Clearing"

Receive Data 0 = Not Full " Read ReceiveRegister Full 1 = Full Data Register

Transmit Data 0 = Not Empty Write TransmitRegister Empty 1 = Empty Data Register

0 = 1::.Q Lowl

DCD1 = DCD High

RliAeRceissegge

State

0 = DSR LowNotl Resent:Rile

DSR1 = 6-ST:1 High Stale0 = No Interrupt ReadIRO1 = Interrupt Status Register

so INTERRUPT GENERATED FOR DAM CONDITIONSCLEARED AUTOMADCALLY AFTER A REPO OF RORANIDTHE NEXT ERROR -FREE RECEWT OF DATA

11SRDWARERESET

PRPORASIRESET

7 6 5 4 2 I 0

Status RegisterFormat

080 SAMPLE & HOLD FOR ANALOGUE SIGNALS

Conventional analogue sample and hold circuits arenotorious for their tendency to drift, a phenomenonunknown in digital memories. It is, therefore, in-teresting to study the use of a digital memory el-ement for storing an analogue signal.The present circuit is based on intermediate storageof digitized analogue information, and therefore re-quires an analogue -to -digital converter (ADC) at theinput, and a digital -to -analogue converter (DAC) atthe output. Unfortunately, DACs and ADCs aretypically expensive components, and the presentcircuit is therefore set up with a DAC only, driven

by an up/down counter-see Fig. 1. The counter isessentially an ADC, since the output voltage of theR -2R based DAC is continuously compared to theinput voltage with the aid of a window comparator.The error signal produced by the comparator ar-ranges for the counter to count up or down, de-pending on the magnitude of the difference be-tween the input and output voltage. The up/downcounter is corrected until the input and outputvoltage are equal. The digitized result of the A -Dconversion is available at the counter outputs.The extensions for converting the basic set-up into

107

1

hold

windowcomparator

E

OSC

U/D-counter

Uref

D/AR -2R

2 40mA-Ie.

T1

BC557B

OR6

(TTL)hold (Th

OAl , A2 = IC2 = LM393NI , N2 = 1/21C3 = 74LS00

I = digitalII = analogue

a sample & hold circuit are relatively simple. Thecurrent count is retained by activating theHOLD input, which enables halting the U/Dcounter. Evidently, the counter state is not subjectto drift, so that the analogue output signal isavailable unaffected for as long as the circuit is pow -

R3

13

C2 C1

70p 70n

A = Orel = 2.55V

87511-1

C3mim

87511-2

100n

0 MSB00

ooI

000 LSB

0

O

ered. The converter used here is the Type ZN435ADC/DAC from Ferranti. This chip containseverything shown in the dashed box of Fig. 1. Withreference to the practical circuit diagram, Fig. 2, theinternal voltage reference and the oscillator are ad-justed with R1 -C1 and R2 -C2 respectively. The lat-

108

ter are dimensioned for 400 kHz, i.e., nearly themaximum oscillator operating frequency. The inter-nal counter is controlled via inputs up , down andmode. The logic level applied to the mode inputdetermines whether the counter continues or haltsupon reaching state 0 or the maximum value, 255.In the present application, the counter is halted.Gates Ni and N2 are added to enable blocking theU/D counter. Opamps A, -A2 form the windowcomparator. Current source Ti -R7 and R6 arrangefor the toggle threshold of Al to be 20 mV higherthan that of A2. This off -set creates the window, or

inactive span, needed to suppress oscillation of thecounter's LS bit, and to prevent unwanted effectsarising from the comparators' offset voltages.Decoupling capacitor C3 is fitted for suppressingspikes that occur during state changes on thecounter outputs. The conversion time of this designis about 640 pis, as determined by the oscillator fre-quency (400 kHz), the resolution (8 bits) and the in-put voltage change (2.55 Vpp max.). This cor-responds to a slew rate of 4 mV/µs at the input. Fi-nally, bear in mind that the output impedance (IC1,pin 11) is relatively high at about 4 kg.

081 SERIAL DATA CONVERTER

1

N5 N6

2

EIMI 30110.

RxD

D

4V7

MIR 0

Cl

4709

N8

3

D3

1N41485V

0FF1

El 0

2

4

T a 6 jus

RST

IF -C2

680P

1N4148

Tc 6 ps

IC2a, IC2b = 4520FF1, FF2 = IC3 = 4013N1...N4 = IC4 = 4093N5...N10 = IC5 = 40106N11 = 1/4 1C7 = 75188/1488

120

CK

C51I,mim

680p

11 14

11

CY .9" CLK

IC14017

CLK EN

15

12

7

3

FF2

CLK a

10

5V

0CK

64r 20OD E CLK

p IC2a

CLK

N QI P

5 4 6 14

OE OF OG OH

IC674HCT4060

12

11 10

".-M71-*

C3* -1 Ca

X1

4.9152MHz

33P 33p

11

Tx

00

I

5V

C d.rF

B 136

NA

-c04

IC2b

CLK

L -M = 9600Bd

L -N = 4800Bd

L.O = 2400Bd

L -P = 1200Bd

87516 1

109

Some computers and communication programs areunable to output serial data composed of 7 data bitsand a parity bit. The present circuit has been de-signed to output this data format when it is drivenwith serial data organized as 1 start bit, 8 data bits,no parity bit, and 1 stop bit. This format is widelyused for accessing bulletin boards, data banks, andthe like with the aid of a modem, and should beavailable on most computers equipped with anRS232 port. The converter has a built-in clock gen-erator which can be set to the baud rates shown thecircuit diagram, Fig. 1. Both odd and even paritycan be generated, and no handshaking is requiredwith the computer or console.The basic operation of the converter is as follows(also refer to the timing diagram in Fig. 2). The ris-ing edge of the start bit in the incoming 10 -bit wordclocks bistable FF-I, whose output Q goes low andso enables counters IC1, IC2a and IC2b, which werepreviously blocked by the high level of RST. Binarycounter IC2a starts counting the clock pulses pro-vided by baud rate generator IC6. The frequency ofthis clock signal is 16 times the bit rate on the serialinput and output line. Bistable FF2 and counter ICIare clocked with signal CK, whose period cor-responds to that of the bits in the data stream. Thereceived start bit and the next seven data bits arepassed through FF2, while IC, keeps count of thenumber of transmitted bits, and actuates output 9during the reception of the ninth bit (i.e., databit 7).

The rising edge of the counter output pulse is dif-ferentiated in C5 -R6 and then applied to NANDgates N1 -N2. These make it possible for FF2 to beset or reset, depending on the state of parity counterIC2b, which keeps count of the logic high bits inthe serial word applied to the converter. Its outputQA indicates whether the number of detected highbits is odd (QA =1) or even (QA = 0), and causesFF2 to toggle when the differentiated pulse fromIC, makes the output of N8 or N6 go high for a veryshort period. When QA is low, the parity bit at Qof FF2 is high because in that case the S (set) inputis driven high. Similarly, the parity bit is low whenQA is high because the R (reset) input on FF2 isthen driven high. These two situations can occurwhen even parity is selected by fitting wire linksA -D and B -C as shown in the circuit diagram. Oddparity is obtained by fitting links A -C and B -D, andpermanently low parity by fitting C -E and D -F(note that a "low" parity level means that the rel-evant bit is logic high in the RS232 convention).After transmission of the parity bit, the circuit isprepared for the next word by the carry (CY) outputof IC, providing a high level to differentiator C2 -R4.

This resets FF1, which in response drives the RSTline high to reset the counters.The convention adopted for the logic high and lowlevels of the data bits in the proposed converter re-quires that this is inserted in an RS232 or RS432data line. Line driver Nil may be omitted, and theserial output signal taken from Q on FF2, if thedriven input can operate with pulse levels of 0 and+5 V.Finally, Fig. 3 shows a suitable alternative for thecrystal -operated clock generator, which may be con-sidered too extensive if the circuit is to work at afixed baudrate of 1200. Multiturn preset Pi is setfor an output frequency of 19,200 Hz.

20

RAI

CA

AIM

Q J FF2

CU, IC2b

OA 1626

5/.FF2

R/221

AR

35V

87516-3

110

082 SERIAL LINE DRIVER AND RECEIVER

This circuit owes its existence to the need for datacommunication over relatively long distances (up to100 metres), inexpensively, reliably, and suitable forspeeds up to 2400 bauds. At the distances con-sidered, the main expense is normally the cable, sohere a readily available 60 Q coaxial cable is used.Because of its relative immunity to noise, currentdrive is employed.In the line driver - figure 1 - transistor T1, di-ode D4, and resistors R3 and R4 form a currentsource that can be fed direct from a non -regulatedsupply of 8 . . .10 V. The transistor should bemounted on a heat sink. The current level of 40 mAensures an adequate input signal to the line receiver.Transistor T2 is a current switch that short-circuitsthe current source and the cable to earth of the in-put to the driver is logic high: only when that inputis logic low, is the current of 40 mA fed into thecable. Diodes D2 and D3 protect the driver against

1

_FIT(TTL)L

BC516

C 5 B0

TZ

noise emanating from the cable, while capacitor Cidecouples the supply line.The line receiver is based on a type LM 311 com-parator. Matching of the input is effected by a wirelink at a relevant tap of resistive divider R5-R6-R7/R8 (in our case: 60 Q). Resistors R9 and Rio, anddiode D5 protect the LM 311 against noiseemanating from the cable. The sensitivity of the re-ceiver is set with Pt Resistor R14 provides somehysteresis. Pull-up resistor R15 ensures that IC,provides at its pin 7 a TTL output signal that is inphase with the input signal to the line driver.The circuit is best calibrated with the aid of an os-cilloscope once it has been installed in its final pos-ition. The level of input to the receiver is then com-pared with the voltage at the wiper of Pt The set-ting of Pi is optimum when the voltage at its wiper(wave form A in figure 3) is exactly opposing the in-put voltage (wave form B in figure 3).

5V

J-LJLITTL)

85443-3

111

083 SIDEWAY RAM FOR BBC AND ELECTRONPLUS ONE

As already reported on numerous occasions inElektor Electronics, the BBC micro ranges amongthe most widely used types of personal computercurrently available. To newcomers in the computerfield, the amount of commercially available ROM -supplied software is truly staggering, and thereseem to be programs to suit almost any requirementand budget.However, the number of ROMs that may be locatedin the BBC computer is limited to four in the basicversion and sixteen when it is equipped with asideway ROM expansion card. Users in posession ofa good many ROMs and EPROMs are, therefore,often forced to exchange these before a program canbe run; a method that is both cumbersome andpossibly bad for the ICs and their sockets.

26

14

IC3

*see text

NWDS

413VCC

27128

VCCAO

AlA2

A3A4

45A6

A7

A8

SOCKET AiA9

o

All412

DO

D1

D2

D3

D4

D5

D6

D7

A way of getting round this problem is to installRAM rather than ROM or EPROM chips on thesideway board, so that software may be readilymoved about between ROMs, direct access memory,disk and RAM, since many of the originally ROM -based programs may also be run from RAM, it hasappeared.Since it was thought convenient to plug 16 Kbytesof static RAM into any one vacant ROM socket,the circuit was constructed in all-SMD technologyon a ready-made PCB of very small size.The circuit diagram shows two 8 Kbyte, low -powerstatic RAMs Type 6264FP-15 as a replacement fora 16 Kbyte EPROM Type 27128; a single inverterselects the relevant 8 Kbyte block when the (for-merly) ROM socket is addressed.

Ni = 1/6 IC3 = HCTO4/LSO4

A21C

28 28

10 10

9 9

8

7 7

6 6

5 5

4 4

3 3

25 2524 2421 21

23 232 2

11 11

12 12

13 13

15 15

16 16

17 17

18 18

19 19

27

OE

GND

CS

22 22

14 14

20

CS2

VCCAO

AlA2

A3

A4A5

A6

A7

A8

A9

A10

AllAl2DO

D1

D2

D3

D4

D5

D6

D7

WE

OE

GND

IC2*6264

CS1

20

9

26C52

28 VCC10 AO

AlA2

A3

A4

A5

A6

A7

A8 IC1

A9 6264A10

AllAl2DO

D

D2

D3

D4

05

D6

D7

WE

OE

9

8

7

6

5

4

3

25

24

21

232

11

12

13

15

16

17

18

19

27

22

14 GND

CS1

20

86452

112

Working with SMD parts to achieve a trulyminiature ROM replacement should be based onthe necessary skills in soldering and handling thesenew parts, and the construction of the proposed ex-tension therefore requires to be done as follows.It should be noted that the through -plated PCB forthis project comes together with the SMD die.Fit 28 short (1 cm) pins at the sides of the PCB toenable it to be received in an IC socket.

The SMD RAMs are mounted piggy -back onto thePCB, with the exception of pin 26 of the topmounted RAM; this terminal should be wired tosocket pin 28. The SMD parts 74HC04 (IC3) andR I may now be fitted to conclude the PCB con-struction. Once the unit has been plugged into aROM socket, a short wire is run from pin 8 of IC77on the BBC main board to the NWDS input on theSMD board.

rps.as B6452 gm

1

Parts list

Ri =100 kICi;IC2=6264FP-15IC3=74HCO4PCB 86425Miniature switch for write protection, if

required

Finally, although not mentioned so far, the ElectronPlus One computer may also benefit from the pro-posed sideway RAM circuit which, as will be readi-ly understood, need not necessarily be constructedwith SMD parts; a veroboard and normal sizedcomponents, along with a bit of wiring, will also doin many cases, although it may be hard to surpassthe elegance of the plug-in unit.

084 SIMPLE D -A CONVERTER

Two simple to build 4 -bit digital -to -analogue con-verters are described here. One translates a 4 -bitBCD code into 10 analogue voltage levels, the otheraccepts a 4 -bit binary code and outputs 16 voltagelevels. Both circuits comprise a digital decoder withopen collector outputs for controlling a resistanceladder. The analogue voltage is obtained by con-trolled connection to ground of a particular sectionof the ladder, and buffering the drop so obtainedwith a transistor.Notwithstanding their relatively low resolution (10or 16 steps), the circuit should have many possibleapplications, including driving digitally controlledpower supplies, triangular wave and sawtoothgenerators, and A -D converters.Table 1 lists the relative values of the resistors in the

ladder network, starting from R, =1KO. Threevalues are given for each resistor: the left-hand col-umn shows the theoretical value, while the nearestequivalent from the E24 and E96 series appears inthe centre and right hand column, respectively.Note that the starting value can be changed to in-dividual requirements, provided all other resistorsare dimensioned accordingly, i.e., their valuesshould be multiplied with the same factor withrespect to 1KO.It is a relatively simple matter to add an 11th or17th output level by driving the decoder such thatnone of its output transistors is enabled. This resultsin an output voltage which is 0.6 V lower than thesupply for the ladder network. In the case of the74LS145, this condition is obtained by applying a

113


non -valid code to the inputs, i.e., one greater than910 (10012). Similarly, on the 74159, enable input GIor G2 can be made logic high.

.34mA

24

Ucc 015

014

01

012

1C274159

01

010

09

2A

02

2

B

C0

2

06

0

0

0

0

0

1-0 G1

196 02 .1.

D117

5V

5 15V

IAl

D216

0315

4D4

I

D531

D611

I

D710

089

D98 11

D107

D116

5D12

I

O 134

0143

O 15

O 161

R2

R3

R4

R5

R6

R7

R8

R

R10

R11

R12

R13

R14

R15

A

71

BC550 C

Uo

R17

Dl...016=1N4148

Table 1 Resistor values relative to 1 10

Fir,10 -step

BCD version6 -step

binary version

RI 1000 1K0 1K0 1000 1K0 1K0

R2 111 110 110 66.7 68 66.5

R3 139 130 140 76.3 75 76.8

R4 179 180 178 87 91 86.6

85 238 240 237 103 100 102

Rs 333 330 332 122 120 121

R7 500 510 499 145 150 147

Rs 833 820 825 178 180 178

Rs 1667 1K6 1K69 222 220 221

Rio 5000 5K1 4K99 286 270 287

Ril 381 390 383

R12 533 510 536

R13 800 820 806

R14 1333 1K3 1K33

R15 2667 2K7 2K67

R16 8000 8K2 8K06

D1...0 10= 1N4148

114 87418

085 SIMPLE VIDEO INVERTER FOR ZX81

The inverter must be connected before the TVmodulator in the ZX81. Switch S, enables bypass-ing of the inverter when inversion of the picture isnot required. The composite video signal is invertedby gate Ni. Gates N2 and N3 separate the syncsignal from the input: the sync signal is thenavailable at the output of N3 at a level of 5 Vpp.The inverted video signal and amplified sync signalare then added again, resulting in an inverted videosignal with the sync signal in the correct positionand at the right level. Preset 131 serves to adjust thecontrast.The circuit can be constructed on a piece ofveroboard so small that it can easily be added in theZX81 case. The power supply can be taken fromIC, in the ZX81: +5 V at pin 40 and earth (0 V)at pin 34.

UK1

USA 2Fr2 USA 3 UK2

D9

ONLY ONLY

011. yid.°

CO. inverse video & sync.

O. inverse video out

OtsULA Sla

IPRnrU.6

+5V

OV

USAR31 ONLYI OV

FrciNutFr ONLY

R32' 09'Fr ONLY

R31'

OV

MI C

EAR

R34

47nF

OV

C11 1KR27 R29

C12

R334K7

10nF

OV

I47pF

1

12

RAMC.S.

34

16 TV/TAPE

N1 ... N3 = = 74LS00

1 1 1 1 1 1 1

6 V 7 15 9 81014

CERAM 35OSC

I C 1

33 Kt3D 0 SINCLAIR

31 K BD 1 COMPUTER

29 KBD2 LOGIC

27 KBD3

25 KBD 4DB 3201 30D2 28D3 2604 24D5 23D6 21D 7 19

ROM LS..13

-22 1811 393837 36

013.18-1

R301.USAONLY

OV

ICI

115

086 SYNC INVERTER FOR THE QL

5V-OV-

N1 = Y41C1 = 74HC00

For some unknown reason, the Sinclair QL (andperhaps some other personal computers) providespositive, instead of the usual negative, fieldsynchronizing pulses to the monitor. Inverting thesepulses with a suitably fast NAND gate or inverteris, of course, no problem. What is a problem iswhere to power this gate from: a special supplywould be nonsense. However, in the circuit propos-ed here, the gate is supplied from the sync signalitself. A monitor with TTL input for the sync signaldraws only a very small current at logic 1, so thatthe additional load presented to the input pulse bythe diode and electrolytic capacitor is inconsequen-tial.Instead of the HC-MOS gate shown, it is also poss-ible to use a buffered CMOS gate, for instance, atype HEF4011B. Standard CMOS devices, such asthe 4011, cause a very small delay, which in practicedoes not matter, and certainly not with a field sync

I

85412

- 4V- ov

signal. Note that it is important, as always withCMOS devices, to connect unused pins to earth(pin 7) or to Ub (pin 14).

087 SYNCHRONIZATION SEPARATOR

Many monitor chassis currently offered by com-puter surplus stores have separate inputs forhorizontal and vertical synchronization signals.Most home micros, however, have a compositevideo output, so that some form of interfacing is re-quired to drive these bargain monitors.The Type TBA950-2 is a sync separator chip whichis frequently encountered on TV chassis. In its stan-dard application circuit, it requires to be driven by

a flyback signal derived from the output of the linefrequency oscillator. Without this signal, which isapplied to pin 10, the sync pulse would end upsomewhere among the picture lines. To be able touse the TBA950-2 in the present application, thehorizontal pulse is slightly shifted with the aid of adouble monostable multivibrator, IC2.The operation of the circuit should be clear fromthe accompanying timing diagram. The output

116

12V

10

IC 1TBA950-2

.+11

C

10p 6V

RS

C5mom

470n

13 I14

C6

041

MKP Pt4k7

12 11 10

1P20k C

50p

* see text

4TR1

R12

C10

330p6 14 15

IC 2

4538

01 -TR

t!i" r

pulse from the TBA950 is fairly wide (26 pis), and itspositive edge triggers the first MMV (Q1), whosenegative output pulse transition in turn triggers thesecond MMV in the 4538 package. The line syncpulse for the monitor is available positive and nega-tive at IC2 outputs Q2 and Q2, respectively.Adjust the circuit as follows: set P2 to the centre ofits travel, and adjust the frequency control, Pi , suchthat the image is stable. Next, position the image byadjusting P3. If the correct position can not be ob-tained, the phase control, P2, must be carefullyreadjusted, followed by P3. The vertical sync pulseis available at pin 7 of the TBA950-2. Finally, thedashed resistors and diodes are required if the moni-tor inputs are designed to accept signals with apeak -to -peak amplitude of 5 V.

VIDEO

IC1-2

MMV-01

MMV-02

13

02

02

ill NA ye

D3-r-

Rt*

R1S *

874144

SV

-044-.4ps

87414-2

088 TWIN KEYBOARD FOR APPLE II

The keyboard supplied with computers is for manyapplications not the ne plus ultra it is claimed to be.Unfortunately, deficiencies normally do not becomeapparent until the machine has been in practical usefor a while. Retailers have long since realized this

and often stock improved keyboards that are fullycompatible with the computer in question. It is,however, not always clear how the new keyboardcan be attached to the computer. One possibility is,of course, to open the computer, remove the existing

117

RESET 0 12 V

C>

OIL plug DomApple keyboard

STROBE A

0

12 V

O

5V

AD0 BE DO. DR

O

RESET

5V

STROBE

N1

5V

0 OW IC2is K 74LS76

CL

RE

STROBE B

N2

0

14 Select

N3

N1 ... N3 = 1C1 = 74LSO4DO ... D6

IA

2A 2Y

3A 3y

4A 45

SA 5Y

67

9A 7Y

RA BY

MUX

es

1B

38

48

50

60

78

16.106 00 Apple motherboard

keyboard, install the new keyboard, and put thecomputer together again. It is, however, much bet-ter to use the solution suggested here, which isaimed at the Apple II and compatible machines.The accompanying circuit makes it possible to con-nect the additional keyboard in parallel with theexisting one. Basically, it is just an electronic switch -over unit, designated MUX in the diagram.Both keyboards are connected to the input of MUXby their data lines. Which keyboard data are appliedto the computer is from now on determined byMUX.When a key is struck, the keyboard does not onlygenerate data bits, but also a strobe pulse. Depend-ing on whether the strobe pulse emanates from theoriginal or from the additional keyboard, the Ty out-put (pin 14) of bistable IC2 is set or reset. This pulse,therefore, serves as a select signal for the MUX. Theelectronic switch consists of two type 74LS157 ICs.Each of these ICs contains four 2 -to -1 multiplexers,so that all eight input data are available at the out-put. If the select input of both ICs is logic 0, out-puts IY . 8Y contain the data present at inputs

STROBE 51_

0 0 01C1 f C2 1C3

c?

1A

18

2A

2B

3A

3B

4A

48

Select

5A

5B

6A

6B

7A

7B

8A

8B

IC4

Q 5 V

Cl C4=r=MI

BV

085478

MUX

3

6

11

10

14

13

1A

1B

2A 1Y

213 2Y

3A 3Y

3B 4Y

4A

4BSE L

4

7

IC374LS157

1Y

9

12

1

1

5

11

12

14

13

1A

1B

2A

2B

3A

3B

4A

4B

SE L

1Y

2Y

3Y

4Y

4

7

9

2Y

3Y

4Y

IC474LS157

5Y

12

15

6Y

7Y

8Y

118

IA . . . 8A. If, however, the input to the ICs islogic 1, the data from 1B . . 8B is available at1Y . . . 8Y.

The Apple II requires a positive strobe pulse, and in-

verters N2 and N3 are, therefore, provided to ensurethat this condition is met whatever the strobe pulsefrom the additional keyboard.

089 TWO -FREQUENCY CLOCK

Many computer systems use one clock signal, fromwhich all other timing signals are derived. The fre-quency of the clock signal determines, amongothers, the maximum number of characters per linethe video controller can display on the monitorscreen. This is normally 32 or 40. If more charactersper line are required, the clock frequency has to beincreased. The clock generator described heremakes it possible to switch between frequencieswhich are related in a ratio of 2:3. The switching iscarried out synchronously, so that no bits are lost.The clock oscillator, Ti, is controlled by an inex-pensive 3rd overtone 27 MHz crystal, XL,. The LCcircuit connected to the collector of T, is tuned to54 MHz. The 54 MHz signal is converted to logicbits by field-effect transistor T2 which are then ap-plied to the Q inputs of dual J -K bistable IC,( = FF1/FF2). The ring counter formed by thesebistables can be changed over by T3.When T3 is on, the J input of FF, is logic high, andthe 54 MHz signal is divided by 2. When T3 is off,the J input of FF, is connected to the TY output of

R2 R3

0µH47 T : r:270po 22n

C1

-1112p

BF 451

R1X1Rom

27 MHz3 rd overtone

FF2 and the 54 MHz signal is then divided by 3.The output frequency can thus be switchedsynchronously between 18 MHz and 27 MHz.If a fundamental crystal is used in the XL, position,the oscillator can be modified as shown inset.

5V< 50 mA

00

BC 557BR7 +5 V:2/3fxi

o v: fx,

FF1, FF2 = 1C1 =745112

85458O

119

090 6 -WAY CHANNEL SELECTOR

A

a)

1<amA

09

b)

D2iN

4148

IC1/Q... IC1/0...

l<100mA

?1BC 557 B

Re

This design proves that a latching 6 -way channelselector with debounced switch inputs need notalways be based on the use of special integrated cir-cuits.When none of the break -type SPDT push buttonsis pressed, the data inputs of ICi are held at + 5 V,while input CLK is held low via R8. When a switchis operated, the associated input of IC, goes low,while CLK goes high, so that the logic state of theDo -Ds lines is latched and transferred to outputsQo-Q5. Each of these can drive a LED or relaybased output circuit as shown.When more than six switches are required, a74LS174 may be added, whose clock input is con-nected to IC,.

3

4

6

-212

VO

0

11

130

Bztk

4

I9 1:11

01

K CLD00

IC 174

LSD3 174 03

oa

015 X

02 7 X

X

I

OS 15

16

ICI

5V

Note that the LS chip may be replaced by a corre-sponding version from the HC or HCT family. Thiswill reduce the current consumption from about20 mA to 6 mA. The maximum output current sup-plied by IC, is 8 mA in all cases.

091 ANALOGUE & DIGITAL

Leafing through some electronics magazines pub-lished over the past few years, it is surprising howfast and vigorous digital techniques have come tothe fore. Even audio, until recently virtually un-touched, is now becoming digitalized at a rapidpace. What are the consequences of these changesto us engineers, technicians, and hobbyists alike?As long as a circuit is totally analogue or totallydigital, all is well. But as soon as these two tech-niques become mixed strange things sometimeshappen. Well-known examples are analogue -to -digital converters that will not give a stable reading:the last few digits do not match and it appears as ifthere is a certain regularity in the deviations.Another example is an otherwise good amplifierthat generates whistles in perfect rhythm with thedigital clock oscillator. And so on . . .

Often, these flaws can be traced to faulty earth con-nections, i.e. the zero supply line, or common

analogue

analogue GND

A/D

digital GND

logic circuits

+15V

15V

ov

ov

+5V

86436-1

120

ground. Because of that, here are a few tips thatmay prevent these annoying defects. Avoid earth loops. Keep the analogue and the digital earths

separated.III Interconnect the analogue and digital earths at

one point only, for instance, at the analogue -to -digital converter, but NOT at the power supply. If there are more earths, connect these to the

same common point.II At high frequencies, the impedances of earth

lines are not negligible: short, thick wires should,therefore, be used.An example that gives good results is shown in the

accompanying drawing. All sensitive parts of thecircuits have been isolated from those parts thatcarry (large) earth currents. Most converters have,therefore, two earth terminals, or an earth terminaland a differential input (which is the same thing).In audio amplifiers most of us do not dream of wir-ing the power supply to the output amplifier via thepreamplifier. In mixed analogue -digital circuits,such considerations are not so self-evident,although the principle is the same.Note that in the accompanying drawing the systemneeds several electrically isolated power supplies:that is unfortunately the price often to be paid fornew techniques.

092 BAND -GAP VOLTAGE REFERENCE

It is generally known that the accuracy ofmeasurements in electronic circuits is mainly afunction of the stability and reliability of the refer-ence against which the unknown quantity is com-pared. Therefore, everything feasible should be doneto maintain the stability of the reference, i.e.,

counteract the adverse effects of variations in theambient temperature, supply voltage, and load cur-rent. The zenerdiode in Fig. I is a usable referencedevice for applications where the above threeparameters are not subject to appreciable variation.The "super zener" in Fig. 2 features excellent stab-ility and is hardly affected by variations in thesupply voltage and the load current. Although thetemperature coefficient of the super zener circuitcan be optimized by careful dimensioning of thecomponents, there exists a still better way for mak-ing a precision voltage reference.The term band gap refers to the difference betweentwo discrete energies of the outer four electrons ina semiconductor atom. Electrons in the highestenergy band contribute to the conduction of thematerial. As the temperature is increased, someelectrons gain enough thermal energy to escapefrom the valence (non-conductive) band, cross theband gap, and enter the conduction band, leavingthe valence band unfilled. Thus, conductivity is afunction of temperature.With reference to Fig. 3, the temperature coeffi-cient of current mirror T, -T2 is compensated bythat of T3. The following conditions should be metif the circuit is to function optimally: (1): R2 10R,;(2): R3 is dimensioned such that VR -= 1.204 V; and(3): the transistors are exactly matched. The latter

2 Ub

1

121

condition is probably best satisfied by using tran-sistors on one and the same chip carrier, e.g. thosein a transistor array such as the Type CA3083. Thevalue of R depends on the supply voltage and themaximum output current. It should be noted thatT3 carries the output current if the circuit is notloaded, so that the resulting dissipation may giverise to temperature differences on the chip. It is,

therefore, recommended to permanently load theband -gap reference. The accompanying calculationsprove that the output voltage of the circuit is not af-fected by temperature variations.

Band -gap reference.

The reference voltage, UR, is obtained fromUR = UBEIT3) + 12 T2.

IR.1 and R2 are dimensioned such that h = 1012, so that R3 drops I UBEIT11- UBEIT2II volts.When the current amplification of T2 is sufficiently high, R3 carries virtually all current12:

12 = UBE(T 1) - UBE(T2)/R3 whence

UR = UBEIT 3) + (UBEIT1 I - UBE(T 2)) R2 /R 3 .

For identical transistors UBE is given for different values of IBE asUBEIT 1) - UBEIT 2 ) = k T/q10ge(11 /12 )

UBE of T3 is also expressed asUBEIT 3) = UBG( 1 - T/To) + UBEO( T/To)

so that UR can be written asUR = URG( 1 - T/To) + UREol T/To) + R2/R3k T/qloge(11/12).

Differentiating this to the temperature domain yieldsdUR/d T= -UBolT 0 +UBEolT 0 + R2/133k/qloge(11/12)if R2, R3 and 11 are dimensioned such thatR2/R3loge(11/12)= (URG-UsEo1T31)Cwhere C=q/kTowhich results indUR/dT= 0 (QED).

k= Boltzman's constant (1.3805 x 10-23 J/K).T= absolute temperature [K].q= charge of an electron (1.6021 x 101° C).UBG = band -gap potential (1.204 V) .

UBE0 = base -emitter voltage at T= To.e= the base of natural logarithms (2.71828).

122

093 BUZZER DRIVER

Piezoelectric resonators, also referred to as buzzers,are frequently used for providing audible signals inall sorts of electronic equipment. Buzzers are small,light, simple to use, and yet provide a loud outputsignal. They are either of the passive or of the activetype. The former are driven by an AF signal source,while the latter feature a built-in oscillator, and re-quire a direct voltage only.This circuit is a double AF oscillator for drivingpassive buzzers. It ensures a richer output soundthan normally obtainable from a piezo buzzer dueto the use of two oscillators, Ni and N2, whose out-put signal lies between 1 and 10 kHz. Gates N3 -N4form an S -R bistable which is controlled by the out-puts of N1 -N2, and drives the buzzer direct. Thespectral composition of the output signal is fairlycomplex, due to the presence of both the fundamen-tal notes and the difference and sum frequency. Thetimbre so obtained varies as a function of the ratiobetween the oscillator frequencies, which are ad-justable with the aid of presets Pt -P2. Note thatdiodes D1 -D2 reduce the duty factor of the oscillatorsignals to about 25%. Optimum effects are achievedwhen a simple ratio is set between the oscillator fre-quencies, e.g. 3:4. The resulting waveform is alwayscomposed of rectangular signals, but these differ in

"ron

14

IC1

®5V

eon

N1...N4 = IC1= 74NCT132

87456

respect of their period to ensure that the buzzer pro-duces a rather agreeable sound.The buzzer driver is controlled by a logic level ap-plied to point X. The quiescent current consump-tion is virtually negligible, while about 10 mA isdrawn in the actuated state.

094 COMBINING DIGITAL CIRCUITS

Many electronics hobbyists combine all sorts ofdigital circuits into works to be marvelled at. How-ever, even they sometimes have that uncertain feel-ing: must they all be powered by one unit or shouldthere be more or can there be more? And in whatsequence should they be switched on? Printer first,or computer first?In digital engineering, which by definition embracescomputers, inputs are driven by outputs: infor-mation is being transferred. When the IC thatdrives has a power supply, but the receiving one hasnot, a current will ensue, whether the circuits areTTL or CMOS. This is an undesirable situation,although it does not normally lead to damage. Butthe ensuing current may be so large that the IC pro-viding the current does not operate efficiently anymore, because its output voltage, owing to the large

current, becomes too low. Particularly bistables canbecome disorganized by this. It is, therefore, poss-ible that a certain equipment does not work prop-erly because another circuit connected to it does nothave a power supply.That situation can become really critical whenseveral outputs of an IC are terminated in that man-ner. Normally, an IC can withstand a short at oneof its outputs, but if that happens at several out-puts, the IC will probably give up the ghost. Thismay happen, for instance, in the case of a Cen-tronics interface, of which the eight data lines arenormally driven by one IC.And what happens to the IC that is provided withthe current? CMOS circuits are generally well pro-tected against this, and TTL devices normally standup well to them also. But other types may not take

123

1

HCMOS

0 V

85420-1

so kindly to these currents.Semiconductor manufacturers have, of course, alsobeen confronted with these problems and havefound solutions to them. Anyone designing andbuilding his own circuits should, therefore, heedtheir experiences and observe the following rules.

2

II Driver ICs, whether TTL or CMOS, musthave an open -collector output. All inputs should be provided with additionalresistance (pull-up resistors) to the positive supplyline.If these rules are adhered to, current can only flowfrom input to output (see figure 2). This does notmatter, because the collector of transistor Ti canstand quite a high voltage and nothing will,therefore, go wrong. Make sure that the pull-up re-sistor is connected at the input side, otherwise it hasno effect.As to the question at the beginning: it does not mat-ter which unit is switched on first, because the ICmanufacturers have made sure that the input andoutput circuits are protected.

095 CURRENT DRIVE FOR STEPPER MOTORS

Stepper motors have either unipolar or bipolarstators. In unipolar models, each stator winding hasa centre tap, which enables the magnetic field to beinverted by switching from one to the other half ofthe winding. Bipolar types have a single statorwinding, so that the direction of the currentthrough it must be changed to attain inversion ofthe magnetic field. From this, it is clear that, giventhat the two motors are of similar size, the bipolartype will provide a larger couple than the unipolar

model. There is, however, a price to be paid for thislarger couple: the drive of a bipolar motor is morecomplex than that of a unipolar type.The drive for bipolar motors may, in principle, beobtained by means of a full bridge circuit. i.e. four transistors per stator

winding; half bridge circuit and dual power supply, i.e. two

transistors per stator winding; half bridge circuit with large output capacitor.

124

1

C1,C2 >..,2000pF/A

N1,N2=1/31C1=7407

R1,112

33n / 0.5W18o / 1 W608 / 2 W3n3 / 4 W

I100mA200mA500mA

1 A

The last method is totally unsuitable for low stepp-ing frequencies or stand -still. Of the other two, thehalf bridge is to be preferred in most cases, in spite 2of the requirement for a dual power supply. In thiscontext, it should be noted that the supply need notbe regulated, since constancy of current is

guaranteed by a zener diode and emitter resistor,even with variable input voltage. The value of thesmoothing capacitors in the power supply is deter-mined by the total stator current, and is a minimumof 2000 µF/A.Values of Ri and R2 are given for various values ofstator current in the table below.

R, & R2 Is

33 52;LI W 100 mA18 52;1 W 200 mA60;2 W 500 mA3S-23;4 W 1A

Current drive ensures a higher pull -in rate,i.e. permissible starting frequency, becausecommutation is quicker with an inductivestator winding.The higher the supply voltage, the more ef-fective the drive, but also, unfortunately,the dissipation in Ti and T2. In practice, a2 x 12 V or 2 x 18 V mains transformer hasproved very satisfactory. Note thatfreewheeling diodes have been included inthe darlington circuit to give a goodmeasure of protection against high inducedvoltages caused by switching.

10...30V

BD679

10...30VO

86517.1

86517-2

The prototype was used in the first in-stance for the control of four -phase steppermotors via an eight -bit output port of a mi-croprocessor system. The interface used toobtain TTL levels was a Type 7407 whichhas 30 V open -collector outputs. The con-trol instructions may be generated astrot instructions may be generated as follows:

Phase 1 2 3 4Bit 7 6 5 4 3 2 1 0Output byte 1 0 1 0 1 0 1 0 initial positionAuxiliary byte 0 0 0 0 0 0 1 1 XOR with output byteNew OtP byte 1 0 1 0 1 0 0 1 made one stepRotate aux. bytetwice*

0 0 0 0 1 1 0 0 preset for next step

*Direction of turning determines rotational directionof motor.

125

If the stepper motor is required to be used on its TEA1012. The latter is dealt with in Circuit 119own, this may be done with the aid of commercially (p. 146) and may be connected as shown in Fig. 2.available control ICs such as the SAA1027 or the

096 DC OPERATED 50 HZ TIMEBASE

Many clocks, both of the digital and the analoguetype, make use of a 50 Hz timebase signal which isusually derived from the mains. In order that theseclocks may also work in places where there is nomains supply available, as in cars, on boats, or, say,on a camping site, this one -chip circuit provides anaccurate 50 Hz square wave output signal, while be-ing fed off any DC supply voltage between 6 and15 V (battery, solar cell array, etc.). Current con-sumption of the circuit is only 3 mA (max.).The Type SAF0300 by ITT Semiconductors merelyrequires a crystal to perform the above task, whilealso offering the possibility to adjust the exact out-put frequency by means of seven active low bits aslisted in the pin assignment table.If a 64 Hz output frequency is desired rather than50 Hz, the crystal may be replaced with a4.194812 MHz type.Finally, the 50 (64) Hz output pulse has a voltageswing of nearly the IC supply voltage, and a dutyfactor of 0.5.

6...15V

86498-1

1 Output 1 (50Hz)2 Adjustment pin 122 ppm3 Adjustment pin 61 ppm4 Adjustment pin 30.5 ppm5 Adjustment pin 15 ppm6 Adjustment pin 7.6 ppm7 Adjustment pin 3.8 ppm8 Adjustment pin 1.9 ppm9 Test pin M (ix/4)10 Cristal connection11 Cristal connection12 Bridge output13 Bridge output14 Ground, 015 Leave vacant!16 Supply voltage

50Hz

097 DECOUPLING IN LOGIC CIRCUITS

Failing to heed the importance of adequatelydecoupled supply rails is one of the most seriousmistakes a constructor of digital circuits can make.Two important facts necessitate a reappraisal of theeffectiveness of decoupling: the introduction of thefast HC and HCT series of CMOS chips, and thegeneral availability of ever larger dynamic RAM(DRAM) devices. The 41256 256Kbit DRAM and6264 CMOS SRAM, for instance, have becomecommonly used integrated circuits, available at rela-tively low cost. The fast spreading use of the newCMOS series of logic circuits has created the widelyheard misunderstanding that these devices can beused without paying the least attention to decoup-ling of the supply lines. However, a reduced currentconsumption relative to TTL devices is by no meansa carte blanche for designers to skimp on decoup-

1

87440 -1

126

ling provisions, as will be seen below.Why does a logic circuit draw current? The currentconsumption of TTL chips goes mainly on accountof indispensable, internal, resistors. CMOS struc-tures are complementary, and theoretically con-sume no current at all in the static mode. As soonas any kind of switching is to be done, both by TTLand CMOS circuits, the charge of the capacitanceat the output must be reversed as illustrated inFig. 1. The switch currents internal to the IC areonly a fraction of those required for the loadcapacitance, and can, therefore, be disregarded, ex-cept in the case of counters.TTL and CMOS circuits thus consume an equalpeak current during switch operations. Decouplingcapacitors are fitted direct to the IC supply ter-minals to prevent the instantaneous supply voltagefrom briefly dropping to an unacceptable levelwhen the switching takes place. The graph in Fig. 2is reproduced from a Texas Instruments databook,and shows the correlation between thecapacitor -to -package distance and the peak ampli-tude of the spikes on the supply line to a typicalHCMOS gate. This shows beyond doubt thatdecoupling capacitors must be fitted as close aspossible to the IC supply terminals, to rule out thestray inductance of supply tracks on the PCB, how-ever neatly these may run in parallel. Often, tunedcircuits are designed with long supply tracks and awrongly placed decoupling capacitor. Any spike isthen subject to ringing effects, which further de-teriorate the operation of the logic circuit in ques-tion. Not surprisingly, Mullard recommend a multi -path supply track when it is impossible to fit thedecoupling capacitor close to the IC. This solutionis called a grid structure, and is definitely preferableto creating relatively wide, single tracks-see Fig. 3.The value of the decoupling capacitor must bebased on the foreseeable number of IC outputs thatare simultaneously active. A conventional startingpoint is 20 to 100 nanofarad for every three ICs.Further reflection on this theme leads to the con-clusion that the supply for a 256Kbit DRAM is farmore difficult to decouple than that for, say, a16 Kbit DRAM. Fortunately, the problems are notas serious as one would expect. In practice, the sizeof the chip carrier, and hence the parasitic capaci-tance, is constantly reduced by the manufacturers,whose foremost aim is to ensure optimum responseof the device at high operating frequencies. CertainDRAM manufacturers recommend the use of 330ndecoupling capacitors (see Fig. 4), but in practice noproblems evolved from the use of the standard valueof 100n.

2

1.0

0.9

OA

0.7

0.6

0.5

0.4

0.3

0.2

0.10 15 30 45 60 7.5 90

Distance From Packages

INCHES

TYPICAL POWER SUPPLY DECOUPLING

VCC = 5 V

C = 0.01 of- TA = 25°C

87440 -2

03

4 130 0.068 old

120A1 pfd

116100 0.061uld

90pfd

° 022 old-.0.33 old70

60

50.05 0.20 0.35 0.50 0.65

0.22 pfd

0.33 yid0.47 pfd

87440-3

256-K DRAM1.0 aid

1.0 pfd

64-K DRAM

CAPACITANCE-1dd

0.80 0.95

89440-4

127

098 DEGLIT'CHER

Extremely short, unwanted, pulses with a period inthe nanosecond range are often referred to as glit-ches, and occur in most, if not all, digital circuits.Whilst the circuit in question can be designed andbuilt with due attention paid to effective suppres-sion of glitches, it is not always possible to foreseethe effects of external noise on, for instance, a clocksignal. The filter presented here effectively rules outthe presence of glitches in a serial data link.Assuming that counter ICI is at state nought, andthat the data input is logic high, IC2 is configuredas an AND gate. Output Q4 of ICI, and hence theoutput of the deglitcher, goes high after 8 clockpulses. A short negative pulse at the data inputmerely results in a few more clock pulses being re-quired before Q4 is activated. After another 8 clockpulses, the counter state is 15. This causes the CI(CARRY IN) input of ICI to be driven high, sothat the clock signal remains blocked as long as thedata input is logic high. When it goes low, IC2 isconfigured as a NOR gate, enabling the clock tran-sitions to be counted down in ICI. Output Q4 goeslow again after 8 clock pulses, and the counter isblocked after another 8 pulses. Therefore, thefiltered output data is delayed by 8 clock periods,

IDATA

but this is insignificant in the proposed application.The data frequency, $p], depends on the clock fre-quency, flu.):

f[D] = f[CL]/16

The maximum usable clock frequency is about 8MHz. The current consumption of this circuit isless than 1 mA.

099 DESIGNING A LOW NOISE AMPLIFIER

To design a low noise amplifier, it does not sufficeto choose a low noise opamp, because the compo-nents associated with the opamp, particularlyresistors, are themselves sources of noise. The noisein a resistor, which is caused by random movementof electrons, increases by the square root of the in-crease in resistance.Figure 1 shows a very convenient characteristic fordetermining optimum values of input resistance.The y-axis gives the square of the sum total of noisevoltage produced in a circuit (in nV over the band-width considered), while the x-axis gives the valueof the source resistance.For instance, a noisy opamp like the 741, which pro-duces some 70 nV of noise over its bandwidth, cancope with an input impedance of some 200 k(higher values would cause the input impedance togenerate more noise than the opamp!). On the otherhand, the less noisy TCA 520, which generates

1

e2neq 104

nV

Vin7f

o6

co'

2

70 nV/N5:7

102 30 nV/V-AT

10

10

100 1k 10k 100k 1M 10M

R5 -ow (1)85455-1

about 30 nV of noise over its bandwidth, shouldhave an input impedance not greater than about50 k.It is not always convenient to use such relatively

128

2a

low values of resistance. For example, the audioamplifier in figure 2a is intended to operate down to0.3 Hz; because of that, the time constant, r= RC,must be fairly long. The input (= source) im-pedance of the opamp is determined primarily byRi. Lower values of this resistor would require ahigher value of C1 and this is not acceptable on costgrounds. The solution to this problem is shown infigure 2b, where both the DC and AC amplification

are the same as in la, but because I?, is 10 times assmall, its noise voltage is reduced by 1/10.

SourcesFigure 1: intuitive IC opamps(T M Frederiksen - NationalSemiconductor)

Figure 2: technical note 068(Philips)

100 DISPLAY INTENSITY CONTROL

This is a light dependent voltage source thatregulates the supply to 7 -segment displays in accor-dance with the intensity of ambient light. Theregulating action is positive, i.e., a higher ambientlight intensity results in the circuit raising thesupply voltage to the displays.Phototransistor Ti does not conduct when it detectsdarkness, and the base of T2 is therefore groundedvia R2 and Pt This causes the voltage at the emitterof this pnp darlington transistor to be about 1.2 V.The voltage across R5 is the reference potential,1.25 V of the Type LM3I7 regulator, so that Iks isabout 5.7 mA, and the output voltage, Uo, of thecircuit is

Uo =1.2 + [5.7 x 10-3(R5+ R3)]

=1.2+1.823 volt

when Ti detects darkness. When it detects a rela-tively high light intensity, the base and emittervoltage of T2 increase. When the base voltage of T2exceeds 2.7 V, R4 limits the emitter voltage to 3.9 Vdue to the constant current of 5.7 mA. T2 no longer

conducts and the output voltage of the circuit is5.7 V, because the total resistance between the regu-lator output and ground is R5 + R3 + R4 =

1,000 Q, and the current through it is still 5.7 mA.The sensitivity of the regulator is adjustable withPt The maximum output current is of the order of700 mA when IC1 is adequately cooled. The inputvoltage range of the circuit is 8 to 15 V.

129

101 DUTY FACTOR ANALYSER

Applications of this duty factor meter include ad-justing and setting up ignition systems, switchmode power supplies, PD modulators, and sensorsignal converters. The circuit itself requires no ad-justment, and has a duty factor resolution of 1%,or 1° in terms of the dwell angle. The duty factorrange is I% to 99% in the frequency range from1.5 Hz to 10 kHz. The analyser is fed from 12 Vand consumes only 50 mA, so that it can be readilyused in a car.The measuring principle is straightforward. A PLL,ICS, is used to multiply the input signal by a factor100 and to clock counter 106-1C7, whose BCD out -

12V

CD °

® 8 8C6

8c7

()ICE,C5IC2 RIM WMIC1 1C4 ICSIC3 - 1C7 =me

220n 100n 100n

®0

ea=

R22

12V

O

C4111/16V

RI

Cl

2 12014!15

7 R/C C C R C-C it

-TR TR TR =194

MMV 1 811113'-._13811113'-._13MMV 2 060 _ 9

Q O-

W 41 12 11

D1

1N4148Ak

R20

10

R21

D2

puts are applied to display' drivers 1C2 -1C3. Thecarry output of 1C7 is fed back to the phase com-parator in the PLL. The counter state is only latch-ed and displayed upon the falling edge of the inputsignal. Since the counter always counts up to 100(leading edge of the input signal); the output statethat exists upon detecting the trailing edge cor-responds to the percentage of the pulse duration inrelation to the period. Example: assuming that theduty factor of the input signal is 60%, the counteris started at state 00 on the leading edge of the inputsignal, and is at state 60 when the trailing edge com-mences, so that '60' is latched and displayed. The

14

R2

BC547B9

13

10

10

11 12

7

13

11

14

2

15

LD 2MIP11=11.

LD

C:I.

9

113101 8I14 x 4700

R9

R8 12V

0abcdef g

IC 24543

LD A B C

11 5 3 2 4

abcdeIC 34543

LD A B C

018 .

-rR16

14

13

c6 kriCext

signal IC 5

(PLL)VCOout

ViCnO 4046 phasecont of

6 11 14 2

4 1

9

R19

C3

5

10 16V

R17

11

0

3

CE

00O 0 woIC 6 064029

PL

12V

ry'L.LLO

7 15

5 3 4 12 13

9B/D

CE

5

5

6

3

11

2

14

U/D 0C7

4029

3 4 12

4

2

13

12V

0

MMV1, MMV2 = IC1 = HEF4528BN1...N4= IC4 = 4093 87448

130

SN

,

151

C4

c

eE

PS

.87448IC

EI

a.

.4as

V)

-,t)

a =À

>8

ct00

0on

..__.

,--ca

...-.0 -Lo

4v) - . -

)..,ct .0

AD

NZ

0 4) 00i..)

f..)' -4 .7

Co

Z ...2

c'*Z<

)

aia..) 5 .0 =

cu-I

iEb.°a>

4.).....

-0c§

<.....,

...-05

t).c-)

o.cn

(a'--

cu -=,i,

I-A

-.ctl

I._

a...,

0 z co 0cO

oal

excce..=

Ea. co

-0 77:,

ccn ,I.

.,..,a)

'15,'c-)

a)C

,)0 E

mi

ca' 5 E 7J

oaco

fi)T

t0)C

V0

iii2?

r,c,t

.0, cc:

cc:=

co,--'4

0' r4co 14 co

cc, Tit

o-0

0 V.

0.)5

001- 2 t3

II 2 8 11c,

..c,

...0.,0.

,,, ,_, ,t

0 .1---d-

,..-;c) uS

E=

a - cs1,

iiii

ii- ,...,

- ---P E

11(=

:,

6 6 IL 6 C

5 6 6 C5

r`N

0co -=

co-c.....', ,

c..r-o

,._.

....c,

._.

±)t7,m

' : : i :5 : ...,":;:'ctl.

::::0

cl, cr .- o - to-

0..

c ,L, 0 -

.. _ei

0(..)

..-.4.,-._,

0 .',.t'-'

',:l

E A

i :e)6=

Eg

J.-,0. 0

.

),,..! 7- ,:,,,,:.7.8 1.L

.'-'°. ti

-'-o- o2 .I'fa. 8

:4rJ z 04

....t, .

C) 0 4.)

..-, .-.0 0

-,

....-'..--

&.) 0 0 a) c .-

.)a)

--o..,

-to2 o

tsA

,as

-.....

,,,,C

.);" t Ie.?,0.,

,,,-

0.,5. - - ,,,,

E kl 9

,,_.,

a.)0. L, 0 0 -0 0 cb 0

00

:61 ) I 11a 30i:" 4

.0 >,

"0.4.=

C;3.

c9 52 L610 -8

,,)

,ct . . -

...,i.)

..0 0 .0

i.-.),....,---- - 0.,

(...)0

,,,,0, -0 ..,..9

.53!- 0,Li4 ,,, 0, f-, ...

en

Naamloos-6.indd 9 28-08-2008 10:05:45

102 ELECTRONIC ROTARY SWITCH

Sooner or later, most types of frequently used multi -way rotary switches develop contact resistance in-stability or other malfunctions, either caused by in-ternal oxidation or wear and tear of the rotarymechanism. Broadly speaking, the same goes formulti -contact relays. It is, therefore, hardly surpris-ing to encounter the electronic, free -of -wearequivalents of the above devices; n -way electronicswitches and solid-state relays are at presentavailable in a wide variety of contact arrangements.The circuit diagram shows the electronic counter-part of a 16 -way rotary switch whose pole is connec-ted to earth. Two push buttons have been providedto enable the switch to be "turned" clockwise (up)or anticlockwise (down).Debouncing bistables Ns -Ns and N7 -N8 supply astable low logic level to monostables Ni-N2 and N3 -

N4 respectively in order that these can output ap-proximately 3.5 pis long pulses to the relevant inputof up/down counter IC,. The rising edges of theup/down pulse(s) cause this IC to generate thecorresponding binary code at its QA . QD Out -

puts, which are connected direct to the D1... D4 in-puts of latching 4 -to -16 decoder IC2 which, in turn,activates the next lower or higher output So . .S15if the relevant control button was activated. Pro-vision has been made to "stop" the switch if thisreaches its first or sixteenth position, which con-ditions cause the down or up monostable respect-ively to be disabled. Other switch configurationsmay be defined by using the correct active -low out-puts to block gates N2 and N4 when the desired stoppositions are reached.Finally, push button S3 resets the counter IC andconsequently causes IC2 to activate its So output,which is also the default switch position at power -on.

S3 1

141RESET T

R7

3

100n

14

15

10

16

reset A

°BA 40193B oc

B 0DC IC1

up down .6

1 24

2 3

6

2

5 4 8

1

132

04

4514B(74HC)

IC2 15

INH121 23

=NI

11

0

15

C2

R6 470p

13

N1...N4=IC3 = 4011N5...N8 = IC4 = 4011

R2

R4

86428

5...15V

132

103 FAST OPTO-COUPLER

The opto-coupler in the normal common emittercircuit at the output of a phototransistor is in-variably too slow for use in data communication.Its great advantage remains, of course, the excellentisolation between transmitter and receiver.To retain the advantage, the phototransistor hasbeen integrated into a cascode circuit, as shown infigure 1. The photograph illustrates data transfer ina conventional circuit (top) and in the cascode cir-cuit - the fast opto-coupler - (bottom) at a fre- u,quency of about 30 kHz.The cascode circuit's faster operation is due to thetransistor's internal Miller capacitance being of noconsequence as the collector voltage remains con-stant. The result is a faster transistor.The base of T2 is biased at about 1.5 V by voltagedivider R1/R2. Capacitor C1 ensures that, even withrapid fluctuations in current, this voltage remainsstable. If you consider 12 as an emitter follower, itis clear that the collector of T1 is always providedwith a constant (direct) voltage, and this causes theMiller (base -to -collector) capacitance to be inactive.A disadvantage of the fast opto-coupler is that itsoutput signal does not go down to 0 V but at bestto I V. TTL devices like this just as little as they doa supply voltage of 12 V. Basically, the circuit canoperate from 5 V, provided R, is altered suitably,but it is better to use CMOS devices.Take care during experimenting not to exceed themaximum LED current (in the TIL 111) of 100 mA(this is the reason for dropping resistor Rv). Thevalue of Rv is calculated from

Al

62

TIL 111

12 V

16 25 mA

0

Cl=I110p16 V

NEM85411-1

Rv =[(Uln-1.5)//LED]4where Uin is in volts and /LED in amperes.

104 FAST OPTO-ISOLATOR

When a computer drives external equipment, it isoften required that the earths between them areelectrically isolated from one another. The simplestway of effecting this is by an isolating transformer.When, however, the system works at high fre-quencies, it is much better to use an opto-isolator asproposed here because that is capable of followingthe fast data transfer.The opto-isolator is driven via a TTL gate. The tran-sistor in the opto-isolator drives comparator IC,.The trigger threshold of this device is set with PtLow-pass filter R2 -C1 prevents spurious triggeringof the comparator by noise pulses.

133


105 HC -BASED OSCILLATORS

Two inverters, one resistor and one capacitor are allthat is required to make a HC(T)-based oscillatorthat gives reliable operation up to about 10 MHz.This sort of circuit is well-known, and appears inFig. la.The use of two HC inverters gives fairly good sym-metry of the rectangular output signal. In the samecircuit, HCT inverters give a duty factor of about25%, rather than about 50%, since the toggle pointof an HC and an HCT inverter is 'AV., andslightly less than 2 V, respectively.When the supply voltage for the oscillator is

switched on, C initially has no charge, and the out-put of Ni and N2 are at the same logic level. Ca-pacitor C is then charged via R, until it has acquireda charge voltage that corresponds to the toggle volt-age, Us, of Ni. Assuming the output of N2 initiallyto be logic low, the waveform of the signal at the in-put of Ni is essentially as shown in Fig. 2. When Cis charged up to level 1 , the output of Ni toggles,and so does that of N2. This causes the voltage atthe input of Ni to rise, via C, to about 1.5Vcx, sothat C is reverse charged to level 3 . From there on,the amplitude changes in a mirror -inverted way toreach the initial state again (level 5 is identical to1 ), and the circuit oscillates. In practice, the curve

in Fig. 2 is slightly flatter, because the peaks atlevels 2 and 4 are clamped to + 5 V and 0 V bythe protective circuits internal to the inverters.If the oscillator is to operate above 10 MHz, the re-sistor is replaced with a small inductor, as shown inFig. lb.The output frequency of the circuit in Fig. la isgiven as about 1/1.8RC, and can be made variableby connecting a 100K preset in series with R. Thesolution adopted for the oscillator in Fig. lb is evensimpler: C is a 50 pF trimmer capacitor.

R>2k7

10pH

87458-1

87458-2

<10MHz

5...20MHz

87458-3

106 HCMOS VCO

Crafty designers are forever trying to use ICs for ap-plications they were never intended for. In this cir-cuit a member of the newish HCMOS family isused as a voltage -controlled oscillator (VCO). Thisis achieved by using the characteristic of theHCMOS family of operating from a 2 to 6 volt

supply. However, at 6 V these ICs are faster than at2 V.In the present circuit, a "supply voltage" variablebetween 1.5 and 5 V is used as the input signal ofthe oscillator, which consists of three cascadedNAND gates. The VCO operates as follows: a logic

134

1 Ni...N4 = IC1 = 74HC00N5 = 1/6 74HC04

VCO 1.5 - 5V 10mA

C)

1 at pin 2 causes a logic 0 at pin 3; this becomes a1 at pin 6, and a 0 at pin 8. Pin 8 is, however, con-nected to pin 2, which, therefore, is no longer 1 butbecomes 0. This 0, because of the delay times of thegates, appears a little later at pin 2 as a logic 1. Andso on: the oscillator works! Gate N4 functions as abuffer for the oscillator output.Since the peak output voltage cannot be greaterthan the supply voltage, i.e. the input voltage to theoscillator, its level must be adapted to those at theremainder of the circuit, which normally will be5 V. This is ensured by inverter Ns, which is pow-ered by a genuine 5 V supply. Because of feedbackresistor 121, the inverter is arranged as a linearamplifier. It is, therefore, sufficiently sensitive toamplify positive signals between 2 and 5 V ade-quately.The characteristic in Fig. 2 shows that the VCO isreasonably linear. Other output frequencies are notpossible with the circuit of Fig. 1, unless the

R1

ion7 C2

mmr

1100 n

2 3°

MHz 20

10

00

5V

C3

700nO

86434-1

TA=25°C

2 3 4

UVCO ("It)

5

number of gates in the oscillator proper is extendedby an even number of identical gates, which in-creases the total delay times, so that the frequencyis lowered. It is also possible to add dividers to theoutput circuit.

107 HCU/HCT-BASED OSCILLATOR

When frequency stability is not of prime import-ance, a simple, yet reliable, digital clock oscillatorcan be made with the aid of relatively few compo-nents.High-speed CMOS (HCU/HCT) inverters or gateswith an inverter function are eminently suitable tomake such oscillators, thanks to their low powerconsumption, good output signal definition and ex-tensive frequency range.The circuit as shown uses two inverters in a74HCT04 or 74HCU04. The basic design equa-tions are

for HCU: f =1/T; T=2.2RC; 3V<Vcc<6V;L =13 mA

-N1, N2 = 1/3 IC1 = 74HCT04, 74HCU0487437

for HCT: f =1/T; T=2.4RC; 4.5V<Vcc<5.5V;lc =2.25 mA

Rs 2R; 1K52.5..R.s.1M52; nF.With Rs and R calculated for a given frequencyand value of C, both resistors can be realized as

135

presets to enable precise setting of the output fre-quency and the duty factor. Do not forget, however,to fit small series resistors in series with the presets,in observance of the minimum values for R and R5as given in the design equations. The values quoted

for Ic are only valid if the inputs of the remaininggates are grounded.Source: Philips CMOS Designers Guide, January1986, p. 105 ff.

108 HEART BEAT MONITOR

The proposed circuit is based on the fact that thedegree of translucence of parts of a mammal's bodydepends, among others, on the flow of blood.Because the blood supply pulsates at the frequencyof the heartbeat, this may be monitored in a simpleway without the need for an electrical connectionbetween the mammal and the measuring equip-ment.

In the proposed circuit, the flow of blood througha finger is monitored. To obviate errors caused bythe position of the finger, the receiver diode is in-cluded in a loop.

The positive input (terminal 3) of IC1 is held atabout 2.5 V. The gain of the device is determined bythe ratio R5:R4. Network R6 -D2 ensures that thecircuit stabilizes rapidly. The amplified signal is rec-tified by IC2. Time constants R8 -C4 and R2 -C4 arechosen such that the potential at pin 2 of IC2 has

1

a sawtooth shape. The CA3130 in the IC3 positionfunctions as a trigger. The output signal may, for in-stance, be applied to the input port of a computer.If a computer is not available or deemed necessary,the beat is made audible by a piezo-electric buzzeroperated by gates Ni and N2.Circuit IC5 provides a WAIT indication that showswhen the circuit has stabilized and is ready for use.The programme is compiled as follows: wait for atrailing edge, then count until the next trailing edgeappears. The count is converted into a number perminute, and this is displayed on the monitor screen.However, the heart beat is not constant, which isquite clear from listening to the buzzer or observingthe monitor screen. It is, therefore, advisable tocalculate an average over, say, sixty seconds. It isthen possible to display the instantaneous value, theaverage value over 60 seconds, and the trend (rise orfall).

T1G

R4

820n

135250

R 1315

ICI

CA3130

2 g 4

C3

820n

820n

IC4

0000G

132...D5 = 1N4148

270n

89

A

270n

N1...N4 = IC4 =4093BIC5 = 4538B

2 IC3

A3130 6 4

-TR1vcc

IC5

CD4538B

RC

+TR1

R18

2

C8

1r)

7 72

= BC550CR17

U

aD6 IV%

WAIT a

N2

R14

005V

C7

81,2

1=1-r0 B7

1-186453.1 0

136

Once the programme is known to work satisfac- analogue -to -digital converter, the output signal oftorily, it becomes interesting to display the actual IC, may be used for the display.signal on the screen. If the computer used has an

2

r

3 zi 6

R1

C4 C5

4

CC

U

r Uin

T1

S 2

z

86453-2

137

109 HEAT SINK MONITOR

In almost any equipment in which a reasonableamount of energy is consumed, there is bound to beat least one heat sink that enables power semicon-ductors to get rid of their excess heat. The rating ofa heat sink is normally determined on the basis ofthe maximum allowable temperature of the siliconchip: a rather haphazard method.The heat sink monitor described here constantlymonitors the temperature of the heat sink. Whenthat temperature stays below 50...60°C, the greenLED lights; between those temperatures and70...80°C, the yellow (orange) LED lights; andabove 70 .. .80°C, the red LED lights. There is alsothe possibility of providing a relay with which, forinstance, the load can be disconnected.The circuit is, in essence, a window comparator, inwhich sensor Di provides a control voltage thatrises 10 mV per degree Celsius. If the sensor voltageis lower than the voltage at the wipers of Pi and PA

LM 335

LM 335

Al, A2= IC1 = TL 082, TL 072

the outputs of opamps Al and A2 are low, and D2lights. When the voltage across Di lies above thatat the wiper of Pi, but below that at the wiper of P4the output of Al is high, so that D2 goes out andD3 lights. When the sensor voltage rises above thatat the wiper of P2 also, the output of both opampsis high: only Ds then lights and transistor Ti isswitched on. Zener D4 ensures that D5 lights bright-ly and that Ti conducts hard.To calibrate the unit, place the sensor, together witha calibrated thermometer, in a tray of water, whichis then heated. Set Pl to minimum and P2 to maxi-mum resistance. Set the cross over from green toyellow (orange) between 50 and 60 degrees Celsiuswith Pi Next, set the cross over from yellow(orange) to red between 70 and 80 degrees Celsiuswith P2 The sensor can then be fitted permanentlyonto the heat sink.

85405

138

110 LOGIC FAMILIES

The introduction of new, faster, CMOS techniqueshas given rise to a considerable increase in thenumber of available logic families. Understandably,this may cause confusion on the part of designersand users of logic circuits. Up until a few years, 3families were commonly known: the CMOS 4xxxseries; the TTL 74xx series; and the 74LSxx low -power Schottky series. TTL and LS chips aremutually interchangable, but TTL consumes con-siderably more current at the same switching speed.The 4xxx series is about 10 times slower than theTTL family, but is more economic as regards cur-rent consumption. In many cases, TTL chips are nolonger considered suitable for new design.The new HC and HCT CMOS families are just asfast as TTL and LSTTL, and have a greatly reducedcurrent consumption. HCT chips can work in LSbased circuits, provided they are not driven fromTTL or LS. This is because of the differently defin-ed switching levels. It is, however, possible to useHCT for driving HC. With this in mind, it is poss-ible to replace the LS family by the HC family. Thisis preferable since the HC family offers the highestnoise immunity.Figure 1 shows the current consumption of aHCMOS gate as a function of the input voltage.The shaded area represents the (logic high) outputvoltage of an LS chip. From this, two conclusionscan be drawn. Firstly, the noise margin is very nar-row: the HC gate sees 2.7 V as a logic high levelalready. Secondly, the current consumption of thegate is a few mA higher than necessary. Althoughusable in practice, driving HC with LS is, therefore,not recommended.Another new logic family was recently introduced:FACT (Fairchild Advanced CMOS Technology),also referred to as ACL (Advanced CMOS Logic) byother chip manufacturers. There are 2 versions: ACand ACT. ACT, like HCT, is fully LS compatible,while AC gives the same drive problems as HC.Both series are typically 2 to 3 times as fast as LSor HC.

Figure 2 shows the correlation between the propa-gation delay, tp, and the power consumption, P, ofvarious logic families. It will be noted that themodern CMOS families are almost as fast as theECL series, hitherto renowned for its unbeatablespeed. It is expected, therefore, that a CMOSequivalent will soon be available for ECL, and thatECL will gradually become obsolete.Replacing bipolar chips in existing circuits withCMOS types is not very useful if relatively high fre-quencies are involved. Finally, a rule of thumb forworking with chips of different families in a singlecircuit: HCT can replace LS, unless driven by LS.

For further reading:RCA CMOS DatabookFairchild FACT Logic Data Book

t

0 U

tp tI.[ns]

11 510

1

5

UIN [VI

I I I

I

MIPEI 1111HC/HCT Ill

ACACT

I=EMIAS .111

MaONIEMII

0 2 6 10 12 14 16 18 20 22 24 26

P

2

1 1 1 LOW VOLTAGE DROP REGULATORS

The fast spreading incorporation of CMOS, HCand HCT chips has created a need for voltageregulators with a very low internal drop to enable

powering CMOS-based equipment from a set of bat-teries delivering 6 V. The recently introduced TypesLP2951 and LP2950 from National Semiconductor

139

la)u,

IC 1 SENSE

LP 5V TAP

2951 FBSHUTDOWN ERROR

1p/10V

are micropower voltage regulators with a variableoutput voltage of 1.24-29 V and a fixed outputvoltage of 5 V, respectively. The former features aninternal voltage divider with a 5 V tap bonded outto a pin, a logic compatible shutdown input, and anopen -collector ERROR output which warns of alow output voltage, often due to an insufficient bat-tery voltage at the input. The ERROR output is ex-tremely useful for an early warning system that ar-ranges for a microprocessor to be reset properlybefore the supply voltage falls to a level that wouldupset the operation of the system it controls.The voltage drop across the LP2951 is only 0.4 Vat a load current of 100 mA, so a 6 V battery packcan be used to power a 5 V circuit. The quiescentcurrent drain of the regulator is about 12 mA at anoutput current of 100 mA. This is fairly high as

LP 2950

LP 2951

87428-1

R 112 13

Tip 7/2511111000/25V

compared with a conventional regulator from the78XX family, and mainly due to the internal seriesregulator transistor being driven into saturation,which causes it to have a relatively low current am-plification factor (the base current flows into theground return line, instead of into the output load,as with the typical 78XX regulator).The application circuit shown in Fig. la should befed from an input voltage of more than 5.4 V, whileits maximum output current is 100 mA. Note thatboth the LP2950 and LP2951 feature internal cur-rent and thermal limiting circuits. The decouplingcapacitor at the output of the regulator should bea good quality tantalum type, fitted as close as poss-ible to pins 1 and 4. At relatively low output cur-rents, less capacitance is required in this location.For currents below 10 mA, 0.33 tiF is satisfactory,

140

while the minimum value is 0.1 1.4F for currentsbelow 1 mA. These values apply to an outputvoltage of 5 V; for lower voltages, more outputcapacitance is needed.The circuit in Fig. lb is a 2 A low dropout regulatorbased on the LP2951. The output voltage is calcu-lated from

Vo= (1 + RA/RB)1.23V

where 1.23 stands for the voltage at the feedback in-

put, pin 7. For an output of 5 V, RA and Rs maybe omitted, and the feedback input pin 7 can beconnected direct to the 5 V tap (pin 6) output. Thesense input, pin 2, is then connected to the Vo rail.In this application, Vin must be at least 0.5 Vhigher than Vo.

National Semiconductor applications.

112 MAINS ZERO -CROSSING DETECTOR

Both safe and remarkably simple to construct, thiscircuit detects the zero crossing moments of themains voltage, in order to provide other circuitrywith timing information about the correct instantfor switching mains -connected loads; in otherwords, when the least possible switching dissipationis involved, and, therefore, least interference is in-duced on the mains lines.The proposed circuit operates direct off the mains,while comprising no more than two opto-couplersand two resistors. It is seen that photodiodes Diand D2 are connected in antiparallel while beingfed with the mains voltage via a resistor, whichlimits the current through the relevant diode toabout 2 mA as it conducts (i.e. lights) during thenegative or the positive half wave (D2 or Di re-spectively) of the mains sinewave; in either case, thecircuit output voltage is low, since the associatedphototransistor conducts and draws current from+Ub via Rz.However, at the moment of zero crossing, neitherone of the diodes conducts, and the voltage at thecircuit output rises to near + Ub level, whence the100 Hz pulse train.

R1

1W

+Ub O 5V...15V

ICI , IC2 = 2xT11_111

0- - -10ms10 ms'

86433

The value of R2 may be adapted to suit the level of+ Ub and the manufacturer -specified typical collec-tor current through the phototransistor. For theType TIL111, the current should not exceed about50 mA. The type of optocoupler used in the circuitshould not be very critical, but the value of RI hadbest be left at the indicated 100 k so as not to runinto excessive diode dissipation.

113 OPAMP-BASED CURRENT SOURCE

A current source based on an operational amplifieralone is likely to be less known than the combi-nation of an opamp and a transistor. This latter cir-cuit can, however, only supply a unidirectional cur-rent, and must incorporate a stable referencecapable of sourcing the required current. The cir-cuit proposed here is different from the usual design

for a current source, because it has a real differen-tial, high impedance, input.In spite of the small number of components in thiscircuit, its operation may not be apparent at aglance. An example calculation example may helpto clarify how the current source works.Assuming that 10 V is applied to input 2, and 4.5 V

141

to the output, the voltage drop across R2 is 0.5 V,and that across R4 is 5 V. It will be recalled that theoutput voltage of a current source is determined bythe value of the external resistance. The currentpassed through this gives rise to a voltage drop thatneed not be constant.When input 1 is 1 V more positive than input 2, thefollowing circuit potentials can be deduced:The + input of the opamp is at + 9.5 V, becauseR2 drops 0.5 V. The operational amplifier startsregulating its output voltage until it detects equalvoltages at its + and - input. The voltage dropacross Ri thus rises from 0.5 V to 1.5 V, while thatacross R3 is increased tenfold, i.e., amounts to 15 V.The output voltage of the opamp is then 11-1.5-15 = -5.5 V. When it is recalled that the outputvoltage of the circuit is + 4.5 V, the drop across R5amounts to 4.5-1-5.5) =10 V. Since Rs =100R, the current is 10/100=100 mA.It is also possible to establish the output current ofthe circuit as follows. The amplification is 10

(R3/121), and the output voltage is available acrossR5, which therefore carries a current ofU. x 10/100, or U./10.

R3

87447

This circuit is probably best operated on the basis ofpower opamps, such as the Types L149 and L150from SGS-Ates, which can handle currents ofseveral amperes. The Type OP50 stated in the cir-cuit diagram is suitable for relatively low outputcurrents (Linz 50 mA), and features excellentstability and precision. Its manufacturer, PMI,states that this application of the opamp is capableof handling resistive, capacitive or inductive loadsequally well.

Source: PMI, Analalog Applications Seminar 1986:Current transmitter (Howland current pump).

114 PIERCE OSCILLATOR

In addition to the description elsewhere inthis chapter of HC and HCT based R-C/L-Coscillators for use up to 20 MHz, this design briefconcentrates on quartz -controlled oscillators whichfind applications in digital equipment and micropro-cessor systems. Such oscillators can only be madewith HCU gates, because HC and HCT ones havebuffered outputs that make them unsuitable for useas analogue amplifiers.The circuit diagram shows a Pierce oscillator set uparound a single gate in a Type 74HCU04 package.The inverter functions as an inverting amplifierwith a phase shift of 180°. The circuit can bemodified into a Collpits oscillator by replacing thequartz crystal with an inductor. It should be noted,however, that the use of a quartz crystal is more ap-propriate because it ensures minimum current con-sumption and adequate suppression of the thirdharmonic frequency. Finally, R2 must be replacedwith a 33p capacitor if the oscillator is operatedabove 4 MHz.

78p

R1

78p

N1 = 1/6 IC1 = 74HCU04

87407

IC1

5V

142

115 POWER SUPPLY SEQUENCING FOR OPAMPS

Most designers know that many problems may arisebetween the paper design and the practical realiza-tion of that design. We are, of course, no exception,and one incident that we experienced recently il-lustrates a problem that is of interest to pass on.Measurements were being carried out on a circuitthat contained some type NE 5532 opamps whichwere powered from a ± 12 V symmetrical supply.When the circuit was switched on, it did not func-tion correctly. Measuring the supply lines revealedthat the positive supply was -0.6 V instead of+ 12 V. When the +12 V line only was switchedoff and immediately on again, the malfunctiondisappeared. Switching off the mains and immedi-ately on again made the defect reappear. Using newopamps made no difference.After some research in relevant literature, it ap-peared that on switching symmetrical power sup-plies temporary polarity reversal may occur.Because of the complex internal structure of inte-grated circuits, it may happen that this polarity re-versal causes parasitic components on the chip to beactuated which places the IC in a stable butmalfunctioning state.The book we consulted, Intuitive IC Opamps,suggests that the malfunction we experienced wasprobably caused by a parasitic thyristor being trig -

85465

gered owing to the negative supply not rising fastenough. The remedy proposed was to connect twodiodes across the supply lines as shown in the ac-companying figure: these diodes effectively preventpolarity reversal.This simple remedy certainly cured the malfunctionin our circuit and is probably the simplest protec-tion circuit in this issue.

Literature:

Intuitive IC Opampsby Thomas M FrederiksenNational Semiconductor Corporation

116 PRECISION CRYSTAL OSCILLATOR

When designing crystal oscillators, it is good prac-tice to ensure minimum capacitance of the active el-ement(s), since any parasitic loading of the crystalis bound to derate the overall stability to some ex-tent. This forms the underlying principle of thedesign described here, albeit that good results arealso obtainable when an additional loadcapacitance is connected in parallel with the exist-ing parasitic capacitance, but only if the former isknown to possess a low loss factor and a low tem-perature coefficient, i.e., if it is a very high qualitycapacitor (and possibly difficult to obtain).The oscillator proposed here is a Pierce type, inwhich the crystal operates in parallel mode. The in-put is formed by a bootstrapped source follower, DGMOSFET T,, which has a parasitic capacitance of

only I pF. RF transistors T2 -T3 are set up as acascode amplifier. A type BF494 transistor is usedin the T2 position because of its low B -Ecapacitance (0.15 pF typ.), which ensures a low out-put capacitance. The oscillator signal is taken fromthe source of Ti, buffered in Ta, and made logiccompatible with the aid of gates Ni-N3. The opti-mum inductance of Li is approximated withLl =1/f, where the inductance and frequency are inmilli -henries and megahertz respectively. Example:for f =10 MHz, Li works out at 100 µH. TrimmerC2 serves to accurately tune the crystal oscillator tothe required frequency. The oscillator works well upto about 20 MHz.Finally, although the dissipation of the crystal is notexpected to give rise to instability, it is still a good

143

... N3= 1/2 IC 1 =741.6 04, 74 HC(U) 04

T3BF 167

idea to keep an eye on its output amplitude so as topreclude the protective diodes in T, being activatedand causing unacceptable instability. If required, R7

<151nA4-mA (74 HCU 04)

<15mA (74LS 04) 5V

6.4

0

is altered until the signal amplitude at the emitterof T4 is less than 1 Vpp.

117 SMART LED SELECTOR

A = green or yellow

19 A

C

B = identical coloured LEDs

In this tiny circuit, for use in, for instance, a two -lights model railway signal, one of two LEDs maybe selected with either a single pole switch or aseries transistor, as shown in the circuit diagrams.Note that the LEDs are fed via a common current

limiter resistor, while a switch is connected in serieswith one of the LEDs.Why do not both light simultaneously when theswitch is closed? Because, apart from their colours,the two LEDs also differ as regards their forward

144

voltage drop; when connected in parallel, therefore,the LED having the lower voltage drop should befitted with the series switch; this arrangementcauses the high voltage drop LED to light when theswitch is open and to go out when the switch isclosed, at which moment the other LED takes over.Two of the accompanying four small circuits showthe use of a series switching transistor rather thana real switch, but the difference hardly requiresfurther detailing, since applying sufficient drive to

the base is in fact the same as closing the switch.Two LEDs of identical colour may also be used asshown, and the additional series diode is seen tocreate the necessary voltage drop difference todistinguish between the LEDs, which, of course,have roughly the same on/off voltage characteristic.Finally, the value of R is established from thesupply voltage level and the typical operating cur-rent of the LEDs, which is usually of the order of20 mA for maximum allowable brightness.

118 SPEED CONTROL FOR DC MOTORS

Simple DC operated motors with a permanent mag-netic stator behave as an independently energizedmotor. The speed of an ideal motor with an infinite-ly low internal resistance is in direct proportion tothe voltage applied, irrespective of the torque. Themotor thus runs at a speed at which its reverse elec-tromotive force (e.m.f.) equals the supply voltage.The reverse e.m.f. is directly proportional to theforce of the (constant) magnetic field, and the motorspeed. In theory, therefore, the motor speed can beheld constant with a constant supply voltage. Thespeed reduction observed in practice arises from thevoltage drop across the internal resistance, R, of

tab connected to pin 3

R6

R7

+VsOUTPUT-VSINVERTING INPUTNON INVERTING INPUT

D1, D2= 1N4001

1

IC3*

"720n

87427-2

18V max.

*see text

18V max.

145

the armature winding. Thus, when the motor isloaded, its current consumption, and hence VR1, in-creases, reducing the effective supply voltage. Thiseffect can be eliminated by means of R, compensa-tion, which essentially entails measuring themotor's current consumption, relating this to themotor's instantaneous drop across Ri, and increas-ing the supply voltage accordingly. In fact, this callsfor a voltage source with a negative output im-pedance, since it caters for a higher output voltagewhen the load is increased.The basic set-up of the supply required here is

shown in Fig. I. The load current is measured as thedrop across sensing resistor Ra. The DC transferfunction of this amplifier is written as

U2 = U1+ ILR2R3/R1

which accounts for the negative output impedancebecause then

Rout = -R2R3/Ri

For optimum results, this impedance must be keptabout equal to that of the motor.Figure 2 shows the practical circuit of the motordriver based on a power operational amplifier. TheType L165 from SGS can supply up to 3 A at amaximum supply voltage of 36 V, and is thereforeeminently suitable for the present application.Capacitors Ci and C2 suppress noise on the reversee.m.f. from the motor. Due care should be taken,however, in so extending the circuit, because thisreadily leads to instability. The motor itself alreadyforms a fairly complex load, since the revolvingrotor winding is mainly inductive, and the rotoritself represents a fairly large capacitance. Noisesuppression components such as R4 and C3 add tothe complexity of the load and may result in control

instability, which becomes manifest in the motor'stendency to alternately reverse its direction at a rela-tively low rate. Also, the response to a fast changein the torque may be impaired, and high -frequencyoscillation may occur (noticeable as exessiveheating of IC) and/or Ra). When the circuit wastested with a small PCB drill, best results were ob-tained by omitting Ra-C3 and including C2. If themotor has a noise suppression network, C2 must beomitted, and R5 added to protect the opamp inputsagainst too high differential voltages as a result ofcommutation voltage peaks. Clearly, Di and D2have been included with this in mind.Preset P, is adjusted until the motor remains stable.Over -compensation of the motor will give rise to ap-parently uncontrolled movement. The adjustmentof P, should be carried out when the motor has notyet reached its normal operating temperature,because its self -heating gives rise to an increase inthe internal resistance.The use of a symmetrical supply (± 18 V max.)enables twoquadrant operation of the motor(cw/ccw rotation), which can then be used to powermodel trains and the like. The motor is halted whenP2 is set to the centre position. The ground rail maybe connected to the negative supply rail if only onedirection of revolution is required (PCB drills). Themaximum supply is then 36 V, making a greatervoltage available for the motor, so that 24 V typescan be controlled also, although it is not possible tocompletely halt these.The motor can be protected against overloading byselecting a supply voltage that causes the opamp toclip when it outputs the maximum motor current.Finally, IC, is capable of supplying considerablecurrent, and must, therefore, be fitted with a fairlylarge heat -sink. The quiescent current of the circuitis about 50 mA.

119 STEPPER MOTOR CONTROL

The control of stepper motors is not simple, particu-larly when no specially designed control circuit isused. The Type TEA1012 is an integrated steppermotor controller that can cope with most if not allsituations. In addition to controlling the phases forwhole and half steps, it also sets the current withthe aid of these phases.The TEA1012 was specially designed for the con-trol of unipolar stepper motors, in which the cur-

rent passes through the stator windings in one direc-tion. Because the windings behave inductively, thecurrent through them will become too large whenthe stepping speed is low. The reason for this is thatin that situation only the ohmic resistance, which isfairly small, determines the value of the current. Tolimit the current, a limiting circuit is connected inseries with the windings. In the diagram, the cur-rent through Ll and L2 is restricted to 0.3/R4, and

146

211631]

11

3

0

05

610 D1_08= 1N4001T1...T4 = 8C639

4x 10k

5

4

Cl C2mm

T n2

Parts list

S1 02

lotf 1

lotl 2

Qa 03 S204

ICI STOPTEA 1012 ccw/u-N

12 A

4

C3

TO n

5i

dsMin

1

110n

O5V

TTL Input

86451 1

ra

T

rU

a cc

03 W P In CU

0

Capacitors: Semiconductors: PCB 86451

Ci;C2=2n2 T1;T2;T3;T4=BC639Resistors: C3;C5=10 n Di to D8 incl.=1N4001R1;R2=10 k C4=10 11;16 V D9 = zener 25 V;400 mWR3;1;t5;R6:R7;R5:R1o= I k IC1=TEA1012RI;R5=1Q8

®

147


that through L3 and L4 to 0.3/R9. This enables thecurrent through the stator windings to be adaptedto any type of motor.The table shows in what sequence the variousphases are driven with full and half step control, aswell as for clockwise and anticlockwise control. Thestepper motor is arrested in the position it occupieswith the STOP input. CL is the clock input: foreach pulse, the motor turns one step forwards orone step backwards.Because inputs CL, STOP, CCW/CW, and F/H allare 'TTL compatible, it is not difficult to connectthese controls to a computer. Resistors R to R14incl. and the associated switches, enable the circuitto be manually provided with control data.The maximum stepping speed depends on the typeof motor and on switch -off time -constants Tom)and Toff(2).

Letters CW and CCW signify clockwise and an-ticlockwise respectively, while input F/H enableschoosing whole (F) or half (H) steps. A double resol-ution is, therefore, possible.The supply voltage of the IC may be between 4.5 Vand 15 V. The outputs of the TEA1012 are open -collector, so that the operating voltage of the step-per motor may be made independent of the supplyvoltage to the IC.

Table

CL -

inputs outputs

FIR CCW/CW S 10P Q I Q2 Q3 Q4

half clockwise run

I

2

3

4

5

6

7

8

0

0

0

0

0

0

0

0

o

o

o

0

o

I

I

I

0

I

I

I

0

0

0

0 0

0

0

halfcounterclockwise run

I

2

3

4

5

6

7

8

0

0

0

0

0

0

0

0

0

0

0

0

0

full clockwise run

I

2

3

4

I

I

I

I

I

o

o

0

0

I

I

0

I

I

0

I

0

0

I

fullcounterclockwise run

I

2

3

4

I

1

I

I

of

120 SYMMETRICAL CASCODE OSCILLATOR

Free running as well as crystal controlled clockgenerators in many digital designs are most fre-quently based upon the use of one or more invertergates. However easy it may seem to use thesedevices for the construction of reliable oscillators,the resultant frequency stability is generally notsuch as might be expected from a look at the rel-evant quartz crystal data, and this is mainly on ac-count of the rather poorly defined capacitive and/orinductive loading of the crystal at resonance.Stability, however, may be improved by a factor 3 to5 by using cascode type inverters in a symmetricalconfiguration, as can be seen in the accompanyingcircuit diagram. Two sets of two n- and p -channelMOSFETs, contained in the Type 4007UB IC, have

been connected to form a highly stable oscillatorcircuit capable of operation at frequencies up to10 MHz, as determined by quartz crystal Xi, whichshould be a series resonant type.As the output impedance of the proposed cascodeoscillator is relatively high, buffer stage T, has beenadded to minimize drift with low impedance loadssuch as (LS)TTL circuits. Furthermore, MOSFETTi ensures well-defined logic high and low levels tointerface with (HC)MOS and (LS)TTL. The valuesof R4 and R5 depend on the supply voltage level(Ub), while the voltage at gate 2 should be between4 and 6 V to achieve a 5 V output level swing. Incase the oscillator is to operate from a 5 V supply,gate 2 of Ti must be connected direct to + Ub.

148

86484

121 THRIFTY LED INDICATOR

It is often necessary that the current consumptionof an essential status indicator is minimal. In thecircuit shown, dependent on the level of the supplyvoltage, a number of LEDs drawing a current ofonly 10...15 mA may be switched on or off asdesired. Moreover, the entire indicator may beswitched off if none of the LEDs lights.The circuit is based on switched current source

The base current of this transistor is set at c. 15 mAwith Rx. The value of this resistor is calculatedfrom

Rx = [4 x 106/(Ub-0.7)] Q

where LA is the supply voltage in volts.Transistor 12 conducts when the input to inverterNi is logic 0: when this becomes a logic 1, the cur -

149

rent source and, consequently, the indicator are 0switched off.If the input to one of the buffers N2 . . N4 is a logic1, the associated LED is switched on.More LED-FET combinations may be added to the N1 ... N4= IC1

circuit as long as the supply voltage permits this. T2 ... T5 = BS 170/VN 10

Also, the dissipation of Ti has to be kept within cer-tain limits. A BC557B can be used for Ti over thesupply voltage range of 5 . .18 V.The circuit is intended for CMOS ICs; if devices ofother logic families are used, remember to take ac-count of the different logic threshold levels.Note that the buffers must be powered from thesame supply as the current source.

1N4148 1 0 0 0= 4049

16 N2

BC 5575

N1 T2 T4 N3

1: 0 RO ' 1 5

01

81

85482

IC1

122 TIME STRETCHER

Anyone with a fascinating hobby must have felt atone time or another that there is not enough timeavailable for his hobby. Any circuit that can stretchthose few hours once or twice a week must,therefore, appeal to many.The time stretcher is a small circuit that can be builtinto almost any digital clock and makes the hobbyevening(s) last an hour longer. The three diodes,Di . . . D3, together with Ri, form an AND gate. 131is connected to segment g of the tens -of -hours dis-play, and D2 and D3 to segments e and g of thehours display respectively.When the clock shows 22.00 h, the common line ofDi . . D3 becomes logic 1, because the threesegments to which the diodes are connected are"on". This means that Ti conducts and the clocksignal of the digital clock is divided by two. Theclock then runs at half speed only so that it will taketwo hours before it shows 23.00 h.For the circuit to work correctly, it is essential thatthe clock signal is divided by two exactly, and thismeans that resistors R2 and R3 must be I per centtypes. This is also the reason that a BS 170 is usedas the switching gate; this MOSFET has no satura-tion voltage. Using a normal transistor with a cer-tain saturation voltage would not cause the clocksignal to be divided by two exactly, so that the clockwould be fast or slow by minutes within a few days!The circuit as drawn is intended for common -anodedisplays; if it is to be used with common -cathodedisplays, simply reverse the connections of diodesDi... D3.

A

fclock

3x1N4148 85503

clock

A = tens of hours displayB = hours display

150

123 TRACKING WINDOW COMPARATOR

The use of comparator circuits in many different ap-pearances and practical realizations is common in awide variety of electronic control and measurementsystems. Usually, the voltage from a sensor device isfed to a comparator which, as its name implies,compares the measured level, Ui,,, with a fixed ref-erence, Uref, and produces a negative output (0)or positive output (1) when Um<Uref andUin>Urer, respectively. A window comparator canbe made by connecting two comparators with dif-ferent reference levels, which define the upper andlower limit of the switching range.In practice, these references are usually adjustedwith presets to dimension the window as required.This arrangement makes it impossible, however, toautomatically shift the window up or down in ac-cordance with, say, ambient light conditions to bemeasured with a light dependent resistor.This circuit has no fixed threshold levels, butderives its reference from the measured signal, sothat slow changes in this cause the window to trackalong.Capacitors C, at the inverting input of Al, and C2at the non -inverting input of A2 store the inputvoltage. When the voltage at the non -inverting in-put of Al rises, this opamp toggles. The associated

la

R1

inverting input lags this change because of thedelay introduced by the capacitor. LED D, lights.The process is similar in the A2 section of the circuitwhen the input voltage drops. This is indicated byLED D2 lighting.Diodes D3 and D4 form an OR function to actuatea simple relay driver set up with T,. The relay is en-ergized when the circuit detects a fast change in theinput voltage. The ability of the circuit to accept avariable input voltage makes it suitable for use inburglar alarms-see Fig. lb. Several break contactarrangements R13 -S1 -R14 may be connected inseries and to the input of the window comparator.Alarm relay Re, is activated when either S, is open-ed or S, -R14 is bypassed. To prevent burglars fromfooling the alarm, R14 must be fitted into Si,because no alarm signal is given when only Si isshorted.The sensitivity of the tracking window comparatoris defined by the ratios R2/R3 and R5/R6. The rel-evant component values indicated in the circuitdiagram give 1:100 ratio, so that, for example, a fastchange of 30 mV is detected when the input voltageis 3 V. The sensitivity also depends on the inputvoltage. Although the circuit can in principlehandle any input between 0 V and the supply level,

Al, A2 = 1/21C1 = LM324D3...D5 = 1N4148

87423 -la

151

the ICs used give reliable operation only whendriven between 1 and Ub-1 volt.The tracking window comparator is preferably fedwith a supply between 5 and 15 V. Its current con-

sumption, inclusive of the LEDs but exclusive ofthe relay, is 10 mA maximum (note that the relaycan be fed separately).

124 TRANSMISSION LINES FOR TTL CIRCUITS

Although cable connections between TTL circuitsare normally not as critical as those for, say, RF ap-plications, it is still worth while to reflect on thissubject because strange things often happen whena TTL transmission line is not correctly terminated.In particular, this discussion is about terminatingcoaxial cable and flat ribbon cable. The latter is fre-quently used for driving Centronics compatible in-puts.A commonly used coaxial cable is RG59B/U, whichhas a characteristic impedance of 75 Q and a propa-gation delay of 5 ns/m. With signal rise and falltimes of 4 ns, the cable may be considered elec-trically long if it exceeds 40 cm. One of the mostcommon terminations used when driving a long co-axial cable with an LSTTL gate is shown in Fig. 1.This set-up is unsuitable for a HCT bus driver, sincethe termination provides a poor impedance match,and requires a current sinking capability of 20 mA.An improved termination circuit is shown in Fig. 2:this ensures reliable signal transmission for cablesup to 15 m. Note that the 1 k52 pull-up resistor isonly required when the driver is an open collectorgate or buffer.Flat ribbon cable often introduces considerablecross -talk between wires, especially when ter-minated in HC(T) gates, which form a high inputimpedance. In general, a flat ribbon cable shouldnot be longer than about 60 cm, but longer runs arepossible when individual wires are separated bygrounded wires (1.8 m max.), or when each wire isterminated with a 11(52 pull-up resistor (1.2 m). A

O

O

=5 87455

combination of these methods makes it possible touse flat ribbon cables with a length up to 2 m, butthis is also attainable without ground wires-seeFig. 3. The combined use of this termination net-work and grounded wires in the flat ribbon cableshould enable a cable length of about 5 m.

125 TUNING AF POWER STAGES

Simple, economically priced audio output stages,such as, for instance, those using the hybrid ICs inthe STK series, may be improved in a simple man-ner as regards distortion, noise, and off -set voltage.To this end, the output amplifier is included in thefeedback loop of an op -amp. Fig. 1 shows the set-up

for inverting output amplifiers, and Fig. 2 that fornon -inverting ones (the normal situation).In the calculations to arrive at the new gain of theoutput amplifier, determined by Ri and R2, it is as-sumed that the LF356 provides an undistortedsignal of 5 Vrms; note also that this type of op -amp

152

must work into a load of not less than 5 kilo -ohmsto prevent distortion.For an output power of 50 W into 4 ohms, the out-put stage must provide a voltage,U = PR =14.2 Vrms. If the amplification of thestage is 3, the op -amp should deliver 4.73 V. For theset-up in Fig. 1, the value of R2 is then R2= 3R1,while for that in Fig. 2, R2 = 2R1. Note that in bothversions only the value of R, should be altered. Thetotal amplification may be calculated from the ratioof RA and Rs as follows: A = (RA + Rs)/Rs.Furthermore, because of the load impedance of theop -amp, R, >10 k (Fig. 1); R2>10 k (Fig. 2);

RA >10 Q; and Rc>10 Q (Fig. 1 and 2).To compensate for the off -set voltage of the outputamplifier, the input capacitor should be replaced bya wire link. The capacitor in series with R, in Fig.2 should also be short-circuited. The lower fre-quency limit of the complete circuit is then deter-mined by Cs =1/2nfsmRs. The off -set voltage isthen smaller than 3 mV, provided both RA and Reare equal to, or greater than, 100 k52. Where greateraccuracy is required, Pl can be used to set the off-set to exactly 0 V.To ensure that there is no direct voltage at the newinput of the amplifier, capacitor Cc should have avalue of Cc =1/fisnRc.Since the amplification of the output stage has beenreduced to 3, its feedback factor has gone up, andthe distortion has gone down. The ad-ditional feedback of the LF356 reduces the distor-tion even further. An overall reduction in the distor-tion from 1 per cent to 0.1 per cent is fairly typical.The altered feedback unfortunately results in achange in stability. If there is a tendency to oscillate,the first thing to do is to bring the upper frequencylimit back to its previous value with the aid ofCy =1/2Trfse,RA. If the tendency persists, capacitors

1

001

Cx must be used: their value lies between 100 pFand 1 nF. Our prototype (using STK ICs) workedsatisfactorily without either Cx or CY.

126 TWO -FREQUENCY OSCILLATOR

Not so long ago, when semiconductors were stillquite expensive, it paid to make a transistor servemore than one function. Although this is no longernecessary because of cost considerations, it is stillfun to do so - and it may even have its uses!The circuit presented here is an LC oscillator thatchanges frequency through reversal of the supplyvoltage.When the supply voltage is positive, Di conducts

and short-circuits LAC,. Oscillations are then main-tained by crystal XL2 and L2C2. The DC operatingpoint is set by 131 in a way which ensures a com-promise between faultless starting of the oscillatorand low distortion of the output signal.When the polarity of the supply voltage is reversed,transistor Ti operates in its inverted mode, i.e., thefunctions of emitter and collector are interchanged.This means that the amplification is reduced, but,

153

of course, an oscillator needs an amplification ofonly just above unity to operate. Crystal XL2 andL2C2 are effectively cut out by D2, and the fre-quency is now determined by crystal XL, and LiCi.The circuit lends itself, for instance, for use as BFOswitched between USB and LSB.The crystals may have values of up to 1 MHz.Current consumption in either mode does not ex-ceed 45 mA.

From an idea in the Master Handbookof 1001 Electronic Circuits.

U= +10 V fx2

U = -10 V -> fxi

C3

mom47n

4BC

Cmom

560C

85438

127 TWO -GATE BISTABLE

Probably unequalled as to its simplicity given thedigital function, this circuit may serve as a single -button on/off control for incorporation in a widevariety of electronic designs. The operation of theproposed bistable is best understood if it assumedthat the input of Schmitt -trigger inverter Ni is atlogic high level; the output of N2 will therefore behigh as well. It is seen that the capacitor is discharg-ed because of the low output level of N,. Therefore,depression of the button pulls the input of Ni tologic low level, causing the bistable to toggle; the ca-pacitor is charged via the 1 M resistor, and the cir-cuit will change state again at the next switch ac-tion. The indicated resistor values have been found

S1 N1...N2 = 1/3 IC1 = 40106

N2

86472-1

to offer optimum stability of the bistable, while theuse of Schmitt -trigger CMOS inverters is essentialto the correct operation.

128 UP/DOWN CLOCK GENERATOR

Various designs of clock generators have appearedin previous Summer Circuits issues of Elektor Elec-tronics, and this tradition is kept up with the pres-ent design which, unlike the other circuits, outputsan up/down indication as well as a rectangularsignal over a wide frequency range; 0 Hz to severalkHz.The output signal and the U/D indication are both

controlled by a single potentiometer. If this is set tothe centre of its travel, nothing happens; turningthe potentiometer in the clockwise direction causesthe U/D output to be at logic high level, and thefrequency of the output signal rises with turningPi further in this direction. The same goes forturning it anti -clockwise, U/D being at low logiclevel.

154

The basic operation of the circuit is as follows. Op-erational amplifiers Ai and Az together constitutea sawtooth/square wave generator. The falling edgeof the sawtooth voltage has a fixed duration ofabout 200 pis, as defined by the current throughD4. The rising edge time, however, depends on thevoltage at the wiper of Pi. The wiper of Pz is ar-ranged to be at a slightly higher voltage than thatat the wiper of Pi, when this is set to the centre ofits travel. The STOP LED will light in this con-dition. If Pi is turned in either direction, thevoltage across RI rises and causes a low current toflow through Rz. This current, and therefore theoutput frequency, is proportional to the position ofthe wiper of Pi, but this only goes for a limited fre-quency range. If the voltage across Rz exceedsabout 0.6 V, Di conducts and connects R3 in paral-lel to R2. D2 and D3 do the same for R4 at about1.2 V; this method causes the oscillator frequencyto be an exponential function of the voltage, setwith Pi; the arrangement ensures a considerableoutput frequency range for the oscillator Ai -A2.Together with one or more universal countermodules (see Elektor Electronics, March 1985), theproposed clock generator may offer a neat replace -

stop

0

down upcount

86 409 -2

ment of the well-known BCD coded thumbwheelswitches; the potentiometer -set value is present atthe Q1 . .Q4 outputs of IC2, as well as visible onthe seven -segment display.The U/D and clock output of the present generatorare connected to the relevant points on themodules, as explained in the above mentionedarticle, but remember to observe the differentsupply voltages of clock generator and countermodule; keep all points marked + 5V at thatvoltage, except the supply pin of the LM324 andR14 and It is, which are connected to the countermodule +12 V supply. Current consumption of thepresent up/down clock generator is modest at about10mA.

129 UP/DOWN COUNTER CONTROL

The up/down binary- or BCD -mode counter is aregularly spotted item in digital circuits of variouslevels of complexity. The up/down counter simply

does what its name indicates; it counts up or down,depending on the logic level applied at the relevantcontrol input, and activates the corresponding out -

155

UP DOWN

put bit pattern at every pulse transition detected atthe chip's clock input.This circuit simplifies the control of up/downcounters in that it allows the user to press one but-ton to increment the counter output state, whileanother decrements it. Each of the changeover typebuttons is connected to a two -gate de-bouncer/bistable (N1 -N2 and N3 -N4), which sup-plies a low pulse at its output when the relevant but-ton is pressed. N8, which Serves as an OR gate,receives the debouncer pulses and, together withN8, provides the output clock pulse to the up/downcounter.Bistable N8 -N6 keeps track of the selected countmode, and provides the relevant logic level to theup/down counter input. It should be noted that thelogic level designation of the up/down input to the

UP

UP/DOWN

N ...N8 =

(f)74LS279

CLOCK

86463-1

86463.2

counter chip may differ from type to type; it maytherefore be necessary to interchange the UP andDOWN keys.The use of counter chips changing output state onthe negative clock transition is to be preferred foruse with the suggested circuit, since bistable NB -N8toggles coincidently with the positive clock pulsetransistion (see Fig. 2). However a minor disadvan-tage of the use of negative -edge clocked up/downcounters lies in the fact that the circuit acts upon re-lease rather than depression of the UP and DOWNbuttons.Finally, the use of the Type 74LS279 is in no waycompulsory; a combination of other types of TTLIC incorporating the necessary NAND gatesshould work equally well, but note the three -inputNAND gate N8!

156

130 VERSATILE TIMER

This simple -looking circuit enables the arbitraryprogramming of seven outputs in a series of notmore than 2048 (21t) steps. The step length may beset as required. The time base is derived from themains voltage. Transistor T produces a squarewave from the mains voltage applied to its base.This square -wave voltage is divided by 10 in ICI, sothat the frequency of the signal at the clock inputof IC2 is 5 Hz. Circuit 1C2 serves as addresscounter for the Type 2716 EPROM. This meansthat 1C2, after a reset, counts upwards from 0 andruns over the successive addresses of the EPROM.Circuit 1C2 has twelve outputs which would enablethe use of a Type 2732 (4096 steps), but, on practi-cal and financial grounds, a Type 2716 is used heresince 2048 steps are normally quite sufficient.The outputs of the EPROM are buffered by a Dar-lington array, 105, so that seven switch outputs areavailable with a sink capacity of 500 mA at a maxi-mum voltage of 50 V. The eighth output containsthe stop -bit that provides the facility of stopping theprogramme if this is shorter than 2048 steps.The start -stop circuit is based on bistable N3 -N4.When the supply is switched on, 1C2 ensures thatthe bistable resets from the stop state. This means

R1

that both divider ICI and counter 1C2 are in pos-ition "zero". The first address in the EPROM must,therefore, have a neutral content, because it is ad-dressed in the stop state and thus appears at the out-put.The bistable is set, and both resets cleared, whenthe start button is pressed. Circuit ICI then com-mences to divide, and 1C2 starts to count. With thepresent time base, the programmed content of suc-cessive addresses will appear at the output of thebuffers at 0.2 s intervals. Counting continues untila stop -bit appears at pin D7 of the EPROM, orstop button Si is pressed. If required, a HOLDfunction may be obtained by connecting a switchacross capacitor CI, which enables the time base tobe switched off.Switching on a specific output a . . . g merely re-quires the corresponding bit position in theEPROM to be left unprogrammed (logic high); pro-gramming a 0 disables the relevant output. Thestop -bit operates with negative logic: a 0 thereforecauses a stop.Finally, the time base may be adapted for the settingof the required step frequency and accuracy.

14

16

C)

CLK CO

_trINH

IC1 4017

J1 RST

16

5Hz12 10

13

6

15 11

AC 50 Hz (10y)

----C1=INTn

O

CLK 1

RST

01

0203

0405

05

IIC2 404009

09

010

9

7 7

6 6

3 4

2 3

4 2

13 1

12 23

14 22

9

012

R

AO10 2

A2 D2 " 313 4

241 2

VPP

4

5

6 IC3 27

8A9

810

.1. CE FE12 18 20

Ds

04

Ds

14 5

15 6

6

0 sv

09148-STOP

0

C2

STARTS281---4

eI= 500 mA max

-11*-1:=1-0

012

4;5c;,

IC5 = ULN 2004

N2

M4 = N1444 = 4093 86413

157

131 VOLTAGE -TO -CURRENT CONVERTER

1;ttv=10=1kk

1C1 = 741/CA3140

Um = max.±10Vlout = max.±2OrnAlow/Um = -1mAN

out

The converter proposed here (also called voltage -controlled current source) is based on just oneopamp, and provides to, or draws from, ground acurrent that is dependent on its input voltage. Theunit can convert negative as well as positivevoltages into negative currents (from ground) andpositive currents (into ground) respectively.When a Type 741 or CA3140 is used in the At pos-ition, Rv =1 k, and R =10 k, Um= ± 10 V max.;Iout = ± 20 mA max.; and gm = -I mS. It is, ofcourse, possible to change any or all of these valuesas required by using a different opamp and alteringthe values of the resistors. The maximum output

100n

1C1C2D.100n

015 V

e15v

current is always dependent on the opamp used. Tomake such changes, the following formulas mayprove useful.U + = U- = (Um-Uout)/2 + Uout

U0= 21(Uin-Uout)/2 + Uout] = Uin + Uout

IRv = Uin/Rv

Iout = IRv + IR = Uin/Rv + (Uin-Uout)/2R

If R> >Rv (the usual case),Lout = Uin/Rv.

158

CHAPTER 5 ELEC PHONICS

132 BLOW THAT SYNTHESIZER!

Circuits for generating electronic music are usuallycontrolled by key switches. Not only do keyboardsoffer the simplest technical solution for producingfast changing, reproducible tones over a wide fre-quency range, but they also enjoy tremendouspopularity because they are considered to be easierto learn to play than string or wind instruments.Because of that, we have not tried to create an elec-tronic oboe, flute, or clarinet with the present cir-cuit. In any case, the technical intricacies associatedwith such instruments would make their elec-trophonic counterpart prohibitively expensive.So, what we have got here is the relatively simple fa-cility of converting breath power into a proportionalanalogue voltage with which the volume of a musicsynthesizer can be controlled; the tones remain con-trolled by the keyboard switches. No doubt, manyof you, ingenious readers, will be able to think ofvarious other applications of the converter.The circuit does not operate direct from the exhaledbreath, but from the noise generated by this. A thin,flexible tube, to which a mouthpiece may be attach-ed, leads into a closed box, in which not only the

C1

-IF470n

R1

P1

circuit, but also an inexpensive microphone havebeen fitted.The noise received by the microphone is amplifiedin IC1, the gain of which can be adjusted with Pi,and subsequently rectified by IC2-D, -D2. An activelow-pass filter removes most of the ripple from theoutput voltage.To keep the circuit as simple as possible, we haveopted for a compromise between input sensitivityand output ripple: the relation between these twoproperties can be adjusted with P2If you have an oscilloscope with slow sweep, cali-bration of the converter should present no prob-lems.First, adjust the value of Pi so that the outputvoltage with hard blowing into the tube just doesnot cause full drive (dependent on the sensitivity ofthe following instrument).Second, adjust P2 so that the output signal is rela-tively free of ripple, while the converter still reactsto normal breathing. A steeper filter would havebeen better here, but that would have increased thecost.

100k

R2

C2

35V

R3

7

1C1

31308

4

R4

R6

D1

1N4148R8

15 V

®100Z5V

C5elm7n

C6

I I --10n

1C3

3130

4

C7

--I

85514

8

0

loolmrsvo

15 V

159

133 DISCO SOUND LIMITER

The environmental nuisance value of discos is indirect proportion to their sound level. The circuitproposed here cannot be disabled by the disc jockey,since it is built into the output amplifiers used in thedisco. Its operation is amazingly effective: if thepreset sound level is exceeded, the input of theamplifier is short-circuited for a few seconds. Anydisc jockey whom that has happened to a couple oftimes soon gives up trying to break the sound bar-rier.The power amplifier output is connected to themetering input of the present circuit (Ci). Thissignal is applied to low-pass filter Ra-C2 via Pi(which sets the maximum volume) and buffer IC,.In case of line inputs, this opamp can be given again of 20 dB by the omission of the wire link acrossR2.

The signal from the low-pass filter is rectified (halfwave) by IC2 and IC3. The resulting direct voltageis applied to Al and A2 which compare it with tworeference voltages derived from potential dividerRe -Rs -Rio. When the first threshold is exceeded, D5lights to warn that maximum sound level has

110

C9

100n

P2

71(1

50k

1C1,1C2,IC3 = 741A1,A2 =1C4 = LM393

BC547B

C8 R15CI6V

614

MEM

almost been reached. When the sound level then in-creases by 6 dB, Al also toggles, which triggersmonostable IC5.The input signal to the power amplifier (via C9,Rio, and P2) is then short-circuited to ground viaResistors R14 and R15, and capacitor C8, obviateany "plops" from the loudspeakers.Power for the present circuit may be derived fromthe output amplifier. The normally quite highsupply voltage there is reduced to ± 15 V by twocomplementary power transistors. Current con-sumption of the circuit is about 40 mA.

20...50V 50139 15V

R7

4 8I el 71

612

IC5

555

R11 C6

C7

0D5

C5 ;5V

1113

C4

T1OOP6V

15V

Al

86505.1

R8

BD lag -100

15V

6

R9

0

R10

86505-1

0

160

134 GUITAR FUZZ UNIT

The fuzzbox, fuzzer, tube screamer, or whateverother name there may exist for the controlled guitarsound distortion unit, is a well-known item in theelectrophonic field, which is of common interest toboth musicians and electronics enthusiasts.The majority of fuzz units are simply opamp con-figurations with some form of maximum input levelcontrol, which determines the degree of overdriveby the guitar input signal, and, consequently, theamount of audible distortion, generally referred toas the object "sound" the player has in mind as hisvery own musical visiting card.This is probably one of the few fuzz units to featurecontrollable symmetrical clipping facilities, whichmeans that the limit for distortion -free amplifi-cation may be separately defined for both thenegative and positive portions of the inputsinewave(s), the peaks of which may be clipped bymeans of shunt transistors Ti and T2 respectively,each with its own clipping level control poten-tiometer (Pi; P2). The transistors, when driven, passthe signal from input opamp IC, to the positivesupply or to the ground rail, before buffer 1C2 can

pass the "fuzzy" guitar sound to the connectedamplifier.Preset Pa determines the minimum gain of the fuzzunit; the desired level may be set with P4 turned toits minimum resistance position. Next, P4 is ad-justed to suit the maximum input level that can beexpected from the guitar. P3 and P4 may then bealternately adjusted to hit the correct compromisebetween these two signal levels.Finally, note the three -pole changeover switchwhich allows easy bypassing of the fuzzer whilesimultaneously switching it off to preserve batterypower.

<10mA 006.SIC Sib

Bp

161

P3

W

a

a a

C Ln

P2

r5i: siNe1

135 LIMITER FOR GUITARS

The basic dynamic characteristic of a chord can beanalysed as a fast rising, needle -shaped pulse witha virtually exponential decay-see Fig. 1. Thistypical amplitude characteristic can only befaithfully reproduced by an amplifier if this is oper-ated well below its overload margin, and that, manyguitar players know, generally results in too low anaverage sound level. Also, when it is desired to usea high volume setting, the distortion soon rises toan unacceptable level. Although the above diffi-culty is widely remedied by means of a tightly setcompressor or limiter, the sound may then lack therequired agressiveness. This circuit is expected togive better results than most other limiters, becauseit is only active in the upper range of the dynamiccharacteristic.The gain of the preamplifier set up around IC, isadjustable with Pt The inverting input of theopamp is grounded via the drain -source junction ofn -channel FET T,, which operates as a voltage -controlled resistance here, and is driven with anegative gate voltage derived from the limiter's out-put signal. The gain of the opamp is there-fore inversely proportional to the gate voltage of theFET, whose drain -source resistance is reduced asthe gate voltage becomes more negative. NetworkR5 -C4 effectively reduces the distortion incurred bythe regulating action of the FET. It may benecessary to redimension R5 and C4 to compensatefor the tolerance on the FET-use an oscilloscopeand a function generator to find the optimumvalues for these components while the circuit is be-ing arranged to operate at maximum compression.The limiter is fairly simple to align. Apply a 1 kHz,

162

150 mV input signal to the input, and monitor theoutput signal with an oscilloscope. Adjust Pi suchthat maximum amplification is obtained with virtu-ally no distortion. Increase the input amplitude to300 mV: this is likely to make some distortionnoticable. Carefully turn Pi back until the distor-tion is reduced to an acceptable level. In some in-stances, when the distortion remains too highwhatever the setting of 111, it may be necessary toreplace Ti , since the Type BF256C is manufacturedwith a relatively loose tolerance.The proposed limiter leaves the lower dynamicrange unaffected, while slightly compressing thepeak amplitudes in the input signal. Optimallyaligned, it suffers none of the notorious side -effectssuch as "noise breathing" and clipping commonlyassociated with other units, while it enables guitaramplifiers to be driven 3 dB harder without produc-ing appreciable distortion.


2<5mA 9V

131

SEM

R3

C1

47n

IC 1TL071

CS

R10

C2CI

171

AA119

D2

AA119C5 - Cl

Enia

22n urn

P250 klog.

C8=Ea

78p

IC1

136 MELODIC SAWTOOTH

Even in this era of programmable, polyphonic syn-thesizers, interest in simple, monophonic keyboardinstruments remains. Many FORMANT ownersare still proud of their, probably first, home -builtsynthesizer and are still on the look -out for new cir-cuits for the generation of exotic sounds. For allthose, here is an easy -to -build circuit that can con-vert a sawtooth signal at its input into an output ofdouble the frequency and half the peak value of theinput signal (figure 1).Comparator IC, transforms the sawtooth signalinto a rectangular signal (see figure 2). Adder 1C2combines the original input signal and the rec-tangular signal.An additional LFO (low frequency oscillator) con-nected as shown provides pulse -width modulationof the rectangular signal, which has a greatlybeneficial effect on the output signal.When switch Si is set to position b, it is possible toinject a rectangular signal whose frequency is in-dependent of the sawtooth frequency, which greatlyincreases the number of melodic variations, asanyone acquainted with synthesizers knows.Power requirements can be met direct by theFORMANT or any other ± 15 V symmetricalsupply. Current consumption is not higher than10 mA.

2

/vv

LFO

////////

P2=0

P2 max.

15V

15V

163

137 METAL PERCUSSION GENERATOR

The objective of this circuit is to obtain asynthesizer -controlled equivalent sound as pro-duced by such metal indefinite pitch percussion in-struments as cymbals, gong, and anvil. Fig. 1 showsthat the generator comprises four independentlytuneable VCOs which supply rectangular outputsignals to a combination of XOR gates.One of four identical KOV (keyboard outputvoltage) driven VCOs is shown in Fig. 2. The use offast opamp types ensures linear VCO operation wellup to 4 kHz, while FET T, improves upon thelinearity of the voltage -frequency curve relevant tothe combination of integrator and comparator.With the VCO constructed four times over and con-nected as shown in Fig. 1, drive controls P1 . . . P4

allow the user to set the output sound as desired.The outputs of buffer opamps Ai . . A4 (IC,, TypeTL084) should measure 0 V offset with the KOVrail grounded. If this can not be attained, the IC willhave to be exchanged with a more stable type.Linearity of each of the VCO circuits is set with thepreset at the drain of the FET, P5 and Ti respect-ively in Fig. 2. Use a scope to check whether therectangular VCO output signal has a 50% duty fac-tor; if not, adjust the relevant preset.As the four VCOs lack a linear to exponential KOVconverter at their inputs, it is not possible to use thepresent circuit with a keyboard of the I V per oc-tave type. However, many keyboards provide an ex-ponential KOV signal whose frequency doubleswith every octave and which are, therefore, suitablefor use with this generator.

KOV

(exp. )

O

OV (exp. )

A2 ,A3,A4

bbICI 1C2,3

44

CI

Al = 1/4101= TL0134102 ,1C3 = CA3130

SPA

138 PATCH CATCHER

This circuit facilitates switching between pro-grammed settings on synthesizers, expanders, andother electrophonic instruments. Most of thesehave some provision for storing or saving user -defined instrument settings, which are usually re-ferred to as patches in the electrophonics en-thusiasts' jargon. Although this facility is a greatasset to many musicians, a problem arises when pat-ches are to be called up in rapid succession whileplaying. On some instruments, this problem is solv-ed by a pedal that, when pressed, enables the instru-ment to operate with the next patch from the user -

defined file (patch increment pedal). In practice,however, the increment function of the pedal maystill be considered cumbersome. Assuming that therelevant instrument supports the use of eight pat-ches, the pedal needs to be pressed no less thanseven times to switch from, say, patch 3 to 2. Thisis obviously a distracting additional task when thekeyboard is to be played simultaneously.This circuit uses a relay whose contact is connectedto the pedal input on the instrument. The userpresses a key numbered 1-8 to select the relevantpatch, and the circuit arranges for the relay contact

164

9V

61

62

1.51

1.52.1--0

LS3e--01.54

ZS50--0 0

10

5

11 Dl12

13

.71.56

e--0 0ZS7

0-0 0ZS8

e--0 0 -RI ... Re8 x 100k

02

0;19

01

D3 IC 1 02154 45322 05

4

ID8

Ear

7

01

13 02 02

616R9

14 03 IC 2 03

4042

POLCL

5

770n

3

N5

10 5

11 4

02

01

300 IC3_.2 4518036

RESET

CL

Re =Siemens, N1...N 4 = IC4 = 4070V23027-00001 A 101

2 N 5..,N 8 = ICS = 4093

770n

13

to be automatically actuated, simulating thenumber of pedal operations that would be requiredotherwise. With reference to the circuit diagram,IC, is a priority encoder whose outputs Qs -Q2supply the binary code of the pressed key Si -S8. Thepulse at therminal E001 is delayed in Rs -C, and fedto Ns -Ns which serve to clock 4 -bit latch 1C2. Out-puts Q, -Q3 of this chip are applied to the inputs ofXOR gates Ni-N3, together with the outputs ofcounter 1C3, whose binary output state is initiallyassumed equal to that of 1C2. Pressing one ofswitches S, -So causes the output of 1C2 to change,and one of the XOR outputs to go high. Thisenables oscillator N7, so that its output pulses, in-verted in Na and buffered with T, energize the relayand increment the patch number on the instru-ment. The oscillator pulses are also applied tobinary counter 1C3, which is set up to count from0 to 7 because its Q3 output drives the RESET in-put. After a maximum of 7 pulses, the logic levelsapplied to each of the XOR gates are equal again,so that the oscillator is disabled via N4.

. The choice between the make or break contact ofthe relay is governed by the type of pedal this circuit

3x1N

4148

01

H2

D3

R11

4

3Idl

00n

C4IC2

100n

14

IC3 IC4 IC500n

T 087432

is to replace. Preset Pi is adjusted such that the in-strument is just capable of reliably following the ac-tions of the relay. After turning on the equipment,it is necessary to first press Si, then select the firstprogram on the instrument, and finally make theappropriate connection between this and the patchcatcher.The circuit, exclusive of the relay, consumes only afew milliamperes. The prototype, fitted with thestated Siemens relay, drew a mere 50 mA from the9 V supply.

165

139 QUARTZ -CONTROLLED TUNING FORK

Musical instruments are tuned with the aid of asignal source that generates a signal at a frequencyof 440 kHz. An electronic tuning fork is superior toits mechanical counterpart as far as dimensions,weight, and stability with temperature are concern-ed. The stability is obtained by controlling thesignal source by a quartz oscillator. The output ofthe oscillator is frequency -divided and thenamplified. The output may be made audible by, forinstance, a small loudspeaker.In the accompanying diagram, NI, N2, and the

C1

60p

quartz crystal form the oscillator. The precise fre-quency, measured at the Q terminal of FF2 with acalibrated frequency meter, is set with CI. DividerType 4059 is easily programmed to a differentdivisor. A duty factor of 50 per cent is ensured byFF2.The transducer is shunted by a 100 nanofarad ca-pacitor, because most transducers have a much bet-ter high- than low -frequency response, which causesvery shrill sounds.

CLK

FF1

R S

24 10 15 0 21 4 13

VDD

CLK 4059 (-1862)

OUT

Vas

7 a 9 16 17 18 19 22 6 5 3 11 12 2

23

Da

FF2

S

R3

X1 = 3.2768MHzN1...N6 = 1C1 =4049FF1, FF2 = 1C2 =4013

86478-1

140 SOUND -LEVEL INDICATOR

This novel indicator is ideally suitable for use in adiscotheque. It consists of eight equi-distantcolumns of eight LEDs arranged in a starlike pat-tern, so that corresponding LEDs in the eightcolumns form concentric circles, as shown in fig-ure lb. The higher the sound level, the more circleslight, giving the impression of a star of constantlyvarying brightness.As can be seen in figure lb, the eight LEDs in anyone of the eight circles are connected in series. Each

of these series chains is driven by a transistor:Ti .. .T8 in figure la. Dropping resistors are not re-quired: the positive supply voltage provides justover 1.8 V per LED, which is a perfect value for redLEDs to show up nicely.Transistors Tl... T8 are driven by differentialamplifiers Al . . A8, which compare the audio -dependent direct voltage across C2, which is buf-fered by Al2, with the potential determined by Diand R . .RA8. If the result of the comparison is

166

la

R7

15V< 800 mA

15V< 20 mA

P1

2x

C1

560n

3

5V6400 mVV

2

*see text

15 V

R10

R11

812

R13

D

13

814

R15

R16

817

818

Al .. A4 = IC1 LM 324A5 .. . A8 = IC2 = LM 324A9 . Al2 = IC3 = TL 084

positive, the associated driver transistor is switchedon, and the appropriate circle of LEDs lights. TheLED in the centre, Da, is driven by T% and onlylights when the sound level is very low.The direct voltage across C2 results from full -waverectification in Alo and Al i of the input signal afterthis has been amplified in A9. The input sensitivityis about 600 mV for saturation, i.e., to light all sixty-four LEDs; it can be increased by lowering thevalue of R2.

The speed with which variations in sound intensityare indicated depends on the value of C2: if this is10 taF, the light pattern changes slowly, whereaswhen the capacitor is omitted, it reacts instantly todifferent sound levels.

9

6

013

12

0

*I)

6

5

.15V

T1 ... T8= BC 550CT9 = BC 560C

0 1

(0 dB)

02

(- 2.7 dB)

03

1-6.7 dB)

-04

(-11 dB)

05

(-15 dB)

06

(-20 dB)

07

(-30 dB)

08

(-40 di3)

85470-1a

The indicator is constructed on two printed circuitboards (figures 2 and 3). The LED board in figure 3has not been provided with a component layoutbecause of aesthetic considerations. The layout is,however, given on the PCB in figure 4 for thosewho want to use it all the same. The two boards canbe fitted together with the use of spacers: appro-priate holes have been provided for this in a mannerwhich ensures that the 11 terminals for interconnec-tions on the boards are opposite one another.An interesting optical effect arises when a sheet ofred perspex is mounted in front of the LED board.Refraction in this material causes the LEDs to showup as sources of diffused, rather than pinpointed,light.

167

The current consumption of 800 mA at saturationmay be reduced by lowering the supply voltage to,say, 12 V, but this will, of course, reduce thebrightness of the display.

Parts list

Resistors:Ri = 270 kRi*, R14 = 10 kR3 = 100 kRa . . .Rs,Ris. . .R27 =15kR9 = 22 kR19 = 1k8R11,1312 = 27 kR13 = 18kR15 = 8k2R16 = 6k8R17 = 2k2R18= 1kRes = 820

= preset potentiometer,250 k

15 V

D68

44-

0

lb

+15V

D53

D45

D37

D29

D21

67

D60 D52 044 D36 D28 D20 01

059

D6

051

D4 ...D68 = LED red

D43

D35

D27

D19

013

8

D11

DS

010

D18

026

D34

D42

050

D58

+4

Capacitors:= 560 n

C1* = 0...10 µ/16 VC3 = 47 µ/16 VC4 . . . C6 = 100 n

Semiconductors:Ti ...To = BC550CT9 = BC560CDI,Di = 1N4148D3 = zener diode 5V6/400MWD4 . . . D69 = LED red1C1,1Ci = LM324IC3 = TL084

* = see text

PCB 85470-185470-2

D62

D54

D46

D38

D30

D22

D14

06

D9

015

D7

D23

D31

D39

D47

D55

D63

De 016 D24 D32 D40 048 D56

D17

D25

D33

041

049

D57

85470.16

D64

168

,

LIZ,0-,Agsl

ft:A:Amr31 lialirPril"Iimir.7100.11Pir

s N.:jrAtO

CEPS 85470-2

" INaamloos-6.indd 13 28-08-2008 10:07:05

3

141 SWELL PEDAL

Reminiscent of the accelerator pedal in a car, a swellpedal enables musicians to alter the sound volumeby foot, since they invariably need both hands toplay their instrument. Electronic organs have theswell pedal normally built into the front near theother pedals. Guitarists have to buy this almost in-dispensable aid for getting the right blend of accom-paniment and solo voice(s) as an optional extra.From an electronic point of view, such commercial-ly available devices are simplicity itself: normallynothing more than a potentiometer operated by thefoot pedal via a toothed bar. The mechanics, how-ever, make home construction a rather more daun-ting task. The swell pedal described here avoids themechanical intricacies.

170

The circuit is entirely contained in a flat case ofabout the shoe -size of the user - see figure 1. Awedge-shaped, hollowed -out piece of foam rubber isglued onto the lid of the case. A light -emitting di-ode, D5, and a light -dependent resistor, LDR, pro-trude from the lid. A small sheet of metal or plastic,the underside of which is covered with white paperor cardboard, is then glued onto the foam rubber.The top of the metal or plastic sheet may be covered(glued) with a small rubber mat.When the foam rubber is compressed by footpressure, the reflective white paper or cardboardcomes nearer to the LED and LDR, which causesthe resistance of the LDR to diminish. Because ofthe amplifying, inverting, and compensating action


1

171

of IC1, a voltage is applied to IC2 which is used tocontrol the drive current provided by transistor Tifor OTA (operational transconductance amplifier)IC3.After the pedal box has been glued together, so thatthe electro-optical components are in a light -proofchamber, adjust Pi so that with non -operated pedal

the sound volume is just at the right level for ac-companiment. For solo playing, the pedal is

depressed as required to obtain the increased soundvolume. It is advisable to fit P1 in the side of thepedal case as shown, so that it can be re -adjusted ata later date if required.

142 WAH-WAH BOX FOR GUITARS

In this day and age of electrophonics, a wah-wahbox is still a popular means of animating an other-wise tired sounding guitar. Such a box, which isbasically a high -Q low-pass or band-pass filter, canbe designed in various ways. Early designs were in-variably based on active (transistorized) double -Tfilters.The present circuit, using opamps and operationaltransconductance amplifiers (OTAs), is rather morecomplex but also more efficient and more reliable.Three pairs of opamps, each consisting of an OTAand a buffer amplifier, in conjunctions withcapacitors C2, C3, and C4, form a low-pass filter.Since the usual series resistances have been replacedby voltage -controlled current sources (OTAs), theroll -off frequency of the filter is determined by thecurrents flowing into pin 5 of the 3080s. These cur-rents are themselves directly proportional to the in-put control voltage, Uc, which has been converted

,J)IC2

ICI IC3 IC5IC4

CIL CL

15V

221

CB

TAI A4= ICI =TL084 15V

85509

in Ai and T1. This voltage, which is derived from aswell pedal, can have any value between 0 V andabout 12 V.The negative feedback from output to input enablesthe Q of the filter to be set with P2The swell pedal may be constructed as describedelsewhere in this issue: it can actually be installed inone case together with this wah-wah filter!As it is difficult to describe sounds, and we are surethat the guitar players among our readers will inany case experiment themselves, we will not dwellon what to expect from this musical adjunct. Nocalibration is needed: the box works or it does not!

172

143 ELECTRONIC VHF/UHF AERIAL SWITCH

Ant. 1

There are many situations where it is useful, ordownright essential, to be able to switch betweentwo VHF/UHF aerials at the aerial mast withoutintroducing losses in the signal paths. The switchproposed here does all this over the usual coaxialdown lead.The switch and its small associated power supplyare fitted near the relevant receiver. The powersupply, consisting of a small mains transformer, arectifier diode, and a three -pin voltage regulator,provides a direct voltage of 5 V, the polarity ofwhich can be reversed by DPCO (double -polechange -over) switch Si. The poles of the switch areconnected to the coaxial cable via decoupling net-work L3-Cl. Resistor Ri serves as a current limiterfor p-i-n diodes Di and D2. Whichever of thesediodes conducts depends on the polarity of thevoltage across the coaxial cable. The signal from theaerial connected to the conducting diode is passedto the input of tuner or receiver, while the othersignal is blocked.A p-i-n diode is a semiconductor diode that contains

9V100mA 85440

11717IMMMI,.000 0 I

a region of i-type semiconductor between the p -typeand n -type regions. They are invariably used asswitching diodes. Their most important property isa very low self -capacitance, while at high fre-quencies they are virtually purely resistive (seeElektor, June 1983, p. 6-36).Choke L3 is made from four turns enamelled copperwire of 0.3 mm dia. around a ferrite bead. If theaerials have no 75 Q termination, this may be pro-vided by L i and L2 which convert the 300 Q balanc-ed aerial impedance to the asymmetrical 75 Q re-quired by the receiver input. These inductors aremade by winding 7 turns of two -core flat cable ona T50-2, T50-3, or T50-6 toroid as shown in fig-ure 2.If the switch is mounted in the open, it should bewell protected from the elements: potting in aralditeis best.

144 FOUR-WAY AERIAL SWITCH

In many cases it may be necessary to switch be-tween two or more aerials with minimum loss in theRF signal. Though this is not generally a problemat low frequencies, it becomes a serious one whenthe relevant signal is in the VHF/UHF range (50-960 MHz). The electronic switch described herekeeps the switching losses minimal by making useof PIN diodes. PIN diodes are essentially current

controlled resistors with properties that make themsuitable for switching and attenuating RF signals.They differ from most other types of diode in thatrectification of the input signal only occurs below acertain limiting frequency. Above this frequency,the resistance of a typical PIN diode will changefrom IQ to 10,00052 when the control current is re-duced from 100 mA to 1µA.

173

The circuit can switch up to four aerials, and iscomposed of two functional parts: the RFswitching section, mounted onto the aerial mast,and the power supply & control section, kept nearthe receiver. In this way, the cost of setting up amulti -aerial system is reduced to some extentthanks to the use of a single downlead cable, insteadof as many as there are aerials.The required aerial is selected by biasing the corre-sponding PIN diode into conduction. Which of thefour diodes conducts depends on the level and thepolarity of the voltage applied to the switching unitvia the downlead cable to the receiver. When, forexample, input 1 is selected with Si, the voltage onthe core of the downlead cable is + 12.7 V withrespect to the cable screen, and can not reach thecircuit around T3 and T4 because D6 does not con-duct. The level of the positive voltage causes zenerdiode D7 to conduct, and so provides a bias fordriving it into saturation. Ti in turn provides the re-quisite bias for PIN diode Di , and at the sametime prevents T2 from conducting. Input 1 is thusconnected to the common output of the switchingunit, through Di . If Si is set to position 2, thesupply voltage on the downlead cable falls to 8 V,which is insufficient for D7 to conduct. Ti now re-mains switched off, and T2 is driven into saturation,providing the required bias current for the associ-

s,

1

2

3

4

ated PIN diode, D2. Diodes D8 and D9 prevent Difrom being biased through R2 and the base -emitterjunction of T2. Input 2 is thus connected to thecommon output through D2.Similarly, when the voltage on the core of thedownlead cable is negative with respect to thescreen, the circuit around T3 and T4 works asoutlined above, with either D3 or D4 conducting,depending on the level of the voltage (-8 or-12.7 V).Inductors L1 -L6 prevent the RF signal from beingearthed anywhere in the circuit, while L7 preventsit from being short-circuited in the power supply.For VHF applications of the circuit, 5µH inductorsor chokes should be used in the Li -L7 positions,while 21,11-1 types are required for UHF operation.The RF signal from the selected aerial is passed tothe receiver input through C19, which serves toblock the direct voltage. In case balanced aerials areto be switched, their outputs must first be made un-balanced and, if necessary, transformed to 7552,using a balun.

Up

.12V78V

- 8V-12V7

174

145 FRONT-END FOR FM RECEIVER

Among the most important technical characteris-tics of a VHF preamplifier are the noise figure, andthe large signal handling capability. Although theseare in principle conflicting requirements, a compro-mise can be found in the use of high -quality RFcomponents. The receiver's ability to withstandhigh input levels can be enhanced by providing suf-ficient selectivity ahead of the active element(s).This is especially important for the mixer, since itgenerates most intermodulation products.In this FM tunerhead, the aerial signal is firstpassed through a slightly overcritically coupledband filter, amplified with the aid of low noise UHFtransistor Ti, and again filtered. The overall gain be-tween the aerial input and the mixer input is about12 dB at 87 MHz, and 17 dB at 108 MHz. The dif-ference is caused by the adopted method of filtercoupling. A wideband Schottky DBM (doublebalanced mixer) is used for the mixer in this design.The Type SBL-1 (LO = + 7 dBm) is probably the

best available of the 3 DBMs stated. Tuneable localoscillator T2 produces very little phase noise, andDG MOSFET T3 provides a LO power of 50 to100 mW at a drain current of about 25 mA. FETT4 enables driving a prescaler or a synthesizer withthe LO signal. Series network R9 -C20 is fitted at theinput of the IF amplifier because any passive DBMshould be correctly terminated on at least two of itsports. To compensate for the 6 dB conversion loss inthe DBM, and to ensure some spare IF gain,medium power RF J-FET T5 is dimensioned to pro-vide a gain of about 12 dB at a drain current of25 mA.The proposed front-end gives fairly good results: itsthird -order intercept point is better than 0 dB whena mixer is used with IP = + 20 dBm, while thenoise figure is about 4 dB. This sort of performanceshould enable the reception of quite weak transmis-sions even with a powerful transmitter within a fewmiles from the receiver.

15V

175

Finally, due account should be taken of the factthat the IF output easily delivers 10 mW, whichmay well give problems if the IF amplifier is notproperly dimensioned.Inductor data for this project:Li .. .L5 incl.= E526HNA10014 (Toko).L6= E526HNA10013 (Toko).

L7 ...L9;L14 = 6 turns 36SWG (0 0.2 mm) enam-elled copper wire through a ferrite bead.L,1= 9 turns 24SWG (0 0.6 mm) enamelled cop-per wire on a T25-12 ferrite core; tap at 3 turns fromC35 -R15 -R16.

146 FRONT-END FOR SW RECEIVER

There are many conflicting technical requirementsfor a good -quality front-end in an SW receiver. Thenoise figure and the intermodulation level should below, the RF insulation between ports LO, RF andIF should be high, and some amplification is

desirable. The Type SL6440 high level RF mixerfrom Plessey ensures a noise figure of around 10 dB,and offers sufficient suppression of the LO signal.The signal applied to the RF input (B) of the front-end is passed through a low-pass filter with a cut-offfrequency of 32 MHz and an output impedance of500 Q. The open collector output of mixer IC, hasa relatively high impedance, which necessitates theuse of Tr, and R5 for correct matching to 48 MHzcrystal filter FL,. The fixed impedance of this filterfor signals outside its pass -band helps to keep the in -

CTCA14440

2

Tr2

111 FL1YF 48D20

termodulation distortion low. Trimmers C13 and C14are aligned for a maximum flat pass -band atminimum loss. The mixer's intermodulation charac-teristics can be optimized by careful dimensioningof R, and R2, provided the amplitude of the localoscillator signal is stable. A third -order interceptpoint of 33 dBm was achieved in a prototype. Themixer IC gets fairly warm, and should be cooledwith a heat -sink.The RF transformers are wound as follows (use30SWG enamelled wire):Tr,: the primary winding is 10 +10 bifilar turns, thesecondary is 10 turns, on a Type T50-12 ferrite core.Tr2: the primary winding is 2 turns, the secondary18 turns, on a Type T50-12 ferrite core.Ls: 6 turns through a ferrite bead.

C10

AO®R3

47n

BOO

II1n

5

* see text 15V 0R4 pm

L6

14 3

R2 Ri g11

C11 C12

ism um22n 470n

A = Local oscillator inputB = Octave filter input

IC1SL 6440

13 12

L1 L2 L3 L4 Ion la .= MI

3pH3 3p 113 3 013 3 pH3

L51C4 C6

1p5 1p5

C2INN

7pC3 C5mIm mom

713p 7pC7 C8

Tilap 720n

C9Imo

Ion

176

147 HIGH LEVEL PASSIVE DBM

The mixer is one of the most important sections inany good -quality SW receiver, since it determines toa large extent the sensitivity and the dynamic range.The so-called switching mixer is often used, becauseit has none of the technical imperfections of activemixers. The most commonly found switching mixeris the diode -based double balanced type (DBM),which is, unfortunately, a notoriously expensivecomponent, especially when a high intercept pointis required to s-nsure low levels of intermodulation.The application of active devices, such as bipolartransistors and J-FETs, in a passive mixer is less wellestablished. And yet, these components enable themixer to remain relatively simple, since the RF in-put signal can be thought of as electrically insulatedfrom the local oscillator output. The present designis based on a pair of well-known UHF transistors,which require no supply voltage or bias circuits.The input and output transformers are wound ontwo -hole ferrite cores (Baluns). The primary of Tr2is 8 turns with a centre tap for the RF input, the

LO

T1, T2 = BFR 91, BFR 96 87983

secondary is 4 turns. Tr, is wound such that the in-dicated LO amplitude is available at the secondary.Only the RF input or the IF output requires correcttermination on 50 Q, the other connections arethen fairly uncritical. The input intercept point ofthis mixer is excellent at between 31 and 36 dBm,while the noise figure and conversion loss are ac-ceptable at about 6 dB. The ID rejection is roughly25 dB, and depends mainly on the construction.The mixer is suitable for RF and IF signals up to30 and 50 MHz respectively.

148 HIGH LEVEL WIDEBAND RF PREAMPLIFIER

A linear RF amplifier can be made in two ways: (1)with the aid of a linear active element, or (2) witha non-linear element operating with negative feed-back. This circuit is of the second kind, using an RFpower transistor as the active element. Feedback is

C2also required to ensure correct termination (50 Q) of inimi

the aerial, since bipolar transistors normally exhibit 470n

a low input impedance. Also, the noise figure is notincreased because virtually no signal is lost.The common -base amplifier is based on a UHFclass A power transistor Type 2N5109 fromMotorola. The feedback circuit is formed by RFtransformer Tr:. The input and output impedanceof the preamplifier is 50 Q for optimum perform-ance. Network R3-05 may have to be added topreclude oscillation outside the pass -band, whichranges from about 100 kHz to 50 MHz. The gain isapproximately 9.5 dB, the noise figure is between 2and 3 dB, and the third -order output intercept pointis at least 50 dBm.The input/output transformer is wound on a Type

D1,D2,D3 = 1N4148

* see text

87477

FT37-75 ferrite core from Micrometals. The inputwinding is 1 turn, the output winding 5 turns witha tap at 3 turns.

177

149 LOW NOISE AERIAL BOOSTER

After having read the design essentials relevant towideband amplifiers, RF filtering, intermodula-tion/crossmodulation characteristics, etc., as givenin the articles listed at the end of this article, therewould seem to be little need for us to dwell on func-tional and electronical aspects of the present ultralow -noise, wideband preamplifier incorporating thewonderful Type BFG65 transistor, which, althoughalready introduced in [3], deserves to be put in theRF limelight as it offers an exceptionally low noisefigure at more than satisfactory strong signalresponse, thanks to the relatively high collector cur-rent (Fa = 0.8 dB at 5 mA, for instance).Since the important points to observe in RF con-struction have been covered in [1] and [2], the largeearth plane on the component side of the ready-made PCB Type 86504 need not cause any wonder;

BFG65

all parts are soldered direct onto the relevant copperfields; the holes merely serve to aid in locating theparts correctly. The hole for Ti should be drilled todia 5 mm for the transistor to be seated andsoldered with the shortest possible lead length.Additional holes have been provided to enable theinput and output coax cables to be secured bymeans of screw -on clamps, although solderingscreen and core should also be possible.It is seen that the ready-made PCB consists of anRF and a supply section, which may have to beseparated by cutting if it is desirable to fit the unitsat different locations, as is the case with a mastheadmounted amplifier and the supply located at thenearest mains outlet, e.g. on the attic. On the otherhand, if is more convenient to cut the downleadcable immediately as it appears indoors, amplifier

12V 5...15mA

50...750 50...75n

12V

F1 50mA

D1...D4 = 1N4001

178

L.ro Jw

lei'

*.

and supply may be left to form one unit for inser-tion in the coax cable. As in that case the amplifiermay be fed direct rather than via the coax cablecore, L4, L5, C5 and C6 are rendered unnecessaryand may be removed; the free lead of R5 shouldthen be connected to the +12 V terminal on thesupply section of the board.The optimum collector current for Ti is adjusted bymeans of P which should be set for a value be-tween 5 and 7 mA if the amplifier is to handle rela-tively weak signals, such as may be received infringe areas. The indicated collector current cor-responds to 2.3 to 3 V voltage drop across R5;

higher values (10 to 15 mA; 4.6 to 6.1 V respect-ively) should be set when receiving two or morestrong (local) transmissions in the 80 .. . 800 MHzband.If masthead -mounted, the amplifier should be fittedin a waterproof enclosure, carefully treated withsilicone spray to preclude corrosion of the soldercontacts.Finally, the coils are wound as follows, using dia0.3 mm (30 SWG) enamelled copper wire:Li: 8 turns, closewound, internal dia 3 mm.L2: 4 turns, closewound, internal dia 3 mm.L3: 5 turns on R4.

L4;L5: 4 turns through 3 mm ferrite bead.

Literature references:[1] VHF filters

(EE, March 1986, p. 50 ff).[2] VHF amplifier

(EE, April 1986, p. 40 ff).[3] Wide band amplifier for satellite

TV receivers(EE, April 1985, p. 66 ff).

[4] Aerial amplifiers(EE, February 1980, p. 27 ff).

Parts list

Resistors:

Ri =1k8R2 =18 kR3 = 330R4 = 820R5 = 470 52131=5 k preset

Capacitors:

Ci;Ca;C5= 68 pC2;C3= 680 pC6 =1 nC7=1/416 V; electrolyticC8 = 470 11;25 V; electrolyticCs;Cio =47 n

Semiconductors:Di...D4=1N4001ICi =78L12

= BFG65 (Philips/Mullard)

Miscellaneous:Li ... 1_5= see text.Tri = 12 V;50 mA.Fi = 50 mA; fast.PCB Type 86405.4 soldering pins.

179

150 MORSE FILTERS

Morse, or CW (continuous wave), is still widelyused thanks to the fact that the necessary equip-ment can be kept relatively simple, and therefore in-expensive, if the operator is sufficiently trained inselective listening. A morse decoding computer,however, requires an adequately filtered inputsignal, because it lacks the noise discriminatingcapability of the human ear. Some receivers can beupgraded with a 250 Hz IF filter for this purpose,but such an extension is usually well beyond thefinancial reach of most radio amateurs. The filtersdiscussed here operate in the audible frequencyrange, and compare favourably with far more ex-pensive types for 455 kHz. Figures 1 and 2 showthe circuit diagram and the typical response of aneighth -order inverse Chebishev filter which hasbeen optimized for non -computer using listeners.The filter of Fig. 3 is less complex, and intended fordriving a computer. The associated frequencyresponse is shown in Fig. 4. Both filters were de-signed with Eldesign Ile, an advanced filter designprogram for the BBC micro. The inverse Chebishevresponse gives a smooth pass -band, while the

1

671

C7INNEN 1.0

Mr; n

17

0

R6

6R20

154

R22

characteristic ripple ends up in the stop band. Thisensures the required phase stability in the pass -band, which is a must for processing burst -likesignals such as morse.Prototypes of the filters gave excellent results: nor-mally hardly audible signals could be recovered forreliable decoding. The supply for the filters is

preferably a symmetrical 15 V type to ensure an op-timum dynamic range. Do not use any other opampthan the LM324, since types with a higher cut-offfrequency may give rise to oscillation. Note that Clin Figs. 1 and 3, and C2 in Fig. 1, is a parallel combi-nation of two capacitors from the E12 range ofvalues, while all resistors used are from the E96range. Should any of the filter sections persist in itstendency to oscillate, either one of the even-num-' bered opamps may have to be dimensioned for aslightly different roll -off point by connecting a100 pF capacitor across the output and the -input, and a 390 Q resistor between the

- input and junction C3 -R5 -(-A1) (example refers toopamp A2).

R24

1C2 1C1

Al... A4= 1C1 =Lt4 324A 5 ...A8 =1C2 =UN 324

87489,

15V

15V

180

260=18.3211

r 194285.7 4 hzi257 8.9

229 ...... 8.8

288 1.7

172 1.6

443

L8.5 ...

114 I., i;85.8

57.2 8.2

8.1.........

.. .

3

'441 ; Hz.

= 1% = 5%, MKT A1...A4 =IC 1=LM32 4

A5 =IC 2 =LM 312

481/ J688 488 .1888

15V

1C2 1C1

15V

0

151 MULTI -MODE µP -CONTROLLED IF MODULE

The intermediate frequency (IF) module shown inFig. 1 accepts 48 MHz, and is suitable for receivingAM, FM and SSB transmissions. CW receptionshould also be possible in the SSB mode when a suf-ficiently narrow bandfilter is included(BW<500 Hz). For radio -teletype (RTTY), it is bestto drive a comparator from the FM detector output.There is no need for a high level mixer to convertthe input down to 455 kHz, since the 48 MHzsignal has already been filtered and occupies abandwidth of only 12 kHz. The RF and mixerstages in the TCA440 operate up to 50 MHz, whilethe built-in AGC has a dynamic range of about100 dB. The mixer output is fed to diode switchesto enable digital selection of the appropriate band-width.The proposed selection circuit ensures a filterseparation of the order of 80 dB. The choice of the455 kHz filters is governed by the particular appli-

cation and the financial means available. The CLF-D12 and CLF-D2 are for FM/AM and SSB respect-ively: the number in the type indication stands forthe bandwidth. The Type CLF-D4 or CLF-D6 canbe used equally well for communication qualityAM. Unfortunately, narrow -band filters for CWand RTTY are difficult to obtain, but "add-on"500 Hz or 250 Hz filters for commercially availablereceivers and transceivers (Yaesu, Kenwood) can beused here with excellent results.The IF output from IC2 is rectified for the AM andAGC sections, and inductively fed to FM detectorIC4 as well as to product detector IC3. Note that ingeneral no AGC action is required in the FM andRTTY mode. The BFO for the product detector isbased on USB and LSB ceramic resonators, whichare found in most SW receivers of Far Easternorigin, but may be difficult to obtain as a one off.The circuit around T4 is a voltage -controlled

181

1

2 kHz

7

12 kHz

USB

C) 0LSB

0

r

FM

)0Squelch

) 0AM

.)0

AGC

0A = D -A converter

1N 4148D2 D3

1914148 j

D2' 03'1 04148

KV1213

TCA 440

NOM

472

C53

682

1N 4148

470211 7002

71

BC 549 C

C30

-I115p

1118

R20

C28

4 2/3v

PLO a 453.5001FL9 a 458.5081

LAICS 4102A

702

=OmC49

039 100n270 I77E-2

2z AA 119

'Cl

D6

10 4148 A

ES4

551...554 = IC1 = 406613

5 042 P

BF 494

0

1N4148

C 8

220p

COO

2202

C42

OP

100k

ES2

XI = 48MHz (3rd overtone)//20p

LM 38648/2,4

87494

182

NI N6 = IC I = 0009BNI NI2 = IC 2 = 0049 8

11 T13, T15, T17, 119, T21,T14, T16, T18, T20, T22, T20

123= BC 50788C 557 87488

48 MHz crystal oscillator (VCXO) that operates inthe parallel mode, requiring due attention to be paidto the correct output frequency if a common, series -resonant crystal is used. The synthesizer for tuningthe proposed receiver outputs 1 kHz steps, so thata D -A converter is required for driving the VCXOinput. A resolution of 10 Hz should be adequate toensure smooth and reliable tuning.The computer interface for controlling the receiveris shown in Fig. 2. This is essentially a 5 V to 15 Vlogic level converter with TTLICMOS compatiblecontrol inputs. The remote control of the receiverobviates the need for this to be housed in a neatenclosure. Albeit that the receiver therefore neednot have a "desktop" appearance with all the con-trols fitted on a front panel, it is, of course, still

necessary to provide for adequate screening andthermal stability. Sufficient AF power is availablefrom 105 to drive a relatively long cable to the loud-speaker enclosure, which is located near the com-puter.The receiver can be controlled from any computerthat has three 8 -bit output ports based on, for in-stance, Type 74LS374 octal latches. A receiver func-tion is enabled when a logic 1 is written to therespective input. Example: 4 kHz bandwidth isselected by driving BANDWIDTH input 4 high, and theremaining five low. Writing a computer program forcontrolling the receiver should not be too difficultif the following sequence is observed: 1. actuate thesquelch; 2. reset all bits on the relevant control port;3. set the required bit; 4. turn off the squelch.

183

152 NOISE BLANKER

A noise blanker is indispensable for improving thereception of very weak signals on the SW bands. Inmost communication receivers, the selectivity of in-termediate frequency (IF) filters cause interferingpulses to be widened, blotting out the wantedsignal. It is useful, therefore, to suppress inter-ference before this can wreak havoc in the IF ec-tions of the receiver.The 455 kHz IF signal is first buffered in Tz, andthen processed separately in two circuits.The lower section of the circuit is a TCA440 basedreceiver for the interfering pulses. The TCA440 is initself a virtually complete receiver, since it com-prises an RF amplifier, a mixer, and an IF amplifier.All stages in the latter are used since pin 4 isgrounded here. The pulse receiver has its own AGC(automatic gain control) to ensure effective suppres-sion of relatively weak interference also. Preset Ptand potentiometer P2 enable precise adjustment ofthe noise blanker for various levels of interference.

ES 1 ...ES4 = IC3 = 4066 B

017

R1t1

TCA 440

C13mimmem

100n

FL I

LF .13

C14MEN

700n

L1

LMCS 4102 A

AA 119

DI

The circuit can be controlled digitally via R23; alogic high level renders the noise blanker ineffective.The interfering pulses are made logic compatiblewith the aid of opamp IC2. LED D3 lights whennoise is detected.In the upper section of the circuit, the IF signal isfirst delayed in FLi to compensate for the process-ing time in the pulse receiver. ESt is opened whena sufficienly strong interfering pulse is recognized,so that the IF signal is no longer applied to outputbuffer Tz. Also, the gate of this FET is thengrounded for RF signals via ES3-C4, while ES2 isclosed to maintain correct termination of FLi.Properly constructed, this circuit achieves noisesuppression of the order of 85 dB. Alterations tosuit operation at an IF other than 455 kHz involveLi and FLi, although due account should be takenof the parasitic capacitance of the electronicswitches at relatively high frequencies.

ES1

C4 015V

770n 100n 100n

R25

1022

C16

4024

CI7

10$] 3V

IC 2

CA3130

15V

R14

10n

Cl

560n

730n03

87492

184

153 NARROW -BAND IF FILTER

680p 220p

-111.see text

3300*

1

n21 1.680p

TI TTSince good crystal filters are expensive, there is aconstant search for (less expensive) alternatives.One of these is the ceramic filter, now widely usedas IF filter in short-wave receivers. The somewhatpoorer temperature characteristics of ceramic filters(as compared with those of crystal filters) are nor-mally not of much consequence.Numerous experiments have finally led to the cir-cuit of Fig. 1, which uses five 455 kHz ceramicfilters. As computer crystals can be obtained cheap-ly nowadays, it would also be possible to constructa similar filter with a number of such crystals.

680p

IT

560p 680p

II -11 II-

*

86442-1

3300

The result of our experiments is a 3 dB filter band-width of about 800 Hz; the attenuation outside thepass -band is of the order of 60 dB.A possible application is its use in a receiver withvariable bandwidth for SSB, AM, and FM oper-ation.Another application is as input filter in a receiverwhose dynamic frequency range is inadequate (butthe IF should then not be 455 kHz).Finally, note that correct matching of both the in-put and output impedance (330 ohms) is imperative.

154 NAVTEX RECEIVER

NAVTEX, the international maritime service thatprovides navigational and meteorological infor-mation via RTTY (radio teletype) on 518 kHz,makes use of FECTOR. This is a system in whichthe information is transmitted twice, with a par-ticular interval between the first character and therepeat. FECTOR is decoded automatically by a mi-croprocessor that is coupled to the ship's mediumwave receiver.It is, of course, not desirable that the decoder istaking up the medium wave receiver continuously.On the other hand, navigational officers, and manyamateur radio listeners, do not want to miss oneiota of NAVTEX information. Obviously, a secondreceiver is the answer, and this can, of course, becoupled to the decoder night and day. Since onlyone frequency, 518 kHz, and one type of trans-mission, FSK (frequency shift keying), needs to bereceived, the circuit can be kept quite simple.The circuit is based on a type ICA440. The AGC(automatic gain control) provided by this IC is not

used because the IF amplifier, due to its internalsymmetry, is already an excellent limiter for FSKsignals.The internal oscillator is not used either: it is re-placed by a crystal oscillator, T,, operating on5185 kHz, that is followed by a decade scaler, IC2.The exact frequency of the crystal depends on therequirements of the decoder; trimmer C3 enables itto be varied by a few kHz, i.e., a few hundreds ofHertz at the output.Thanks to the TCA440, the remainder of the re-ceiver is fairly simple without the need of specialcomponents. Standard chokes can be used in theL2 . . .L4 positions; Li consists of 6 turns enamelledcopper wire of 0.3 mm dia. on a ferrite bead.Sensitivity of the receiver is good at a few AN.Calibration is very simple: adjust input trimmers Ciand C2 for maximum output, and then turn C3 un-til the output frequency matches the decoder.The crystal should be suitable for parallel resonancewith a capacitance of 30 pF.

185

C

.10010%1

73 p

mom1y16 V

L:a-47 gmH 47 mH

L1*

C2i

16 1272" Its /14 6

i C1

TCA 440

631000

CJ 4]u

C13

*- 5185 kHz

Current consumption is not greater than 10 mA.The supply voltage may be 4 . . .15 V.It is, of course, a fairly simple matter to make the

co 1C2 c40 76

13? 6 11

C12

100n

67 do

BF 494

72 0 P

® 4 ... 15 VCel < 10 mA

=900n

85501

11.=.1 5185 kHz

40p

O

receiver suitable for use on other maritime mediumwavelengths.

155 RTTY CALIBRATION INDICATOR

To calibrate an RTTY (radio teletype) decoder cor-rectly in accordance with the marks and spaces, anoscilloscope is needed. The mark and space signalsare applied to the X and Y inputs of the instrumentrespectively, when, on correct calibration, the

MARK

screen of the oscilloscope displays the well-knownRTTY cross.If an oscilloscope is not available, the circuit shownhere can be used. It consists of two amplifiers withhigh -impedance input, T1 and TS that are followed

BF 245CD N74

1111 El] BC 550C s-.11,

O 12 V60 mA

BF 245C

G SD

C2R130

SPACE

7.1

T ®85495

186

by driver stages T2 . . T4 and T5 . . . T7. The driverstages control three LEDs, Di . . D3 direct. Diode131 (red) is the mark indicator, D2 (green) is the spaceindicator, and D3 (amber) indicates whether thedecoder has been calibrated symmetrically.Preset potentiometers Pi and P2 determine the am-plification of the field-effect transistors. Proper set-ting of these components enables the indicator tobe matched with the filter outputs of any RTTYdecoder.After the indicator has been coupled to the RTTY

decoder, that unit can be calibrated as follows: tune the short-wave receiver to the marks; the

BFO knob must be adjusted until the red andamber LEDs both flash brightly; the RTTY decoder is then adjusted to the correct

frequency deviation, indicated by the flashing ofthe green LED. If the amber LED lights continu-ously, the decoder has been calibrated correctly.Otherwise, the above procedure should be repeatedcarefully.

156 RTTY/CW FILTER

An appreciable part of short-wave radio traffic takesplace via morse and radio teletype transmission. Toensure optimum reception of these types of trans-mission, a practical bandwidth of about 300 Hz isrequired in the receiver. Such a bandwidth allowsfor some drift of both transmitter and receiver, andalso for the frequency shift of RTTY signals. Ascommercially available filters meeting these re-quirements are still rather expensive, it pays to buildyour own: a suitable one is shown in the accompa-nying diagram.The crystals used are inexpensive types, commonlyfound in computer systems.Inductor Li is made by winding 2 times 20 turns en-amelled copper wire of 0.3 mm diameter onto a

xl X4 = 2AS-76 MHz

I II j 111

L1

C1 C2

100p

C

CZ CA CS CS10o .... min MIMI

6 SP 68P 685 335

T50/2 RF toroid (available from Cirkit).Some parameters of the filter are: bandwidth at -6 dB points : 300 Hz bandwidth at -60 dB points : 1100 HzII insertion loss : 7 dBII ripple in pass -band : 1 dB

lkn

157 S METER

Since many amateur receivers are fitted with an Smeter that functions far from logarithmically, theproposed circuit should be a welcome extension ofsuch receivers.Although ICs such as the CA3089 or the CA3189are not in common use any more, they serve auseful purpose in the meter circuit, because, apartfrom a symmetric limiter, a coincidence detector,and an AFC amplifier, they contain a very goodlogarithmic amplifier -detector.As is seen, the circuit is fairly simple, but rememberthat these ICs operate up to about 30 MHz, so thatthe wiring of the meter, and also its connections inthe receiver, should be kept as short as possible.

12V< 40mA

024251

187

Note further that the input of the CA3189 must be terminated by

50 Q; the connection to the input of the CA3189

should be in screened cable; if it is not possible to obtain the input signal from

a low -impedance source, a source followershould be used between it and the meter circuit. 7114.7726,-

100 1 k 1 Ok

NPUTSIGNAL

158 SEND/RECEIVE IDENT

12 V

PTT

transmit/receive

12 V

12 V

Some radio amateurs like to give an identificationsignal at the beginning and end of a message; othersfrown upon this practice which they find disturb-ing. If you belong to the first group, you may findthis circuit useful as it gives an ident signal auto-matically when the transmit/receive key is pressed

R5 C5 R6

CELT220n

1:15

1N4148

12V

mike

and just after this has been released again. The twosignals are identifiable by being slightly different infrequency.XOR gate N, functions as a monostable, whoseoutput is high for a short time after its inputs eitherchange from high to low (at the onset of a trans -

188

mission), or from low to high (at the end of a trans-mission). Its output is applied to an oscillator,N2/1\13, and to the transmit/receive switching sec-tion. When the input pin 6 of N2 is high, this XORgate functions as an inverter, so that the oscillatorgenerates a short tone in the medium audio rangewhich is fed to the microphone via limiter D4/D5.The frequency determining network is earthed viaCl and Di or via CI and R1, depending on whetherthe transmit/receive key is pressed or has just beenreleased.

During transmission, the rx/tx output is low: thisoutput is intended to be connected to the corre-sponding input of the transceiver. Transistor T2 ison, so that relay Re, is actuated: its contact(s) maybe used, for instance, to disconnect the loudspeakerduring transmissions.Current consumption, ignoring the relay current,amounts to about 15 mA.

159 SIMPLE FIELD STRENGTH INDICATOR

A practically proven small circuit that is verypopular with many model fliers, as it enables themto verify that their remote control transmitter isactually transmitting. Any doubt as to whether afault lies in the receiver or transmitter is alsoquickly resolved.The only active element in the circuit is a transistorthat is used as a controlled resistance in one of thearms of a metering bridge. The base of the transistoris connected to the wire or rod aerial. The increas-ing HF voltage at the base of the aerial drives thetransistor so that the bridge is brought out ofequilibrium. A current then flows through R2, the

0.5 1 in

mA meter, and the collector -emitter junction of thetransistor. The meter should be zeroed with P1before the transmitter is switched on.

160 SPOT FREQUENCY RECEIVER

Monitoring a number of frequencies in the short-wave band, such as the international shippingdistress frequency, is a fascinating pastime. Sinceonly a limited number of stations is normallymonitored, and their frequency is invariably fixedby international treaty, the receiver needs only to becapable of being switched between those spot fre-quencies.The receiver works on the direct conversion prin-ciple, i.e., the oscillator frequency is equal to the re-ceived frequency, so that the intermediate fre-

quency is zero.The aerial signal is fed to tuned RF amplifiers Tiand T2 via a switched preselector. The RFamplifiers are coupled to an S042P type mixer.There are three crystal -controlled local oscillators,which are switched into circuit in accordance with

the preselector.The output of the mixer is the audio signal, whichis fed to AF amplifier IC2 via low-pass filterRil. .R13 -C28. C30. The gain of IC2 is about60 dB.Part of the output of IC2 is rectified in 131 and D2and used for AGC (automatic gain control) of Tiand T2The output of IC2 is fed to power amplifier IC3which drives a loudspeaker or headphones. There isalso a tape output. Volume control is provided by13%

Inductors Li . .L3 are each wound on a T50/2toroid as follows: L = 115 turns enamelled copper wire of

0.15 mm dia. with tap at 11 turns; L2,L3 = 90 turns enamelled copper wire of

189

12V 12 V 12 V

3042P

CZ C9

?TOP 409

0.2 mm dia. with tap at 9 turns.If different frequencies from those shown are re-quired, one or more of the crystals must, of course,be replaced, but at the same time Lt, L2, or L3, asappropriate, must also be modified. The change inthe number of turns and the tap is directly pro-portional to the change in frequency. If, for in-stance, a frequency of 2600 kHz instead of2182 kHz is wanted in position 1 of switch Si, thenumber of turns, n, of Li should becomen= 115(2182/2600) = 97 turns and the tap should beatn =11(2182/2600) = 9 turns.

C29 C20 Cal

" 74""

T1,12 = BF 90013 T5 = BF 49401,02 = AA 119 12V 24VD3... D5 = 1N4140

12 V

XI 21629N.2 52.45

6690411.

26491

Oscillator capacitors Ca, Cs, and Cs should have ahigher value if the frequencies are chosen at the lowend of the short-wave band.When the receiver has been built correctly in accor-dance with HF requirements (short connections,ample decoupling), it should work up to about18 MHz. The dashed lines in the circuit diagramrepresent earthed screens between the various sec-tions.The receiver is calibrated by adjusting Cs, C7, andCs for zero beat, and then adjusting CI, C2, and C3for maximum audio output.

161 SWITCHABLE BANDSELECTOR

In many older types of SW receiver, intermodula-tion in the mixer was generally avoided by includinga tuneable, often automatically tracking, pre -

selector. In a computercontrolled preselector, theuse of varactor diodes for tuning the inductors oftenleads to considerable intermodulation distortion. Adifferent approach is therefore used in this design.

The circuit diagram shows the use of PIN diodesType BA244 for selecting one of 5 bandfilters fol-lowed by a low-pass section. Selection of a filter iseffected by having the computer drive the associ-ated input high. An impedance transformer is pro-vided at the input to enable connection of 50 Q aswell as 500 Q aerials. For most purposes, the 500 Q

190

R1 15V

5000

5012

0(7

16 .32MHz

8.. 16MHz

4 8MHz

2...4 MHz

1 2MHz

cl BA244

L2510mH

770n

LO I5286 1 Op

BA 244

C3 CS L3

L2 I18715 6p8 151016

04 I C10 27p / C12 27p 14

15 I L7 I 11915pH I 15p 4pH7 I 569 15p 115pH

BA 244 01911

1888I c1a 47p I 021

L12 II 39 6pH8 I1200

BA 244 L16

115 I33pH 1 82p

7 C29 L18

L171088

C22 I5114

330 12201

C30 i1L19

82p 533pH

BC 5478

BC 547 B

BC 547 B

R13 615

BA 244

BA 244

BA 244

BA 244

10011

012 A.

BC 547 B

C.C32

017 18 020C3

770n 7170n

BA 244 35 C36 ;a.OP 1 037

L20 I L22 I 112410088 I 1320 2208 14709 82p 11000

BC 5478

622 2 R S40 41

70n 7170n

BA244 VI.1"..21...

6880 120071 68pH1016 043 C44 C45i5 701, 560p 5

one608

46 I1129I lOrnH

BC 547 B

R27 626 0C47

470n

BA 244

BA 244

<1MHz I

1130110mH

O

0

470n

87479191

input is preferable, since it allows short aerials to becorrectly terminated. Input transformer Tri iswound on a ferrite core Type FT37-75 fromMicrometals. The total number of turns is 19, with

a tap at 2 /2 turns from the ground connection. Theinput should be provided with an overvoltage pro-tection if the aerial is a large, outside mounted, ar-ray.

162 SYNTHESIZER FOR SW RECEIVER

The synthesizer shown in Fig. I is computer con-trolled, and outputs a local oscillator signal (LO) be-tween 48 and 78 MHz for driving the mixer in theSW receiver proposed on page 00. The circuit isbased on the Type MC145156 synthesizer fromMotorola. This IC is relatively inexpensive, and en-sures good LO suppression in the receiver whenused in combination with a good mixer. Also of in-terest is its serial control input, which enables theoutput frequency to be programmed from a com-puter.The internal reference frequency, 1200 Hz, is ob-tained by dividing the signal from oscillator T5 -T6by 2048. The DAC connected to the output of thefirst ID gives a resolution of 1200/255 5 Hz. Thedivider composed of IC,, IC2, ICs and Ni has aprescale factor of 128/129. Opamp IC6 is connected

1 i)o

-Ct

RBI

as a simple loop filter with a reference signal rejec-tion of about 60 dB. An alternative filter that en-sures a rejection of 80 dB, but has a slightly longersettling time, is shown in Fig. 2. This circuit is

driven from the phase detector output of the syn-thesizer chip. Opamp ICi is used in a speed-up cir-cuit that may be included to equal the settling timeof the filter with IC6. Diodes Dl -D2 also serve toshorten the lock -in period of the synthesizer. Theuse of the Type E420 (Ti) is not obligatory: othertypes of AF double FET should also work in thisapplication. The power supply for the synthesizer isshown in Fig. 3. The L -C filter in the + 5 V railsuppresses noise on the synthesizer supply, and D2has been included to compensate for the drop acrosschoke L2.The data format for programming the MVC145156

NI= 105 591560

IC.869

IC

596604

'P

1

1C3

5794IC 4

14C 1451.

159

B1. DB:BB .9

IN1174145

10

BF 982

T -

494

94

16

81,195, I

192

5V

4

2

1C4 pin 6

r)0

0

C7

8 560pIC 1

3130

3r-

muftitum

15V

111BC 560 C

mi

5510

2

550CBC

613 614

=Ow MxCl C2 C3RS

g C4

68T 25T 25T 470T

P2250k 50k

E 420

Ti

R3

BC 550 C

is shown in Fig. 4. Bits SW, and SW2 control theswitching outputs, and are not used here. The syn-thesizer divides by 128N + A: when counter Areaches state 127, N is increased by 1, and Abecomes 0. Data is latched into the synthesizer onthe trailing edge of the clock signal. When the con-trol word is complete, the enable signal is brieflymade high to transfer the data from the shiftregister to the programmable dividers. The squelchis then enabled to suppress locking and tuningnoises.The construction of this synthesizer requires someexperience in building RF circuits. The ECLdividers and the synthesizer chip should lie upsidedown on an unetched board to enable effectivegrounding and cooling. The chips are interconnec-ted with the shortest possible wires. Great careshould be taken in the construction of the VCO andthe TXO. These sections should be screened and

C

1p5

2x1N 4148

C

10p

87495.2

VCO

built such that mechanical stability is ensured at alltimes. VCO inductor Li is especially critical in thisrespect: make sure that the wire turns are secure onthe core.Finally, the winding data for the home-made induc-tors in this circuit: (use enamelled copper wire): Li(VCO): 14 turns 22SWG (0 0.8 mm) on a T50-12core, tap at 4 turns from ground; L3 ( + 5 V rail): 8turns 30 SWG (0 0.3 mm) through a ferrite bead.

CD

* Z Z > >

or)co

cnou wco

cow

4- A counter bits N counter bits

T Last data bit in (Bit no. 16) First bit in (Bit no. 1)

193

163 TUNEABLE ACTIVE AERIAL FOR SW

Many of the modern, synthesizer -tuned, generalcoverage SW receivers incorporate the latest typesof high dynamic range RF prestages and mixerdevices, while the good old tuneable preselectorstage seems to have been eradicated in all but themost expensive and sophisticated types of multi -mode receiver. It would seem as if manufacturers as-sociate a simple tuning control with an attack onuser friendliness of the receiver, while a well -designed, tracked or individually controllable inputattenuator would have been a better solution to theproblems caused by the worldwide escalation of SWtransmitter output levels.A likewise argued plea for reestablishing the tuningcontrol could be entered for the active aerial which,while not able to offer the performance of a longwire or multi -band beam aerial, is none the less gen-erally recognized as a satisfactory means for receiv-ing broadcast programmes in the SW bands up toabout 15 MHz.As generally known, an active aerial is composed ofan aerial proper and associated amplifier. As to thelatter, the ciruit diagram shows that the design hasa varactor-tuned, symmetrical input using twoFETs Type BF256C which are fed over the coaxcable to the receiver. Opamp IC, functions as a fastsymmetrical to asymmetrical converter capable ofoperation up to about 30 MHz. Note that thevaricap diode set is tuned over a separate cable;twin -lead 75 Q coax cable is, of course, ideal forthe present purpose. The indicated varicap set en-sures a tuning ratio of about 1:2 to 1:3.When constructing the aerial to this design, itshould be noted that neither the circumference of

411110.,

0.5...2m

86473-2

the loop aerial nor the total length of the dipolemust be in excess of one tenth of the relevantwavelength in order to ensure the correct directivitycharacteristics, especially in the case of the loop

194

aerial; the dipole will typically fail to match theamplifier input impedance and thus cause problemsin getting the device tuned properly.Table I summarizes the aerial construction data,given a number of possible operating frequencies.The aerial should be mounted in such a position asto receive a minimum amount of man-made, shortrange interference; the amplifier's symmetrical in-put should ensure sufficient aerial directivity to finda dip for the interfering source.The loop aerial is uncritical as to the height aboveground, but not so the dipole, which is bound to actas a vertical rather than horizontal aerial whenmounted at less than a quarter wavelength aboveground.

Table 1.

Fmin

[kHz]

L1

1µH1

turns

n

I

[ml

150 2200 32 1

51 0.5

350 390 13 1

20 0.5

1000 47 4 1

6 0.5

2000 12 2 1

3 0.5

4000 3.9 1 0.5

164 TUNEABLE FM BOOSTER

This FM band (88-108 MHz) preamplifier has beendesigned to come round the problems associatedwith wideband as well as narrowband aerialboosters. Most commercially available boosters arewideband types with relatively poor selectivity and

1

BB106

75 Op

R1

U

R2

BF961BF981

C5

7n

BB106

R6

adjacent station rejection, while the (more expens-ive) narrowband types are rather impracticablewhen it comes to receiving stations well removedfrom the (fixed) frequency of peak amplification.This proposed design is the best of both worlds,

BD139T2re 1N4002

BC557A,B

3 1 BC547A,BR8

400mW

400mW

D4,D5 = 1N4148

L1 = 9t , tapped 1t from earth { 22SWGL3 = 9t , tapped 3t from earth closewound on pencilL2A = 6t 26SWGon ferrite ring 1cm dia (egT37 - 12)L2B = 3t 26SWG on ferrite ring 1cm dia (egT37 - 12)

C11

momIn

195

2 D8

1N4002

since it features good selectivity and strong signalhandling, as well as a relatively low noise figure andsufficient amplification over the entire FM band.Tuning the preamplifier is done in the living room,by means of a simple potentiometer mounted in anenclosure which is conveniently located next to theFM tuner as part of the hifi set.The unit can also be made to function as a 2 metresamateur band (144-146 MHz) preamplifier by modi-fying the tuned circuits to suit the higher frequency.The circuit diagram of the tuneable booster-Fig.I-shows that two remote tuned circuits, alongwith a MOSFET tetrode have been incorporated tominimize the chances of running into cross- and/orintermodulation caused by strong local signals.Varicap diodes Di and D2 form the variablecapacitance to coils Ll and L3 respectively. The tun-ed circuits are set to the desired frequency by meansof the voltage applied to the varicap diodes (3 to24 V, reverse bias). The RF gain offered by Tishould be of the order of 25 dB, while the noise fig-ure is expected to be about 2 dB.The amplifier supply/tuning voltage and superim-posed RF output signal are connected to the coaxcable core which is run to the power supply/tuningunit, shown in Fig. 2. Tuning control potentiometerP1 constitutes the feedback loop to the voltage regu-lator composed of T7, T8 and T9. Turning Pi thusvaries the voltage to the mast -mounted boosterform 15 to 36 volts. Regulator T2 -T3 -T4 (Fig. 1) pro-vides MOSFET T1 with a fixed voltage of 11.4 V,irrespective of the DC level on the coax core. Sub -

TO RECEIVER

86483-2

traction of 12 V from the 15-36 V input voltage isby means of zener D6 and current source Ts. RFoutput voltage and DC supply are coupled to thedownlead cable through C1, and L4 respectively.Cia and L5 (Fig. 2) have the same function in thePSU. D13 prevents the PSU output voltage from ris-ing above 37 V in case of any breakdown in thesupply unit, while D7 protects the booster fromaccepting a reverse voltage in case coax core andscreen are accidentally reversed. T6 limits thesupply short circuit current to a safe 60 mA.The following are important points to observe inconstructing the masthead amplifier and associatedindoor control unit:1. Use a copper -clad board of maximum earth plane

surface (the Type 85000 RF prototyping board isideal).2. Mount a metal screen across the MOSFET case

to suppress any tendency to parasitic oscillation.3. Keep the source lead as short as possible; solder

it direct to the copper surface.4. Keep the leads of G2 decoupling capacitor C4 as

short as possible; a ceramic disc capacitor is idealfor this purpose.5. Keep all coil connections as short as possible to

avoid amplifier tuning over the wrong frequencyrange.6. Fit T9 with a small heatsink.7. Mount a screen between amplifier and DC

supply section.

After the construction of RF head and PSU has

196

been completed, the latter is tested by verifying thepresence of the variable (15.6 to 36.6 V) supply andtuning voltage on the coax cable core. The voltageacross R14 should be lower than 0.4 V with theamplifier connected at the far end of the cable.Turning Pi should cause the voltage at the collectorof T5 to vary between 3 and 24 V.The voltage at the emitter of T2 should be constantat 11.4 V with respect to ground, irrespective of thetuning voltage set with Pt Drain resistor R4 shoulddrop between 0.7 and 2 V. Set Pi to the centre ofits travel.Optimum RF performance of the booster can beachieved by carefully stretching or compressing L3for maximum amplification at about 95 MHz; tunethe receiver to a weak transmission at this fre-quency and align for maximum S meter deflectionor optimum audibility of the signal above the noiselevel. Do the same for signals at either extreme endof the band and set Pi accordingly. Ensure that thetuning potentiometer can be set to give optimum

amplification for every frequency in the 88 to108 MHz band and mark the tuning scale on the in-door unit in steps of 1 MHz. In case it is not poss-ible to obtain equal amplification across the band,L3 may be adapted carefully by increasing ordecreasing the number of turns. The tap, however,should remain at 3 turns from ground.Those constructors striving for utmost perfectionmay fit a 40 pF trimmer capacitor instead of Ci, inorder that the amplifier may be tuned for optimum(i.e. lowest) noise figure, which is not the same astuning for optimum amplification.Finally, the coil data for the tuneable booster are asfollows:

= 9 turns 22 SWG (0.7 mm dia) enamelled wire,close wound, coil diameter 7 mm. Tap at 1 turnfrom ground.L3 = the same, tap at 3 turns from earth.L2a;L2b = 6 and 3 turns respectively, 26 SWG(0.5 mm dia) enamelled copper wire on dia 10 mmferrite ring Type T37 -I2.

165 VLF CONVERTER

Strictly speaking, the VLF (very low frequency)band stretches from 3 kHz to 30 kHz, and the LF(low frequency) band, often called the long -wave-band, from 30 kHz to 300 kHz. The converter de-scribed here covers the frequency range

R1

10... 150 kHz

L2 L4

8mH2 8mH2

L3

LIE C1 C2 10mH

100 mH mow

33073907 390T

4 MHz

xt

C5

100r1

10 . .150 kHz and falls, therefore, half -way be-tween being a VLF and an LF converter.Frequencies between 10 kHz and 150 kHz are con-verted to 4.01. . . 4.15 MHz which can be fed to anyshort-wave receiver capable of accepting those fre-

L5

OOpH

C Ct4101

1007 nOp

6...12V5... 10 mA

1100n

L6

12pH ci2

NM 10n

4010 . 4150 kHz

085484

197

quencies. The converter is connected to the aerialinput of the receiver via coaxial cable.Many converters suffer from break -through of themixer/oscillator frequency in the output signal,which is normally caused by the mixer being asym-metrical. Because of that, the present converter usesthe well-known S042P frequency changer, the sym-metry of which can be set accurately with a 1 k

preset potentiometer connected between pins 10and 12.To prevent reception of image frequencies, the aerialsignal is first applied to an LC band-pass filter,before it is fed to the frequency changer.The output of the frequency changer (pin 2) is ap-plied to an LC circuit that is tuned to the frequencyrange 4.01. . 4.15 MHz. This circuit, consisting ofa 100 pH inductor in parallel with a 100 n capacitor

and a 60 p trimmer, effectively suppresses anyspurious signals produced in the frequency changer.The 60 p trimmer is used to tune in to the desiredtransmitter in the 10 . . .150 kHz range (loudestreception!). The symmetry of the frequencychanger is set by tuning the short-wave receiver tothe frequency of the quartz oscillator, i.e.,4.00 MHz, and then adjusting the 1 k preset forminimum output from the converter, that is,minimum deflection of the S meter, or other fieldstrength indicator, on the receiver. During this cali-bration, the input of the frequency changer, point Ain the diagram, should be short-circuited to earth.All inductors are standard RF chokes. The value ofthe output inductor, 121.4H, is not critical.The aerial should be as long a wire as possible.

166 WEATHER SATELLITE INTERFACE

D1...03 = 1144148 A1...A4 = IC 1 = TL074A5 = IC 2 = CA3130

An increasing number of electronics enthusiasts isbecoming interested in weather satellite reception.Most non-geostationary weather satellites, likethose in the NOAA series, operate in the 138 MHzcarrier. For optimum reception, the detector shouldfeature a relatively high carrier suppression.It is assumed here that a picture signal is availableon a cassette tape. Opamp Ai has an amplificationof 48, while A2 -A3 form a precision two-phase rec-tifier. The 2,400 Hz ripple arising from the slightlydifferent specification of the opamps amounts to nomore than 0.2%. For commonly used A -D con -

C1.1

6V

0

R10

1N4148

87521

verters, this corresponds to an error smaller than1/2 (LSB).

The main ripple signal is 4,800 Hz. This is readilyremoved by a double 71 filter set up around Li andLz. At 2500 Hz, the attenuation is baout 3 dB, at4,500 Hz about 45 dB. The parallel R -C and L -Cnetworks at the + input of As compensate for theohmic resistance of the inductors in the 71 filter. Li,L2 and L3 are preferably ferrite -encapsulatedchokes from the Toko 1ORB series, available fromCirkit PLC (Li & Lz: 181LY-473. L3: 181LY-104).The interface is suitable for processing carrier fre-

198

quencies up to 4,800 Hz, so that it is possible toplay the tape at double speed for reading into thecomputer (provided, of course, the program canhandle this). Components RH, D4 and Ds protectthe A -D converter against voltages higher than 5and lower than 0 volt. The use of the Type CA3130

BiMos opamp ensures an output voltage swing of5 V when a ±6 V supply is used. The maximumsupply level and current consumption are ± 9 Vand 15 mA respectively. The input signal amplitudeshould be greater than 68 mVims for a 5 Vpp out-put.

199

167 AUTO FOCUS FOR SLIDE PROJECTOR

This circuit is intended as a replacement for theelectronics in a partly or wholly defective autofocusdriver in a slide projector. The mechanical parts inthe autofocus system are assumed to be still func-tional.Most automatic focusing systems in slide projectorsare based on the use of an optical module, whichcomprises a small lamp, a few lenses and mirrors,and a light sensor made from two series -connec-ted light dependent resistors (LDRs), which func-tion as a potential divider. As shown in Fig. I, lampLa projects a narrow beam onto the centre of theslide, A, whose surface reflects it onto the LDRs.When the slide surface bulges inside or outside, theprojected image on the screen is blurred, and thebeam from L is received on the surface of one of theLDRs (point 2 or 3). This is detected by a motordriver circuit, which ensures that the focal distancebetween the objective, 0, and the slide surface iscorrected to maintain a sharp image, i.e., the objec-tive is moved until the circuit detects that thereflected beam from L falls exactly in between theLDRs (point I).

2

* see text

200

1

2

..........

\

2

87517-1

87517-2

The circuit is based on the use of an existing set ofLDRs as part of the optical module in the slide pro-jector. The symmetrical supply shown to the left,and the motor plus decoupling capacitor, are alsopart of the projector. The inverting input of opampIC, is at ground potential when the above men-tioned test beam falls in between the LDRs. Theoutput of the opamp keeps the non -inverting inputat 0 V as well, so that no motor voltage is availableat the emitters of power drivers Ti -T2. Should thereflected beam illuminate either one of the LDRs,

the circuit arranges for the motor to move the objec-tive glass towards the correct focal position, until novoltage difference between the LDRs is detected.The feedback gain of the circuit has been kept rela-tively low to keep the motor from continuouslymoving the objective glass past the target position,causing the system to oscillate slowly. Resistors R3and R4 may have to be dimensioned differentlythan shown to achieve optimum response as regardsspeed and stability.

168 DIGITAL JOYSTICK INTERFACE

The BBC and Electron computers produced byAcorn have a joystick port to which only analoguejoysticks can be connected. For many purposes, adigital joystick, i.e., one with four contacts, is much

,S 8 8 0 09 8 7 6

=NC.

Cl

T1i0 m0 V

IC1

O

FIRE -0- if6)1 351 4)f 1)0'

RI

R6

R2

more suitable. The interface suggested here enablesa digital joystick to be used with the two computersmentioned.The joystick port is provided with a voltage of 1.8 V

S5

2ff12

R3

ES

11

O

ES2 8

fY VREF

13

R4 .v00

R5

ES3

3

R8

O

ES4 4

P250k

T1

BC 5476

ES1 . ES4 = IC1 = 4066 1185477

5V

201

when the analogue joystick is set to the left or toppositions, 0 V with the joystick in the right or bot-tom positions, and 0.9 V with the joystick inneutral. The 1.8 V is the reference voltage of theanalogue -to -digital converter in the computer.As can be seen from the circuit diagram in figure 1,the various voltages can simply be provided by foursets of contacts or switches. Each of the sets of con-tacts controls an electronic switch. The 0.9 V forneutral is obtained from a potential divider. Theelectronic switches are required because the con-tacts in the joystick have a common connection andcan, therefore, not be used direct for shortingresistors in the potential divider. The fire button isconnected to the + 5 V line by a junction in thejoystick, and thus produces a logic 1 when it is

pressed, whereas the computer expects a 0. Thesignal is, therefore, inverted by transistor Ti.The interface is calibrated with the aid of a small

auxiliary program: REPEAT PRINT ADVAL(1)ADVAL(2): UNTILO Potentiometers 131 and P2should be set to the centre of their travel.Connect the joystick and the interface to the com-puter, start the auxiliary program, and adjust thepotentiometers so that the two numbers on thescreen are as near as possible 32768.

Table 1.

Interconnections interface tocomputerterminal joystick 1 joystick 2

A 8 (gnd) 8 (gnd)B 7 (ch.1) 4 (ch.3)C 11 (Uref) 11 (Uref)D 1 (+5 V) 1 (+5 V)E 15 (ch.0) 12 (ch.2)F 13 (PBO) 10 (PB1)

169 ELECTRONIC TOSS-UP

The electronic version of the well-known coin totoss up prior to commencing a football match-orany other sports event where there is a generally es-tablished formality on part of the referee-consistsof a row of seven LEDs, the centre one being green,the others red. After having reset the circuit, theodds are exactly equal for either one red LEDlocated next to the green one to light when the toss-up key is pressed; we have, therefore, a left/right de-cision circuit operating on the basis of pure ar-bitrariness.As to the operation of the circuit, button Si may bepressed at any time to preset counter IC1, whichresponds with outputting the binary code for 0 atits Qe, Ql and Q2 outputs, causing BCD -to -decimaldecoder IC2 to light the corresponding LED, i.e. thegreen one-D4-at the centre of the row. The presetcode for the initial state of the circuit is determinedwith preset inputs P. . . P3 being tied to ground,causing IC, to load 0000 as the binary start-upvalue when Si is pressed.Depression of button S2 causes the bistable com-posed of Ni and N2 to toggle, providing a singlepulse transition at the clock input of IC,. Depend-ing on the logic level at the UP/DOWN input ofIC,, the one -of -eight decoder will light either D5(right) or Da (left), since counting up from 0000gives the next higher binary code 0001, while coun-

ting down gives 1111. The latter value causes IC2 tolight D3 at the Q7 output, since the most significantbit input-D-has been tied to ground.The arbitrariness of the toss-up circuit is ensured bythe speed at which oscillator Na-N4 applies pulsesto the counter UP/DOWN input. The odds are 1 to1, theoretically, while the circuit can not bebribed . .

Seven of the eight active -high outputs of IC2 havebeen wired direct to the corresponding LED, whileQ4 serves to inhibit the counter via the CARRY INterminal. It is readily seen that counter inhibitingoccurs automatically when IC, counts up from out-put state 3, or down from state 5; both conditionscause Q4, and therefore CARRY IN, to go high,disabling further counting until the reset button ispressed.Finally, repeatedly depressing S2 without resettingthe circuit will cause any other, random, LED in therow to light, and this facility may be put to good usein any other, random decision based game or seriousapplication you have in mind.

202

Q5 Q6 07 QO Q1

CARRY IN

UP/DOWN

10

11

12

10

C1

10n

R3

11

02 01

IC.1

4029

8

PO P1 P2 P3 CLOCK

4 12 13 3 15

D1 ... D3, D5 ... D7 = LED 5mm redD4 = LED 5mm green

D7 N1 ... N4 = IC 3 = 4001

9V 10mA

JamRESET

80

86444

0

IC 3

170 FLASHING LIGHTS

This application of the well-known Type 555 timeris intended for model railway enthusiasts wishing toconstruct a two -lamp flashing beacon with aminimum of components.With reference to the circuit diagram, the numberof LEDs need not be restricted to two: several maybe connected in parallel to achieve a higher lightintensity, but a total current consumption of200 mA should not be exceeded to prevent thedestruction of the output stages in the 555. EachLED added should have its own current limiting re-sistor, similar to D1 -R3 or D2 -R4.The flashing rate is defined with C1, The statedvalue of this component is likely to be optimum for

applications in model railways. The supply voltagefor the circuit is not critical, but should remainwithin the range from 5 to 10 V. With two LEDsfitted and a 5 V supply, the flashing circuit shouldconsume less than 50 mA. The intensity of theLEDs can be adapted to individual preference bychanging R3 and R4, but too low resistance valuesshould be avoided to prevent the destruction of theLEDs.

203

5V15 ... 50mA

171 HALOGEN LAMP DIMMER

The circuit proposed here is suitable for fitting intoslide projectors without a dimmer facility (24 V ACfed halogen lamps). With a few small alterations, itcan also be used for dimming 12 V halogen lamps,but not those in a car, because these are fed from aDC source. The circuit shown in Fig. 1 is intendedfor operation from a 24 V AC supply, and canhandle a lamp load of up to 150 W For loads up to

I24V AC

250 W, the TIC236 should be replaced by aTIC246.The illumination of the halogen lamp is controlledby applying a direct voltage to pin 5 of dimmer chipIC, . A voltage of + 2.5 V gives maximum illumi-nation, while +5 V results in the lamp beingturned off completely. The lamp intensity controlrange -2.5 V to 5 V-can be extended upwards by

1-1=1C1= TCA 280 A

TIC 236*

a,a2g

Tri 1

20487452-1

2

.4%, +11 liC1

Tr'

decreasing the value of C2.The TIC246 should be used when the circuit is tocontrol a 12 V lamp that consumes more than50 W. Figure 2 shows details of the connection of apotentiometer to the intensity contro! input of theTCA280A. Voltage divider Rio -Pi -R11 is fitted ex-ternally and can be fed from the stabilized voltageavailable at pin 11 of IC1. The minimum and maxi-mum intensity of the lamp are determined by Rioand Rii, respectively, so that the control range canbe dimensioned to individual preference. When po-

3 0e -

O

O

wort:.

Parts list

Resistors ( ± 5%1:

Ri =470R; 0.5 WR2;R7=100KR3 = 22KR4 = 330KR5 = 150KR6=270KRo=82KRo = 150R

Capacitors:Ci =470µ; 16 V; axialC2 = 1;4; 16 V; axial*Cu = 1n5

Semiconductors:DI =1N4001Trii =TIC236 or TIC246*IC, =TCA280A

87452-3

Miscellaneous:

PCB Type 87452Heatsink for Irk.

* See text

tentiometer control is used, C2 must always be100n.Figure 3 shows the signal waveforms at variouspoints in the circuit.The halogen lamp dimmer is constructed on aprinted circuit board as shown in Fig. 4. When thelamp power is greater than 15 W, the triac should befitted onto a heatsink, and the tracks to the al anda2 terminals should either be covered with solder, orstrengthened with short lengths of wire.

205

172 MODEL AIRCRAFT MONITOR

Older, i.e., not using a computer, radio controlledmodel aircraft are highly vulnerable to breaks inradio communication, which can lead to a crash orthe model landing out of reach, or both. Owing tothe allocated radio frequency range being usurpedby pirates, its is essential for every model flyer tomake sure that the channel to be used is free. Evenif it is, it is advisable to continue monitoring it.In combination with a short-wave receiver, the cir-cuit presented here enables monitoring the 27 MHzradio control band. The aerial signal is filtered(26 . . . 41 MHz) and applied to the input of differen-

BF 494 BF 2569

CE

L3 C3a me

4µH75p6

CILL

73.2p

ImoT1,T2 = BF 494T3 . . T5 = BF 2566D2 ... D4 = AA 119

3

0

D2

L21

I

OpH22

1220 µHD3

C9 C14

71.00p7rOn

R5

X1:7845 kHzX2: 13515 kHz } 50pF par.X3: 3050 kHz

D1

400 mVV

tial amplifier Ti -T2. Since the current source of thisstage consists of an oscillator, the amplifier func-tions as a mixer. The crystal oscillator can operatewith almost any crystal between 2 and 32 MHz.The output circuit, La -Cs -Cs, is tuned to about27.2 MHz. This frequency is inversely proportionalto the values of the coil and capacitors.The values of the crystals are based on a 40 -channelset-up. In switch position A, the circuit functions asan aerial amplifier; in position B, channels 38...49are converted to 8 . . .19; in position C, channels50 . . 53 are converted to 20 . . . 23; and in position

100 µ1-1

100P 110n

9 ... 12 V

<5 mA

C7 C4 C6

TDnmini

Elm Nom

73n 78p

OS1

33p

®0

D

S

II 470 pH

C131 C16

100p 10n

R7

85505

R8

206

D, channels 61. . .79 are converted to 21. . .39. ceiver can work on FM by detuning the monitor aThe receiver into which the monitor is coupled few kHz.need not be suitable for FM reception: an AM re -

173 MODEL RAILWAY MONITOR PANEL

Many railway modellers would love to have a track however, not too difficult to make one yourself.monitor panel, but, unfortunately, the few commer- The reproducing of the track diagram and thecially available types do not justify their cost. It is, mounting of the monitor lights on the panel can be

1

Tr.1

* see text

1C27805

C7

70n0 ®

D10... 013= 4V7 - 400 mVV

D1

C D3

R112 4+

D1 ... D4 = 1N4148

N1 ... N4 = 1C1 = 74LS28

R3

it R9

if

it

R10

5V

85493

207

2

ICE

S S

WryeEPS.B5493

"Ri.c-0-2,414-.

A B J. C D

accomplished without too much trouble. There is aproblem, however, in indicating the position of turn-outs and colour -light signals, because theseelements are normally operated by spring -loadedswitches to prevent the burning out of thesolenoids. After the push-button on the controlpanel has been released, the supply line is no longerlive and can, therefore, not be used for lighting anindicator. This problem can, fortunately, be solvedby a couple of R -S bistables (NOR gate latches).The push-button switches and solenoid coils shownin figure 1 are those already contained in therailway set-up. Note that the system is assumed tooperate from a 9 . . .15 V AC supply.Each signal normally requires three lines: one foreach of the two coils and a common line. TerminalsA, B, C, and D in figure 1 are connected to the rel-evant outputs of the control panel. The circuit asshown is suitable for monitoring two turnouts ortwo colour -light signals via A -B and C -D respect-ively, but can be extended as required.The voltages used to energize the coils are rectifiedand applied to an R -S bistable. This NOR gate latchis set or reset, depending on the nature of the inputsignal, and this causes the relevant LED to light. If,

Parts list

Resistors:

.. .R4 = 4k7Rs...R8 = 2k2Rs,R1,0 = 330 Q

Capacitors:

C.1 ...C4,C6...C8 = 100 nCs = 220 µ/40 V

I

Semiconductors:131 ...D4 = 1N4148D5 . . . D8 = LED (red or green, as required)Do = 1N4001Dio .. .D13 = zener diode, 4V7/400 mW

= 74LS28IC2 = 7805

Miscellaneous:Tn = mains transformer, 9...15 V secondary

(if not already available from the existingsystem)

PCB 85493

208

for instance, pin 8 of N2 is high, pin 10 of this gateis low, and D6 lights.The circuit as shown has a current consumption ofabout 30 mA per R -S bistable branch.Not all monitor LEDs will correctly show the pos-ition of the relevant turnout or light signal immedi-ately upon switching on the supply. Briefly pressingone of the two push -buttons of each turnout orlight signal will correct this situation.The circuit is most conveniently built on the printed

circuit board shown in figure 2. This board can ac-commodate two monitor channels as shown in fig-ure 1. If more are required, these can be built on ad-ditional PCBs. The section containing C5, 1C2 andD9 may be cut off subsequent boards, but if manyadditional PCBs are used, make sure that the powerrequirements are still met! The + 5 V and 0 V ter-minals on all boards should be interconnected.

174 "ON THE AIR " INDICATOR

In radio and television studios it is customary to in-dicate to all concerned when the microphone orcamera is "on the air". This is normally done witha red light at or near the relevant camera or micro-phone. The circuit described here is intended as anauxiliary for a DIY mixer unit.To make the circuit automatic in action, stereo slidepotentiometers are used at the audio inputs. Whenone section of these potentiometers is connected tothe +15 V line, the potential at the wiper of thissection is a measure of the potentiometer setting.

This potential is amplified in opamp A, and appliedto the inverting input of A2. The latter opamptoggles as soon as the level at its inverting input ex-ceeds that at its + input, which has been set withpreset Pt

Al ,A2 = IC1 = TL 082 85410

209

The slide potentiometers for this purpose are alwayslogarithmic types, so that the voltage rise at the be-ginning of their travel is always pretty small. To en-sure correct operation of the circuit even at thesesettings of the potentiometers, the gain of Al hasbeen arranged fairly high, about 26 dB.Opamp Ai also serves as a summing amplifier thatmonitors a row of audio inputs. If it is required thateach audio input has its own monitor, the twoopamps must be repeated for each input, but P ofcourse, continues to provide the non -invertingpotential for all opamps in the A2 position.

The output of the indicator is provided by a typeBC 547B transistor, which can switch up to100 mA. This current is sufficient to light a signallamp or light -emitting diode (LED) with bias re-sistor, or to drive a relay.Current consumption with the BC 547B offamounts to not more than 10 mA.If low -resistance stereo potentiometers are used, thedirect current through the "indicator section" maybe too high; if that is so, it is advisable to use a drop-ping resistor in series with the section.

175 PWM DRIVER FOR DC MOTORS

The speed of DC motors is relatively simple to con-trol. For independently energized motors, the speedis, in principle, a linear funtion of the supplyvoltage. Motors with a permanent magnet are a sub-category of independently energized motors, andthey are often used in toys and models. In this cir-cuit, the motor supply voltage is carried by meansof pulse width modulation (PWM), which ensuresgood efficiency as well as a relatively high torque atlow motor speeds.A single control voltage between 0 and +10 Venables the motor speed to be reversed and variedfrom nought to maximum in both directions.Astable multivibrator ICI is set up as an 80 Hz os-cillator, and determines the frequency of the PWMsignal. Current source Ticharges C3. Thesawtooth voltage across this capacitor is comparedwith the control voltage in IC2, which outputs thePWM signal to buffer Ni-N3 or N4 -N6. Thedarlington-based motor driver is a bridge circuitcapable of driving loads up to 4 A, provided therun-in current stays below 6 A, and sufficient cool-ing is provided for the power transistors T2 -T5.Diodes D2 -D5 afford protection against inductivesurges from the motor winding. Switch Simakes itpossible to reverse the motor direction instantly. BD 680

A =11_11 80Hz

B = PWM OUT

1C3

BD 680

NI...N6=IC3=404987462

210

176 SECTION INDICATION FOR MODEL RAILWAY

This section indication system may be just the thingyou have been looking for when you own a fairlylarge model railway with tunnels and tracks atseveral levels, and are sometimes at a loss to find thewhereabouts of a particular train. This circuit usesLEDs to indicate the train's position. Each trackblock is split up into 8 sections, whose startingpoints are marked with reed contacts (S1 -S8). Aninth reed contact is fitted at the end of the block,to enable turning off the indication for the relevantlength of the track.The circuit is composed of 8 set -reset (S -R) bistables,which drive a LED each. All SET inputs are com-bined in a NOR gate, Ni, which drives a pulseshaper and buffer to reset the bistables with a brief

pulse to ensure that only the LED for the lastpassed track section is lit. The reed contacts are ac-tuated with the aid of a small magnet fitted to theunderside of the engine. Depending on the mostsuitable location of the magnet, the reed contactsare fitted in between the tracks or alongside the leftor right hand rail.Several of these section indication systems may befitted in series to enable making a control panelwith many lights to indicate the train positions.Observing the direction of travel of the trains, sec-tion junctions are fitted with S9 (end of previoussection) and Si (begin of section) located next toeach other.

12V

Si

S2

C

S3

5

0A

R

R10

2

9R11

R

S414

0

R

Z_.10

S5

S6

S7

58

R1.. .R44x100k

DR

15

11

1 C1

5

A a2

3

12

B

10

R5.. -R84 x 100k

X111

R

RRD

D3

R12 HD4

R13 10k.D5

R14 ND6

R15 ".R16 11/41

f=3 D8R17 vklt.

D9

1C24043

15

N 1 = IC 3. 4078N2= V6 IC 4 = 4049

tb 14 16 16

1C4 1C3 1C2 1C1

TTTT 0=I=

87402

211

177 SERVO -ROBOT DRIVER

Over the past few years, robotics and cyberneticshave become new fields of interest for many anowner of a personal micro, equipped with thenecessary add-on boards to effect peripheral con-trol. However easy it may seem to write robot con-trol programs and to build the associated computerhardware, the construction of accurately operatingmechanical parts (or, if you like, limbs) often posesunsurmountable problems, since a miniature set ofgears, ball bearings, spindles and cog -wheels are inno way readily made items for those skilled in pro-gramming and soldering.Despite their limitations as to precision of move-ment, servo -motors used in model aircraft or boatconstruction may offer an interesting alternative tomore complicated mechanical constructions; appli-cations such as robot arms and sorting machinescan be made quite easily with the use of cleverlymounted servo -motors.A number of wait loops need to be pro-grammed to supply active low (0) pulses lastingabout 0.5 ms, while the interval length between thefirst and second pulse determines the position ofservo 1, while servo 2 is positioned by means of theinterval between the second and third pulse, and so

4

on. The repeat rate of the control process should beabout 50 Hz (20 ms); see the inset timing diagram.The synchronization interval is generated with Di -R i-Ci, which reset the Type 4017 counter when nonegative pulse has been received for about 3 ms.The control inputs of the servo -motors may be con-nected direct to the counter outputs.Finally, interested robot constructors are advised toconsult the excellent Robot book, published by WH Smith & Son Ltd, Leicester, under ISBN number0-7112-0414-4.

178 SOLID-STATE DARK -ROOM LIGHT

Light -emitting diodes are perfectly suitable for dark-room light, because they (a) obviate the need offilters; (b) emit cold light; (c) have a life that is notshortened by continuous on -off switching; and (d)do not radiate infra -rays. The types used must, ofcourse, have a high light output; fortunately, thereare nowadays LEDs with a luminous intensity ofhundreds of millicandela.The sensitivity of photographic paper lies betweenwavelengths 300 nm and about 550 nm, whereasthe wavelength of the light emitted by green LEDsis about 565 nm; that by amber types around585 nm; and that by red LEDs about 640 nm. Fromthis, it is clear that all three types of LED may beused with impunity. None the less, in practice, it isbest not to use green ones. Because of the specialcomposition and high sensitivity of colour negativepaper, only yellow LEDs with reduced light output

should be used when processing this paper. The pro-posed light, therefore, has provision for reducing theemitted light. Note that since colour reversal paperis sensitive to all colours, it can only be processedin total darkness. When working with or-thochromatic paper, only red LEDs should be used.With reference to the diagram, each group of threeLEDs is fed from a current source, Ti to T6 respect-ively. The current level, and consequently the lightoutput of the LEDs, is determined by the setting ofPt Zener diode D19 provides the reference voltagefor the current sources, ensuring that the light out-put of the lighting unit remains virtually constantover the life of the PP3 battery.Maximum light output is set with the aid of P2. Tothis end, both Pi and P2 are first set to maximumresistance; after this, P2 is adjusted until a potentialof 0.2 V is measured at point A. The maximum cur -

212

oil6

D3

8 x BC 5478

0see iext

Dia D18b LED red CQV 51H, CQX 54, COW 24 2Dib D18b LED yellow CQV 53H, COX 74

rent through the LEDs is then about 20 mA.As the photograph shows, the unit has been con-structed so that Si is easily operated. Since thisswitch is a press -to -make type, the light will switchoff as soon as it is put aside, thus preserving the bat-tery. It is possible to have the light on continuouslyby connecting an external battery to Uext. In thatcase, Rio must be matched to this source accordingto

Rio = (Uext-9)10 [k52]

Re

3 V3400mw

D20

9V

1060

86466

but only if NiCd batteries are used. If standard cellsare used, D20 and Rio must be omitted.If a variety of photographic paper is processed, itmay be useful to be able to switch between red andamber LEDs. For that purpose, each of the eighteenoriginal yellow LEDs is duplicated by a red LED,shown in dashed lines. Switch S2 may be used to sel-ect the relevant bank of LEDs (red or yellow) as re-quired for the specific application.

179 SPEED CONTROL FOR R/C MODELS

The speed and direction of rotation of a motor in aradio controlled model aeroplane or boat is gener-ally controlled by pulse width modulation of thesupply voltage to the motor driver stage.In the present circuit, shown in Fig. 1, bistable FFiis set up rather unconventionally to function as amonostable multivibrator, whose period is set with

This period determines the toggle point atwhich the motor's direction of rotation is reversed.Output Q of FF2 goes high when the pulse at theD input (PWM signal) is shorter than that at theCLK input (signal from FFi). This causes Ti to ac-tuate Rei, so that the motor direction is reversed.The PWM control signal applied to the circuit isalso fed to N2, whose output pulse width is the dif-ference between that of the input signal and that

from FFi. The pulse width at the output of N2therefore decreases as the relevant control handleon the transmitter is moved further towards eitherextreme, and is maximum when the handle is in thecentral position. The output of N2 is integrated byAl to obtain an output voltage proportional to thepulse width. A4 compares this output voltage withthe triangular signal at the wiper of P3, so that avariable duty factor signal is obtained for drivingthe power output stage comprised of T4- T5. Mean-while, A2 compares the proportional voltage fromAl to the level set with P2. When the output of Alis lower than the threshold, i.e., when the motorspeed exceeds the preset level, T2 activates Re2. Thiscauses the collector -emitter junction of series regu-lator T5 to be bypassed by the relay contact, and so

213

enables the motor to run at full speed, because theforward drop across T5 is eliminated.The frequency of the triangular signal from A3 is ofthe order of 2 kHz, which is suitable for mostmotors. Capacitor C6 may be increased to lower thefrequency for non-standard motors. Conversely, ifthe frequency is increased, care should be taken toobserve the maximum switching speed of Ts,which is a commonly available, but relatively slowpower transistor.Presets P4 and P3 determine the limits of the in-operative range of the handle, and the point thatcorresponds to maximum motor speed, respectively.More specifically, P3 sets the amplitude of thetriangular signal, while P4 sets the offset level, toenable A4 to output the triangular wave undistortedand with the maximum possible voltage swing.Preset P2 is used to define the point at which the

1

motor is switched to full speed. Some care should betaken in this setting to allow a sufficiently large con-trol range for the handle, and also to avoid the riskof Reg clattering or being blocked.Be sure to fit the 470n capacitor across the motorterminals, and the 47n capacitor between one ofthese and the motor body-see Fig. 2. The coilvoltage of the relays should be equal to the voltagefor the battery that powers the motor, while thecontacts must be capable of handling the currentdemand of the motor. Transistors T4 and T5 shouldbe fitted with a heatsink. Note that although theType 2N3055 can handle currents up to 10 A, itmay be a good idea to fit two in parallel with OR1emitter resistors for equal current distribution whenheavy loads are to be controlled. The current ratingof D6 and D7 must also be observed: for the stated1N5401s, Inman = 3 A, and two may have to be

u

D1

o

FF1CLI4, 0

4

s

FF2CO -co, 4

10

12

02N 4001

ea

cz

1071

BC 140

ioo T1C3

A 1...44 sr-- ICI = LM 324FF 1, FF 2 IC 2 = CD 4013

N 1...N 3 = 3/4 IC 3 = C04070

U.

4V810...15 rn4

bb

87426-1

214

connected in parallel when this current is approach-ed. Finally, U + is the model's battery voltage(4.8 V), and + Lice is the supply voltage for themotor.

2

180 STARTING -PISTOL SIMULATOR

In this circuit, a discarded power loudspeaker servesto simulate the loud bang from a starting -pistol.Power transistors Ti -T2 and mains transformer TRIform a power oscillator, which can be started bypressing Si. Zenerdiodes Di -D2 protect Ti -T2against inductive voltage surges. The operating fre-quency of the oscillator depends on the corematerial of Tri, and the current through its 240 Vwinding. While Si is kept actuated, the oscillation

47V400mW

E

frequency is lowered from several kHz to about 50Hz as the charge voltage across flash capacitors C2 -C3 rises. The charge current is limited by R5, whileD3 -D4 form a voltage doubler, so that several hun-dred volts are available across the contact of Rel.LED D6 lights to indicate that the flash capacitorsare fully charged, and that Si may be released.When the fire button, S2, is pressed, Re, is ener-gized, and a short pulse of 50-100 Ap is passed

o

SI

C21000...2200v400V

cad1000...2200p400V

<3A* 6...9V*

0

300V1A

4...8n*60W

?rc

215

through the loudspeaker's voice coil. Make surethat this can handle the current surge, on penaltyof once and for all destroying the cone suspension.The current consumption of the starting -pistolsimulator is about 3 A immediately after pressingSI, and gradually falls to 0.5-0.8 A when the reser-voir capacitors contain their nominal charge. Theloudness of the bang can be increased by raising the

supply voltage to a maximum of 12 V, provided theloudspeaker can handle a peak power of more than1,000 WIt is recommended to test the circuit at a supplyvoltage of 3 V. Finally, bear in mind that thegenerated high voltage is extremely dangerous, andrequires due attention to be paid to proper insu-lation of all components in the high voltage section.

181 TIME-LAPSE UNIT

Which amateur film maker has not wishedsometime that he could experiment with time ex-posures? Fortunately, it is now possible with asimple electronic circuit to see the grass grow with-out having to sacrifice a night's sleep.The circuit consists of a clock oscillator, N1 -N2, a12 -stage binary counter, IC2, and a monostable, N3 -N4. Preset P1 is adjusted to make the highest oscil-lator frequency about 16 Hz. When switch S2 con-,nects pin 4 of N2 to pin 8 of N3, the circuit worksin "real time". Each successive step of S2 from thisposition doubles the time between exposures. At

the minimum oscillator frequency of 0.5 Hz, and S2connected to the Q11 output of IC2, intervals of upto two hours are possible between exposures.As the signal at the wiper of S2 is a square wave,which is - by definition - logic 1 for half the time,it is essential that it is shaped in a monostable. Theduration of the consequent pulses is determined byPa Their width should, of course, not exceed theperiod of the clock oscillator.Many film cameras are provided with a miniaturesocket via which they can be operated for singleframe exposures and film transport. Contacts X and

N1 ... N4 = IC1 = 4093

216

Y of relay Re, should be connected to this socketvia a suitable cable. If you have any trouble withthis, or are not sure of the socket connections, it is

best to seek advice from your local photographicdealers.

182 TIMER FOR FIXING BATH

When, after developing, photographs are immersedin the fixing bath at irregular intervals, it becomesdifficult to observe the correct fixing time for eachof these. This problem is solved by the present timer,which is capable of "remembering" up to 32 immer-sion times, and automatically provides a signalwhen a photograph is to be taken out of the bath.Any time a photograph is immersed in the fixer, theuser presses the start key on the timer, whichresponds by lighting a LED. When the fixing inter-val has lapsed, the timer provides a short beep.The circuit is composed of a 64 -stage shift registerwhich is loaded with zeroes on power up, becauseit lacks a reset input. Electronic switch ES, con-nects the frequency determining capacitor to the in-put of clock oscillator 1\11 . The logic level that exists

S1

Reset

16 14 14

T T T

1C3 1C2

:di

ES1 ... ES3 = 1/4 IC 1 =40661111 ... N4 = 1C2 =4093

60

S2Start

ES1

13

O

at the Dm terminal of IC3 is shifted towards outputQ at a speed that is defined by Pl , which enablesdefining fixing times between roughly 1 and 10minutes, 9 minutes being a commonly used value.When the START button is pressed, S -R (set -reset)bistable N2 -N3 toggles, and LED Di lights. A logic1 is written into the shift register with the aid of apositive pulse transition applied to terminal CP.After 64 clock pulses from Ni , the logic high levelis available at the output of the shift register, andenables oscillator N4 to drive piezoelectric buzzer

. The LED is turned off shortly after the START

button is pressed, because the bistable is reset by theCL. OUT pulse from IC3.The timer is conveniently fed from a 9 V battery,and should not consume more than about 10 mA.

IC 34031

CP

Di S

1 10

CS

470r

RB

5

1,14

9V

Bat(PB2720)

87454

217

183 7 -DIGIT CODE LOCK

This code lock provides a high degree of securitywhilst being a very simple design. At the heart ofthe circuit is the Type 4022 octal counter. In thenon -active state, C2 is charged via R5, so that thereset (R) input of the counter is kept logic high. Thiscauses output Qo to be actuated, while all other out-puts are logic low. When Si is pressed, T, is

switched on via debouncing network R2 -C,, andIC, receives a clock pulse. Also, C2 is discharged viaR4-1)1, ending the reset state of the counter and en-abling it to advance. The time required for R5 torecharge C2, i.e., to reset the counter, is the maxi-mum time that can lapse before the next key ispressed. The above cycle is therefore repeated onlywhen S7 at the Ql output is pressed in time. Whenall keys have been pressed in time and in the correct

R1

order, Q7 goes high for about 4 seconds to enabledriving the unlock circuitry, e.g. a relay driver for anautomatic door opener. The code for the lockshown in the circuit diagram is 1704570: this is butan example, however, and the combination code isreadily altered by swapping connections betweenthe counter outputs and the switches. When the 7 -digit code is considered too simple to crack, the4022 can be replaced by a 4017, which makes itpossible to add two keys. This means that thenumber of combinations is 109 instead of 107.The quiescent current consumption of the codelock is negligible at 0.5 µA, so that battery oper-ation is feasible. The circuit works well from anysupply between 6 and 15 V. The accompanyingphotograph shows that the code lock can be built as

R311 R5EI

1N4148

R4

BS 170T1

C1

100n

D15

4

*unlockcode = 1704570

16

0

2

IC 13

402204

CLK

6

7EN Nilo

C2 130

Tu716V

8

11

5

10

R6

R7

R8

R9

R10

R11

R12

-0k

sais20 0

30 0 0_.

40 0 0_.

555

0 0

6-0-0 0-0

757

-0-0 0-.

0 0-.

Sal9

S

0 0 0-

6...15V

0+ 0

C3

87463

218

a very compact unit thanks to the use of a printedcircuit board that holds the 10 keys also.

Parts list

Resistors 1± 5%1:

IThR3= 10KR2= 220KR4 = 100RR5 = 1 MO

R6 . . 1312 incl. = 4K7

Capacitors:Cl;C3 = 100nC2=4µ7; 16 V; axial

Semiconductors:

=1N4148ICi =4022Ti = BS170

Miscellaneous:

So... So Digitast momentary actionbutton (SE or S version, ITT Schadowl.

PCB Type 87436

ar

a

7............_:.,....4, cr..- -s.....:-.-nk.-..,- 'mir IL Cr

ID

,741.44:74,4,;054eliw' ,,, C4.4IWIL,4:, CI /7

moo , arier.,,,airoari g"ISM, max - -rag -t al 2-1

Amu - -,-Psodz-i Jaw 4 41

1/111e411141WllOSA 4

18188 8 8

184 ABSORPTION -TYPE METAL DETECTOR

The action of the detector, which indicates thepresence of ferrous as well as non-ferrous metals,depends on the absorption of magnetic energy. Aninductor, which forms part of a tuned oscillator cir-cuit, radiates a magnetic field. When a metal object

is introduced into this field, enough magneticenergy is absorbed to cause the oscillator to stopworking,The Colpitt's oscillator in figure 1 operates at a fre-quency of around 70 kHz. Inductor Li also serves

219

as the sensor. Because of the high value of the emit-ter resistor, R,, the oscillator only just operates.This is desirable, otherwise any losses in the tunedcircuit would easily be replenished by the transistor.The oscillator output is rectified by 131 and D2, andthe resulting direct voltage is applied to the invert-ing input of Schmitt trigger IC). If that voltagedrops below the level at pin 3 (preset by Pl), the out-put becomes logic high, and the relay is energized.The detector is best constructed on the printed cir-cuit board shown in figure 2 (this is, unfortunately,not available ready made). Inductor Li is not in-tended to be fitted on the board. This is a standardnon -screened choke of 100 mH.

12 V

Ref

12 V

<50 mA

If the oscillator does not readily start at any settingof PI, the value of R, must be reduced. If, on theother hand, the oscillator does not stop workingwhen a metal object is held near L,, the value of R,must be increased. The stated value of R, has beenfound right when Li is a Toko type.Starting with the wiper of 131 to earth, adjust thepreset so that the relay just does not operate. If alower sensitivity is required, advance the wiperslightly further.Current consumption is determined primarily bywhether the relay is energized or not; in any case itis not greater than 50 mA.

185 ALTERNATING FLASHER

The proposed circuit is intended for use bymodellers at railway crossings, work in progress,advertising boards, and many others. It can be builtquite quickly from but a handful of components.In the accompanying diagram, Al determines theflashing frequency, which may be altered by chang-ing the values of R2 and, particularly, R3. The latter

may be replaced by a suitable preset if the frequencyneeds to be varied often.Inverters A2 and A3 function as fixed delayelements, while A4 inverts the drive to D2. Thealternately flashing LEDs may, of course, be of anycolour suited to the application.

220

88430

186 AUTOMATIC SLIDING DOOR

Nobody pays much attention to automaticallyopening and closing sliding doors nowadays. Inview of the complex mechanics involved, not toomany people have so far attempted to fit an auto-matic sliding door in their living room. If you arehappy with a relatively slow movement, such a doorcan, however, easily be realized with the aid of a DCmotor and a small electronic control unit.A suitable length of stranded nylon wire is attachedto the left- and right-hand sides of the door andstrung across four nylon roller guides as shown. Thewire is attached to the spindle of a DC motor, therotational direction of which depends on its po-larity. Such motors are available in variety in manymodel building shops or from electrical suppliers,and should be suitable for operation from 6 . . .18 V.

1

EO

9 ... 18 V

83

9 ... 18 V

6

25 V

00k

..44 T414

5

It will be sufficient to loop the nylon wire a coupleof times round the motor spindle. Correct tension isobtained by incorporating a tensile spring in thewire.A small push button switch is fitted in the left- andright-hand door frames so that when the door isfully open or closed, a switch contact is closed.You also need a light barrier or similar device thattransmits a positive pulse of suitable length on theapproach of a person. Such devices have been pub-lished in Elektor Electronics before, and there is alsoone elsewhere in this book.The diagram in figure 1 contains a bridge circuit,consisting of transistors T, . . . T4 which, dependingon the logic level at the bases of T1 -T3 or T2 -T4,determines whether the motor is at standstill,

9 .. 18 V

N1,N2,N3 = 37< IC2 = 4030N4,N5=h 1C3 = 4011FF1,FF2 = ICI = 4013

TD

On25 V

T4

CI) D3 ... D6 = 1N4001T1.T2 = BD 239.T3,T4 = BD 240.

IC3

0IC2 1C1

85492-2

221

rotates clockwise, or turns anti -clockwise. When thecircuit is being tested, the motor may be replaced byDi and D2 (with limiting resistors R8 and R9 re-spectively).The choice of transistors depends on the currentdrawn by the motor, which should not exceed500 mA. T1 -T3 and T2 -T4 form complementarypairs, for instance, BD239-BD240,A short pulse at pin 6 of bistable FF, sets the doorin motion: the first time, it may be necessary to re-verse the connections to the motor! When the door

is fully open, it touches switch S2. It does not mat-ter whether it is just a touch or whether the doorkeeps the switch depressed: the motor stands stillfor a short time, which is adjustable with Pi, andthen rotates in the opposite direction so that thedoor closes. If, while the door is closing, the lightbarrier is actuated, the motor changes directionagain, and the operation repeats itself. When thedoor is closed, switch Si provides a pulse whichcauses the motor to be switched off until the nexttime the light barrier is actuated.

187 BURGLAR DETERRENT

Most burglar deterrent systems are based on thesame principle: once the presence of an unwantedor suspicious individual has been detected (by elec-tronic or other means), some action ensues whichmakes it clear to passers-by or neighbours thatsomething is amiss. It is often overlooked that theunwanted visitor first had to ascertain that there isnobody at home. The majority of burglars whooperate by daylight just ring the bell. Once theyhave repeatedly rung without anyone answering thedoor, they go about their nefarious ways. Once in-side, they may well set off a conventional alarm, butby then it is already too late. The circuit proposedhere was designed to prevent the intruder gettingthat far. When the bell is rung, a number ofmonostables is actuated, which, after a suitabledelay, switches on a cassette player that generatesan awesome sound. This can vary from the barkingof a large dog to the roar of a lion, depending on the

L ----------J

D1

premises. Sometimes a simple "sorry, no canvassers"may be adequate.The circuit consists basically of two monostables.The delay between the ringing of the bell and thecassette player being switched on is preset with 131between 0.22 and 2.4 seconds. The time the cassetteplayer operates is set with P2 between 47 s and 8 m37 s. The cassette player is switched on via the relaycontacts.The circuit is powered by_ the bell transformer. Inthe circuit it is assumed that this is a 6 V type, andthe relay is, therefore, also a 6 V type (which heredraws a current of 50 mA).

85464

222

188 CALL COUNTER

A telephone has the unfortunate disadvantage thatyou have to be near it to be able to make use of itscommunication possibilities. If you have a tele-phone answering unit, you know at least who hascalled and how many callers there were. If you can-not, or do not want to, hire or buy such a unit, thepresent low -budget one may be of interest. Low -budget involves the limitation, however, that the in-coming calls are merely counted: who has called, orwhat the message was, can only be guessed.Moreover, to avoid problems with British Telecom(or whoever your PTT authority is), the unit isacoustically coupled to the telephone. Such a design

1

must, of course, have excellent pulse suppression,since extraneous sounds must not be inter-preted as an incoming call. Finally, the counter'scurrent consumption should be (very) low to enableits operation from a battery.A small, inexpensive loudspeaker is used as thedetector, the output of which is applied to windowcomparator Al -A2. In the absence of a signal fromthe detector, the output at interconnected pins 1

and 7 is logic high. When the loudspeaker picks upa sound from the telephone, the output consists ofnegative -going pulses. Monostable MMV1 is trig-gered by the leading edge of the first pulse, and sup -

Al,A2 =1C1 = LM393

9V

T

S1 onfolf

0IC1

0

R5

9V

2

4m716V

MMV2

TR

I S3C3 hiMilim 9 Fri

i RESET

R7

9

R18

R

IC3 u0

CLK

0 15C2

,o, 16V

O1C2

6

S2

0DISPLAY

R8

14 2

2 6

IC4B

4511

e

d

b

a

9

MMV1,MMV2= IC2 = 4538

IC3

R10

9-- saoa ALL g

12 1313 - 1415

14 2

R16

IC4

86429-1

'see text

R17

T1

BC547©

MAN4740

a

fidLi 9

DP IDP

4

CS

I100n

2

® 9V

0

9V

O

223

presses the pulse. Only after a time lapse of 0.4 sdoes MMVI enable a second monostable, MMV2.If the sound is still being detected by the loud-speaker, MMV2 is then also triggered. This ar-rangement ensures that noise pulses of less than0.4 s duration are effectively suppressed. SinceMMV2 is retriggerable, and its mono time is about5 seconds, the intermittent ringing of the telephoneis converted into a single pulse.The decimal point of the display is switched on viaR is and Ti indicating that the circuit is in a trig-gered state.The remainder of the circuit is straightforward: adecimal counter, IC3, with switch -on reset (R7 andC3, and a BCD 7 -segment decoder, IC4. In thequiescent state, the display is not energized in orderto keep the current consumption low. Pressing S2will indicate how many telephone calls there were.The circuit is reset by briefly switching it off, andthen on again, but could be arranged by a simpleswitch across C3. Current consumption in thequiescent state amounts to about 0.6 mA, so that areasonably long life may be expected from the PP3battery.If the input sensitivity is poor, it may be improvedby lowering the value of RI and R2 to 10 Q. If thisis still not sufficient, a simple input amplifier asshown in Fig. 2 should be added. The LM393 isthen replaced by an LM324, which has foursuitable opamps. One of these is then used as input

A1',A2',A3',=3/41C1'=LM324

amplifier, and two of the remaining three as thewindow comparator. Diodes Di and D2 arenecessary in this case, because the outputs of theLM324, in contrast to those of the LM393, are notopen -collector. The value of R21 is established bytrial and error to find optimum input sensitivity.Adding the input amplifier has the small disadvan-tage of increasing the current consumption toaround 1 mA.

189 CENTRAL HEATING CONTROL

1This circuit is used for optimum regulation of theflow of hot water in a central heating system. Itmeasures the water temperature, and arranges for aparticular valve or pump in the system to beswitched on to achieve a user -defined temperaturedistribution in the home. Residual heat in the cen-tral heating system can thus be used to lower thecost of fuel. Fig. 1 shows that water in temperaturerange I can be used for the central heating and thestorage vessel, while that in range II is also suitablefor directing to the boiler. In most cases, it is not rec-ommended to re -use water with a temperaturebelow 30 °C. The circuit arranges for an alarm tobe activated when the water temperature falls below5 °C, or exceeds 95 °C.The circuit diagram of the central heating controlappears in Fig. 2. Relays Re, and Res are activated

95 Re1

90

50

30

T CC)

5 -IP. Re587412-1

224

2

Al, A2 = IC2 = TLC272A3, A4 =1C3 = TLC272A5, A6 = IC4 = TLC272N1...N6 = IC5 = ULN2003T1...73 = BC5478

08V

IC1

LM 35

3

Fl

R2

ULN2003

2

Al

R1

0

Dl. .D4 = B40C1000

87412-2

8V

Relay PresetTemperaturerange

1 Pi 93-103 °C(upper limitalarm)

2 P2 77-93 °C3 P3 33-77 °C4 P4 11-33 °C5 Ps 5-17 °C

(lower limitalarm)

(hysteresis on a I toggle points:2 °C).

225

upon measuring the maximum and minimum per-missible temperature, respectively. The temperaturesensor is a Type LM35, which has a scale factor of+10 mV/°C. Its output voltage is amplified in Aland fed to the non -inverting inputs of comparatorsA2 -A6. The presets at the inverting input of each ofthese is used to set the toggle voltage, i.e., the tem-perature at which the relevant relay is switched onor off. The relay drivers are open -collector powerbuffers with built-in freewheeling diodes to affordprotection against inductive surges. The use of theType ULN2003 makes it possible to use relays witha coil voltage of upto 50 V without the need for ad-ditional interfacing.

Each temperature setting has a hysteresis of about2 °C. Transistors Ti -T3 serve to disable thepreviously energized pump or valve upon detectinga water temperature that falls within another,predefined, range. In this manner, only one relay isactivated at a time.It stand to reason that the temperature sensor, IC,,must be mounted such that it is in thermal contactwith the water in the heating system. Make surethat the device is well -insulated, and that it does notcause leakage.

190 CH BOILER CONTROL

If you still alter your central heating system's boilerthermostat according to the season (many peoplenowadays leave it at the same - fairly high - set-ting throughout the year), this may cause the boilerto be switched on and off too frequently when theweather is unseasonly cold (see figure la). Thisproblem may be resolved by the present circuitwhich prevents the boiler being switched on forsome time, td, after the switch -off temperature, T2,has been reached. After td has lapsed, the boilertemperature, T, should have dropped well below theswitch -on temperature, Ti (see figure lb).The circuit in figure 2 is an extension of the centralheating monitor (Elektor Electronics, July/August1984, p. 7-30) and formed part of the electronic gasmeter (Elektor Electronics, November 1984, p. 11-59). The make contacts of a relay are inserted intothe 24 V boiler circuit. The state of bistable N2 -N3determines whether transistor Ti is on or off, i.e.,whether the 24 V circuit is open or closed. As soonas the bistable is set, Ti conducts and delay time tdcommences. At the same time, the reset of counterIC2 is cancelled. After some time, IC2 has reachedmaximum count: the consequent change of logiclevel at the output selected by Si causes the reset ofthe bistable via N4. This is the end of td.The set input of the bistable is connected to the col-lector of T3 in the central heating monitor via Niand R3 -C2 and R2 -G. That transistor drives theLED that indicates the interrupted heat request ofthe boiler thermostat.Delay time to can be set within wide limits with P1and Si. A period of 10 minutes is probably a good

la

lb

Toff

Ton

Toff

T..,

T weer

226

2

5C mtCoul-leczornoi:ois3

Ole room thermostat

Oa CH monitor

R2

ICE D.15V

R4

O

a

C3

R3

C2

15 V

0

N2

IC1

BC 549C

R9

R113

002 Re1

SiemensV23027 A2-4101

12V)

1N4148

71 72

5 4

47n6

IMO

1N 4148

=I 22,,25 V

8

R5

N3

N4

.a

R11

BD 139

U

12

16

Cfrood

1.02

rouge

15V

2

3

33nR6

Si

20'

0

11

IC22 4060 0

R7

220k

24 V.

2

R8

EMI /AlC5 PI

50k

N1 ... N4 = IC1 = 4093

starting point. It is then impossible for the boiler tobe switched on and off more than six times perhour. Based on the number of times the relevantLEDs in the central heating monitor light, you canalter the delay time with P1 and Si. Briefly connect

I 85467.2

the junction of R3 -C2 to earth: this causes thebistable to be set; D3 goes out and the delay periodstarts. Adjust P1 so that D3 lights again after 5, 10,or 20 minutes. It is also possible to use periods of4, 8, or 16, or 6, 12, or 24 minutes.

191 COLOUR WHEEL

Coloured light effects enjoy a high popularity as or-naments, eye -characters, etc., and the present circuitproves how an apparently rotating colour effectmay be obtained with a mere handful of commonlyavailable components.The colour wheel is composed of twelve bicolourLEDs, arranged in a circular form. First, a set offour red LEDs lights, followed by a green set, and,finally, amber. The colours are arranged to move ina clockwise direction, and at a speed that givesviewers the illusion that the motion is smooth andcontinuous.The bicolour LEDs consist of anti -parallel connec-

ted green and red diodes in a single transparent case.When both light simultaneously, their compositecolour, i.e. amber, is emitted.Which group of LEDs lights depends on the dutyfactor of the drive signal from gates N5, N6 and N7.Gate N1 is a clock oscillator whose frequency iscontrolled by Pi IC1 has been connected to func-tion. as decade ring counter, which sequentiallyenables oscillators N2, N3 and N4 by a logic highlevel at counter outputs Qo, QI and Q2 respectively.If, for instance, N2 is enabled, it oscillates at about500 Hz with an output duty factor of 50%, causingboth the green and red LEDs contained in Di . . D4

227

to light. At the same time, D5 . . D8 and D9 . . D12

light as red and green, respectively, since the associ-ated driver gates N6 and N7 provide a steady highand low level, again, respectively.

IC4IC5 IC3 IC2

R1

ClNom

12V

050mA

0

R

00

ICI

4017

0301

CLK

13 8

02

N13

++ ++ ++ ++

+4,

G

4+

N1...N4 = IC2 = 4093N5...N8 = IC3 = 4030N9...N 14 = IC4 = 4050N15... N20 = IC5 = 4049D1...D12 = LED*

++

* see text

86510

N16

N18

N20

192 CURRENT MONITOR AND ALARM

These circuits are intended for remote monitoringof the current consumption on the domestic mainsline.The circuit in Fig. 1 lights the signal lamp upondetecting a mains current consumption of morethan 5 mA, and handles currents of several ampereswith appropriate diodes fitted in the Di and D2 pos-itions. Transistor T, is switched on when the dropacross Di -D2 exceeds a certain level. Diodes fromthe well-known 1N400x series can be used for cur-

rents of up to I A, while 1N540x types are rated forup to 3 A. Fuse Fi should, of course, be dimen-sioned to suit the particular application.A number of possible transistor types have beenstated for use in the T, position. Should you con-sider using a type not listed, be sure that it can copewith surges up to 700 V. As long as T, does not con-duct, the gate of the triac is at mains potential viaCI, protective resistor R2 and diode D3, whichkeeps Ci charged. When T, conducts, alternating

228

1

TXC 02A50BUX 82BUY 69BU 205

*see text

87433-1

current can flow through the capacitor, and thetriac is triggered, so that La, lights.The circuit in Fig. 2 is a current triggered alarm.Rectifier bridge D4 -D7 can only provide the coilvoltage for Re, when the current through D1 -D2 ex-ceeds a certain level, because then series capacitorC, passes the alternating mains current. CapacitorCi may need to be dimensioned otherwise than

2

3AR1 BU 246A

D1, D2 1N5401D3...D7 = 1N4004

Cl 220n/400V

-1

8 C E

Ret

24V200(1

87460-1

shown to suit the sensitivity of the relay coil. Thisis readily effected by connecting capacitors in paral-lel until the coil voltage is high enough for the relayto operate reliably.

Finally, an important point:Many points in these circuits are at mains potentialand therefore extremely dangerous to touch.

193 DECEPTIVE LOCK

This circuit offers a means to fool all but thecleverest burglar, but, although it is a clever design,it has been kept simple, as a glance at the diagramwill show. On the surface it looks like a simpleoperating panel with ten push buttons. However,anyone trying to open it illegally is in for a surprise!It is not just a matter of keying in the correct code,it is also necessary to keep one of the keys depressedfor about 10 to 15 seconds.The circuit is based on a single Type CD4093,which contains four NAND gates with Schmitttrigger inputs. Gates N, and N2 form a bistable thatcontains the status of the lock.Assuming that the circuit has been off for sometime, switching it on causes network RI -C2 to setthe bistable to the "lock" position, that is, the out-put of N2 is logic low. Capacitor C3 is discharged,and the only way the circuit can toggle is by

recharging this capacitor. This is done by pressingkey Sx long enough for the trigger threshold of N,to be reached. When that happens, the bistable isset to the "open" position, that is, the output of N2is logic high.Capacitor C3 remains charged via R4 -D2, even afterSx has been released. In other words, the bistableremains in the "open" position. The lock is closedagain when one of the other keys is pressed, or, ifrequired, by means of a special lock key. This causesC3 to discharge rapidly via Dl -R3, which returns thecircuit to the "lock" condition.When the lock is "open", relay Re will also be openin the present circuit. It is, however, possible to havethe relay energized in this condition by connectingthe remaining free gate in the 4093 in series with R5as an inverter.

229

l<0,1mA

R1

"1Q0

ISO S8 I lock

NMI

D3

A1N4001 Re

T1R5

N1-N4AC1:4093

80139

to

a C)86410

5-15V

194 DESICCATOR

Many of the smaller working areas available to hob-byists suffer from humidity, which in no time causesa number of tools to be covered in a thin layer ofrust. Humidity does not do most test equipment orbooks and the like any good either. The onlysolution to this is to try to keep the area drier by in-creasing the temperature.

IC2

A couple of 100 W light bulbs or a 100-200 Wheating element work wonders in this respect, wereit not for the increases in the electricity bill. Andthat is where the present circuit can help.With reference to the diagram, the twoHEF4001Bs, in conjunction with humidity sensorH, generate a voltage across R5 that is directly pro-

N1...N4 =ICI = 4001N5 = 1C2 = 4001

A1,A2 =1C3 = LM358

471

C11100n6000

1W

a2

TIC206D

GLa1

1W

Co

DY

1N4007

Tag,A.1070n

23086506-1

portional to the degree of humidity. The function ofopamp Al is merely to present a high impedance toUR5. The voltage is then applied to the invertinginput of comparator A2, which has an hysteresis ofabout 15 per cent. The reference voltage at the non -inverting input of A2 can be set between 0.6 V and3.0 V with PI, which corresponds to a humidity be-tween 20 per cent and 100 per cent. As soon as theambient humidity exceeds the value set by thecomparator toggles, the triac conducts and switcheson the heating element. Current consumption ofthe circuit is a modest 13 mA.If light bulbs are used, they should be shielded with

a metal hood to prevent the likelihood of a fire.Calibration is carried out with the aid of a solutionof cooking salt in some water; placed in areasonably small, closed space, this will soon raisethe humidity to 75 per cent. Adjust Ci to obtain apotential difference across R5 of 2.25 V. Next, ad-just Pi so that the triac just does not conduct. Inpractice, the circuit will then come on at a humidityof about 80 per cent.

For further details of this circuit, see also the April1981 (p.19) and July & August 1981 (p.72) issues ofElektor Electronics.

195 ELECTRONIC BELL -PULL

The simplest circuit in this issue of Elektor Elec-tronics consists of a single transistor and resistor,which, when put together as shown, constitute theelectronic equivalent of an old fashioned, stylishbell -pull used in conjunction with a chime or bellcircuit of any relative complexity offering whirlingmelodies, buzzing or ringing sounds, or chime im-itations to prompt the houseowner to open the frontdoor.The bell -pull is made from a TO39-style NPN tran-sistor which is taped or isolated by means of alength of heat shrink sleeving, after the emitter andcollector leads have been fitted with wires for theelectrical connection to the bell circuit. A small,conductive plate is secured onto the isolated transis-tor head, and the base lead is joined to this plateover a series resistor which is dimensioned accord-

UbR = -(kri)5

86446 -1

ing to R = Ub/5 [1(521. The completed assembly mayalso be cast into epoxy resin to make a nice compactunit that can handle the treatment of even theroughest caller at the door!

196 ELECTRONIC DOG

To produce a faithful reproduction of the voice ofman's best friend, we have borrowed several ideasfrom our music synthesizer. When push buttonswitch S2 is pressed, the frequency of voltage -controlled oscillator (VCO) Al -A2 changes in aboutan eigth of a second from almost 0 Hz to a preset -table value of 100 . . .1000 Hz. That signal is passedthrough band-pass filter As -A6, the centre fre-quency of which corresponds with the highest VCOfrequency. Voltage -controlled amplifier (VCA) Tiensures that the single pulse generated by the VCOwhen S2 is open cannot be heard.Gates Ni and N2 form a monostable relaxation os-

1

cillator. When S2 is closed, a short pulse appears atthe output of N2 that charges capacitor C2. Because

231

of R3, the pulse shape will be as shown in figure 1.This pulse controls the output frequency of theVCO as also shown in figure I. Potentiometer Pidetermines the highest frequency: its settingdepends on whether you want the sound of a yapp-ing poodle or the deep bark of an alsatian.The function of C4 is similar to that of C2: it shapesthe pulse applied to the VCA. This transistorbehaves as an electronic potentiometer, i.e., itoperates as a voltage -controlled resistor. Adjustingpotentiometer P2 influences the manner in whichthe tone decays after the switch has been released.

29V

R1

P2

50k lin.C4

v - 1N4148

N1 ... N4 = IC1 = 4001

Al ... A4 = IC2 TL 084A54,..A7 = % IC3 = TL 084

Instantaneous dying of the tone would sound justas unreal as its lingering on. With a little care, andafter some practice, it will be possible to create avariety of canine dialects.The centre frequency of band-pass filter As -A6 is setwith Pa Correct setting of this is important, buthere again, trial and error is probably the best way.With the output connected to a power amplifier, thecombination can be used as an alarm installation:even dyed-in-the-wool burglars think twice beforethey risk entering a house that is obviously guardedby a fierce dog!

0

BC 547

R11

DI

HC3 1N4148

10n

R10

P3

10n

R6

C7

100p16V

0

C8

100p168

85448-2

9V

IC2IC3

197 FLASHING LIGHT WITH TWILIGHT SWITCH

The special feature of this flashing light is the op-tical switch, which automatically switches the lighton when it gets dark, and switches it off again atdawn. This makes the light ideal as a warning lightnear obstructions. It may also be used for educa-

tional purposes to show the operation of transistorsin conjunction with optoelectronics.Assuming that it is light, the LDR (light dependentresistor) has a low value so that there is sufficientbase current through Ti for the transistor to con -

232

6 ... 10 V

85401

BS170*

/ I \G S

duct. Its collector voltage is then small, so that T2,an n -p -n darlington, is off, and lamp Li stays out.When the ambient light reduces, the resistance ofthe LDR increases until the base current in Tibecomes insufficient and the transistor switches off.Its collector voltage then rises, T2 conducts, and Lilights. This process takes place quite quickly,because when the collector voltage of T2 suddenlybecomes nearly 0 V, this potential is immediately

applied to the base of Ti via capacitor Cl, whichreally cuts off Tl. The capacitor then charges via P1and the LDR that is now being illuminated by thelighted lamp. Owing to the optical feedback, thevalue of the LDR diminishes, the voltage across R2

increases, and Ti conducts again. The darlingtonswitches off, and the lamp goes out: a new cycle hasstarted.The flashing frequency is primarily dependent onthe value of with 47 fAF, it is rather low; reduc-ing the capacitance increases the frequency.The BC 517 darlington may be replaced by twoBC 547B transistors or a VNIOKM MOSFET. Theonly thing that needs watching is the currentthrough Li: the maximum permissible with twoBC 547Bs is 100 mA; with a BC 517 it is 400 mA;and with a VNIOKM it is 500 mA. Current con-sumption of the circuit, with lamp Li out, is about6 mA at 6 V and around 10 mA at 10 V.The light -dependent resistor may be one of theusually available types: LDR 03, 05, 07. To ensurethat the optical feedback works, the LDR must befitted near lamp Li. The onset of operation is setwith Pi

198 FOUR POSITION TOUCH DIMMER

Any electric light may be adjusted with this dimmerto very low, low, medium and maximum, which inmost cases will be sufficient. After all, it is all verywell to be able to control an electric light over the

R5 R4

TIC206D/226D

Al A2 G

whole range of its brightness, but how often is thatfacility really used? Moreover, in everyday use, pos-ition control has practical advantages: setting, forinstance, takes a second or two.

1N4148 220 n400 V

TIC206D/226D

Al ClImomin

Tril 150 n400 V

I L140pH

O

La

max, 400 W

85444

233

The circuit is based on an LS7237 and some discretecomponents. The dimmer may also be used as anelectronic on/off switch, in which case the mode sel-ect pin (7) must be connected to earth (pin 1). Sucha switch does not produce sparks and consequentnoise in nearby electronic equipment. Anotherpossibility is leaving pin 2 open, whereby a threeposition dimmer ensues: low, medium, and maxi-mum.The LS7237 has all the necessary facilities to drivesilicon -controlled rectifier (SCR) Trii. Resistor R2and capacitor C4 filter a 50 Hz signal from themains that serves to synchronize the on -chip phaselocked loop.

Network R1, C2, and D2 provide the supply for theLS7237, while filter Li/Ci prevents excessive noisefrom reaching the mains supply.Different types of triac may be used, as long asthese can provide the required current, and aresuitable for operating voltages of not less than400 V. For safety's sake, no deviations from thestated voltage ratings of the various componentsshould be tolerated. The two 4M7 resistors provideample safety for the user: under no circumstancesshould these be replaced by a single 10 MQ resistor!The complete circuit is small enough to be accom-modated in the pattress or plaster box of a lightswitch.

199 HOTEL SWITCH

It is often required to switch an electric light or ap-paratus from various positions in a building. Atypical example of this is the hotel switch, whichmakes it possible to control lights from a number ofpositions. With some electronics and electric wiring,the number of switching positions may be extendedad infinitum.The actual switching is effected by a relay that iscontrolled by an R -S bistable, N2/N4, via transistorsTi and Tz The state of the bistable is of import tothe position of logic switches Ni and N3. A triggerpulse at the junction of R-1 and CI is only appliedto that input of the bistable which causes thebistable to toggle. In other words, a train of triggerpulses, 0;1;0;1;0 , with a minimum interval be-tween pulses of a few seconds, results in a series of

5...15V

85454

logic level changes which causes the relay to be ac-tuated and de -energized alternately.The trigger pulses arise when one of the push but-tons, Si . ..Sn, is pressed briefly. The push buttonsare all connected in parallel, so that they can be in-terlinked by a two -wire system.It would be possible to fit an LED at every switchposition, but this would entail an additional wire.Such LEDs would, of course, also be in parallel, sothat it is advisable to use similar types.The value of resistor Rio is calculated fromRio = ((U - 2)/IDn) Qwhere U is the supply voltage in volts; ID is thecurrent through each LED in A; and n is thenumber of LEDs.

234

200 INDUSTRIAL -CLOCK CONTROLLER

Nowadays, most clocks and watches are quartz con-trolled and, therefore, accurate to within a fewseconds a year. Older type electric clocks, particu-larly those used in large groups in warehouses,department stores, factories, railway stations, and soon, were centrally controlled and synchronized.This synchronization was effected by pulses derivedfrom the mains and sent to each clock via a separatecable network. Many people have such a clock as acuriosity, but have not the means of driving it. Thecircuit described here will help . .

With reference to the diagram, pulse shaper T, trig-gers monostable IC2 at the mains frequency of50 Hz. Counter IC3 is reset automatically afterevery 3000 pulses by IC4 and T2. At the same time,bistable IC5 toggles and causes the bridge circuitcomposed of T2 . . . T8 to reverse the motor polarityevery 60 seconds.

Fl 4x1N4001

* see text

1N4148

C1.-C4.4.100n

470171087630n

(7)555

Depending on the type of clock you have, the trans-former secondary voltage may have to be selected tosupply about 0.7 times the normal operatingvoltage of the clock motor. Furthermore, the bridgecircuit as shown should not be made to operate atvoltages in excess of 30 V, while the maximum cur-rent is about 250 mA.There is only one adjustment point in the circuit,namely Pi, which should be set to achieve maxi-mum suppression of mains borne noise; if this cannot be checked, the preset may be turned to itscentre position. Should the clock be slow, Pi may beadjusted to give a slightly lower resistance, but careshould be taken to avoid setting a monostable timelonger than 20 ms, as in that case only half thenumber of 50 Hz periods can reach the counter.

041C3 05

CLK 07

BC547

105

86470

BC

547

810

235

201 INFRA -RED LIGHT BARRIER

Infra -red light barriers enjoy great popularity astiming devices at sports venues, as detectors inalarm installations, as optoelectronic switches incounting equipment, and many others, because oftheir low cost and immunity to electrical inter-ference.The present light barrier consists of a transmitterand a receiver.The transmitter, shown in figure 1, consists of anastable multivibrator (AMV), lC3. The output ofthe AMV, pin 3, consists of a pulse stream with aduty factor of about 30 per cent. The output is con-nected to a constant -current source, T2. This sourceprovides infra -red transmit diodes D7 and D8 witha current of just over 20 mA, which pulsates inrhythm with the output signal of the AMV. Theinfra -red light is, therefore, transmitted in rhythmwith the pulse stream also.The receiver, shown in figure 2, is based on anSL486 demodulator, IC). The output of the

2

-

C5mum

:On

6

BP 104 15

C7

7n

C9

4n7

R1

47

12 13

1

IC2

13

C6 C12 C11vim miniton Tn2 7n

N1 ... N4 = IC2 = 4093

10 mA

2x1N4148

D3

9V0 C)

dF ReA B

D2 1:23

PBrs 2720

85449-2

BC 550C

* see text

236

Parts list (receiver)

Resistors:

Ri = 47 QR2,Ra = 470 kR3,138 = 1 kR5 = 4k7R6,Ra = 22 kR7 = 1 0 k

Capacitors:

Cl,Cio = 10 µ/16 VC2 = 10012/16 VC3 = 33 nC4, Ca = 4n7Ca = 330 nCa = 150 nC7 = 15 nCa = 2211/16 VCli = 47 nCl2 = 2n2

Semiconductors:D,... D3 = 1N4148D4 = BP104Ti = BC550CIC, = SL486 (Plessey)IC2 = 4093

Miscellaneous:

Rel = PCB type relay, 6 VBz = piezoelectric buzzer

(Toko type PB2720 or equivalent)Si = spring -loaded push -to -make switch

Parts list (transmitter)

Resistors:

Rio = 39 kRil = 82 kR12 = 3k9Rio = 3Q9

Capacitors:

C13 = 100 nCia,Cla = 1 n

Semiconductors:

D5,D6 = 1N4148D7,Da = LD271HTi = BC560IG = 555

Miscellaneous:Reflectors for D7 and Da (optional)PCB 85449

demodulator, pin 11, also consists of a 10 kHz pulsetrain with a duty factor of around 30 per cent. Thispulse stream is applied to integrator R2 -C12. Thelogic level at the input of N, remains low as long asD4 receives the pulsating infra -red light. Because of

237

this, monostable N2 is disabled, and oscillator N4,which drives a piezoelectric buzzer, is switched off.Relay Rei is, however, energized via N4 and transis-tor Ti.When the pulse stream between D7 -D8 and D4 isbroken, the logic level at pin 11 of IC, goes high, sothat the output of Ni becomes logic 0, which trig-gers monostable N2 via Di. Oscillator N4 is thenswitched on and actuates the buzzer. At the sametime, N3 ensures that Ti is switched off, so that therelay returns to its quiescent state.When the monostable pulse decays, which with thestated values of R4 and Cio is after about 5 seconds,oscillator N4 stops and the alarm tone ceases. DiodeD3 ensures that the relay remains in its rest state,however, by transferring the high voltage level ofthe collector of Ti to the input of N3 whose conse-

quent low logic output continues to hold the tran-sistor off.The equipment switched by the relay contacts,therefore, does not only indicate when the light bar-rier has been interrupted, but also when the supplyvoltage has failed. The relay is re -energized whenreset switch Si is operated. If D3 and Si are omitted,the relay is re -energized when the monostable pulsehas decayed.Current consumption of the transmitter is about50 mA; that of the receiver around 10 mA.The printed circuit board shown in figure 3 is in-tended to be cut into three along the dashed lines,although it may not be necessary in some situationsto cut the relay section from the receiver section. Ifthe latter two are separated, they should on comple-tion be interconnected by a suitable cable.

202 JUMBO DIMMER

The name jumbo dimmer points to its associationwith the Jumbo Display (see EE, July & August1985), but it can, of course, also be used with otherappliances such as lamps, pumps, ventilators: inshort for all applications where a direct voltage is tobe controlled by pulse duration modulation.With reference to the diagram, A i is a rectangular -wave generator: a useful by-product of this stage isthe (quasi) triangular voltage at its inverting input.This signal is applied to the non -inverting input ofcomparator A2. The reference voltage for this stageis derived from preset Pi. The output of the com-parator is a rectangular voltage with a frequency ofaround 200 Hz and a pulse duration that is variablebetween nought and 100 per cent. The onset pointof the pulses is determined by the setting of PI.The actual control function is provided by transistorTi, which switches the relatively large display cur-rent of up to 5 A.

5.30V

A1...A2 =IC1= CA3240

The supply voltage must lie between 5 V and 30 V:note that the efficiency of the circuit is directly pro-portional to the supply voltage.

203 JUMBO DISPLAYS

Although this project will not be of interest toeverybody, it has many possible applications. Thename refers to the respectable dimensions of theseven -segment displays: 280 x 140 mm. These sizesimmediately indicate that the displays are intendedto make alphanumeric information legible at a

distance. This is of import, for instance, for scoreboards, speed indicators, lap counters, digitalchurch clocks, etc.These displays have a number of advantages: they are entirely solid state, which prevents

segment failure since the life of LEDs is much

238

la+5

Clint

1100n

R3

5 V 0 0

16

a13

R412

IC1AO

115135:f B

C11co

R674LS

10

24830

LTLTso

RBI 81RBI0

4cRBO 0 WOW

R24

+24

111111113

14

T5

T7

R S0 0

2

1-1...1.7=7x13C517

Figure 1. Circuits for the control of (a) a seven -segment display; (b) a "1" display, and (c) a ":" dis-play.

b

8541.11,

longer than, for instance, that of incandescentlamps; they do not need intricate reflector construc-

tions; if any one LED fails, they remain fully legible by

virtue of the special segment construction; they are easily arranged in a variety of colours -

red, green, blue, yellow, orange; they work from 24 V with relative high ef-

ficiency, which keeps heat dissipation low.It may be said that the large number of LEDs re-quired is a disadvantage, but, in our opinion, this islargely negated by the advantages.

C

24 V

85413-1a

The seven -segment display, shown in figure la, is

based on a type 74LS248 decoder, which has thesame features as the well-known type 74LS471247,but has in addition internal pull-up resistors and in -

239

2 I -ejjc n I- IJ O i= I_ lb

0 I 2 3 I. 6 6 3 8 9 10 11 12 9 18 M

856134

Figure 2. Correlation between the input and outputsignals of a 74LS248 decoder and a seven -segmentdisplay.

Figure 3. Printed circuit boards for the " :" display;the seven -segment display; and the "1" display.

31-...0"../azor..."....c-...,"...- ......e......EN....-:%.,TIT. .- - . . ..i......, . ' . -- %, . . I

' 1 1 \\

c-+1 li1111111 1= 1,,,,,..............., i----...., -- .......,......,----

r111

\gui \\

k\4/

i/ %rk;4)i

\\T 0\\ ii\\ ii

240

verted output signals, so that external transistorscan be used to cope with the large currents drawnby the segments. The inputs and outputs to thedecoder, the read-outs, and the additional functionsare correlated in figure 2.All input and output controls have been arrangedexternal to the decoder, so that they can be used inthe same way as with normal displays. Wire link R-

S serves to interconnect the earths of the + 5 V and+ 24 V supplies.At the output of the decoder there is a switchingstage for each segment that switches the relevantsegment on or off.Each segment consists of four parallel groups ofeight or nine LEDs in series with a current limitingresistor.

241

Parts list

Seven -segment display:

...R7 = 100 k198...Fill (7X) = 270 Q (with 9 LEDs)

= 330 Q (with 8 red LEDs)= 390 Q (with 8 green LEDs)

= 74LS248Ti,. .T7 = BC517Ci = 100 n232 LEDs, 5 mm, colour as required

"1" display: = 47 kR2 = 1 MR3,R4 = 470 kR5 . R8 (2X) = 270 QDi,D2 = 1N4148

= BC517BC547B

72 LEDs, 5 mm, colour as required

":" display:19-1,R2 = 270 Q18 LEDs, 5 mm, colour as required

PCB 85413-1PCB 85413-2PCB 85413-3

The displays can be powered from a non -stabilized20...24 V supply. The current drawn per segmentvaries from 50 mA to 100 mA.Figures lb and lc give the diagrams for displayswith a "1" and a ":" respectively. Both can be usedfor a 12 -hour clock. The "1" display has provisionfor a lamp test (LT); open inputs are considered ac-tive, i.e., the display lights. This is in contrast to theseven -segment display which treats inputs that arenot connected as logic high, that is, inactive.As mentioned earlier, read-out boards consisting ofseveral figures may be composed by mounting anumber of displays side by side on a flat base. Thewhole may be protected by translucent red perspex:this also acts as a light filter, which improves thelegibility considerably.As you need a large number of LEDs, shop aroundfor these because many dealers are prepared toallow a quantity discount. Uniformity of brightnessof these diodes is not so important for this appli-cation, because at the distances for which these dis-plays are intended, differences in brightness do notshow up.

242

204 LED DIRECTION INDICATOR

An LED indicator with a difference: three alter-nately lighting LEDs indicate a direction, for in-stance, in a model railway, or to an emergency exit,or to a door on badly lit stairways, and so on.When the supply voltage is switched on, the inputsof gates I14. . . N6 are logic 1, their outputs logic 0,and all LEDs light. One of the RC networks(R, + R2/C2; R3/C3) will reach the triggerthreshold first. Let us assume it is R, + P,/C,. Theoutput of N1 then goes low, the output of N4 goeshigh, and DI goes out. There is then no voltage forR2/C2, the output of N2 remains logic high, and N5remains logic low: D2 then lights. Subsequently, theoutput of N3 goes low, the output of N6 becomes 1,and D3 goes out. The logic 0 of N3 is, after a delayin R, + P,/G, again at the input of Ni. The out-put of Ni goes high, that of N4 goes low, and D1lights. This process repeats itself, so that first one,then two, and then one LED again lights. At everystep, the light pattern shifts one place to give the im-pression of a running, flashing light. The runningspeed is set with PtIt does not really matter whether you use invertinggates (4049) or non -inverting ones (4050) in the IC2position, as long as you connect the unused gates tothe positive or negative supply rail. The RC net-works may also be modified to taste or if special ef-fects are desired.

If you want to make the circuit even smaller, forgetIC2 and use the three remaining inverters in IC, asLED drivers, provided you are using a type 40106.The LED currents are then only 5 . . .10 mA, soyou have to use high output LEDs (that are brightat low currents).The current consumption of the circuit withoutLEDs and operating from 15 V is about 100 µA.With LEDs, it depends very much on the LEDs andthe supply voltage: with standard LEDs and at15 V, each LED draws up to 30 mA.

Ni N3 = ICI ='340106.14409315 V N4 N6 = IC2 ='54049, '54050

LJ

C1 IC2

0 0

205 LIGHT-SENSITIVE SWITCH

This switch is energized by light and can, therefore,be used, for instance, to switch on the aquariumlighting in the morning. Both the sensitivity and thehysteresis of the circuit can be preset; Re is ener-

T1

TIL818P103

Ni2k

A = onB = autoC = off

gized in the presence of sufficient light.The sensor is an n -p -n phototransistor Type TIL81or BP103, which conducts when light falls upon it.The consequent current is divided between T2 and

86500

243

R4-Cl. Since T2 is connected as a current source, nocurrent will, however, flow through Ra-C, as long asthe current in T1 is smaller than that through T2 asdetermined by PI. When the current in T1 is largeenough, some will flow through R4 and charge Cl.As soon as the resulting potential across Ci isgreater than half the supply voltage, the CA3130toggles. A current then flows through R8, P2, andR3, which will cause a small reduction in the cur-rent through T2. This means that even if the cur-rent in T1 drops slightly, the circuit will not revertto its original state. The magnitude of this hysteresisis dependent on the setting of P2. Note that thehysteresis prevents the circuit oscillating around thestarting level.

The sensor may also be a photodiode or light -

dependent resistor (LDR), but a phototransistorgives better performance, particularly when the dif-ference between the on and off states of the circuitis small.Resistor R4 and capacitor Ci could be omitted, butthey augment the hysteresis by delaying the inputsignal from reaching the CA3130.The current consumption of the circuit is deter-mined primarily by the requirements of the relay. Ig-noring the relay, the circuit consumes about 10 mA,which makes it possible to use a Type 78L12 asvoltage regulator.

206 LIGHT-SENSITIVE TRIGGER

This circuit activates a relay upon detecting theabsence of light on an LDR (light dependent re-sistor). It is particularly well suited to control out-side lighting as used for driveways and garage en-trances.Contrary to its normal use as an astable ormonostable multivibrator, the Type 555 IC in thiscircuit functions as a comparator. To explain thisrather unusual application, it is neccessary to notethat the operation of a 555 is normally as follows:the output goes high upon receipt of a trigger (start)pulse on input pin 2. This pulse is a voltage whoselevel is lower than IA of the supply voltage. The out-put goes low again when the voltage at the secondinput, pin 6, has briefly exceeded 2/3 of the supplylevel. In the present design, the second input is notused, but the output of the chip can none the lessrevert to the low state, since pin 6 is connecteddirect to the positive supply rail. This set-up is ac-counted for by the accompanying Table, taken fromthe 555's data sheets.In principle, the supply voltage for the circuit mustequal the coil voltage of the relay. Do not applymore than 16 V, however, as this may damage the555. The current consumption of the circuit is

4 mA, exclusive of the relay, at a supply level of12 V. Components R2 and Cl ensure a delay ofabout 10 s before the relay is energized, so that thecircuit is rendered insensitive to rapid changes in thelight intensity.Basically, the circuit has no hysteresis effect. How-ever, when the supply is not regulated, the actuationof the relay will lower the supply level somewhat.This lowers the internal threshold of the IC, since

<16V

.1111111D I g)

aTRESH ACC

TRIG OUT

IC1 DIS555

GND CONT

<100mA

87439

the trigger point is defined as 2/3 of the supply level(pin 2). Therefore, the hysteresis of the circuit can bedimensioned as required by fitting a resistor in serieswith the supply. It is also possible to fit a resistor be-tween pins 5 and 7 of the 555, as shown in the cir-cuit diagram. The amount of hysteresis is inverselyproportional to the value of the resistor, and 100Kis a reasonable starting point for experiments.The sensitivity of the trigger circuit can be con-trolled if RI is replaced with a IMO potentiometeror preset.

NE555

FUNCTION TABLE

RESET(4)

TRIGGERVOLTAGE (2)

THRESHOLDVOLTAGE (6)

OUTPUT(3)

DISCHARGESWITCH

Low Irrelevant Irrelevant Low On

High < 'A VDD Irrelevant High Off

High > Y3 Voo > 34 VDD Low On

High > 'A VDD < % VDD As previously established

244

207 LONG INTERVAL TIMER

This low-cost timer circuit can offer switching inter-vals up to about 24 hours and may, therefore, beuseful for a variety of domestic as well as electronicapplications.Depression of S, causes Re, to be energized and thetimer to be started; the position of P, determinesthe duration of the timing interval - the givenvalue for C2 allows a maximum of 12 hours. Doub-ling the capacitance of C2 lengthens the timing in-terval accordingly; the timer may thus be employedto control a NiCd battery charger. Depressing Si atany time during the interval causes the timer to bereset and Re, to be deactivated.

0 10 - 12V

C1 I100916V

81C1 1C2

®

.51,eo-.

tl 82

1N4148

01

BC547B

eJ S0

CK FF1

K R P4

T1

011

Re1

The funtion of FF, is that of a debouncer circuitfor Si which, when actuated, causes FF, to applya logic high pulse to the clock input of FF2, whichtoggles. IC2 starts counting, since its reset conditionis ended. At the same time, T, is driven with apositive logic level, and Re, is energized. After thetiming interval has lapsed, i.e. when counter outputQ13 goes high, FF2 is reset and Re, deactivated inconsequence.Setting the exact duration of the timing interval isreadily accomplished by temporarily using counteroutput Q3 rather than Q13 to reset FF2; with thecomponent values as indicated, the interval shouldbe adjustable between 3 and 45 seconds. Divide thedesired relay -on time by 1024 and set P, according-ly; connect the FF2 R input to Q13 again, depressSi and have Re, power the relevant equipment foras long as set.

R6

10

13

J

CK FF2

0

Ro

9 12

15

0

R4

83

12

100P1

R5

= C2Rim

226 16V

11 10 9

1'1 470 4)0 3 S2

1C20 7

FF1,FF2 = 1C1 = 40271C2 = 4060

Re 1 =12V86459 1

208 MAINS -BASED REMOTE CONTROLLER

This combination of transmitter and receiver is

based upon the use of the mains network in thehome for remote control of mains -operated dom-estic appliances.Figure 1 shows the transmitter, which merelysuperimposes a 36 kHz signal on the 50 Hz mainsvoltage if Si is operated. It is noted that ICI is feddirect off the mains voltage by means of a rectifier

circuit composed of Di, D2, zener diodes D3, D4,and smoothing capacitor C4; the proposed con-figuration is to supply + 20 V with respect to themains neutral (0) line. The 36 kHz output signal ofthe opamp is fed to the mains by means of couplingcapacitor C3. R2 is a bleeder resistor to dischargeCI and C2 after the circuit has been unpluggedfrom the mains outlet.

245

The receiver, shown in Fig. 2, is fed with an inex-pensive door bell transformer, although any othertype supplying 6 to 8 V AC at about 300 mAshould do just as well.Apart from being used to power Tr!, the mainsvoltage with the 36 kHz carrier is filtered by paralleltuned circuit Li -C6 to detect the presence of thesuperimposed 36 kHz carrier, which is passed to

2

amplifier ICI via R7. Subsequent rectification byD9 enables the relay driver circuit composed of Tiand T2 to energize Rei. Preset Pi is adjusted to findthe right compromise between receiver sensitivityand noise immunity. R14 should be dimensioned tosuit the relay coil current.As to the construction of the receiver and transmit-ter, it should be made quite clear that the presenceof the mains voltage necessitates the use of soundand safe con-ABS enclosures to prevent accidentalcontact with the live wires. Do not take any risk inthis respect, neither while experimenting with thecircuits as shown nor while setting up and testing.The transmitter, then, is readily fitted in a salvagedmains adaptor case with a small hole drilled into itfor Si.The receiver ABS enclosure is likely to be of largersize if a mains socket is incorporated for easy con-nection to the appliance to be controlled. The con-tact rating of Rei should be duly observed in caseheavy loads, such as a coffee machine (4 A), are tobe switched.

O

86413-2

209 MAINS FAILURE ALARM

This circuit was originally developed to detect aridsignal interruptions of the mains supply to artificialrespiration systems. The signalling is done in twoways: a buzzer is sounded, and a small lamp is quen-ched.The supply current to the monitored equipment in-duces a variable flux in a small transformer thatserves to keep the relays actuated, so that Lai lights

and Bz is off when the mains voltage is available.When a mains failure occurs, apparatus X nolonger draws current, so that both Rei and Reg arede -actuated, resulting in the lamp being turned off,and the battery -operated buzzer being activated.Transformer Tr2 is a modified 3 VA mains typewhich functions as a current transducer: theoriginal primary winding functions in this appli-

246

* see text

= IN 4146

87413

cation as the secondary, while the original second-ary winding is replaced by about 7 turns of 20SWG(0 1 mm) enamelled copper wire. Every precautionshould be taken to ensure that the new winding iscapable of safely handling the current demand ofX. Thanks to the so created high turns ratio in thetransformer, a relatively small current suffices tokeep the relays actuated and the smoothingcapacitors C, -C2 charged. Push-button Si makes itpossible to test the alarm by simulating the absenceof induced current. Tr, can be a small bell type, orone salvaged from a mains adapter for a pocketcalculator. Switch Sz, finally, is used to turn off thebuzzer when apparatus X is disconnected orswitched off.

210 MAINS VOLTAGE MONITOR

It is often desirable to know at a glance whether themains voltage is at the low side; for instance, whenyou are about to work on a computer program. Thedanger is, of course, that when it is already low,further loads may cause the mains to drop below anacceptable level.The supply for the present circuit is taken directfrom the mains, which exists across R1 and Pi The15 V stabilized voltage produced by Rz, C,, Cz, Di,and D2 provides two reference voltages. Thesevoltages are compared in Al and A2 with a fixedproportion of the mains. If the mains is below

R1 C1=RI1W 400 V

D3

Min1N4001

R3

O

680 n

R25W

R5

R6

210 V, D7 lights, and when it is higher than 250 V,D8 lights.When neither D7 nor D8 lights, Ti switches on andcauses D4 to light, indicating that the mains voltageis within acceptable limits. The mains voltage limitsare set with P1 with the aid of a multimeter and avariac; where perfectionism is not required, thepreset may be set to roughly the centre of its travel.Remember that this circuit is not isolated from themains and it must, therefore, be housed in a man-made fibre case.

red

Al ,A2 = IC2 = LM 358

D7

05

R9

141N4148

D4

85510

.of green

247

211 MAINS WIRING LOCATOR

The accompanying circuit shows a simple means oflocating current -carrying conductors. The detectorcoil is a telephone pick-up with suction pad. Themagnetic field of a current -carrying conductor in-duces a very small voltage in LA that is amplified inopamps A and A2. Capacitors C2. . a have avalue which ensures maximum amplification in Aiand A2 of signals around 50 Hz. Diode Di will lightduring positive halve -waves of the mains current.

9V

R2

9V

R7

A1 ,A2A3 = 3/4 1C1 = LM 324

1C1

85496

9V

212 METAL DETECTOR

In contrast to the other metal detector in this issue,the present one works on the principle that the fre-quency of an LC oscillator changes when the in-ductance is altered. Any metal object brought nearthe inductor will modify the inductance.The degree by which the frequency changesdepends on the nature of the metal and on the fre-quency. If the frequency is very high, a metal objectwill act as a shorted turn, which lowers the induct-

ance, so that the frequency increases. If the fre-quency is low enough for eddy -current losses to beignored, it is possible to distinguish ferrous fromnon-ferrous metals.The inductance required for an oscillator frequencyof not greater than 200 Hz would be pretty difficultto make, and the oscillator in the present circuit,therefore, works at about 300 kHz. The inductancethen needed is quite easy to make and consists of a

248

single turn of coaxial cable as shown in the accom-panying diagram.The circuit consists of oscillator T,, frequency -to -voltage converter IC1, and BiMOS operationalamplifier IC2. With a detector coil diameter of c.440 mm, the values of capacitors Cl and C2 ensurean oscillator frequency of around 300 kHz. Smallerdiameter coils need more turns.The level of the oscillator signal should be at least500 mVitit to be able to drive the 4046B satisfac-torily. At that level, the phase comparator ensuresthat the internal phase -locked loop always locks.The source follower output at pin 10 is fed to aCA3130 where it is amplified substantially.The centre frequency of the phase -locked loop, and,

therefore, the zero of the centre -zero microammeter,is set with Pl; fine adjustment with P2 may benecessary if the sensitivity of the opamp is high.That sensitivity is set with P3 which is connected inthe negative feedback loop to the inverting input.There is also positive feedback via the microam-meter and R10 to the non -inverting input. If,therefore, a meter with a different resistance is used,it may be necessary to alter the values of R9, Rio,and Ril accordingly.Note that in treasure hunts the size of the objectssought should have some relation to the diameter ofthe detector coil: looking for coins with a 440 mm(17.5 in) diameter coil is a fruitless task!

85490

<10 mA

213 METAL PIPE DETECTOR

Water and gas pipes, as well as electrical conduit,embedded in walls are not easy to trace, althoughthis is essential when work is to be carried out tothe wall. This handy little unit will be a godsend atsuch times.The principle of the detector is based on the prop-erty of metals of absorbing magnetic energy whenthey are brought into a magnetic field.Transistor T1 in figure 1 is a simple LC oscillator, ofwhich the sensor, L1, forms a part. The oscillator

frequency is around 15 kHz. When energy is

withdrawn from the magnetic field around L1 by ametal object, the alternating voltage across the LCcircuit will diminish. By rectifying that voltage inIC1, and applying the resultant direct voltage to adifferential amplifier, IC2, which compares it with avoltage preset with 134 an on/off indication is ob-tained. When Li is brought in the vicinity of metal,D4 goes out. The sensitivity of the detector is setwith P1 and Pa

249

The unit is powered by a 9 V battery (PP3).To calibrate the detector, adjust Pi for maximum re-sistance and connect an oscilloscope to the collectorof T,. Adjust the peak value of the oscillator signalwith P2 so that the oscillator just does not stopworking. This is checked by adjusting Pa so that theLED just lights. If then a coin is held near the fer-rite rod, the LED should go out, indicating that the

Li I

oscillator has ceased working.At the start of the search, use the smallest peakvalue of the oscillator signal (Pi at maximum resist-ance), combined with the lowest trigger level (wiperof P3 to earth). After the location of the pipes hasbeen ascertained roughly, the peak value of the os-cillator signal and the trigger level can be increaseduntil the required accuracy is obtained.

Di

4V7400 mW

C11

4pl3V

R3

R2 Cie

7n T1n

D2,D3 = 1N4148

1:1 =500 turns enamelled copper wire,0.2 ... 0.3 ram dia.on ferrite rod 200 mm long and 10 mm dia.

Ci1p 16V

C5

Num100n

< 30 mA

D4-0-

ses. IC1 = CA 3130 85473IC2 = CA 3130, CA 3140

9V

214 MINIATURE RUNNING LIGHTS

The type UAA170 integrated circuit is normallyused to drive up to sixteen LEDs, and the presentcircuit is no exception, as can be seen from figure 1.The 555 is used as an astable multivibrator, butnote that its output is not connected to theUAA170. Instead, the driver is fed from the junc-tion of an RC network with a triangular voltage, theperiod of which is set with Pa It is advisable to use

a tantalum capacitor in the C7 position to keep theleakage current down.The voltage at the input of IC2 must not exceed6 V. To ensure that the triangular voltage remainsbelow that value, the supply voltage of IC, islimited to 9.1 V by D17. If necessary, this zener di-ode may be replaced by an 8.2 V or even 6.8 V type.The voltages on pins 12 and 13 determine the

250

2

1016 V

R1

017

R2

P1

1Mlin.

330Q

12 V

11

10

13

C1

mom 411710 VTant.

9V1400 mW

IC1555

R4 S R3

16

14

12

R5

12 RR

voltage range swept by the LEDs.The reference voltage for D-16 is provided via pin 5of IC, and amounts to about 3/4 of the supplyvoltage to the 555. The reference voltage for D, isdetermined by the potentional at the junction of

15

B

A9E

IC2 F

UAA 170

C

01

D2

ft

3

D3

D4

05Deo

rca.DB-

09r

D10

114

D11

D12

013

D14

015

016

1414

14

++

85451

R4 -R5 ( = pin 12 of IC2), which with values shownamounts to about 3 V.Current consumption is around 30 mA, so that bat-tery supply is only possible with two PP3s in seriesand a 12 V regulator.

215 MUSICAL GREETING CARDS

The designer of this circuit will readily admit thatit is literally not much to make a song or danceabout, since what is shown as the circuit diagram

Table

speaks (sings) for itself.Available in about 30 different song versions, theType UM3166-xx is a fully autonomous melody

TYPE MELODY TYPE MELODY

UM3166- 1 JINGLE BELLS + SANTA CLAUS IS COMING UM3166-16 TOMORROWTO TOWN + WE WISH YOU A MERRY UM3166-17 WE WISH YOU A MERRY XMAS +XMAS SILENT NIGHT

UM3166- 2 JINGLE BELLS UM3166-18 WEDDING MARCH (WAGNER)UM3166- 3 SILENT NIGHT UM3166-19 FOR ELISEUM3166- 4 JINGLE BELLS + RUDOLPH, THE RED -NOSED UM3166-20 WHEN THE SAINTS GO MARCHING IN

REINDEER + JOY TO THE WORLD UM3166-21 CONGRATULATION + HAPPY BIRTHDAYUM3166- 5 HOME SWEET HOME UM3166-22 JINGLE BELLS (NEW VERSION)UM3166- 6 LET ME CALL YOU SWEET HEART UM3166-23 IF YOU LOVE MEUM3166- 7 CONGRATULATIONS UM3166-24 TWINKLE TWINKLE LITTLE STARUM3166- 8 HAPPY BIRTHDAY TO YOU UM3166-25 MARCH OF TOY SOLDIERUM3166- 9 WEDDING MARCH (MENDELSSOHN) UM3166-26 ROCKABYE BABYUM3166-10 I WILL FOLLOW HIM UM3166-27 CHORAL SYMPHONY (BEETHOVENUM3166-11 LOVE ME TENDER, LOVE ME TRUE SYMPHONY NO. 9)UM3166-12 SUCH A WONDERFUL DAY UM3166-28 HAPPY BIRTHDAY TO YOUUM3166-13 EASTER PARADE (NEW VERSION)UM3166-14 GRADUATION MARCH UM3166-29 BLUE BELLS OF SCOTLANDUM3166-15 ALOHA OE UM3166-31 LULLABY (SCHUBERT)

251

4,12V

generator chip which operates at extremely low bat-tery voltages (1.3. . .3 V), while capable of directlydriving a small piezo-buzzer from antiphase outputterminals 2 and 4. If you wish, you may connect anAF amplifier to either of these pins in order thatmore listeners may be captured by the melody selec-ted from the accompanying table. The melody maybe played continuously by connecting terminal 3 to7 rather than 1. (St)

86456-1

216 RANDOM LIGHTS CONTROLLER

Unfortunately, we are all well aware that the annualholiday season is an anxious time for many people,since they worry about leaving the home unattend-ed and therefore liable to be visited by burglarsand/or hooligans. Right now is, therefore, an idealtime to construct this circuit before you leave yourhome and all of your highly -valued property.It goes without saying that simulating one'spresence in the home may be accomplished byhaving some electronic or mechanical timer deviceswitch on a number of lights when it grows dark,

07 IC2,1C3,M4

N1... N4 = IC1= 4093N5... N8 = IC544071 o

N9... N12 = IC64 4081N13... N16 =1C7= 4081

START

01D

0121

013

I

1N4142

-617812

C4 2 CB

10000

' Re1...Re4:Nt 12V,max 89mA

71-74:8054703-06:1N4001

10340175

4016110

I

3 12 12.,

C1i 2 9

merely keeping them on until a fixed time intervalhas lapsed. The potential housebreaker, however,may soon detect the regular pattern that occursevery evening, encouraging him to embark on hisnefarious activities, since he realizes he is dealingwith a harmless timer rather than persons in thehome.This circuit, while also being a timer, offers a bettersimulation of human activity, since it automaticallyarranges for a number of lights to be switched onand off in an apparently random manner, which

IC4 :4060

13

86495-1

252

gives the burglar the impression that there arepeople at home. In actual fact, the lights pattern ispseudo -random, but 16 possible configurations arebound to ensure sufficient diversity to keep yourmind at ease and that of the attentive burglar quitepuzzled for at least a few weeks.And now for the operational principles of this easy -to -build circuit. The evening's specific lights con-figuration is determined by the four -bit logic codesupplied by counter IC2 at the moment it becomesdark. Since this never happens at precisely the sametime every evening, IC2 may be considered as afour -bit (1 of 16) random code generator. Wheneverthe LDR fails to detect the presence of daylight, theoutput of N2 goes high, and Di charges Ci. Mean-while, Ni constantly applies 100 Hz pulses to theinput of counter IC2. When the voltage across Ciand R2 has risen to a level, sufficiently high to berecognized as a logic one by the clock input of quadlatch IC3, the four -bit counter code is latched andtransferred to the Qo... Q3 outputs of IC3. In ad-dition, N3 simultaneously enables IC4 to startcounting and dividing its on -chip generated clocksignal.The latch (IC3) and counter (IC4) outputs are com-bined in AND gates N9.. .N16. The oscillator

parts to IC4 R4 -Pi -Rs -C3 (the latter is a bipolar typewhich may be substituted by two series -connectedelectrolytic capacitors) have been dimensioned suchthat output Qio produces 15 -minute long, 50%duty factor pulses; this interval may be set accurate-ly by means of Pi. Since IC4 is a binary (2n) divider,outputs QI2, QI3 and Q14 provide pulse periodtimes of 60, 120 and 240 minutes respectively.Whether or not these pulses can appear at the out-puts of N9. . .N16 depends on the current logiclevel of each of the associated latch outputsQo . Q3. The AND gate outputs have been pairedin four OR gates Ns. Ns; therefore Ns and N7may supply either 15, 60, or 75 -minute intervals,while N6 and Ns cater for relay -on times of either60, 120, or 180 minutes; longer times (e.g. 360minutes) are not feasible since N4 resets IC3, fivehours (Q12 AND Q14 = 60+240 = 300 min.)after it fell dark at the LDR mounting position.It is seen that Rei and Rea are therefore best usedfor those lights that can be expected to go on andoff for relatively short periods during the evening,while Reg and Rea are energized for longer times atlater hours that same night.Finally, the inset timing diagram illustrates thepulse sequence relevant to the four relay outputs.

217 REMOTE CONTROL FOR LIGHT SWITCHES- PART 1

We all sometimes wish that some of the switchesaround the home were just a little easier to locateand operate, notably so in the dark and with less fre-quently used light switches, such as those for thecellar or garage light. For the physically handicap-ped, some switch locations present a real hindranceto their mobility in the home; for them, it would bevery convenient to be able to operate the switchfrom a distance.The proposed wireless control system differs from,say, an IR-based set-up in that it requires no line -of -sight path between transmitter and relevant re-

ceiver, while the practicable operating range is ofthe order of a few metres.The circuit diagram of the control transmittershows an oscillator composed of Ti, T3 and T2, thelatter transistor merely functioning as a switchingdevice. The oscillator frequency is set at about30 kHz by means of C5, C6 and Ll: the latter con-sists of about 200 turns of 36 SWG (0 0.2 mm) en-amelled copper wire on a paxolin former to suit the

diameter of 10 to 20 cm long ferrite rod, which maybe salvaged from a discarded MW/LW pocket radio.The tap on the coil is made at 20 turns from theearth connection.In order to compensate for the relatively low radi-ation efficiency of the proposed transmitter aerial,the peak pulse voltage across C6 amounts to some150 Vpp when the oscillator is turned on for 8 msby T2, which is driven with an 18 Hz signal fromIC,.The pulsed mode operation of the oscillator ensuresa relatively low mean power consumption of thebattery -operated transmitter when a receiver unit isto be activated.Testing the transmitter is readily done with a scope;observe the pulsed 30 kHz carrier, which shouldlook as indicated by the inset signal waveform draw-ing; the pulse -on time of 8 ms is determined by C4 -

R4, and their values had better not be changed,since they are the optimum compromise betweentransmitter current consumption and output power.

253

R1

R2

C3

IC17555

88884711

101/C>

R3

Cl C2

T2n 10

.....T20n

14-kl8ms

30 kHz

55ms

C4

100n

R4

Li: A = 180 turnsB = 20 turns

R8

T3

C7

C.BC560C

400V

//l/IIIl3....4cm

10....20cm

4117101/

< 5mA

9V

-Imo

86476-1

4

218 REMOTE CONTROL FOR LIGHT SWITCHES- PART 2

Just like the associated hand-held transmitter (seeprevious article), the receiver is simple to construct.As can be seen from the circuit diagram, paralleltuned circuit Li -C, receives the transmitter signal,which is first buffered by means of a dual gateMOSFET - Ti - in order to prevent excessiveloading of the tuned circuit. Further amplificationis performed by T2, before rectifier circuit Di -C6can provide a pulsating voltage to T3 which drivesPLL detector IC, with a sawtooth-like signal. Thelock output-pin 8-of IC, controls Rei via relaydriver circuit T4 -TsAs to a few details concerning the receiver circuit,the PLL chip signals the lock condition by pullingpin 8 low; C14 is charged and functions as a bufferdevice in case the PLL input voltage disappearsbecause of the fact that the transmitter coil is nolonger held steady for optimum reception (directiveeffect of the ferrite rod). At the receiver input, Rshould be mounted direct at the relevant MOSFET

gate so as to prevent possible oscillation tendency of

Like the transmitter coil, Li is wound on a 4 cmlong paxolin former, which can be slid over the fer-rite rod to find the position that gives optimumreception. Use 210 turns of 36 SWG (0 0.2 mm)enamelled copper wire; the coil length should beabout 3 cm. Li and L2 should be separated fromeach other with a metal screen to preclude straycoupling.

The receiver is readily tested and adjusted by plac-ing an operative transmitter at a distance of about4 metres. The optimum position of the coil on theferrite rod can now be found by connecting a scopeto the drain of Ti and sliding Li for maximum re-ceived signal. In the absence of an oscilloscope, thesignal at the PLL input (pin 3) may be connected toa loudspeaker to position Li for maximum voicecoil movement at 18 Hz. After it has been pos-

254

itioned correctly, L1 may be glued into place on therod.Adjusting the PLL is done with PI, which should beturned carefully across its travel to establish thepoints at which the PLL fails to lock on the incom-ing signal (Re, is deactivated and the lock indicationLED, if fitted, goes out). Now set Pl to the position

R4

C4

ol

Iti

R8

T3

47n

SC550C

C9

820n

2

P1

in between the no -lock points. Carefully manoeuvrethe transmitter to a place where reception is worse,i.e. where Re, is observed to go off. Careful adjust-ment of Pl and further trial and error will enablethe user to establish the preset position that cor-responds to optimum receiver sensitivity and re-liability under less than favourable circumstances.

C141 R13

inn* 6473PV

A 50k

1C1567

614

BC560C14

RIS

7

0

ILi

C110 C12

/111/////13....4cm ol

10....20cm

74.1176V 6V

er

616

BC550C

R17

TS

618

002

5V

C) 5V<50 mA

86475-1 (:)

219 RODENTS DETERRENT

There are a number of well-founded argumentsagainst the use of poison to get rid of mice, rats andother rodents in and around the home. From anecological point of view, the undesirable side effectsare mainly the disturbance of the natural foodchain of animals we do not wish any harm what-soever; most poisonous substances devised to exter-minate mice are, unfortunately, quite difficult tobreak down compounds, which may, in the end, be-come manifest as dangerous to our own health.The ecologically accepted method of getting rid ofa population of mice is, therefore, based on the con-trolled introduction of such predators as cats andowls, causing a high degree of stress on part of the

mice, which are then quite quick to leave the rel-evant premises or area.Another method of bringing about a high degree ofstress is to produce a high -pitch, frequency -sweptsignal just above the audible range for human be-ings. The signal is swept rather than of constant fre-quency in order to prevent mice from becoming im-mune to the sound.The proposed rodents deterrent is based upon theType 555 timer chip, which is configured to producea 20 to 40 kHz output signal, swept at a 50 Hz rate.The latter frequency is obtained from the mains bymeans of C4 and R3, which pass the modulatingsignal to input pin 5. The output of the swept oscil-

255

lator is connected direct to a high -efficiency piezo-ceramic horn tweeter, which ensures a sufficientlyhigh sound pressure level to keep rodents out ofreasonably sized areas, such as attics and garages.The completed rodents deterrent circuit, along withthe tweeter, may be mounted in a simple ABSenclosure, but care should be taken to observe thedirectivity of the loudspeaker when fitting the unitin its final position.

Parts list

Resistors:

Ri=1 kR2;R3=15 k

Capacitors:

C1 =1 nC2=1µ;16 V electrolyticC3=10 nC4=220 nC5 = 1000µ;16 V electrolytic

Semiconductors:

Di . . = 1N4001IC1 =555

Miscellaneous:

Tri =6 V;200 mA.TD1 = piezo horn tweeter.Fi =50 mA, fuse, slow.Fuseholder, PCB type, for Fi.PCB Type 86490ABS enclosure for wall mounting.

50mFlA

TR1

6V/200mA

e EPS.86490

D1...D4=1N4001

86490

in

220 SET POINTER

Aneroid barometers invariably have two pointers:one that is operated by the mechanics, and one thatis set manually. The manually set pointer is reallynothing but a mechanical memory that enablesvariations in barometric pressure to be ascertained.The set pointer can, of course, be made electronic,for which a slide potentiometer is ideal. Such apointer is not restricted to a barometer: it can alsobe used with a thermometer, a hygrometer, a batterythat needs to be monitored; in short, with anysensor that delivers a slowly varying voltage.

256

The circuit consists of an amplifier, IC), and a dis-play stage, IC2. The display consists of between 3and 9 LEDs, the centre one of which, D5, is yellowand represents the point of origin. Potentiometer Pican be adjusted to make this LED light. When theinput voltage rises slightly, D6 (the colour of whichdepends on the application) lights; when it drops,D4 (again, the colour depends on the application,but it should be different from D6 . . D9) lights.Greater variations in input signal cause D7 . . D9 orD3 . . . Do respectively to light. It is at all times poss-


ible to adjust 131 in a manner which causes thecentre LED to light.The potentiometer could be provided with agraduated scale to enable the input voltage to beread direct. It is not difficult to produce such a scale.Apply voltages of 0.1 V, 0.2 V, and so on in steps of0.1 V, and for each voltage turn Pi till the centreLED lights. At each of the positions of 131 so found,draw a thin line.The sensitivity of the circuit is of some import,because about 1 V is necessary at pin 5 of IC2 tomake Di and D9 light. As the amplification of ICIis unity (R4/R3), about 1 V is, therefore, also neededat the input of the circuit for these LEDs to operate.Opamp IC, deducts the voltage at the wiper of 131from the input signal, and adds the potential at thejunction of R5 and Re to the result.Since 131 is connected to the reference voltage(1.28 V), only this voltage can be compensated for.Strictly speaking, there is no reason why Pl shouldnot be connected to the positive supply line in serieswith a suitable resistor. In that case, the display isonly stable if the supply line is well regulated.If the input sensitivity is too low, the values of R4and R2 may be increased; note, however, that thesevalues should always be the same.Current consumption is determined primarily bythe current through the LEDs, and that in itself isabout ten times the current through R5 and R6. Thelatter current is equal to the on -chip reference

1

2r

red ye I ow

---] 1r

854511

green

0 0 0 0 0 0 0 0 0

B5461-2

voltage of 1.28 V divided by the total resistance ofR5 +R6. The maximum current through the LEDsis about 40 mA (the current via pin 7 must not ex-ceed 4 mA!) so that the total current does not ex-ceed 50 mA.

221 SIREN

In spite of its modest configuration, the circuitshown here is capable of generating quite a sound.This is made possible by the n -channel MOSFET,

which drives the loudspeaker.Such a MOSFET can be driven direct by CMOSlogic circuits, and the type chosen here has an out-put ( = drain -source) resistance of only three ohms.Moreover, its drain current can be as high as 1.7 A,while the maximum drain -source voltage is 40 V.These parameters are independent of the polarity ofthe applied voltage, since the device has internal di-ode protection.Since the MOSFET is virtually indestructible, it isperfectly all right to load it with just a loudspeaker.The circuit can be controlled simply from a com-puter, and is operated by making the ENABLE in-put logic high (which can also be done with asimple switch instead of a computer). When the in-put at pin 5 of gate N2 is high, the pulses from

Schmitt trigger N, cause N2 to oscillate. The outputof N2 is applied to the MOSFET via buffer N3. Thefrequency of N2 can be adjusted with PiAs to applications, this siren is particularly suitablefor use in alarm installations.

N1 ... N3 ='%IC1 = 4093

1N4001-

8541]

9 12 V

257

222 SMD DIE

"Alea iacta est" (the die is cast, freely) someone saidquite a few years ago, and promptly engaged in sun-dry military actions that are generally reported ashaving been decisive for global history. Whateverthe relative importance of this notorious person'sdecision at that time, he is not likely to haveemployed a SMD die as described here, since heused the verbal form cast rather than a clausal con-struction (in Latin, of course) to indicate thepresence of clock pulses from a Schmitt -trigger gateoscillator, at the relevant input of a Type 4029binary counter which is preset to state 9 by meansof jam (preset) inputs Jo . . . J3 while itsQo . . Q2 outputs may represent 1 of 6 pseudo-random states 9 . . .15 after removing one's fingersfrom the touch -sensitive contacts between oscillatorand counter clock input.Counter output states 9 . . .15 were chosen ratherthan 1. . . 6 with the corresponding preset 1, inorder that the CO (carry out) could be connected toPE (preset enable) via inverter N2. This arrange-ment causes the binary value at the Qe . . Q2 out-puts to vary between 1 and 6, since Q3 is left unus-ed. CO goes low any time the counter reaches out-put state 16, which can not be represented by meansof the four binary outputs to the IC (24=16).

9V

258

Consequently, the counter loads the preset value 1(9), since PE goes high.LEDs Di.. . D7 are arranged in the form as usual

N1...N6 = IC2 =40106

D2.

D5eD7

D306

Die D4 86454


Parts list (all parts SMD)

Resistors:

Ri;R2=100 kR3 . . R6=560 52

Capacitor:C1=12 n

Semiconductors:... D7 = LED Type CQV231 or LSS210D0(Siemens)

IC1=4029IC2=40106

Miscellaneous:battery clips for PCB mountingPCB Type 864549 V battery PP3

on the "six" face of a die, and the random numberis, of course, displayed as an imitation of the spot(s)seen on the cube faces.As to the construction of the SMD die, the tinyparts are fitted onto ready-made, through platedPCB Type 86454, which comes together with theType 86452 (sideway RAM for BBC and Electron,also a SMD project in this issue).

gag.86,84'4374{' 4

It is is noted that the 9 V battery is clipped direct ontothe circuit board to make a compact unit with theLEDs facing up. The "cast" contacts are fourlengths of stripped wire at the LED side of the PCB,mounted at all four sides. Placing your fingers ontoeither two of these wires facing one another causesall seven LEDs to light, while on release a pseudo-random value is displayed.

223 SMOKE AND GAS DETECTOR

This circuit is intended for use as a preventivedevice. We all know about accidents that occurthrough the accumulation of gas or of people over-come by smoke. The preventive character manifestsitself by timely warnings in case of high gas concen-trations in a manner that does not cause the gas toexplode.The circuit is based on sensor type TGS109 whichis sensitive to gases enumerated in the accompany-ing table.Power is provided by an 8 -volt bell transformerwhich is tapped at 5 V. The voltage developedacross the 5 V winding is rectified by D3, smoothedby C,, and regulated by R2, D4, and C2. Theresulting direct voltage of about 5.6 V is used tosupply IC,. The 3 V alternating voltage is used tooperate the sensor, which needs 1 V at about 0.5 A.Resistor Ri provides the necessary voltage drop.The mutual inductance between the two windingsof the sensor increases with rising gas concentra-

Table Hydrocarbons: iso-butaneCH3CHICH51CH3

n -butaneCH3CH2CH2CH3

ethane CH3CH3propane

CH3H8,CH3CH2CH3methane CH4

Inorganic gases:

Organicsolvents:

hydrogen Hammonia NH3carbon monoxideCO

ethanol CH3CH2OHacetone

C3HBO, CH3COCH3n -hexane

CH3(CH214CH3benzene C6H6

259


* see text

TGS 109

700v35V

TGS 109

2x

1N4148

D1

C3 CL

5V6 63V 16ValF78

400 mVV

tions. Note that there is no difference in the twowindings: the sensor may therefore be inserted intothe socket in any way it fits. In practice, a rising gas

R7

220 n

'Cl14

Cr)

10n

N1 ... N4= IC1 =4011

85421

concentration will cause an increased alternatingvoltage in the secondary winding of the sensor. Thisvoltage is rectified in Di and smoothed by C3; itslevel ( = sensitivity) is preset with Pi Diode D2 pro-tects one of the inputs of N, against too high inputlevels. Gates N, -N2 and N3-Na are astablemultivibrators which cause the buzzer to operatewhen there is too high a concentration of gas.Resistor R3 serves to counteract changes in sensi-tivity caused by temperature variations.The detector can be built into a small case, but bearin mind the heat dissipation in Rl.Finally, in case of an alarm, be careful in the inspec-tion of the relevant room or space for which thealarm is sounded.

224 STAIRCASE LIGHT CONTROLLER

This circuit has been designed to function as anautomatic switch -off facility on the lines of the well-known hotel switch circuit, i.e. the combination oftwo switches and a single light. While not exactly areplacement of any of the two changeover switchesat the top and the bottom of the stairs, the proposedcontroller may be fitted into one of the relevantjunction boxes in which a live mains line is

available.The circuit diagram shows that the controller is feddirect off the mains. C3 and R21 create a suitableseries impedance which charges C4 to 6.8 V by

means of rectifier D6 and zener diode D7. Set -resetbistable T3 -T4 keeps track of the position of S2,which determines which of the two triacs is to bedriven so as to turn the light on. Any time S2 is op-erated, timer ICI is started by means of CI -R17, C2-Rts, NI, N2 and N3; the output of the latter goeshigh in this condition, resetting ICI and causing itto pull all of its counter outputs low. Note that thereset condition can also be forced by depressing Si.FET Ts is turned off at reset, and 50 Hz clockpulses are applied to the b (clock input) terminalof ICI. Any one of the five timer outpus Qs...Q13

260

Tri Tri2

Tr i1,Tri2 =TIC206D

R3

a

BC547B

D3V

R5 R6

T1 U U 12

CO ODBC547B

0 0

D2

R4

a

R7 R8 D4V

D1... D4 .1N4148

11

N4

5s f,J 52r-208105 0 0109405 0 57.0-, 011

80s 0 01,01216050 )::P- 013

.a 3

I C14060

RST

S1

2

Cl

R17

0

Ri R14

0

D8 2x D91N4004

0 P<400W

R11 R12

100n VII4 TI

R15 R16

a aBC547B

R

BC547B

N12 5

93

69START

4;1

N3

R19

11R20

T5

ClOi

O

C2

100n

C3

u220n

R21

I = 10mA(0.5 mA stand-by) 06

6

D7 6V8

In TN°

0N2

16

C5IC2 ICI

100n

01W

05

D5,136 .1N4004

may be wired to the inputs of gate N4 to select thedesired on -time for the light; longer intervals maybe realized by adding a further counter.When the selected light -on interval has lapsed, Tsconducts and disables ICI from receiving clockpulses; the counter state is thus frozen until a resetpulse is applied at terminal 12. Finally, Ti and T2provide DC control of the relevant triac, whileAND gate simulators Di -D3 -R3 and D2 -D4 -R4 en-sure the correct selection of Trii or Tri2 to powerthe bulb.The circuit is readily constructed on a piece of

N1...N4 =4093 (IC2)

86449

630V

veroboard and fitted into an ABS mains wiringjunction box, as a replacement of one of theswitches in the hotel circuit.As many points in the circuit are at mains potential,due precautions should be taken in the constructionand wiring of the controller. Note that Si shouldbe rated at 240 V AC, in view of the necessary iso-lation with respect to the mains voltage.Trii and Tri2 require no heatsinks if the bulb israted at 100 W or less, while the maximum powerrating for the triacs is about 400 W.

225 SUPER DIMMER

Most dimmers use a silicon -controlled rectifier(triac or thyristor) which is triggered at a fixedphase angle and then conducts until the next zerocrossing of the mains voltage. This method is

simple, but at thecontrolling smallflickering). The cafact that owing to

same time it gives problems inor inductive loads (hysteresis;

use of these problems lies in thethe small load the current sup -

261

1

A1A2G

TIC 206 D

1N4005

1/2W

CI I470n630V

PI P2

C4MEI

10n

TIC206 D

A2C5

22n 63 0V

R1 4A

P max = 600W

R2w q = 0....98%

plied to the bases is insufficient to allow conduction 2to continue. This means that a region of the controlcharacteristic is not used. The effect is even worsewhen the load is inductive.The proposed circuit offers a solution by providingthe SCR continuously with gate current, so thateven loads of 1 watt can be controlled. To keep thecircuit as small and simple as possible, it makes useof the well-known timer -buffer Type 555.The output of the 555, which is normally activehigh, is made active low with the aid of a negativesupply voltage. The supply is provided by networkCI -R3, rectifier Di -D2, and stabilizer D3 -C2. Tran-sistors TI to T3 provide a start pulse at the triggerinput of the 555 during the zero crossings of themains. For a period determined by the setting of Piand P2, the output of the timer is high, and thereis, therefore, virtually no potential difference be-tween pins 3 and 8, i.e. the SCR is turned off. Whenthe set period has lapsed, pin 3 goes low and theSCR is triggered. For the remainder of the halfperiod, a gate current flows which keeps the SCR inconduction. The minimum position at which, forinstance, a light bulb should just not light, is setwith Pi.Filter R7 -Cs -Li provides the requisite decoupling ofthe SCR.Finally, note that the maximum power that can becontrolled is of the order of 600 watts.

3

P 20%

0

In

O

86439 - 1

86439 -2

262

226 TELEPHONE -BELL SIMULATOR

This circuit is intended for use in a small privatetelephone installation. The ringing tone sequence is400 ms on, 200 ms off, 400 ms on, 2 s off.In the accompanying diagram, Ni and N2 form anoscillator that operates at a frequency of 5 Hz,which gives a period of 200 ms. The oscillatorsignal is fed to two decade scalers, which are con-nected in such a manner (by N3 and N4) that the in-put signal is divided by 15.The second input of N4 may be used to switch thedivider on and off by logic levels. If this facility isnot used, the two inputs of N4 should be intercon-nected.Resistors R3 to R6 incl. form an OR gate that con-trols a relay via T1 and T2 which are connected ina darlington circuit.Outputs 5 to 9 of IC2 go high sequentially, so thatthe relay is energized for 400 ms (when 5 and 6 arehigh), then off for 200 ms (output 7 is not connec-ted), and then energized again for 400 ms (when 8and 9 are high). After that, the relay is off for 10periods = 2 s, and then the cycle repeats itself.

NI ...N4= IC1=4011

86502

12V

227 TEMPERATURE REGULATOR WITH ZEROCROSSING SWITCH

This temperature regulator can be built withoutspecial ICs and may be used with powers up to3.5 kVA.The circuit is based on a two -point regulator witha thermistor as the temperature sensor. As the loadcurrent is switched only during zero crossing of themains, no additional interference suppression is

necessary.The series combination RIC, serves to lower themains voltage to a level suitable as supply voltagefor trigger T,. As R is small compared with thereactance of C,, the current leads the voltage bynearly 90°.If the ambient temperature is higher than a givenvalue, determined by potentiometer Pi, the resist-ance of Rth is low enough to cause Ti to conduct.Silicon controlled rectifier Thl is supplied with gatecurrent and switches on during the negative half cy-cle of the mains, because the currennt throughRiCi leads the voltage. When Thi is on, thyristors

Thl = BRX49 1 A)Th2, Th3 = BT 120)< 25 A)T1 = BC 160 85437

Th2 and Th3 remain in the blocked state, so that nocurrent flows through heating element REWhen the temperature drops below the value deter -

263

mined by P; transistor Ti and thyristor Thl remainoff, so that Th2 conducts. As the voltage acrosszener diode Di leads the mains voltage, Th2switches on when the remains crosses zero. At the

onset of the negative half cycle, Th3 switches on.During the positive half cycle, C2 is charged via R7and D2, and so provides the gate current to switchon Th3 at the onset of the negative half cycle.

228 TEMPERATURE SENSOR

The LM35 is a temperature sensor which providesan output voltage that is directly proportional to thetemperature being measured in degrees Celsius.This means that if the temperature is 0 °C, the out-put voltage is 0 V. The output voltage increases by10 mV for every degree Celsius, i.e., at 19.8 °C, theoutput voltage is 0.198 V.This is an important advantage over other tempera-ture sensors that are calibrated in kelvin. Usingsuch sensors to measure in degrees Celsius requiresa very stable reference voltage that must bededucted from the reading.Another advantage of the LM35 is its very low cur-rent consumption of less than 60 µA. This means along battery life and small internal power dissi-pation, so that errors caused by internal heat are

10 mV/°C4 ... 20 V ici * 0°C = 0 mV

minimal: 0.1 °C with a battery voltage of 4 V.The sensor can be connected direct to an analogueor digital multimeter, or, more interestingly, to acomputer which can then process and store the in-formation. A suitable interface for this purpose isdescribed in direct reading digitizer elsewhere in thisissue.

The accuracy of the LM35/LM35C is typically0.4 °C at 25 °C.To keep the self -heat minimal, the load should benot smaller than 5 k52.If a long screened cable is used between the sensorand indicator, an RC network (10 Q in series withI µF) should be connected between the output ofthe sensor and earth to prevent any oscillations.

85479

229 THERMOMETER

At the heart of this simple circuit is the well-knownType KTY10 temperature sensor from Siemens.This silicon sensor is essentially a temperature -de-pendent resistor, which is connected as one arm ina bridge circuit here. Preset PI functions to balance

the bridge at 0°C. At that temperature, moving coilmeter Mi should not deflect, i.e., the needle is in thecentre position. Temperature variations cause thebridge to be unbalanced, and hence produce a pro-portional indication on the meter. Calibration at,

264

say, 20°C is carried out with the aid of P2.The bridge is fed from a stabilized 5.1 V supply,based on a temperature -compensated zenerdiode. Itis also possible to feed the thermometer from a 9 Vbattery, provided D1 -D3, 111 and CI are replacedwith a Type 78L05 voltage regulator, because this ismore economic as regards current consumption.

1N4001

86503.1

230 THERMOSTAT -CONTROLLED SOIL HEATING

1

Elm Tri 1 = TIC 206; TIC 216; TIC 226; TIC 236

Many people with a keen interest in growing plantsinsist on the fact that many of the more exoticspecies, such as certain species of orchid and fungi,will only thrive in warm soil and relatively highhumidity.Whether or not this is a correct assumption, this cir-cuit offers the possibility to keep the soil tempera -

2a)

b)

TIC ..

Al A2G

71620pV

ture in a miniature hot -house at a constant, adjust-able level.The heating element is made of several loops ofplastic covered steel wire, such as used in gardening.The wire used in the prototype had a diameter of1 mm and a resistivity of about 0.2 Q per metre.The circuit diagram of the soil heater shows that

I

eft

265

the heating element is temperature controlled bymeans of a triac, driven by a Type TDA1024 elec-tronic thermostat which gets the necessary infor-mation as to the soil temperature from R6, an NTCtype sensor.The circuit is fed from the transformer secondaryby means of rectifier Di and series resistor R2.Regulation at 6.5 V is internal to the IC, and C3smoothes this voltage. R3 and R4 provide the ICwith a mains synchronizing signal, while Ci causesa controlled phase shift in order that the relativelylow operating voltage can still ensure the correctzero -crossing synchronization.The temperature sensor circuit is composed of R5,

Rs, and Pi The sensor proper, Rs, must be placedinto the soil at a suitable position, electrically wellisolated, of course. The optimum soil temperature,which should be established by trial and error, is ad-justable with preset Pi; Fig. 2 shows the correlationbetween soil temperature, heating element voltage,and preset temperature.If necessary, a more powerful heating element maybe dug into the soil, but the ratings of the fuse, Tr,and Tril should then be changed accordingly. Thetransformer secondary voltage, however, should re-main at 9 V. With the components as indicated inthe circuit diagram, the heating energy is about 40joules.

231 TWIN BELL -PUSH

It is often desirable for a single doorbell to be oper-ated by two bell -pushes. for instance, one at thefront door and the other at the back -door.The additional bell -push, S2, in series with the breakcontact of relay Re), is connected in parallel withthe original bell -push, Si. When S2 is pressed, thebell voltage is rectified by Di and smoothed by G.After a time T = R1R2C2, the direct voltage acrossC2 has risen to a level where Ti switches on. RelayRe, is then energized and its contact breaks the cir-cuit of Sz, so that the bell stops ringing. After ashort time, Ci and C2 are discharged, the relayreturns to its quiescent state, and the bell ringsagain.In this way, Si will cause the bell to ring continu-ously, while S2 makes it ring in short bursts, so thatit is immediately clear which bell -push is operated.

S2

Trl = bell transformer

232 TWIN DIMMER

Dimmer circuits are always popular and this one of-fers two independent controls in one.Control of each section of the circuit is provided bya type S576 which is an improved version of theS566. This type of IC controls the phase gating byshort or long command pulses emanating from atouch pad. Pulses shorter than 60 ms are treated asnoise.Short pulses between 60 ms and 400 ms cause thelamp to be switched on or off, depending onwhether it was off or on respectively.If the touch pad is touched for more than 400 ms,

the appropriate lamp is dimmed at a certain speed.If the finger is held on the touch pad, the lamp willgo out completely and will then slowly light upagain: when it reaches full brightness (and thefinger is still on the pad), it will begin to dim again,and so on.The 5576 is available in three versions: A, B, and C.With the A and C versions, the lamp is alwaysswitched on or off half -way between maximum andminimum brightness, and it first attains maximumbrightness before it can be dimmed. The B versionis interesting in that it remembers the last

266

1Sensor S"

, 180°-50°-

5 576 A

01

90°-

30°-

180°a 150°

S

0 90°576 8

0

180°

a 150°-

5 576 C

© 90,

30°-

1 2 3 4 5 6 7 8 9 10 11 12

brightness level, so that the lamp is always switchedon or off at the last brightness setting. Thesevarious possibilities are summarized in figure 1.The circuit of the twin dimmer is shown in figure 2.Power for the ICs is provided via R2, Ca, Di, and

2 D2

HF I

0 NMI4A

Lal

C5

100 n400 V

La1

C8 A2mlm

Tri2100 71- Al400 V

Tri1,Tri2 = TAG 226D, TIC 206DD2 . . . D4 = 1N4001

a = current phase angleUL = voltage across lampS = sensor touched

85480-1

D3. The supply is smoothed by C7. Capacitors C3and C6 determine the speed with which the lampsdim or get brighter.The twin dimmer is best built onto the printed cir-cuit board shown in figure 3. This board is intended

*see text

R8

Lal,La2 = 40 ... 400 W

L1,L2 = 30 . .. 50 µH

85480-2

TAG 226DTIC 206D

A2

267

to be fitted into a standard round junction box.Because of this, it is, of course, important that thecomponents used are of the correct size as shownon the board.The board is connected to the lighting system viathree terminals: L to the live wire, and Si and S2 tothe switching wires of the lamps. The junction ofthe lamps is (already) connected to neutral.Note that the dimmer cannot be used with neontubes.

Parts list

Resistors:

Ri,R6 = 1M5R2 = 1 k/1 WR3 . . R5,R7 . . R9 = 4M7

Capacitors:Ci = 47 µ/16 VC2,C7 = 470 pCa,C6 = 47 n ceramicC4 = 220 n/400 VC5,Cs = 100 n/400 V

Semiconductors:Di = zener diode 15 V/400 mWD2 . . . D4 = .1N4001IC1,1C2 = S576 (see text for which version)Trii,Tri2 = TAG226D or TIC206D

Miscellaneous:Li,L2 = 30...501.4F1/2 A

= fuse, 4 A, delayed action and associ-ated PCB holder

1 three-way ceramic terminal block (5 A)PCB 85480

L2

0

0

4

I

D

233 TWO -TONE -CHIME

This electronic chime is easily built from commonlyavailable, inexpensive parts.Depression of the door bell button, S2, causes in-verter Ti to pass a logic low level to NAND gateNi, which responds with a logic high level atits output, enabling the oscillator composed of N2and N3 to toggle at about 1 Hz. Since buffer ca-pacitor Ci remains charged for some time after S2has been released, the oscillator will continue toprovide 1 Hz pulses to C4 and C5, as well as to asecond oscillator section, composed of N4 and as-sociated parts via R6.A logic high level at pin 10 of inverter N3 enables

268

T2 to connect preset P2 in parallel with frequencydetermining parts R7 -Pi, which arrange the fre-quency of N4 to toggle at a few hertz. The twosuperimposed frequencies may be adjusted to in-dividual taste with Pi and P2.In addition to controlling the tone frequencies ofthe chime, the 1 Hz pulses also determine theenvelope shape of the resultant chime sound bymeans of T4 -Ts and associated parts. Preset P3 isused to define the desired decay characteristic, whileemitter follower T6 functions as a very simplevoltage -controlled amplifier, driving one -chip AFoutput amplifier Type LM386.


N1....N4 =IC1 = 4093IC2 = LM386

234 VENTILATOR CONTROL

Many toilets have a ventilator, which is energizedalong with the toilet light. However, since not everyvisit of the toilet requires the ventilator to startturning, this circuit offers an improved controlmethod, which is still based upon the use of thelight switch.The circuit configuration marked I in Fig. 1 may beused in case the toilet ventilator is powered from thesame mains lines as the light. Bridge rectifier Biand opto-coupler Type TIL113 serve to detectwhether or not the toilet light is on. The ventilatoris arranged to start turning after the light switch hasbeen operated twice. If this is the case, the outputof NI will go high twice; the first time, C4 ischarged, the second time will cause pin 6 of N2 tobe logic high, while the output of this NANDSchmitt trigger gate will supply a logic low pulse toN3 when the voltage at point 3 reaches the logicone level (see timing diagram Fig. 2). N3, then,charges Cs which, along with P2 and R9, deter-mines the ventilator "on" interval, while Pi, C4and R8 establish the maximum interval betweenthe reception of first and second trigger pulse.The circuit option with T2 may be used if it is lessdesirable to run an additional wire to the light forthe purpose of obtaining the trigger pulses; theLDR should be located as close as possible to thebulb in order to preclude erroneous triggering dueto the presence of daylight. The use of the LDRdoes not change the basic operation of the circuit,of course, and the indirect method of triggering is in

2

2

4

r

fact to be preferred in view of the risk associatedwith direct mains connection in the case of the firstmentioned circuit option.Another interesting use of the circuit option whichincorporates SI, T3 and T4 is a semi -intelligent doorbell arrangement; bell I will sound any time Si isdepressed, while bell 2 will only do so if the buttonis operated twice within the given interval; it is notdifficult to come up with a number of useful appli-cations for this circuit when used in and around thehome. However, note that the timer parts Ca, Piand R8 will have to change places with Cs, P2 andR9 respectively, if the second bell is to be kept fromsounding for about 50 seconds after Si has beenoperated twice.The power supply for the circuit may be of conven-tional design, incorporating the ubiquitous 78xxtype of regulator. Current consumption of the cir-cuit is mainly dependent on the type of relay, but 50to 180 mA would appear to be a typical value.

269

Dl...D5 = 1N4148D6...D9 = 1N4004N1...N4 = IC1= 4093

-S

Re 12V - 80mA max

D -S

86441-1

235 WATCHDOG

This timer automatically switches off equipmentleft operating unattended for more than aboutthirty minutes.The circuit operation is readily understood by fol-lowing its power -on and time-out functions. Almostimmediately after S4 has been depressed, relay con-tact rely closes to power Tri and the equipmentconnected to the mains outlet. This happensbecause the initial presence of the +12 V supplyvoltage in the circuit causes counter -oscillator ICIand set/reset (S/R) bistable NI -N2 to be reset bymeans of a short, logic high pulse at the junction ofRi and CI. The outputs of Ni and N3 go high andlow respectively and Ta can energize Rei. So far forthe power -on automatic hold function of contactrela.

After being reset, ICI starts counting down its on -chip generated clock pulses which have a frequencyof about 2 Hz. LED Di flashes at this rate to indi-cate the countdown condition. Note that S2 has

been provided to reset, i.e. disable the timer perma-nently, in which case Di lights steadily. The LED,therefore, has a threefold indicator function in thepresent circuit: timer on (flashing), timer and equip-ment off (off) and timer off while the equipment ison (steady light).As long as counter output Q12 remains at logic lowlevel, the voltage at the collector of Qs inverter T2can not cause the relay coil current to be inter-rupted by T4. If, however, some 34 minutes(T(Q12) = x 212= 2048 s) have lapsed since ICIand NI -N2 were reset, Q12 goes high, causing thetwo -gate bistable to toggle; the output of NI goeslow, but Rei remains energized by Ta, since theother input of NOR gate N3 is still high, i.e.

counter output Qs has not been set as yet. Theselfoscillating buzzer starts sounding at a 2 Hz rate,however, since T3 is driven by NOR gate Na whichreceives two logic low levels at its inputs. The useris thus notified that the has another

270

+12V #

52On

IC1

'CO .t 092

C2

Di

101 1

R2

1N4148

1N4148

BC547

EDI, ESE

Bel 120

11 EmAmax

1N4148

IC1 = 4060N1...N4 = IC2 = 4001

Buzzer

BC547

15 seconds or so left to depress Si for another 34 -minute interval. If no such action is taken to resetthe timer before Q5 goes high, N3 disables the relaydriver transistor, and contact rela consequently cutsthe mains voltage to Tri and the connected equip-ment.The foregoing outline of the circuit operationmakes clear that depressing Si or switching on Sz isthe only way to keep the buzzer from sounding and

BC557

+12V

86492.1

the mains relay from switching off both equipmentand timer circuit. If desired, push -to -break switchS3 may be operated to break the mains supplywithin the half hour interval, and without the an-noying sound of the buzzer.Finally, the indicated timing intervals may bechanged to suit individual requirements by usingother counter outputs and/or another clock fre-quency for ICI (adapt the values of Rz-C2).

236 WATER -DIVINER

This little unit may be used to give an audible alarmwhen, for instance, a washing machine hose hasburst, or when it starts to rain so you can get thewashing in, or it can call you to the bathroom toturn the bathwater off. No doubt you will be ableto think of some more uses.The circuit may be powered from a 9 V batterywhich, since the current consumption is very low,will last for at least a year. After a year it should bereplaced because it will then become unreliable ow-ing to its self -discharge.Basically, the unit consists of a sensor, an R -S

bistable, an oscillator, and a driver stage for thealarm buzzer.The sensor consists of a waste piece of wiringboard, about 40 x 20 mm. Connect all odd and alleven tracks together with wire links, that is, I to 3to 5, and 2 to 4 to 6. Tin the tracks to protect themagainst corrosion. When the board is dry, the resist-ance between the two sets of tracks is high, butwhen it is wet, the resistance drops sharply.The sensor is in series with resistor R2 and the twotogether, therefore, form a humidity -dependentvoltage divider, which resets the R -S bistable when

271

input 1 of N2 goes low. Oscillator N3 is thenswitched on, and driver N4 energizes the buzzer.The bistable is set automatically on power up viathe series combination R, andThe circuit can also be used as a lie -detector. Thesensor is then replaced by two lengths of wire ofwhich the ends have been stripped. The bare wiresare then placed in the hands of the person being in-terrogated. If the lies (which causes his hands to be-come damp) the buzzer will sound.The sensitivity of the circuit is determined by thevalue of R2: some experimenting may be necessaryhere.The oscillator (and, therefore, the buzzer) is disabledby closing switch Si.

NI N4 = ICI = 4093 85460

272

237 12 -VOLT NICD BATTERY CHARGER

If you attempt to charge a 12 V NiCd battery froma 12 V lead -acid car battery, you will soon find thatthat is not really possible: the charging voltageshould be somewhat higher than the nominal bat-tery voltage. A 12 V battery should be charged froma source of about 14 V.The present circuit is, therefore, a voltage doublerbased on the well-known 555 IC. The IC oscillates,which means that output 3 is connected alternatelywith earth and the +12 V supply voltage.When pin 3 is logic low, C3 is charged via D2 andD3 to almost 12 V. When pin 3 is logic high, thevoltage at the junction of C3 and D3 becomesalmost 24 V, because the negative terminal of C3 isat + 12 V and the capacitor itself is charged toabout 12 V. Diode D3 is then reverse biased, but D4conducts, so that C4 is charged to just over 20 V,which is ample for our purposes.The 78L05 in the IC2 position functions as a cur-rent source, which tends to keep its output voltage,Uri, appearing across R3, at 5 V. The output cur-rent, in, is therefore easily calculated fromIrl = Un/R3= 5/680 = 7.4 mA.

The 78L05 itself also draws current: the central ter-minal (normally earthed) delivers about 3 mA. Thetotal load current is, therefore, of the order of10 mA, which is a good value for continuouslycharging NiCd batteries. The LED has been incor-porated to indicate that charging current flows.

The characteristic of the charging current versusbattery voltage in figure 2 shows that the circuit isnot perfect: a 12 V battery will be charged with acurrent of only about 5 mA. There are severalcauses for this:

112V p2

0 0+11N4148

91N4148r

cL5

T.,1.O#

N4148

85485-1

III the output voltage of the circuit tends to dropwith increasing current;

the voltage drop across the 78L05 is about 5 Vto which must be added the 2.5 V the IC needs

to operate correctly;II there is a voltage drop of about 1.5 V across the

LED.None the less, a 12 V NiCd battery with a rated ca-pacity of 500 mAh can be charged continuouslywith a current of 5 mA, which is 1 per cent of itscapacity.

2

238 ACTIVE RECTIFIER WITHOUT DIODES

The active rectifier proposed here is based on theproperty of an operational amplifier that its outputcannot become negative if its power supply is asym-metrical. We have used an RCA type CA 3130opamp which is eminently suitable, because it cancope with input voltages down to 0 V, and has aCMOS output stage that can also work down to0V.With a supply voltage of 15 V, the maximum input

level is about 1.2 VI-rm. The frequency range, fornot more than 1 dB change in output, extends fromDC to just over 25 kHz.Negative half cycles at the input of the opamp areinverted and amplified by a factor R2/R1. Positivehalf cycles are also inverted, but, as stated, the out-put of the opamp cannot become negative, and ittherefore remains at 0 V. The positive half cycles arealso applied to the output of the opamp via a

273

resistive divider, R1 -R2 -R3 -P2 The result of all thisis that only positive half cycles are present at theoutput, just as if full -wave rectification had takenplace. If the asymmetry of the supply is set correctlywith Pg the peak values of the inverted negativehalf cycles and the positive half cycles are equal.Preset 131 should be adjusted to give zero outputwhen the input of the opamp is connected to earth.The rectifier has a low -impedance input (source im-pedance should be not greater than 100 52) and ahigh -impedance output (load impedance should benot less than 1 MQ). If these requirements as tosource and load impedance cannot be met, thevalues of Ri and/or R3 should be modified: Ri +source impedance should be about 2k2, while theparallel combination of R3 and the load must bearound 10 k52.

239 BATTERY CHARGE/DISCHARGE INDICATOR

Many of today's cars and motor cycles are equippedwith a meter for monitoring the battery voltage.However, this meter does not provide informationon the battery condition, or whether it is beingcharged at all. When the voltmeter reading is toolow, the battery is generally in such a poor state asto necessitate switching off heavy loads to savepower for use of the starter engine later. Especiallyon motorcycles, the battery capacity is relativelylow, which justifies the need for a reliable monitor-ing system. A standard 30 A ammeter offers too lowresolution, and is rather awkward to fit perma-nently.In this charge/discharge indicator, the measuredcurrent is converted into a potential difference byR., which is either two 1R0 5 W resistors, a fuse,

or a few turns of copper wire. The direction of thecurrent through R. is detected by comparator IC t,which then indicates whether the battery is beingcharged or discharged by lighting the relevant LED.The 100R preset enables shifting the indicationthreshold somewhat. Input terminal + on the indi-cator unit is best connected to a point behind (thatis, electrically behind) the contact switch, althoughit is also possible to fit the circuit with a separateon/off switch. Finally, the circuit is only suitable foruse in or on vehicles having a 12 V battery.

B = remaining loadsC = starter motorDI = chargeD2 = discharge

87474

274

240 BATTERY CHARGING INDICATOR

Sealed 6 V or 12 V lead -acid batteries, under nor-mal charging conditions, are charged at a constantvoltage of 2.3 V per cell. The charging currentreduces during the charging: when it reaches avalue of 10 mA, the battery is deemed fullycharged. To check this, you do not need an expens-ive ammeter. The present circuit uses an LED (light -emitting diode) to indicate when the battery is fullycharged.The green indicator LED is connected in the collec-tor circuit of a p -n -p transistor. As soon as the tran-sistor conducts, the LED lights. This happens whenthe voltage drop across resistor Ri reaches the for-ward bias threshold of the base emitter junction(about 0.6 V). When this resistor has a value of56 Q, a charging current of around 10 mA willcause this drop. To ensure that the charging currentcan exceed 10 mA, Ri is shunted by diode Di whichlimits the voltage drop across the resistor to about0.7 V. The maximum charging current depends onthe diode used and lies between 1 and 3 A.The LED does not light when the charging currentis less than about 10 mA, i.e., when the battery isfully charged, when the battery is connected withwrong polarity, or when the output is short-circuited. The red LED will light when the batteryis connected with reverse polarity.

The indicator should be connected between thecharger and the battery. It may either be built intothe charger housing, or be constructed in a smallcase that can conveniently become part of thecharging cable.

T1

D3

.4-

green

R3

R1

D1 =1N4001 (1 A)1N5401 (3 A)

85433

6V(12 V)

241 BATTERY FITNESS CENTRE

This circuit is designed primarily for maintaininglead -acid batteries that are often not used for longperiods in good working order. It charges the bat-tery, after which the battery discharges slowlythrough its internal resistance and the present cir-cuit. When the state of charge reaches a predeter-mined level, the charger is switched on again, thebattery charges, and so the cycle repeats itself.The circuit is based on Schmitt trigger Ti/T2. Zenerdiode D7 determines the state of charge at whichthe charger is switched off. Resistor R2 provides therequired hysteresis. With the mains disconnectedand no battery connected to the battery terminals,check with voltages (from a regulated power supply)of 13.6 V and 12.5 V applied across the battery ter-minals that the relay switches off and on respect-ively. The "on" threshold may be corrected by, for

instance, connecting a 1N4148 (cathode to + line!)in series with D7. The "off" threshold is correctedby altering the value of R2, for example, by replac-ing this component with a 100 Q preset.It is, of course, possible to replace the mains trans-former and bridge rectifier by a battery charger (see,for instance, Elektor Electronics, June 1984, p. 6-45), in which case the rest of the circuit can be fittedinside the charger.It is not possible to connect a full discharged batteryto the circuit, because the relay would not be ener-gized. Such a battery should first be charged toabove 10 V, but it is also possible to fit a switch inparallel with the relay contact and switch on themains with that.It is possible, of course, to maintain two 12 V bat-teries in condition by doubling the secondary

275

voltage of the mains transformer, the zener voltageof D7, the hysteresis, the rated coil voltage, and con-necting the batteries in series across the terminals.Fuse Fl is necessary to provide protection againstshort circuits. The transformer primary circuit may

10 A T

also be protected by a fuse (like Fl a delayed actiontype) rated at 1 A.The circuit does not need a smoothing capacitorbecause that function is carried out by the battery.

85441

1

12V

242 BATTERY GUARD

This protective circuit is readily incorporated inbattery -powered equipment which is typically in-tended to operate for less than about a minute; poss-ible applications that come to mind include IRremote control units, calculators, etc. Forgetting toswitch off such devices irrevocably causes the built-in batteries to be exhausted after a while, however"low" the standby current.The proposed battery guard automatically switchesoff the supply current to the circuit, either afterabout one minute has lapsed after power -on, orwhen the battery voltage has fallen below the ac-ceptable level for normal operation.Series regulator FET T1 can pass a maximum cur-rent of 150 mA in the circuit as shown, and it is ad-visable to use a more powerful type than the BS250in case more than about 100 mA is expected to beconsumed by the equipment connected to the out-put terminals. The Type BS250 FET drops about0.5 V at a drain current of 100 mA, and 0.8 V at150 mA, whence the foregoing consideration.As T1 is a p -channel FET, it conducts and powersthe equipment when the output of Schmitt trigger

NAND gate N3 is low, i.e. when both gate inputsare high. This is so at power -up, since C2 is stilldischarged and the inputs of N4 are kept at logiclow level via R6. Consequently Ti is enabled andcauses C2 to be charged via R3. After about one mi-nute (R -C time), the voltage across R3 is low enoughfor N3 to recognize a logic low level at pin 1, therebyturning off T1. N2 provides a hold function of thisstate, since otherwise N3 might oscillate owing tothe slowly varying voltage across R3.

At power -on, the output of N2 is pulsed high bymeans of R -C network R1-R2-Cl, whereby anyresidual charge in C2 is cleared; the circuit may,therefore, be switched on with S2 -S1 immediatelyafter automatic power down.Battery voltage monitoring is accomplished by D3,

R5, Re and N4. The latter's trigger threshold levelis, as with all Schmitt trigger gates, in direct propor-tion with the supply voltage level to the IC. As longas the supply (i.e. battery) voltage is sufficientlyhigh, N4 will recognize a logic low level at junctionR5 -Re -N4. However, if the battery voltage falls, D3keeps the input voltage to N4 at a fixed level, caus-

276

S2= onS1 = main

S2

Si'0

"Ir9V

-00

° C3--100nIC1

O

R1

C1 R2

mos1n

229G BS250 16V

8 9 12 13

D

R50

1 D2

2x COMn1N4148

R3 R6

3

+ (TA 9V0V-ri<150mA

6V8

N1...N4 = IC1= 4093

ing the gate to supply a logic low level to N3, whichconsequently turns off the series regulator FET.It should be noted that the exact values of Ra, C2,R5 and R6 may have to be adapted to suit operationwith certain makes of the Type 4093. Also note thatthe interval time of one minute may be changed toindividual requirements by suitable redimensioningof timing elements R3 -C2.Adjustment of the battery guard is carried out bytemporarily exchanging R5 and R6 with a 100 kpreset to determine the correct resistor values for agiven switch -off level. Current consumption of the

86509

BS250

DIUOSG

proposed circuit is mainly determined by the zenerdiode, which has been biased to pass only 1 mA.After automatic power -down, the guard circuitdraws a (negligible) current of less than 1 µA.

EPS. 86509

243 CURRENT INDICATOR

Who has not sometimes wished that the powersupply he was using had a voltmeter AND an am-meter? Unfortunately, the high cost of such unitsprohibits their use in many situations. The proposedcircuit, which does not include the input section ofthe power supply, can be built from standard com-ponents, except for the low -value 5 -watt resistors.We shall not dwell on the well-known Type L200voltage regulator, but shall confine ourselves to thecurrent indicator section.Fig. 1 shows that the circuit contains five LEDs:one (131) to show whether the supply is switched on,and the other four to indicate the current consump-tion in steps of 0.5 A; 0.8 A; 1.3 A; and 1.8 A. As

may be deduced from these figures, the unit iscapable of providing up to 2 A at an output voltageanywhere between 3 V and 30 V. The colour of theLEDs is immaterial, although it would be useful ifthe final one would be red to show that maximumcurrent is being drawn.The non-standard resistors, R4 to R7 incl.,"measure" the actual current consumption. Be-tween point A and the positive output terminalthere exists a potential difference. When this p.d.reaches a value of 0.6 V, T2, and consequently Ts,switch on and this causes D2 to light. In the sameway, when the p.d. between points B, C, and D re-spectively, and the positive output terminal reaches

277


P1

35 -40V

Nr.s.ID1

eo

ICIL 200

R2

Cl

100n

SI . R20

448C547

T6...Y9 = BC 5478= BC 5578

D6...D13 = 1N4148

C2 R3

Imme100563V

10k64 R5 Re R7

OAY'h

5W 5W SW

72

WAY

0.5A 0 0.8A 0 .3A "A

. ¢R13.0.. . 0 .76 T7 T8 79

et ) OP 0:R16 D6 617 D8 P1811 012

07' UD9 DIIU

laD3

D13

Re

03 -30V

C3

70n

about 0.6 V, transistor pairs T3 -T7; T4 -T8; and Ts -Ts switch on, and the associated LED will light.Resistor R2 and capacitor C2 provide a soft start fa-cility at switch -on. Transistor Ti provides anemergency switch -off facility, which in practice hasproved very useful.The input section (not shown) should consist of amains transformer with 24 V; 2.8 A secondary; abridge rectifier (e.g. B80C2200/3300); and a4700 uF; 40 V smoothing capacitor.The L200 regulator should be mounted on asuitable heat sink. This device has internal short-circuit and overload protection; its pin assignment isgiven in Fig. 2.

2

5586448-1 (D

1 = input2 = limiting circuit3 = ground4 = reference voltage5 = output6 = ground

86448 -2

244 CURRENT INDICATOR FOR 723

Although the Type 723 voltage regulator has beenwith us for quite a few years, it is still a favouritecomponent for making simple and good qualitypower supplies. The 723 possesses excellentcharacteristics, including a highly stable outputvoltage, adjustable current control, and short-circuitprotection, but it lacks an output for signalling theactivity of the built-in current limiter.The current limiter in the 723 consists of only onetransistor, whose base and emitter are brought outto chip pins 2 and 3 respectively. When the voltageacross these pins exceeds 0.5 to 0.6 V, the transistoris turned on and cuts the drive to the output tran-

sistors. In most applications, the voltage drop forthe B -E junction of the current sense transistor isdeveloped across an externally fitted resistor. In thethe supply proposed here, this is either Rs, Rd/Rs,or Rd/Rs. A difficulty arises if it is intended to pro-vide an overcurrent indication for the shutdown cir-cuit with the aid of an external transistor fitted inparallel onto pins 2 and 3, since the external and in-ternal transistor are highly unlikely to have pre-cisely the same characteristics. When the internaltransistor has the low B -E voltage of the two, the in-dication will not work, while in the other case theexternal transistor takes away the base current for

111/132RR

278

the internal transistor, so that the current limiter isrendered ineffective.In this design of a power supply, a current overloadindication was realized by fitting the external tran-sistor with a high value base resistor, R7, to ensurethat the current limiter in the 723 is not disabled.A further transistor, T3, has been added to keep thebase current for T2 as low as possible. Since thebase -emitter junction then has a diode character-istic, the associated voltage drop is always lowerthan that of the transistor internal to the 723.The three output voltages from this supply are prob-ably the most commonly used for testing asym-metrically fed designs: 5 volt for many TTL andCMOS circuits, 9 volt for battery operated equip-ment or logic circuits equipped witha 7805 regulator (this requires an input of at least8.5 V), and 12 volt for RS232 drivers, and

F1

miscellaneous opamp or transistor based circuits.The current limiter can be set to 10 mA, 100 mA,or 1 A for safely powering experimental circuits.Power regulator Ti should be fitted with a heat sinksized at least 10 x 10 cm. LEDs D7 (green) and Ds(red) are the power on and current overload indi-cator, respectively.The output voltages of the supply may not be as ac-curate as required, and this is mainly due to the useof resistors from the E12 series. Close tolerance isespecially important in the 5 V range, since thevalue shown for R3 gives a theoretical output of4.9 V. This can be increased readily by fitting a re-sistor in parallel with R3, until the output voltage is5.0 V precisely. Switches Si and S2 are preferablySPDT types with a centre position, but three-wayrotary switches should also do if in both cases thecentre contact is not used.

5/9/ 12V0 /. 10mA/ 100mA/ 1A

270 p

5V " 12e: 061N

1112 4001ca

10y 25V

07403

245 DC/DC CONVERTER

In circuits where two signal paths must be elec-trically isolated, use is often made of an opto-coupler. Unfortunately, these devices require two

power supplies: one for the sender, and the other forthe receiver. In industrial and professional under-takings this requirement is met by a proprietary

279

DC/DC converter. As these are by and large very ex-pensive, they are not of very much interest to theaverage hobbyist. However, the do-it-yourself con-verter presented here is much less expensive and,moreover, easy to build.The circuit diagram in figure 1 shows that the con-verter consists of an oscillator, IC), and a driver,IC2, on the primary side, and of a rectifier,D1 . . D4, and buffer capacitor, C3, at the second-ary.In our prototype, operating from a 12 V battery atthe maximum 74 per cent efficiency, we measureda secondary output voltage of 10.64 V, and a sec-ondary output current of 9 mA (the correspondingprimary current amounted to 10.8 mA). The sec-ondary current should not exceed 10 mA, becausethe secondary output voltage then drops below

12 V

12V

10 V and the efficiency deteriorates. That appliesalso to low -load conditions: when the secondary isopen -circuit, the output voltage is 14 V, but the ef-ficiency is, of course, 0 per cent! In other words: thecircuit works optimally at a secondary load currentof 9 mA.Oscillator IC, operates at a frequency of around100 kHz. Its two output signals are each amplifiedin three parallel -connected buffers contained in IC2,and then applied to the primary of the isolatingtransformer. The voltage induced in the secondarywinding is rectified and smoothed by C3. The statedvalue of that capacitor is more than adequate forthe relatively high secondary frequency of 200 kHz.The isolating transformer is a DIY item: it is woundon a pot core of 22 mm dia. and 13 mm high with0.35 mm dia. enamelled copper wire - 80 turns for

N1 ... N6 = 1C2 = 4049UB

85424-1see text

the primary and 80 turns for the secondary. Thespecific inductance, AL, of the core should be400 nH. The core should not have an air gap. In-sulating foil should be placed between the two win-dings to ensure an isolating voltage of 4 kV.

246 DIGITAL VOLTAGE/CURRENT DISPLAY

This VII display module is eminently suitable forbuilding into an existing DC power supply, where itgives a precise indication of the set voltage or thecurrent consumption of the load.The circuit diagram appears in Fig. 1. The 3 -digitreadout is based on A/D converter Type CA3162

and BCD -to -7 segment decoder Type CA3161, bothfrom RCA. The common anode connections ofLED displays LDi-LD3 are successively connectedto the positive supply line via T,Provision has been made to select the correct pos-ition of the decimal point. In the voltage range, the

280

ti

IC 37805

640C1000 caO1Zop

25V

LD1 . . LD3 = 7750

T1 ... T3 = BC 640

BC 640

C

C3

100n

0 1:1111

0

C1

12

MKT

270n

4

P1

4

NT.CAP

HI LSO

650 3M50 4

Na 3162

16LD la

13 1

a 12 13

11 10IC 2 lo 8

6 2' 3161 I t15 2 d NM14 11 OPpLO 2

2

GAIN13

Ok

C2moo=MN

100n

7 2°

a

8'b

D214

e.I-

- OP

..0

D3 la

/b

- Op

decimal point lights on LD3, and the resolution istherefore 100 mV. Two current ranges are possible:0-9.99 A (link a) or 0-0.999 (.999) A (link b). Thecurrent sensing resistor is therefore either OR1 or1RO-see Fig. 2. It is important that R6 does not af-fect the output voltage of the supply in question. Itmust, therefore, be fitted ahead of the voltage div-ider that controls the output voltage. DPDT switchS1 selects between voltage and current readings.When voltage measurement is selected, P4 -R1 at-tenuates the input voltage by a factor 100. Also,point D is pulled low so that the decimal point onthe IS display, and the "V" LED, are illuminated.When current measurement is selected, the dropacross the sensing resistor is applied direct to theHI -LO inputs of DAC IC). The sensing resistor hassuch a low value as to render the voltage divider in-effective.There are four adjustment points in the module:Pl: current range nulling;P2: full-scale current calibration;P3: voltage range nulling;Pa: full-scale voltage calibration.These points should be adjusted in the above order.Two presets, Pi and P3, are required to ensure cor-rect nulling of the module. P1 compensates for thequiescent current consumption of the regulator cir-cuit in the supply. The resulting small negative devi-ation in the voltage range is compensated by P3.The VII display module is conveniently fed from

87468-1

the unregulated voltage available in the supply(max. 35 V)-see points E and F in Fig. 2; bridgerectifier B1 may then be omitted. The minimum in-put voltage for IC3 is 8 V, and this regulator shouldbe fitted with a heat -sink if the input voltage isgreater than 12 V. It is, of course, also possible topower the module from a separate 8 V; 200 mAmains transformer.The unit can be constructed as a double to obtainsimultaneous V and I readings. It should be noted,however, that the current sensing resistor isshort-circuited via the ground connections whenboth modules are fed from the same supply. There

2

87468-2

281

are two ways to overcome this problem. One is tofeed the V unit from a separate supply, and the Iunit from the "host" supply. The other is moreelegant and entails hard wiring points E to the leftside of the current sensing resistor. Note, however,that the highest V indication then becomes 20.0 V(Rs drops 1 V max.), since the voltage at pin 11 maynot exceed 1.2 V. Higher voltages can be displayedby selecting the lower current resolution, i.e., R6becomes OR I. Example: R6 drops 0.5 Vat a currentconsumption of 5 A, so that 1.2-0.5 = 0.7 V re-mains for the voltage indication, whose maximumreading is then 100 x 0.7 = 70 V. Again, thesecomplications only arise when two of thesemodules are used in a single supply.

Parts list

Resistors ( ± 5%1:

Ri =82KR2;R3 = 82RR4 = 15KR5 = 27KR6 = OR1 or 1R0*Pi = 50K presetP2 = 10K presetP3 = 1 OM presetP4 = 1K0 preset

Capacitors:

Ci =270nC2;C3 = 100nC4 = 470 II ;25 V

Semiconductors:Di; D2 = LED redBi = BC40C1000LDi;LD2;LID3=7750

. . .T3 incl. =BC640T4 = BC547BT5 = BC557BIC, =CA3162IC2=CA3161IC3=7805

Miscellaneous:

Si = miniature DPDT switch.PCB Type 87468

" See text.

- aC)

t\e4 4-1 :\4°01137V2

C 1 4a%

B1C4

2

a

247 DIRECT -CURRENT MONITOR

Many direct -current monitor circuits use a resistorin series with the current -carrying wires, and actu-ate some indicator by the ensuing voltage dropacross that resistor. The drop causes a reduction inthe available load voltage, which at relatively highcurrents can be appreciable. In the present circuit,use is made of a reed relay, around which thecurrent -carrying wire is wound a number of times.The consequent losses are minimal. This methodhas a bonus in that a switch contact is immediately

282

available for a number of applications.One possible application is that of a low loss lampmonitor. As long as the lamp (here represented byRL) is on, the LED lights. The number of turnsdepends on the relay use;., and the load current. Asa guide, most reed relays operate at 50 ampere -turns, so that in the case of, say, a car headlight(60 W at 12 V gives a current of 5 A) about 10turns are required.The more complex circuit diagram shows an elec-


tronic fuse, which also offers overvoltage protection.The state of the circuit is indicated by two LEDs.When the supply is switched on, the thyristor is off,and relay Re will be energized via the 150 -ohm re-sistor. The load will then be connected and thegreen LED lights.If the load current becomes too high, the reed relaywill close and trigger the thyristor via the 470 -ohmresistor. The thyristor then short-circuits relay Re,which causes the load to be disconnected. At thesame time, the green LED will go out and the redone will light. The circuit may be reset with S2,which breaks the current through the thyristor andcauses it to switch off.Overvoltage protection is provided by the zener di-ode across the reed relay. When the input voltagebecomes greater than the zener voltage and thethyristor trigger voltage, the thyristor will be trig-gered and switch on the protection circuit.These two applications are primarily of use in cars,but, no doubt, ingenious readers will think ofothers.

D4 = red D5 = green D6 = green

11-15V

11-15V

86508

248 DIRECT -VOLTAGE DOUBLER

A direct -voltage doubler is particularly useful whenfrom an available supply voltage a higher one hasto be derived. As the current in most such cases ispretty small, the cost of a suitable circuit can bekept down.Astable multivibrator IC, is a rectangular -wave gen-

12 V ®

erator operating at about 8.5 kHz whose outputdrives transistors Ti and Tz When the level at pin 3of IC, is low, Ti is off and T2 conducts. As thenegative terminal of C3 is then connected to earth,the capacitor charges via diode Di. When the out-put of IC, is high, T2 is off and Ti conducts. Ca -

R1

R2

C1 C218n2 7lOn

01 D21N4001 1N4001H *I

C3O

BC 639

40 V1

=i=0 µ l0µ

40V

BC 640

0 20 V

85418-1

IBC 639BC 640

ECB

283

pacitor C3 cannot discharge because of Di, but C4charges to a voltage roughly equivalent to thesupply voltage of +12 V and the p.d. across C3 andDi. In our prototype, this voltage across C4amounted to 20 V approximately. The maximumcurrent should not exceed 70 mA: at that value, theoutput voltage is 18 V, at an efficiency of thirty-twoper cent.We have not tested the circuit with other supplyvoltages, but it can be safely assumed that it can be

used over the whole supply voltage range of theNE 555.Construction is possible on a small piece of pro-totyping board, after which the doubler can be fit-ted inside the power supply unit.If a regulated output is required, it is possible toconnect an appropriate voltage regulator, for in-stance, in the 78LXX series, but in that case thepower requirements of the regulator must, ofcourse, be taken into consideration when the maxi-mum load current is calculated.

249 ECONOMICAL POWER SUPPLY

The power supply described here uses a silicon -controlled rectifier (SCR) that, depending on theload current, selects taps on the secondary of themains transformer. The output voltage of around9 V is eminently suitable as input voltage for a 5 Vregulator, which consequently works with the ab-solutely minimum power dissipation.With low to medium load currents, the SCR is inthe blocking state. Rectification of the secondarytransformer voltage then takes place in 131, D2, D5,and D6 only. The load current flows during thepositive half cycle via Di, load, and D5; during thenegative half wave it flows through D2, load, and

Fl

Si D5,D6 = 1N5401

KAG

D6. The tapped secondary voltage amounts to 8 Vin either case, while a 2 V section remains unused.With increasing load current, the output voltagedrops until no current flows any more through thezener diode. Transistor T7 switches off whichremoves the short circuit from the gate of the SCR,which then conducts. As soon as that happens, thefull secondary transformer voltage is rectified byDi . Da, while diodes D5 and D6 are reverse bias-ed.As the voltage across the zener diode is alwayslowest during the zero crossing of the secondaryvoltage, the SCR always switches on at or near that

85445

284

instant. This prevents high current pulses and othernoise often associated with SCR switching: nofurther suppressors are therefore necessary.To build this supply, you need a mains transformer

with a 12 V secondary that has taps at 2 V steps:2-4-6-8-10-12 V. For load currents up to 1.5 A, a 2 Atransformer will suffice; an output current of up to2 A requires a 3 A transformer.

250 LEAD -ACID BA1TERY CHARGER

Although in electronics more NiCd than PbH2SO4batteries are used (or so we're told), there is still ahealthy demand for good chargers for the lead -acidtypes. The present one enables 6- or 12 -volt types tobe charged rapidly; switches itself off automatically;and is protected against thermal overload, short cir-cuits, and polarity reversal of the battery.If you are not fully acquainted with modern sealedlead -acid batteries, here are some of its more im-portant properties. It may be used in any position,even upside down. The charging voltage should be2.3 V per cell (2.45 V for fast charging): i.e., 6.9 Vfor a 6 V battery and 13.8 V for 12 V types. Thecharging current need not be limited to 0.1 . . .1 C( = capacity in Ah - the actual figure depends onthe manufacturer). The battery is charged when thecharging current has dropped to 1 per cent of thecapacity. Some manufacturers state that it is

preferable that their batteries are charged in ahorizontal position. Never charge these batterieswith a NiCd battery charger!!The circuit of the suggested charger is based on atype L200 voltage regulator which ensures a con-stant charging voltage. The actual level of the charg-ing voltage is set with 131 in the absence of a battery.Resistors R and R2 provide current limiting, but R2is only necessary if a charging current above 0.5 Ais required or to enable the output current more pre-cisely. The current is limited to

12V

- -0

Fl100mA

[0.45(R, +R2)1R1R2] A; its actual value is indicatedby MI.The L200 may be mounted on a small heat sink,but this is not strictly necessary since the device hasinternal thermal protection.Normally, the battery charger works from themains, but it can also operate from a 12 V (car) bat-tery.All possible situations, some of which are highlyundesirable, are enumerated in table 1. The one ex-ception is that when the battery is really flat, thetable does not apply. The battery must then be seento be connected correctly to the charging terminals.

Table 1.

condition LED meter

polaritycorrect

voltageat input lights 400 mA

no voltageat input out 0 mA

polarityreversed

voltageat input out 400 mA Illno voltageat input out 0 mA

batterynot connec-ted

voltageat input lights 0 mA

no voltageat input out 0 mA

.SO0 IBM

T 6V

L200

1 2 34 5

285

Also, the LED indication will then initially notwork.Table 2 gives some examples of 6 V batteries andcircuit variations required for the different types.The charging currents here are limited to 1/10 of thebattery capacity in Ah (ampere -hours): this is a safevalue which is permissible under all circumstances).If the charger is required for n V batteries, themains transformer must have a secondary voltage

Zd3gui

0

Parts list

Resistors:= 1 g

R2 = see textR3 = 820 gR4 = 560 g (see text)R5 = 470 Q

= 500 Q preset (see text)

of at least 18 V, and capacitor Ci must become a35 V type. Furthermore, resistor R4 should be in-creased to 1k8 and preset 131 to 1 k52.

Table 2.battery Trn Dr...D. R, Rz Mr

type 0,066V 4 Ah 12 V; 0.6 A 154001 14 .. 0.5A6V 6 Ah 12 V; 1.OA 1N4001 1 52 24 1.OA6 V 8 Ah 12 V; 1.2 A 1N5401 1 4 1 4 1.0 A6 V 10 Ah 12 V; 1.5 A 1N5401 0.824 0.824 1.0 A

Capacitors:C, = 1000 m/25 V (seeC2 = 330 nC3 = 1 µ/16 V

Semiconductors:D, ... D4,D7,D6 =

(see table 2)D5,D9 = 1N4148D6 = LEDIC, = L200

text)

154001

Miscellaneous:M, = moving coil meter, 500 mAf.s.d.

Tr, = mains transformer, secondary12 V, 600 mA (see text)

S, = DPST mains on/off switch= fuse, 100 mA, delayed action

heat sink for IC, (optional - see text)PCB 85446

251 LOSS -FREE SUPPLY PROTECTOR

Any diode -based circuit that protects against rever-sal of the supply polarity introduces a certainvoltage drop. Also, when relatively high currents areinvolved, the choice of a suitable diode, and its dissi-pation, may become problematic.This circuit utilizes a relay contact to break thepositive supply line when the input voltage has thewrong polarity. The coil voltage of the relay may belower than the input voltage, because Re is ac-tivated within a few milliseconds, and then receivesthe correct coil voltage via T1-131. Since the holdvoltage of a relay is generally lower than the actua-tion voltage, D2 can be dimensioned such that therelay operates reliably with a minimum of zenercurrent taken from the supply.

286

87102 - 3


252 LOW -DROP VOLTAGE REGULATOR

Integrated 3 -pin voltage regulators are not suitablefor use where the input and output voltages arenearly equal. In fact, with most such regulators, theinput voltage is typically 3 V higher than the out-put potential. To cater for situations where the twovoltages are nearly equal, it is necessary to usediscrete components. The series transistor is thenconnected in a common emitter circuit, so that theoutput voltage is lower than the input voltage onlyby the saturation voltage of the transistor. However,it is then difficult to provide short-circuit protectionas is the case in integrated regulators. But, wherethere is a will, there is a way.In Fig. 1, the series transistor obtains its base cur-rent from T2, which together with Ti forms a dif-ferential amplifier. This arrangement ensures thatthe junction of voltage divider Ra-Rs has the samepotential as the cathode of zener Dz. The crux ofthe circuit is that T3 has a certain current amplifi-cation, but 12 can only provide it with as muchbase current as R2 allows. The potential differenceacross R2 has a maximum value of the zenervoltage minus the base -emitter voltage, UBE, of T2,which in practice is about 4 V. The maximum cur -

19..6...15V 9.4V 0.5A

86440 -I

rent through R2 is, therefore, about 11 mA, so that,assuming that T3 has a current amplification of 50,the maximum output current is 0.55 A. If a highercurrent is drawn, the output voltage will drop. If itdrops below the zener voltage of D2, the p.d. acrossR2 will drop also. The result is that the output cur-rent will behave as shown by the fold -back charac-teristic in Fig. 2. It is clear, therefore, that the seriestransistor is protected against high (short-circuit)currents.Diode Di and resistor R: provide a soft start,because the voltage across the diode, which is con-nected to the output of the regulator, is nought atswitch -on. Since the circuit, because of the highgain, has a tendency to oscillate, capacitor C: is in-cluded to improve the stability.The output voltage level, U0, can be freely selected,within the limits of the series transistor, by D2, R3,and R4, and is determined fromUo = Uz(Rs + R4)/Rs.Resistor R2 must be matched with the actual cur-rent amplification of the transistor used. The maxi-mum dissipation of a well -cooled BD140 is of theorder of 5 W. If a noise -free output is required, anadditional 10µF electrolytic capacitor should beconnected in parallel with Dz. The circuit will thenhave a real soft start: there will be no output forabout 0.2 s after switch -on.

2

R4 + R5Uo- R5 Uz

86440-2

253 MAINS POWER SUPPLY WITH PRIMARYREGULATION

The unusual circuit shown in figure 1 has anunusual efficiency: according to SGS, this amountsto no less than 37 per cent at an output voltage of3 V and output current of 2 A. With traditional sec-ondary regulation, an efficiency of about 8 per centwould have been normal. The output voltage can be

varied over the range 1.2 . . .25 V, and the outputcurrent can be 1.5 A at any of these voltages, pro-vided IC2 is mounted on a suitable heat sink.Another advantage of primary regulation is thatthe power supply is protected against variations inthe mains supply. This aspect is normally ignored

287

LI L2

404H 404H3 A C9 3 A R1

NMI 7W

40 uH3A 3A

L120A

CS C

2200V 16V

Trl

with secondary regulation, as it is assumed thatprimary fluctuations have no effect on the second-ary regulation. The present circuit is, therefore, ofparticular importance for use where the mainssupply is subject to large variations.The regulation functions so that the voltage dropacross voltage regulator IC2 is held constant. Thisvoltage drop is transferred by current source T1 intoa current through the LED in the opto-coupler.When the voltage drop diminishes, the currentthrough the LED is smaller. The transistor in theopto-coupler gets less drive, and the voltage at pin 3of IC, drops. Voltage regulator IC, contains a com-plete circuit for phase gating control with silicon -

850C5000/3500

ti R16SW

C2470005

mom100s400V

1C3TIL 119

5

BC 557

IC2LM 317

T

I

`see text

TIC 226 D LM317T

Al A2ad, Or)

0Uo

C11 40V

85419

controlled rectifier Tril. The gating angle of thistriac depends on the comparison between the directvoltage at pin 3 and an on -chip generated sawtoothsignal, the frequency of which is determined by ca-pacitor Ci ( = 100 n). In our example, the triacswitches the mains voltage earlier so that buffer ca-pacitor C2 receives more energy.Noise caused by the phase gating circuit must beprevented of entering the mains supply by a mainsnoise filter as shown.SGS application

254 MAINS ZERO -CROSSING DETECTOR

Both safe and remarkably simple to construct, thiscircuit detects the zero crossing moments of themains voltage, in order to provide other circuitrywith timing information about the correct instantfor switching mains -connected loads; in otherwords, when the least possible switching dissipationis involved, and, therefore, least interference is in-duced on the mains lines.The proposed circuit operates direct off the mains,while comprising no more than two opto-couplersand two resistors. It is seen that photodiodes Di

and D2 are connected in antiparallel while beingfed with the mains voltage via a resistor, whichlimits the current through the relevant diode toabout 2 mA as it conducts (i.e. lights) during thenegative or the positive half wave (D2 or DI re-spectively) of the mains sinewave; in either case, thecircuit output voltage is low, since the associatedphototransistor conducts and draws current from+ Ub via R2.However, at the moment of zero crossing, neitherone of the diodes conducts, and the voltage at the

288

circuit output rises to near + Ub level, whence the100 Hz pulse train.The value of R2 may be adapted to suit the level of+ Ub and the manufacturer -specified typical collec-tor current through the phototransistor. For theType TIL111, the current should not exceed about50 mA. The type of optocoupler used in the circuitshould not be very critical, but the value of RI hadbest be left at the indicated 100 k so as not to runinto excessive diode dissipation.

41

ICI ,IC2=2xTIL111 86433

255 NEGATIVE SUPPLY CONVERTER

It is sometimes required in certain circuits that arepowered from just one battery to derive a negativesupply voltage from the positive battery potential.As the loading of such negative lines is normallypretty minimal, it is possible to use a TL 497A IC

5VU,

RCL

14 13 10

TL 497A

llo

currentiimiting

Oscillator

CT

170p

ae text

85932

to provide the negative voltage. This saves a trans-former, rectifier, and a smoothing capacitor.The TL 497A is a switch -mode IC from Texas In-struments, that may be used as an up-wards/downwards transformer, but also as anegative supply converter.Inductor L makes it all possible, because when theon -chip transistor is switched off, a fairly largeback-e.m.f. is generated across L, which causes anegative potential at the emitter of the transistor.The diode then conducts, and capacitor CFcharges. The output voltage, U0, is determined byU0 = [-Uot./t0]Vwhere Ub is the supply voltage; ti is the time thetransistor is switched on; to is the time the transis-tor is switched off. Period ti is determined by thevalue of Cr.The output voltage is devided across R1 and R2and applied to the inverting input of an on -chipcomparator, whose + input is a 1.2 V referencevoltage. When the actual value of U0 lies below thewanted value, the comparator toggles and switcheson the oscillator, which in turn drives the transistor.The TL 497A also contains a current limiting cir-cuit which ensures that the coil cannot be saturatedand that that transistor is not affected by voltagespikes.Coil L may be any fixed inductor with a value of100...500 µ11.The output voltage is calculated fromU.= -[N +12]Vwhere N is the numerical value of R2 in kilohms.The output current should not exceed 50 mA.

Texas Instruments Application

289

256 NICD BATTERY CHARGERS

Quite arguably, Nickel Cadmium (NiCd) batteries 2are frequently used as replacements for disposabletypes of battery; this is possible because they can beinserted readily in the existing battery compartmentand supply the same voltage as disposable batteries.The fact that one need not go out to purchase (rela-tively expensive) batteries puts the rechargeablecells in an advantageous position.However, one drawback of the use of rechargeablebatteries is the need to remove them from the equip-ment any time their charge is exhausted. It would,therefore, be convenient to leave them where theyare, i.e. in the battery compartment, as they receivethe charge current.Two circuits are suggested for the incorporation inexisting battery -operated equipment. Figure 1shows the absolute minimum in the form of asimple current source. The reference voltage is ob-tained from the forward drop across LED Di(about 1.5 V for a red LED). R2 fixes the currentthrough the LED, and the voltage at the base of Tiis therefore about 1.5 V lower than the positivesupply rail. The voltage across Ri is about 0.85 V,and this value may be used to determine the chargecurrent for the battery, since A= 0.85M, indepen-dent of the circuit supply voltage.The value of Ri is thus readily calculated if it is

known that most NiCd batteries are preferablycharged with a current of one tenth their capacityin amperes per hour (Ah). A number of the morepopular battery types and associated values for Rihave been listed in Table 1.A noteworthy aspect of the circuit is the fact thatLED Di will go out in the absence of a battery,since the voltage across Ri inevitably drops; the

1

D1

R2

T1

BC557/BD140

NiCd

D2

T1R3

BC557/BD140

R4

Table 1.

IN4001

NiCd

86477-2

battery type size capacity[mAh]

chargecurrent[mA]

Ri[Q]

9 V block - 110 11 82lady RI N 180 18 47micro R03 AAA 180 18 47penlight R6

(mignon)

baby R14

AA

C

500

1200

50

120

15

6.8

1800 180 4.7mono R20 D 4000 400 2.2

Table 2.number of

cells

Vi(min.)

R2

N4

R3

N4

2 5 270 223 6 330 27

4 7.5 470 39

5 9 560 476 10 680 56

7 12 820 68

LED current which used to flow through R2 willnow pass through Ri and the base -emitter junctionof Ti.The elaborated version of the NiCd charger, shownin Fig. 2, includes a diode to protect the circuit frombeing damaged by input voltages having the wrong

86477 -1 polarity. R3, R4 and T2 have been incorporated to

290

disable the charger in the absence of a sufficientlyhigh input voltage; Table 2 lists the relevant valuesfor R2 and R3, given the number of 1.2 V cellscontained in the NiCd battery.Almost any type of silicon PNP transistor in the BCseries should work satisfactorily in the Ti positionif the charge current does not exceed about100 mA. Higher input voltages and/or charge cur-rents are, however, better handled by a medium -power transistor from the BD series.The input voltage to the charger need not be

elaborately regulated or smoothed; in fact, any typeof inexpensive adapter providing the necessarydirect output voltage and current may be used. De-pending on the number of cells contained in theNiCd battery, the charge current may also be ob-tained from the 12 V car battery.The circuits as shown are readily fitted on a smallpiece of veroboard to suit incorporation in the rel-evant equipment; the input voltage to the charger isconveniently connected to a small plug or socket fit-ted onto the cabinet.

257 ONE -CHIP DC CONVERTER

This DC step-up circuit may prove useful for the in-corporation in equipment that requires the presenceof a supply voltage in excess of the normal circuitsupply rail of, for instance, + 5 V. Ideal therefore forgenerating the necessary + 8 . . .12 V voltage tofeed RS232 transmitter devices, or the + 25 V pro-gramming voltage for EPROMs, the Type L497 DCconverter requires very few additional passive partsto produce any of the output voltages listed in thetable below.As to the components in support of the converterchip, note Li, which is a small coil, readily made bywinding about 85 turns of 34 SWG (4) 0.2 mm) en-amelled copper wire on a small (11 x 7 mm) pot corehaving an Al rating of 160, e.g. the Siemens Type6531 -L160 -A48. The total inductance of Li shouldbe of the order of 100 µH. Resistor Ri must be di-mensioned as indicated in the table for any of theno-load output voltages. Note that the voltageacross R2 is fixed at 1.2 V, and that the value of Rimay therefore be computed from Rt.-. (Voot----/.2)<1d2>.Finally, the output current may, of course, beboosted by means of a medium power transistor ina suitable configuration at the Vo output.

V, Vo* li (max) Riv

5 10 125 8.85 15 80 13.85 20 60 18.85 25 50 23.8

<V> <V> <mA> <1(52>

* Specifies no-load output voltage. Theoretical value; select nearest E12 or E24 value.

258 PRECISION RECTIFIER

This precision rectifier operates from an asymmetri-cal supply, handles input signals up to 3 Vpp andhas a frequency range that extends from DC toabout 2 kHz. Its amplification is unity, and dependsmainly on the ratio Ra/R3. Opamp Al is connectedas a voltage amplifier (Ao =1), A2 as an invertingamplifier (Ao= -1). Opamp Az, transistor T1 and

diode D2 ensure that the output voltage, U2, isidentical to the positive excursions of the inputvoltage, Ul. When U1 is positive, the output of Aiis held low at about 0.25 V, so that T2 is disabledand can not affect the rectified output signal.Components R2 and Di protect the pnp input stagein A2 against negative voltages, which are effective -

291

ly limited to -0.6 V. For negative excursions of theinput signal, the function of Al, T2 and D2 issimilar to the previously mentioned components.The peak output voltage of the rectifier circuit is de-termined mainly by the maximum output swing ofthe opamps and the voltage drop across the tran-sistors plus D2: this amounts to about 3 V in all.When the circuit is not driven, it consumes about1 mA, and is therefore eminently suitable forbuilding into portable, battery -operated equipment.

U2

A1, A2 =1C1 = LM358T1, T2 = BC 550CD1, D2 = 1N4148

874161

259 SIMPLE NICD CHARGER

Dry batteries have one major disadvantage: they goflat. Rechargeable types, such as NiCd cells, alsosuffer from this drawback, but they can at least berecharged. Sometimes even a fifteen minute chargeis sufficient to give enough life to, say, an electronicflash battery.A NiCd charger is, in essence, nothing but a sophis-ticated current source. The present design contains

B80C

Tr1

1500

fi

I:5+B

CIC3 cL1000700n16V

12V / 1A

50mA

IC17808 -1

D1...D4 = 1N4148T2...T5 = BC160

D5...D8 = LED red

R11

1=1D4

four such sources with a common control switch,but each with a separate LED that lights as soon asa battery is connected to it.In position 1, the sources each provide a current ofabout 90 mA; in positions 2 and 3 values of be-tween 100 and 300 mA as required. Note, however,that with charging currents above about 200 mAthe transistors must be fitted on suitable heat sinks.

1W

R3

1W 1W

BC547BT1

15

RR D8

R10 R8

D3 D2 D1

C* CO COR1

-D7

14

R6

D6

g

T3

D5SCR

R4

=1.4VA

T2

P2

10k

P1

10k

3

R2

86437 - 1

2

292

On stability considerations, it is advisable to mountdiodes Di to D4 incl. in good thermal contact withthe relevant transistors.Terminals +B and -B enable the circuit to operatefrom a 12 V DC source, such as a car battery, insituations where a mains supply is not available.

Modern NiCd cells can be given a fast charge with-out any problems. The present unit can charge TypeAA (= HP7 = R6) cells in about 8 hours (position1=90 mA); Type C (= HPI I =R14) cells in 10-14hours (position 2 =180 mA); and Type D(= HP2 = R20) in 20 hours (position 3 = 200 mA).

260 SIMPLE ZERO CROSSING DETECTOR

Zero crossing detectors are often contained in rathercomplex circuits, or they are part of an integratedcircuit, the rest of which is not required. Basically,such a detector is required to give a pulse every timethe mains voltage passes through the zero potential.The detector proposed here is very simple indeed:the mains voltage is transformed down, rectified inDi, and smoothed by C1 to give a direct voltage of17 V. Part of the mains voltage is taken from acrossR2 and used to drive transistors T, . Ta Duringpositive half cycles Ti conducts and T2 and 13 areoff, whereas negative half cycles switch on T2 andT3, while Ti is off. When the momentary voltageacross R2 lies between + 0.6 V and -0.6 V, noneof the transistors conducts, so that the outputvoltage is high. In this way, a short positive pulse isproduced every time the mains voltage passesthrough zero potential. Since operation is direct

*see text

C>

C>85409-2

from the mains, there is no phase shift caused bythe usual isolating transformer.Where the direct voltage output of the circuit isused for supplying external circuits, attentionshould be paid to the current required by those cir-cuits and the rating of the transformer. It may alsobe necessary to increase the value ofFinally, remember that the circuit and, therefore,any external units are connected direct to themains!

261 SUPPLY PROTECTION

The use of external mains voltage adaptors forcassette recorders, portable radios, home computers,pocket calculators, and so on, is common practicesince the typical enclosure sizes of this type of elec-tronic equipment either does not allow the incor-

poration of a mains supply, or the device has beenprimarily intended for battery operation.Unfortunately, the degree of idiosyncrasy amongmanufacturers of adaptors is rather high; a standardfor adaptor output voltage and output polarity is

293

definitely hard to find. It may, therefore, be quiterisky to power, say, the home computer from anadaptor which is not tailored to this (expensive)piece of electronic equipment.Here is a simple circuit to prevent a lot of trouble;its extremely low cost fully justifies incorporation inany equipment operated from an external DCsupply. The supply protection consists of a merefour parts and a fuse, which may already be incor-porated in the equipment. The zener diode is selec-ted to have a zener voltage of about one volt higherthan the equipment supply voltage.In case the input voltage to the circuit has thewrong polarity, the zener diode conducts and causesthe triac to fire, since its gate is driven positive withrespect to MT2; the current flow through the triacis sufficiently high to look like a short-circuit to thefuse, which duly melts and breaks the supplyvoltage, before damage is inflicted upon equipmentparts.Operation of the circuit in an overvoltage conditionis even simpler, since in that case the zener also sup-plies the gate of the triac with a firing voltage.

MT 1MT2 0

TAG 226

`see text86401

Obviously, if everything is in perfect order, the pro-tection circuit is as if non-existent to the equipmentit is part of, because it introduces no additionalvoltage drop. Finally, the only modification to thecircuit for use in positive -earth equipment involvesinsertion of the fuse in the negative -supply line.

262 SWITCH MODE POWER SUPPLY

One of the major problems in the design of switchmode power supplies is that most available (andsuitable) ICs only offer the absolutely necessaryfacilities and not, for instance, thermal or short-circuit protection.Linear Technology offers a solution to these prob-lems with their LT1070 range of switching ICs.These devices are as easy to use as the familiar 3 -pinregulator ICs. All steps have been taken to make thedesign of a simple, yet efficient, switch mode supplyas easy as possible. The peak output current is 5 to9 A, and a current -limiting circuit is provided.The diagram shows a switch mode DC -to -DC con-verter, whose output voltage may lie between 12and 48 V, provided the input voltage is greater than3 V. The input voltage of the circuit as shown mustalways be lower than the output voltage. It is, how-ever, possible to modify the circuit to obtain an out-put voltage that is lower than the input voltage. Oneof the modifications is replacing Li by a suitabletransformer.The output current is dependent on the value of theinput voltage. For an input voltage of 3 V, the out-put power is 10 watts maximum. Our prototype,operating from 3 V, delivered about 50 mA at 48 V,

BOTTOM VIEW

Usw tic

CASE IS GND

L1* BYX55

*see text86518

FRONT VIEW

1 2 3 4 5

UC t t USW

FB t UIN

294

while at an input voltage of 24 V the output cur-rent was over 1 A.In the construction, account must be taken of highpeak currents: all connections should, therefore, beshort, and the input and output lines should beSWG20 (0.8 mm 0) or thicker. This also applies tothe earth connections.

It should also be noted that spikes are present onthe output voltage. If necessary, these may beeliminated by an LC filter, the inductance of whichhas the same value as Li, and the capacitance isbetween 10 and 100 µF. The quality of the capaci-tor is of importance because of the required lowseries resistance for RF signals.

263 VARIABLE 3 A POWER SUPPLY

As far as construction is concerned, this is a realmini power supply, but it can deliver up to 3 A atan output voltage of 1.25 . . . 25 V. Note, however,that integrated voltage regulator IC, has on -chipoverload protection that comes into operation whenthe dissipation in the device reaches 30 W. TheADJ(ust) pin of the regulator is connected to thejunction of potential divider The outputvoltage, Uo, is calculated fromUo = [1.25 (1 + Pi/R1)] Vwhere Pi and Ri are in ohms (the value of Pi ismeasured between the wiper and the junction withRi, i.e., 0...2.5 k52).Capacitor Ci is a conventional filter capacitor,while C2 and C3 improve the regulation. Protectiondiodes Di and D2 ensure that at switch -off the

potential at the output of IC, is more positive thanthat at its input. The value of R has been chosento ensure that the minimum load current throughIC, is about 3.5 A.It is essential that IC, is mounted on a heat sinkrated at about I K/W - do not scrimp on the heatconducting paste!When only low output voltages are needed, itmakes sense to use a mains transformer with alower secondary voltage (for Uo = 5 V, the second-ary voltage should be 9 V). When a 24 V secondaryis used, and the required output voltage is 1.25 V,the maximum output current is 1 A, otherwise themaximum dissipation of the LM 350 is exceeded,and the internal protection will switch off the regu-lator. When the secondary voltage is 9 V, and Uo= 1.25 V, the maximum load current amounts to2.5 A.

D1

1V25....25 V

264 VISIBLE POWER -ON DELAY

While in the process of repairing or testing elec-tronic equipment, it is often desired to have moretime available for hooking up an oscilloscope probeor test lead to the part in question, after power hasbeen applied.

This circuit gives you plenty of time to reach anycomponent in the circuit, since the equipment isonly switched on after a fixed interval followingdepression of the start button.The basic operation of the circuit is as follows. Ac -

295

8 . 10V

0 if <4; Cr

N1...N4= IC1 = 4093N5...N7= IC2 = 4025

O

tivating start switch Si sets the bistable composedof N2 and N3, causing N3 to provide clock pulsesto decade counter 1C3. The driver transistors at theoutputs of this IC will light the LEDs one afteranother, indicating the countdown function of thecircuit. Rei is energized the moment the last LEDin the row lights; 1C3 is disabled via its CE input,and the N2 -N3 bistable is reset, all at the sametime. The equipment, powered over the relay con-tacts, is turned on, and the user may take thedesired reading.The power -on interval may be restarted or inter-rupted during countdown by depressing S2, which

R8...R17 = 10kR18...827 = 390k

T1...T10 = 8,05478

2

68

E:31610

1411

612

813

0 j..)

IC 1

Ors 1

1418

03

02

D6

07

08

09

4IC 2D5

IC3t1

86488

resets the bistable and counter.During the final three stages of the countdown, awarning buzzer is arranged to sound by means ofN4; this function may be disabled by means of S3.Re I should not consume more than 100 mA of coilcurrent, while its contact(s) should be rated to suitthe load to be switched.

265 VOLTAGE INVERTER

Here is a circuit that produces a negative voltagefrom a positive one, for instance, from +5 V to -10 V. The output voltage, U0, is determined from

U0 = -I .2(Ri/R2 + I)

296

As in other, similar circuits, the maximum outputcurrent depends on the ratio between input and out-put voltage and is calculated from

iolmax) = 500/(Ri/R2 + 1) [mA]

The choke is readily made with a 17.5 mm pot coreon which 85 turns of 0.2 mm2 enamelled copperwire are close -wound.The maximum input voltage to the IC is 15 V. Ef-ficiency is of the order of 60 per cent. ,ee text

297

266 8 -CHANNEL VOLTAGE DISPLAY

Simultaneously monitoring the trends of 8 slowlyvarying voltages is normally very difficult, if not im-possible, even with the aid of 8 analogue or digitalvoltmeters. This circuit turns a common oscillo-scope into a versatile 8 -channel display for directvoltages. The trend of each of the 8 input levels isreadily observed, albeit that the attainable resol-ution is not very high.The circuit diagram shows the use of an 8 -channelanalogue multiplexer, ICI, which is the electronicversion of an 8 -way rotary switch with contactsX0 -X7 and pole Y. The relevant channel is selectedby applying an binary code to the A -B -C inputs.Example: binary code 011 (A -B -C) enables channel7 (X6-.Y). The A -B -C inputs of ICI are driven fromthree successive outputs of binary counter IC2,which is set to oscillate at about 50 kHz with theaid of Pt As the counter is not reset, the binarystate of outputs Q5, Q6 and Q7 steps from 0 to 7 ina cyclic manner. Each of the direct voltages at inputterminals 1 to 8 is therefore briefly connected to theY input of the oscilloscope. All eight input levelscan be seen simultaneously by setting the timebaseof the scope in accordance with the time it takes thecounter to output states 0 through 7 on the Q6 -Q6 -

Q7 outputs. The correct starting time for the oscillo-scope trace is ensured by using the Q8 output of thecounter to supply the trigger pulse. Diodes Di andD2 provide for some space between adjacent bars onthe display, and create a horizontal reference line.The timebase on the scope should be set to0.5 ms/div, and triggering should occur on the

87438

positive edge of the external signal. Set the verticalsensitivity to 1 V/div. The input range of this circuitis from -4 V to +4 V, and connected channels areterminated in about 100K.Adjusting the 8 -channel voltage display is straight-forward. Simply select the previously mentionedscope settings, and adjust PI to make all 8 channelsvisible over the full width of the scope screen-seethe accompanying photograph.The circuit has a modest current demand of lessthan 5 mA from a simple ± 5 V supply, or from two4.5 V flatpack batteries.

267 AUDIO TESTER

A simple millivoltmeter and an equally simple sinewave generator are ideal instruments for checkingand testing audio equipment. The audio tester com-bines the two, as shown in figure I, where Al andA2 form the millivoltmeter circuit, while the sinewave generator is built from A3 and A4.As the audio tester is supplied (asymmetrically)from a 9 V battery, this supply must be halved forthe operational amplifiers. This is essentially doneby zener diode D7. This zener is biased by R6, andthe reference voltage is taken from the junction ofdiodes D8 -D9 via resistor R7. The reference voltageis, therefore, about 5.3 V. The constant voltage dropacross the two diodes is applied across preset P3

which serves to negate the offset voltage of A2 (en-abling the millivoltmeter to be calibrated to zero).The input signal is applied across high pass filterCi/Ri to the non -inverting input of Ai. For all prac-tical purposes, this sets the input impedance at1 M4. Note that the circuit is fully driven with aninput signal of 50 mVrms. Higher inputs necessitatea voltage divider at the input or a reduction in gainin Al by dropping the value of R3. When this re-sistor is reduced to 6k8, for instance, the gain of Aiis 2, and the input sensitivity is 275 mVrms.Full-scale deflection of the meter is set by PtOpamp A, together with diodes D3 . . .D6, functionsas an active full -wave rectifier. The meter is connec-

298

R3

R10

WI 0

08P3S

R.

k

9 VRS

R12

U

A2

R15

C5

Al ... A4 = IC1 = TL 084

EMI

R11

R13

DS

03

4x1N4148

Tln7

00

BC 5478

Parts list

Resistors:

= 1 MR2 = 6k8133 = 68 kR4,R5 = 150 kR6,R9 = 1 k

R7 = 100R. = 15 kRio,Ril = 2k2R12 = 4k7R1a,R14 = 100 k

Ti

R14

BC 547B

R15 = 8k2R16 = 82 kR17 = 470 QPi = 25 k preset (see text)P2 = 1 M stereo preset, log.P3 = 5 k preset

Capacitors:

= 1 p metallized plastic foilC2,C6 = 100 p/10 VC3,C4 = 4n7C7 = 560 nC. = 220 p/16 V

ted in of the diagonals of the diode bridge. To ensurethat even small AC voltages can be measured, thepotentials at both inputs of Az must be absolutelyequal. Because of this, a small offset voltage is ap-plied to the non -inverting input via Rs.The sine wave generator is essentially a Wien bridgeoscillator, A2, whose frequency determining compo-

T2

06

0 M150 pA

D4

R16

9V

O

R8 02

C7mks

560n

1N414801

1N4148

C6

100 itT,.

9V

85423-1

2V

Semiconductors:

Di ...D6,D8,D9 = 1N4148D7 = zener diode 4V7/0,4 WT1,T2 = BC 547BIC/ = TL084

Miscellaneous:

M1 = moving coil meter, 50 pA (see text)PP3 (9 V) battery with dual miniature clipSPST on/off switch (optional - see text)PCB 84923

nents are P4 C3, and C4. To ensure stable operation,an active feedback loop takes part of the outputsignal of buffer amplifier A4, rectifies this (Dl, D2)and applies the consequent DC voltage to the in-verting input of A3 via buffer stages Ti and Tz Theoutput voltage of the generator section is 2 Vpp.The audio tester is best constructed on the printed

299

circuit board of figure 2. The meter can be almostany type from 50 µA to 1 mA. Note, however, thatthe value of 131 in figure 1 applies to a 50 µA instru-ment; for a different f.s.d., this value can bechanged inversely proportional. For instance, if theinstrument is a 500 µA type, Pi should be 2k5.The millivoltmeter is calibrated by tapping the refer-ence voltage with a divider of 820 Q in series with

100 kQ: the voltage at their junction will be 45 mV.Apply that voltage to the non -inverting ( +) inputof Al, and adjust 131 till the meter reads "45".The current consumption of the audio tester issome 10 mA, and it is therefore advisable to incor-porate an on/off switch. The frequency range ex-tends from 150 Hz to 20 kHz.

268 AUTOMATIC SWITCH OFF

If you are one of the many who frequently forget toswitch off their digital multimeter, this circuit is foryou.When this little circuit, which is intended to be in-corporated in the multimeter, is switched on, ca-pacitor Ci is connected to the +9 V line via Di.Since Ci is discharged, the gate of T3 is also at+9 V which causes T3 and T2 to conduct. Themeter is then switched on.Capacitor Ci slowly charges via R2. After about 2or 3 minutes, the potential at the gate of T3becomes too low to keep the FET in conduction.Transistor T2 then also switches off, and the batteryis disconnected from the multimeter.Transistor Ti ensures that when the multimeter isswitched off manually, capacitor Ci is discharged.

300

85474

BS250BS170

IrI1\

DGS


When the maltimeter on/off switch, S1, is opened,a base current will flow to the negative terminal ofCl via R, and R2. Transistor Ti then conducts anddischarges C,. The circuit is thus immediately readyfor use again. Without T,, there would have to bea delay of a few minutes before the circuit could beswitched on again.The circuit is best built on a small piece of veroboard and then fitted between the on/off switch andthe meter itself.

A final tip: T2 could be replaced by a Darlington,such as a BC516, in which case a I MS2 resistorwould have to be inserted in the connection to thedrain of Ta This .arrangement would have the ad-vantage that the BC516 is more easily obtainablethan the BS250, but the disadvantage of causing aslightly larger voltage drop across the circuit: 0.8 Vas compared with less than 0.1 V when a BS250 isused. The current in both cases is 10 mA.

269 CALIBRATION GENERATOR

A calibration generator is of particular use withmany older generation receivers, which have no, ora poor, frequency read-out. However, the RF sec-tion of these receivers is invariably far superior tothat of most modern models, and consequentlythere are still many of these òldies' in use.The circuit in the accompanying diagram providescalibration signals at multiples of 100 kHz and1 MHz, all of which are available simultaneously, sothat no switching is necessary.The output signal of the crystal oscillator, Ti, isdivided by 10 in ICI . Astable Ni operates at a fre-

R2 R3

quency of around 22 Hz, which is low enough toallow zero beat tuning even in SSB operation. The100 kHz harmonics sound (on AM) like a sort ofwoodpecker.Astable N3 operates at about 1.5 kHz and is gatedwith the 22 Hz signal. Consequently, the 1 MHzsignal appears for 22 ms as a carrier wave, which ismodulated with the 1.5 kHz signal during the next22 ms. This signal is also easily tuned for zero beat.The circuit is usable up to 30 MHz when CMOSdevices are used, and up to around 300 MHz withHCMOS ICs.

N1...N4=1C2.409313

N2

56p

86457

0Si

Min9V

sum

O

270 CRYSTAL TESTER

Many electronics hobbyists have crystals lyingabout, but don't know whether these .are still work-ing all right. The crystal tester described here will

quickly show whether a crystal can be used orshould be discarded.Transistor Ti and the crystal under test form an os-

301

cillator. Capacitors Ci and C2 form a voltage dividerin the oscillator circuit. If the crystal is in goodorder, the oscillator will work. Its output voltage isthen rectified and smoothed by Di and C4 respect-ively. The resulting direct voltage at the base of T2is sufficient to switch this transistor on, so that theLED lights.The circuit is suitable for use with crystals of a fre-quency between 100 kHz and 30 MHz. Currentconsumption is about 50 mA.

85457

9V

271 DIVIDER CASCADE

This circuit can be driven either with an analogue,or a digital precision 10 MHz signal for dividingdown to a number of commonly used timebaseperiods. The oscillator proposed in (1) is particularlysuitable for driving the present cascade, since it of-fers excellent stability thanks to the use of I

10 MHz quartz crystal fitted in an electronicallycontrolled oven. It should be noted, however, thatits output is digitally compatible, so that compo-nents R i-CI and R2 at the input of the circuitshown here can be omitted, i.e., Ni is driven direct.Where an analogue, sinusoidal, 10 MHz signal isused, the amplitude must be 750 mVpp. Evidently,R( -C: and R2 are then required to make the signaldigitally compatible for clocking IC3. The circuitdiagram shows that the cascade can be extended byadding further 74HC(T)390s and pairs of bistables.The Type 74HC(T)390 (IC3) holds two counters,the first of which divides by two (1QA), and by

N1 N 3 = '/z IC 1 = 74HC(T)04FF 1, FF 2 = IC2 = 7414C(T)73

:750mV

10 MHz / DCF

0100p :e:*

0

5 (IQc). Bistable FF is driven with the 1QD out-put, and outputs the :10 signal, which is also ap-plied to the CLK inputs of the second counter inIC3. This also divides by 2 and 5, while FF2 givesa total division factor of 100 in the first block of thecascade. The use of decade counters results in out-put periods commonly used for an oscilloscopetimebase. Counters and bistables may be added toobtain relatively long, yet accurately defined,periods for specific applications. The current con-sumption of the circuit as shown is about 12 mA.With two divider blocks added, the total currentdrain is expected to be approximately 25 mA, not36 mA, since HCMOS circuits require less power atlower clock frequencies.

Reference:(1) Oven -compensated oscillator. Elektor Elec-tronics, January 1986.

82

6

K FF1

CLR

a10

O5

J

7

FF

19

100 200 0

20A

1CKB 2CKB 12 - ICK1;

- 1CKA 2CKA 5 CKA

IC2 IC3 Nam

'Clto

IC3 IOC

74HC(T)

390

C210

1CLR 2CLR

T1CLR 2CLR

2 14

L0

0 0 0 0 0 0 0100ns 200ns Ins 5ns lOps 2045

soons 2115

87501

302

272 FAST VOLTAGE -CONTROLLED PULSEGENERATOR

Certain measuring and process control applicationsrequire pulse generator sections which are tooperate over a large frequency range and must,therefore, produce a signal with very low pulsewidth. It is for this reason that the proposed circuituses high-speed complementary MOS (HCMOS)type gates; the prototype typically produced an out-put pulse width of 20 ns over the frequency rangeof several hundred hertz to 25 MHz.The combination ICI -Ti is a voltage -controlled cur-rent source which discharges C2. The fast chargingof this capacitor is effected through the voltage at

0...3V

R2

the output of Schmitt trigger Ni-R3-Di. The lowerfrequency limit of the proposed circuit mainlydepends on the offset voltage of opamp ICI. Inorder to enable setting the lower frequency limit,Ts must be arranged so as not to draw any currentat an input voltage of OV; to this end, offset presetPi should be correctly adjusted. Finally, the outputpulse width may be widened by increasing thecapacitance of C2; this will not alter the attainablesweep range.

Literature: E Abbel, Electronic Design 18 (1984), pp270-271.

P1

25k

R1

5

C1sumsum

1n

8

ICI

CA3130

R3

5V

5V

N1 ,N2 = IC2 = 1/3 741-1C14

O

273 FAULTFINDING PROBE FOR µPs

Anyone who has ever tried to faultfind in a micro-processor system with a test probe will have ex-perienced the uselessness of it. This is because thesignals at the address, data, and control buses areconstantly - and rapidly - changing. This meansthat it is not just the signal level that is important,but also the instant the signals are present. Forfaultfinding properly, you need a logic analyser,which is capable of indicating several signals simul-taneously.If you have no logic analyser, the probe presented

here may provide the solution. Strictly speaking,this is nothing more than a bistable multivibrator(FF1). Data are simply read direct and cause Di tolight or stay out, depeflding on the state of FF1.The bistable only reads at the instant its clock input(pin 3) switches from low to high).The clock signal is thus the key for allmeasurements carried out with the probe and thatmeans it must be chosen with some care for everytest. Suppose you have to check whether a certainportion of memory is all right. The CE signal in the

303

memory is then connected to the QUAL input ofthe probe. Switch S4 must be closed, because CE isactive low. The probe can then only read data dur-ing a CE of the RAM under test. The CLK inputof the probe is connected to the RD signal of thememory. Reading must then be carried out duringthe trailing, i.e., the positive -going, edge. Switch Simust, therefore, be closed. Reading is effected by, forinstance, a PEEK command in BASIC. Diode Diwill then light in accordance with the signalemanating from the RAM during this process.Be careful that this BASIC is not used by the RAMsection being tested, because then there will bemore than one read process and the probe will onlyretain the last of these. There is no easy solution inthat case, but often it will be possible with the aid

DATA ,Cr

CLK

QUAL

5V

0IC2

of a monitor to make the microcomputer executeonly one command in machine language.To keep the probe small, DIL (dual in line) switchesare used in the Si . .S4 positions. Note that onlySi or Sz and S3 or S4 should be closed simul-taneously at any one time.LS type ICs may be used, but as these put a rela-tively high load on the circuit during tests, HCTtypes are better. These are fully compatible with theIS types but have high impedance inputs. HC typesshould only be used where systems are already ex-ecuted entirely in CMOS; the supply voltage canthen be higher than 5 V.Current consumption of the circuit is small: 10 mAfor the LED and and 5 mA for the ICs (if these areTTL).

N1 . . . N4 = IC1 = 74HCTOO74LS0074HC00

F F 1 = IC2 = 74HCT7474LS747414C74

C1I DATA

100n

CE

RD

85447

274 FUNCTION GENERATOR

This is a downright simple design for an AF func-tion generator that supplies a rectangular andtriangular signal, and can be fed from a single 9 Vsupply. The signal generator proper isa Type TLC272 dual CMOS opamp from Texas In-struments. This chip is remarkable for its low cur-rent consumption and wide operating range.The circuit is essentially composed of two func-tional parts. Opamp Al is connected to function asa Schmitt -trigger whose toggle point is set to 4.5 V,

while A2 is an integrator that converts the rec-tangular signal from At into a triangular waveform.The oscillation frequency of the circuit is fixed sole-ly by the ratio RC and can be calculated fromfo = R2/4RR,C. Resistor R may be replaced by thecombination of the 10K resistor and 100K poten-tiometer as shown to effect continuous adjustmentof the output frequency within the AF signal band.The generator should not be terminated in less than10K.

304

MI00k

R

A = 405Et= 4V5C. 4V5

87421-1

275 GHZ PRESCALER

In the 1.2 GHz input stage (February 1985) for themicroprocessor -controlled frequency counter, weused an SB8755 prescaler in the IC7 position. ThisIC, which divides the 100 . . A200 MHz signal atinput C by 512, is perfect for the purpose, but is

5V

rather expensive. Just recently, another prescaler,which is much cheaper and more sensitive, hascome onto the market: the U665B from Telefunken.The U665B is a 1024 prescaler with integral pre-amplifier. Its sensitivity is better than 10 mVrms forfrequencies between 80 MHz and 900 MHz. It isfully usable up to 1200 MHz, but its sensitivitydrops to about 30...40 mVrms at that frequency.To fit the U665B onto the PCB, first remove exist-ing IC7, IC8, and P3. No other components shouldbe removed because, although they may looksuperfluous, they are needed for the interconnec-tion between the component and track sides of theboard.The new IC is fitted so that its pin 1 coincides withpin 8 of the previous IC7. Next, solder capacitors

85513-1

C101

C104 C103 C102

305

C1 0 1, Cl 0 2, C103, and C104 direct to the relevant pinsof the new IC and to the earth plane. Then, solderpins 4 and 6 direct to the earth plane and place awire link between pin 8 of the U665B and the holewhere pin 1 of IC7 used to be (see drawing). Finally,solder a wire link between the holes where pins 1and 11 of 1C8 used to be.So much for the hardware; now something about

the software. The U665B divides the input fre-quency by 1024, while IC7 + IC8 divided by only512. This difference means that one byte in theEPROM must be altered: address $627 reads $09;this should be amended to $0A.

276 INSTRUMENTATION AMPLIFIER

This instrumentation amplifier was originally de-signed for the serial digitizer described in (1), butshould be suitable for many other applications also.The amplifier makes it possible to use a relativelylong, interference -free, connection between thetransducer or sensor and the digitizer input.The theoretical basis for the circuit is summarizedin the accompanying Table. It is seen that the com-mon mode rejection of the amplifier serves to sup-press interference. In practice, however, the lowdrive margins of the inputs and outputs of theopamps impose some limitations. Both suggestedtypes have PNP input transistors capable of hand-ling input voltages between 0 and Ub-1.5 V. Theoutput of the OP -220 can deliver voltages between0 and Ub-1 V, that of the LM358 between 0 andUb-1.5 V.The current consumption of the opamps is low atabout 150 µA for the OP -220, and I mA for theLM358, while the slew rate is about 0.04 V/µs and0.4 V/ms respectively. For optimum accuracy it is

recommended to use high stability (I%) resistors inpositions R1 -R5 inclusive.

Ro

Ub

fl

Uid

/S I1

Al, A2 =ICI = OP220 , LM358R1 = R2 = 2x RXR3 =114 =115 = Rx

PMI Application

A2

j 41

87064

Reference:(ii Universal peripheral equipment (2): SerialDigitizer. Elektor Electronics, September 1986 p. 23ff.

Micropower instrumentation amplifier

Consider an input UCM-1/2Ud at the - input of the circuit, and UCM+ 1/2 Ud atthe + input. This corresponds to a common -mode input UCM, and adifferential input Ud. The currents at the inverting input of each opamp canbe summed to form two equations:

(Ub-UCM+ 1/2Ud) (1/R1)+(lid/Ro)+(lit+ 1/2Ud) (1/R3) =(UCM-Y2Ud) (1/R2) (1)

(Ui =UCm-1/2Ud) (1/R4)+Uo-Y2Ud) (1/R5) = Ud/Ro

When Re =R2= 2R3=2R4 = 2R5= 2Rx, (1) and (2) can be combined to

Uo= 2(1 +Rx/RolUd + Y2 Ub

which shows that the common mode input (UCM) has been rejected. Thedifferential gain, Ax, of the circuit is therefore

Ax = 2+ (2Rx/Ro)

and is adjustable between 0 and 1,000 by varying Ro.

(2)

306

277 LINE BAR GENERATOR

The video signal transmitted by most TV broadcaststations is rather complex. For most tests and ex-periments, however, a fairly simple signal will suf-fice. The circuit presented here provides a small, in-expensive source of line synchronizing pulses andline bar.The first of the three timers in the diagram provides4.7 pis sync pulses. It is arranged as an astable multi -vibrator with a period of 64 his. The rising (here:negative -going) edge of the sync pulse triggers a sec-ond timer. The width of the output pulse of thistimer determines the position of the line bar. Theline bar proper is provided by the third timer. To ob-tain a usable video signal, the sync and bar signalsmust be added, which takes place in Ra-R3-R6. Theresistor network is followed by a buffer that ensuresan output impedance of 75 ohms. The unit can,

P1

therefore, be connected direct to a standard video in-put. The sync and bar signals occupy 40 per centand 60 per cent of the composite signal respectively.Calibration is carried out by connecting the unit toa monitor or, via a modulator, to a normal TV re-ceiver. Presets Pi, P2, and P3 are set to the centre oftheir travel. Turn Pi to obtain a still picture. If thesync pulse is too wide, it will be visible at the left-hand side of the picture. The pulse may be narrowedwith the aid of P2, after which P, may need a smallre -adjustment.Where an oscilloscope is available, P2 can initiallybe set to obtain 4.7 pis pulses at the output (pin 3)of IC,. Then, the total period is set to 64 pis with theaid of PiThe line bar is centred with P3: as its width is fixed,this completes the calibration.

2

C

20k 6

4,7p,

C>_

C2

IC1

C8.44Ton

20p

PglOk4 4

IC2a

3 7

C3 C4MOM

IC1 = 555IC2a,IC2b = 556

TI n

82O

12

(;) C5

56p

06mim7. 7,

IC2b

117

C

85

C112.

0 12V/80 mA

11)08

99

c_1111. c>I L "p

0 12V

BC547

Eno740

86513-9

- - -r60%

40%

278 LOGARITHMIC SWEEP GENERATOR

This circuit outputs a logarithmic signal for drivingthe VCO input on the Elektor Electronics FunctionGenerator (11, but can be used for other generatorsas well. Usually, the exponential function is derivedfrom the (temperature -sensitive) B -E junction in atransistor, but this design uses a simple R -C net-work and an opamp to generate the logarithmicsweep.

With reference to the circuit diagram, Us is appliedto the generator's VCO input, while Usync is usedfor triggering an oscilloscope on the positive signaledge. Contrary to the Elektor Electronics SweepGenerator in (2), the timebase of the scope is usedfor the horizontal deflection, so that the horizontal(frequency) axis has a logarithmic scale. The sweeprange is 1:100 (Uvco =0.1-10 V). Opamp IC2 is di -

307

mensioned for a gain of 2 (R3 -R4) and generates answeep voltage, Us, with the aid of networkP3/134-R1-C1/C2:

Us = Ulexp(t/RiC1) when111:5_Uss.U2.

When Us reaches level U2, the bistable in 1C3 isreset. Capacitor Cl (or C2) is then discharged via R2and the discharge input on the 555 (or 7555) untilUs = U1, causing 1C3 to be set and the next sweepperiod to commence. The output of the monostablesupplies the trigger signal for the oscilloscope.To adjust the circuit, set the function generator to100 Hz; external frequency. Connect the VCO in-put to the wiper of 131 (do not forget the ground con-nection), and adjust this preset for an output fre-quency of 100 Hz. Next, connect the VCO input tothe wiper of P2, and adjust this preset for an outputfrequency of 10 kHz. Proceed with connecting theoscilloscope, set to 10 ms/div., external trigger. Thesweep voltage is applied to the Y input, and the ver-tical sensitivity is adjusted until the maximum ex-cursion of Us reaches the top of the display. Set S,to position A (sweep 0.1 s), and adjust P3 until thepeak of the exponential voltage is displayed in thetop right-hand corner. This is repeated with Si setto position B (sweep 1 s), and the scope set to100 mV/div. (adjust 134). This completes the adjust-ment procedure, and Us can be connected to theVCO input. The current consumption of the circuitis less than 25 mA or 15 mA with a 555 or a 7555fitted, respectively.

References:

Function generator. Elektor Electronics,December 1984.

P3 UB

15V

UswEEP

AA® ""-C)IC2

U2:

U1:

4

R 07 DISCHARGE

TRESHOLD

2

IC3(7)555

TRIGGER

CONT7VOL

0Us15V

Al, A2= 1/2 IC2=1.1.1339

tSWEEP

U SWEEP

UsYNC

C5 OTDP ,0P 25V

UsyNC

0

87950

C.)

(2) Audio sweep generator. Elektor Electronics,November 1985.

279 LOW CURRENT AMMETER

This 7 -range ammeter measures currents between afew pA to 100 µA without using precision resistorswith very high values. The circuit is set up arounda current mirror Tra-Tib. The input current is mir-rorred in this transistor pair, and the currentthrough Tlb is greater than the input current by afactorset with Si. Meter MI is a 100 µA fsd type fordisplaying the measured value. The effective seriesvoltage drop at the input terminals of the instru-ment is only 500 µV because the voltage across the

inputs of A) is forced to virtually nought.The accuracy of the ammeter depends mainly onthe components used. Depending on the requiredprecision, certain components may be replaced bytypes with a better specification. The Type LF411opamp used in the Ai position, for example, can bereplaced with the Type OP -41 to achieve a tenfoldreduction in the input bias current, and hence animprovement in the final accuracy of the instru-ment. Transistor pair Tia-Tib may be replaced by aType MAT -02, and the voltage reference set up with

308

11

Tl...T4 = 1C1 = CA3046Al = 1C2 = LF41 ,0P41R5...R10 = 511 , 1%D1...132= 1N4148

T3 -T4 by a Type LM313. These high -quality partsshould ensure an accuracy of 1% over most of therange. The meter is calibrated in the 1µA range.Preset Pi is adjusted for full scale deflection of MIat an input current of I µA.When it is intended to make a printed circuit boardfor the pico ammeter, it should be borne in mindthat two 2.5 cm long, parallel running, coppertracks spaced 1.25 mm and etched on a high qualityepoxy/glass carrier represent a leakage resistance of about100 GQ. This corresponds to a leakage current of150 pA at a voltage difference of 15 V. Evidently,the PCB for the present ammeter should be

S1 11

a 100 pA

b mAA

C 10 nAd 100 pp

e 1 pA

f 10pA

9 100 pA

87507

100n

thoroughly cleaned to rule out leakage currentthrough residual moisture or resin. Also note thattheinsulation of standard test leads is likely to makereliable measuring of currents smaller than I pA im-possible. The only way to overcome this diffi-culty is to use dry air or FTFE (Teflon).

Source: PMI Linear and Conversion ApplicationsHandbook.

280 MEASURING WITH THE BBC MICRO

The BBC micro, one of the best value -for -moneycomputers on the market, can be used for a varietyof applications thanks to the various interfaces pro-vided as standard. The four analogue inputs, eachwith a resolution of 10 bits, make it particularlysuitable for measuring all kinds of processes.There is unfortunately one drawback: the ratherpoor reference voltage associated with the analogueinputs. That voltage is obtained from three normaldiodes connected in series. The alternative de-scribed here has been in use in our BBC micro forsome time.Diodes D6 . . . D8 in the diagram provide a referencevoltage of 1.8 V, which is fine for use with a joystickinterface, but will not do where absolute values areto be measured. The three diodes are, therefore, re-

placed by one zener diode, a 2.5 V type LM336Z.This diode deviates no more than 1.8 mV over thetemperature range of 0 . . .70 °C; its long-term stab-ility is better than 20 p.p.m. at 25 °C. Its internal re-sistance is 0.4 2, which makes it ideal for our pur-pose. Moreover, it is easily fitted into the microwithout the need for any alterations other than theremoval of D6 . . . D8. The micro remains, of course,fully compatible with existing software.Cut off the adjust terminal from the LM336Z, andunsolder D6 . . D8 from the computer. Solder theanode and cathode of the zener to the cathode con-nection of D6 and the anode connection of D8 re-spectively. A good -quality small soldering iron is in-dispensable here!

309

EOC

R171100R

NRDSN

2415

16

17

18

19

2021

22

7

26

R71

280.4737-C..220

IN HZ EP"2

5V

D8

D7

3xD6 1N4148

RDWRD7

IC73pPD7002 I

VREFC25 C2733n F 10F

1\5C°IDOAlA0EOCCS

ClCH3

2

1CHO9AGND

8

XI GND31

12

,5V - 5V .411-0 '.R71 TVS T B

2 K 7

OV

p2t1 fir, 00 y 418--+03

1N4148xD8

7

9

10

0 :11 PADDLE04

, o:05

I 0.418-;-00 V.48--:-0 6

: 0'07

AND LIGHT12 PEN

CONNECTORSK 63

4

5

:08/

LM 336Z

/1 \adj. +

- adj.

85494

281 METER AMPLIFIER

A meter amplifier is intended for use between asensor or other measuring device, such as a probe,and the indicator. It is characterized by a high inputimpedance, typically 1 MQ, and a differential input.A differential input ensures that the output signalcannot be affected by hum or noise on the meterleads.The input signals are buffered by differentialamplifiers Ai and A2. The 22 pF capacitors in theCI and C2 positions obviate any tendency tooscillate. The output of opamp A3 is a function ofthe difference between the two input signals.Opamp A4 serves to compensate for any offset andalso to set the amplification at exactly 1. The band-width of the circuit as shown is not less than100 kHz, and the phase shift is 0°.As already mentioned, the amplifier may be used

with any sensor, for instance, in computer control ofthe central heating, or to monitor the ambient tem-perature in rooms. It can also be used with amultimeter or oscilloscope.The peak -to -peak level of the input signal shouldnot exceed about 80 per cent of the supply voltage.Current consumption is not greater than 25 mA ata supply voltage of ± 18 V.Calibrate the unit by adjusting P2 under no -signalconditions for zero output, and setting the amplifi-cation to exactly 1 with Pt If you aim at perfection,use 1 per cent resistors.

310

(+)

R1

R3

( )

R2

5 . .. 18 V P2Rut 2k5 R15

R11

R8

A1,A2 = IC1 = TL 082, LM 833A3,A4 = IC2 = TL 082, LM 833

5 .. . 18 V e

5 ... 18 V e

Calm j)4100n 100n

10071007'

R16

IC1 IC2

. 18 V

85481

O

282 METERING SELECTOR

When just one meter is used to measure the voltageof three different sources, it is, of course, possible touse a three position rotary switch to select any oneof the sources. However, care must be taken here,because the switch must break before make, other-wise two sources are interconnected and this is nor-mally highly undesirable.Any electronic equivalent of the rotary switchmust, of course, also break before make. Unfortu-nately, transistors have the property of switching onmuch faster than switching off. For example, a well -driven BC 547 takes a couple of 1.4s to switch off, butfar less than that to switch on.The present circuit circumnavigates these potentialtroubles by using the output level as criterion,whereby a 4028 serves as the referee. The 4028 is aone -of -ten active high decoder which drives one ofthree transistors, T, . . . Ta Let us assume that T1 ison: its collector voltage is low, and so is input A ofthe 4028. The other two collectors are high, and soare inputs B and C of the decoder. The 4028

therefore sees binary code 6 (110) at its input andthis causes pin 6 to go logic high, so that T1 isdriven hard.When in this condition another key, for instance, Sz,is pressed, a wrong code, i.e., 4 (100), ensues. Out-put 4 of the 4028 is, however, not connected, T1switches off, but T2 is not yet driven. Only after T1has actually switched off, and its collector goeshigh, does 5 (101) arise at the input of the 4028: T2will then be driven.In practice, the voltage at the collector may be usedto control a CMOS switch that arranges the changeover of the meter or the sound channel. It is alsopossible to replace the collector resistor by a suitablerelay, but this would, of course, introduce evenlonger delay times (of the order of milliseconds). Inthat case, the feedback to the input must be effectedby a separate contact of the relay, but there is then,of course, absolute certainty that switching is cor-rect!Another variant is including a resistance in each

311

feedback loop and shunting each switch contact by while that of the transistors depends on the value ofa capacitor. This RC network will ensure a the collector resistors. With values as shown, itreasonable delay during the change over. amounts to 18 mA for a supply voltage of 10 V.Current consumption of the 4028 is small (CMOS!),

E::35..15 V

0Cy=00n

16

10

13

12

S1 S3

r? klS2

0 0

0 03A

B IC14028 05

C

06

1

T3

BC 547BT2

BC 547BT1

R6

D3

1,41

D2

5...15V

D1

BC 547B 085453

283 NOISE GENERATOR

Noise is normally defined as unwanted electricalsignals spread over a relatively flat, wide frequencyspectrum. In most equipment, great care is taken toreduce the amount of noise to a minimum, resultingin a low noise factor.None the less, noise is useful for measuring thebehaviour of a circuit under varying input con-ditions. A noise generator is used, for instance, formeasurements on coaxial cables, microwave links,and RTTY (radio teletype) and CW (continuouswave = radio telegraphy) decoders. The present cir-cuit may also be used to imitate the sound of wind,mosquitoes, bees, and other buzzing insects.The circuit consists of a relaxation oscillator, IC,,which is provided with positive and negative feed-back via 131-R2 and P3 -P2 -R3 -C1 respectively. Zenerdiode Di functions as noise source. The amplifi-cation of the noise is determined by the setting of *see text

312

P3 (coarse) and P2 (fine). The setting of 131 deter-mines the noise bandwidth: a small effective valueresults in a narrow band, and increasing valuescause wider bands.Due to the negative feedback, the opamp also func-tions as a low-pass filter: a small feedback factorresults in a low roll -off frequency. The pass band of

the opamp also depends on the value of C2: a valueof 47 n causes a noise similar to the buzzing of amosquito or bee. Diodes D2 and D3 serve as inputlimiters. The output level of the generator can beadjusted with P4Current consumption is not greater than 10 mA at12 V.

284 OPAMP TESTER

N9)

15V/00 mA

L-1.9C21 10g (12,1

25 V

IC1 IC2 IC3

11 11

25 VoT

15V e100 mA

C3

All types of operational amplifier can be functional-ly checked with the tester proposed here.The principle of the tester is quite simple: atriangular voltage is applied to the inverting (-) in-put of the specimen. This voltage is, of course, in-verted. If then the inverted and the original

15V

15V

Al A3 = %IC1 = TL 084 (LM 324)A4 ... A7 = IC2* =

A8 = IC3* =opamps to be tested

85430-1

triangular voltage are added, the result should bezero. Any deviations from this are taken as a mal-function which is indicated by one of two light -emitting diodes (LEDs). The tester has, of course, aself test facility so that the error -free operation of itcan be readily ascertained.

313

Opamps Al and A2 form a triangular pulse gener-ator. Opamp Ai operates as an integrator: capacitorCi is charged, and as soon as the voltage across itreaches the threshold value of Schmitt trigger A2,resistor R4 is connected to earth, and Ci dischargesuntil the voltage across it reaches the secondthreshold of A2, when the process repeats itself.Opamp A3 functions as the summing stage whoseoutput is fed to two transistors that drive LEDs.The specimens are connected as inverters in eitherpositions Api Ap4 or Ap5. In the design it wasassumed that the most frequently encounteredopamps are contained in a 14 -pin DIL housing (as,for instance, the TL 084 used for Ai . . A3), or inan 8 -pin DIL package (such as the LM 355 orLM 387). For different packages, the specimen con-nections in figure 1 should be modified accordingly.When a specimen is defect, the output of A3 con-sists of a triangular voltage superimposed on the(DC) offset. This is sufficient to bias the drive tran-

sistors and one or both LEDs flash in rhythm withthe triangular voltage. The frequency of that signalis about 10 Hz, and this can be altered to some ex-tent by changing the value of R4 and/orIt is clear that the voltage at the output of A3 mustbe greater than ± 0.6 V, otherwise the bias for thetransistors is too small. Preset Pi should thereforebe adjusted so that the LEDs just do not light whenan opamp that is known to work correctly is in-serted in the relevant specimen position.The self test function is easily checked: when P2 isturned from one extreme of its travel to the other,first one LED, then both, and finally the other LEDshould light.In positions 1. . .4 of switch Si, the four opampscontained in, say, a TL 084 can be tested sequen-tially; in position 5, the single opamp contained in,say, an LM 355; and position 6 is the self test set-ting.

285 POCKET FREQUENCY METER

A1,82 =ICS = L14393 (1/2 L/4339)181,102 = 4518

BF256A LM336

OS IM

This easy to construct circuit meets the demand fora simple, yet versatile battery operated frequencymeter which can interpret signals with a minimumrms voltage of 10 mV and a maximum frequency of100 kHz. The quiescent current consumption of themeter circuit is only 4 mA, which ensures a long life

86497-1

for the 9 V battery. Also of interest is the fact thatthe circuit continues to work normally with batteryvoltages down to about 5 V. The meter input is pro-tected up to 250 V AC.From the circuit diagram it is seen that the inputsignal is applied to the gate of Ti via RI and C2. Ci

314

is an additional speed-up capacitor to improve theresponse at higher frequencies, while anti -parallelconnected diodes Di and D2 protect the FET gatefrom high voltage surges. Ti functions as a buffer,preceding Schmitt trigger Ai, which has been di-mensioned for a relatively low hysteresis of about18 mV to prevent the overall sensitivity being toostrongly degraded. The output of A i is fed direct todivide -by -two counter ICia, which is followed bythree cascaded divide -by -ten IC sections. Sz selectsthe divisor and hence the relevant frequency range.Whatever range is selected, a frequency of 50 Hz atthe pole of S2 corresponds to a full scale reading onmoving coil meter M.The signal at the pole of Sz is used to trigger themonostable built around the Type 7555 low -powerprecision timer. The correct operation of this circuitsection can only be achieved if the time period ofthe monostable is less than half that of the full scalefrequency, i.e. 1/2(1/50)s = 10 ms. Therefore, amonostable time of 8 ms is used in the proposedconfiguration.The output signal from IC3 has a duty factor

which is proportional to the input signal frequency.The pulses from 1C3 are levelled at 2.5 Vpp by IC4,before being integrated by Riz and Cs to produce adirect voltage which is proportional to the input fre-quency. The circuit around Az and T2 is a simplevoltage -to -current converter with the 100 µA mov-ing coil meter connected in the collector supply lineto T2. C9 may be added to stabilize the read-out atthe lower end of the scale.Though a Type LM393 opamp has been used, theless expensive Type LM339 also works all right, pro-vided the inputs of the unused opamps contained inthis chip are tied to the positive supply rail tominimize their power consumption.The frequency meter is so sensitive that merelytouching the input terminal with a finger causes themeter to read the mains frequency. This is, inciden-tally, a convenient method of calibration, since Pimay be set to give a reading in accordance with thelocal mains frequency, which is usually stablewithin 1%.

286 PROGRAMMABLE BAUD -RATE GENERATOR

Only some computers, e.g., the Samson 65, enableyou to reprogram the ACIA (asynchronous com-munications interface adapter), or whatever yourserial interface may be, if you want to connect aprinter and a modem to your computer. With mostother micros, you have to use an additional circuitlike the one proposed here.The circuit is based on a presettable, synchronousdown counter, a CMOS IC type 40103. AnotherCMOS IC, type 4060B, serves as a crystal con-trolled clock generator. The crystal frequency, fx, is2.4576 MHz, while the clock, fc, is 153.6 kHz. Theoutput frequency, fo. of the generator is determinedfromfo = [153.6/(N + 1)] kHzwhere N is the decimal equivalent of the numberthat is input to the JO . . J7 terminals of the 40103(see table).The number N is provided by the computer andfrom there written into, and stored by, the 74LS374.The table gives various baud rates (also for RTTY- radio teletype) and the corresponding decimaland hexadecimal numbers. If you want to shift thebaud range upwards, select a higher value output of

baud -rate N (dec) N (hex)

4800 1 01

2400 3 031200 7 07

600 15 OF

300 31 1F

150 63 3F

110 86 56100 95 5F

75 127 7F

57 167 A750 191 BF

45.45 210 D2

the 4060B: for instance, output Q5 for a maximumrate of 9600, output Q6 for a maximum rate of19200, and so on. As for every one of these stepsthe output frequency is doubled, the relevant valueof fo must be used in the formula given above; i.e.,307.2 kHz when Q5 is used. 614.4 kHz when Q6 isused, and so on.The address decoder in the circuit diagram is ar-ranged for a Z80 computer, as can readily be seen

315

5Vfrom the control signals, but this is purely taken asan example. The signal from the data bus is appliedto the 74LS374 at the leading edge of the STROBEpulse at the output of the decoder. The articles ad-dress decoding and memory timing in the Januaryand February 1984 issues of Elektor Electronics re-spectively contain all the necessary information forthe design of an address decoder for any type ofcomputer.The address decoder in the diagram is shown for thedecoding of the hexadecimal address F0. Many ver-sions of BASIC on Z80 computers permit the pro-gramming of the baud -rate generator with the in-structionOUT 240, N

5V

B5605

DI

DE

D3

D4

D5

06

D7

MR0

AO

AI

A2

A3

A4

A6

A6

C2-47

20

00 2 r* IC1 025 6 33 IC2O 8 03 038 33 CO/204.O 13 D4 74 0412

CF: 14 05 LS O5 15 11

O 17 374 0616 12t

O 18 0, 0)19 13 a/C0

0

0

O

00

0O

O

0O

DO COO " APE 04DI 01 8 8 31 CO DI

40 "--°

c1.6 CLK m610

04

IC340608

1061

XI

- C4

245).6 kHz 6-1

N1,N2 IC4 = 40728C3

N3 ='h 105 = 74LS20 7331/ 7317

5 V

287 RECTANGULAR PULSE GENERATOR

The excellent properties of counter/divider IC type properties is the provision of divide ratios anywhere4059 have, so far, not been given the prominence in between 3 and 15 999 depending on the logic levelElektor Electronics they deserve. One of these at inputs Ii ..uis and the setting of switches

TB TAw1D2 131

11 I I

kkk124

N2 = 1/2 IC1 = 4093FF1 = 1/2 IC3 = 4013

S19I5181 31711

ÀAA \7

0icl CL:3

1C3 ""loop16V

R16

8

910

15

16

17

18

19

R9 R8

21

22

J16

IC2

K

K

4059 Kc

14 LE

I

14

13

11

23

5

3

R1

12

R1 ... R16 - 16 100 k smi

2R17 R18 R19

5 .. 15 V

0

FF1 2CLK 5 D-

R S

4 8

65511.1O

316

S17. .S19.

The 4059 is clocked by a relaxation oscillator, 111-N2, which could have been a crystal -controlled typeinstead of the one shown in figure 1.The dual -D bistable type 4013 at the output isessential because the width of the pulses at pin 23of the 4059 is comparable to the clock frequency.The bistable ensures that the pulses emanatingfrom pin 23 are reshaped into rectangular form. TheQ output of the bistable is, of course, half the fre-quency of the wave train at pin 23 of the divider.Inputs .1.1...J16 of IC2 are divided into groups offour. The binary information at these inputs iscalled TA . . To. Inputs .11...J4 are further sub-divided into Di and D2. In total there are, therefore,five data inputs, of which the smallest, D2, is onlyone bit wide. Furthermore, the 4059 has three modecontrol inputs, Ka...Kc, whose compositionresults in a factor K as shown in table 1. If K =10,input Di becomes four bits wide: .14 then forms partof Di! The divisor, n, is then calculated from:n = 103TD + 102Tc + 10Ts + DiIn all other cases,n =K(10132 + 102TD +10Tc + TO+ Di

CLK

J1 J4 J5 J8 J9 J12 J13 ----J18

0-D1

TA

D2

-0 TB TC

ModeControl

0Ka K5 Kc

0 0 0

TD

4059

85511.2

The 4059 can be programmed by computer, byhand, or with the aid of an up -or -down counter.The generator can be used in electrophonics, inmeasuring techniques, as timer, and even as a digitalphase -locked loop frequency synthesizer in FMtuners.The circuit operates from a 4 .. .15 V supply andcurrent consumption is small.

288 RMS-TO-DC CONVERTER

For some obscure reason, establishing the root -mean -square (rms) value of an alternating voltageseems to be among the least familiar procedures formany an electronics hobbyist; measuring the alter-nating voltage may be easy, but deciding on the rel-evant unit expressing quantity - rms, mean, orpeak -to -peak value - is quite another matter.Since the rms value of an alternating voltage is themost frequently used of the above mentioned three,some convenient means of obtaining that valuewithout calculations may be of interest in practicalmeasuring techniques.The rms value of an alternating voltage U across aresistor R equals the direct voltage causing the samedissipation level in R.Example: a 50% duty factor, I Vpp rectangularvoltage across a resistor R. Find the rms level of thisvoltage.The mean dissipation in R, caused by this periodicsignal equalsrz (Upp)2/R = 1/(2R) [W]

The direct voltage causing the same dissipation hasa level of%1/2 V, since P = (Y2 I/2)2/R = 1/(2R)[W].

This is also the conversion factor for obtaining therms value from the peak -to -peak value, sinceUrms = IMUpp2 = 1/2 Upp0.71Upp[V]therefore Upp 1.4IUrms in this example.Although moving coil meters measure the meanvalue of the rectified (pulsating) input voltage, theyare calibrated in terms of rms voltages. Therefore,the calibration is only valid for sinusoidal voltages.The proposed rms-to-DC converter is a relatively

317

simple circuit as it incorporates a dedicated chip, theType AD536 by Analog Devices. Alternatingvoltages applied to input terminal 1 are propor-tionally converted into a direct output current,which causes a direct output voltage across an inter-nal 25 kohms precision resistor; a buffer opamp out-puts the direct voltage equivalent (i.e. rms value) ofthe input alternating voltage.ICI functions as an input buffer in view of the rela-tively low input impedance of the rms converterchip. The maximum permissible peak -to -peak valueof the input voltage to the Type AD536 equals thesymmetrical supply voltage level; Di and D2 havebeen added to protect ICI against accepting inputvoltage levels in excess of the ± supply voltage. S2

functions as a x 1/ x 10 input attenuation selectorto enable high voltage measurements; the functionof Si is to block any DC components in the inputsignal to the converter. It is useful to realize that therms value of a composite (AC + DC) signal iscalculated fromUrrns = j/UDC'-l- UAC2.Preset PI should be turned to obtain 0 V withrespect to ground at terminal 6 of IC2 with no in-put signal applied and S2 set to the x I position.The converter achieves an accuracy of 1% for inputvoltage levels lower than 100 mV and input fre-quencies up to about 6 kHz. For signals up to 1 V,the bandwidth is expected to be of the order of40 kHz, while 100 kHz may be attained with inputvoltages above the 1 V level. Current consumptionof the circuit is about 5 mA.

C2

k40

a a

Parts list

Resistors:

R1 =1M;1%R2= 10k;1 %R3 = 100Q;1%134Rs = 10kP1 =100k preset

7/%14.,

Capacitors:

Ci =4.7/.4;25 V electrolyticC2 =1µ;MKTC3;C4= 100n

Semiconductors:Di;D2=1N4148ICi =CA3140IC2 =AD536J

Miscellaneous:

Si = miniature switchS2= toggle switchPCB Type 86462

289 SERVO -MOTOR TESTER

This circuit is of interest to two categories ofreaders; first, model aircraft/boat/train enthusiastsobjecting to having to leave their radio controltransmitter switched on for extended periods inorder that a servo -motor and associated mechanicalparts can be made to function as desired; secondly,constructors of computer -controlled robots incor-porating servo -motors. The latter field of interest isa typical combination of mechanics and electronicsplus software, and it is sometimes urgently requiredto be able to keep them apart as the specific partsof the robot have been prepared for testing, which,arguably, should be possible to do without having towrite special programs to this end on the computer.The proposed servo -motor test unit is, as can be

318

seen from the circuit diagram, a downright simpledesign based on the use of Type 555 or 7555 timerchips connected in a cascade arrangement whichcan be expanded further to drive more than two ser-vos at a time, if desired; it is readily seen that thesecond and third stages of the circuit are identical.The first timer, ICI, serves as an astablemultivibrator whose output pulse period time is de-termined with T= 0.693(R R2)C. The indicatedvalues for the relevant timing parts therefore pro-vide a pulse period time of about 20 ms at pin 3 ofICI. The rising edge of this square wave triggersmonostable IC2, whose output pulse width may beset with Pi; the given series connection of Pt andR4 ensures a large enough pulse width range for


+4,8V/10mA

0

O

R51W

00 00

100 PA

R2

2

C1

6

l00n1..2ms

1C1 (7)555

R3

P1

R4

25klin.

700n20ms

10 n

C4

most types of servo -motor, which typically requireactivation pulse widths of the order of 1...2 ms.The second servo control stage is identical to theset-up around IC2, and up to eight triggered 555stages, each with its own pulse width control, maybe cascaded in this way.It is suggested to fit each of the servo control poten-tiometers with a simple scale in order to have a rela-tive indication regarding the motor's lateral orangular position.The ammeter in series with the positive supply line

n

1C2 (7)555

Cs 5mem

loon

P2

R6

C6

25 k

2

1C3 (7)555

3

ton

C7

100 n

C8

700 n

86481-1

offers an indication about the total current con-sumption of the servos, and it is thus readilydetected when one or more have got stuck in theirmovement. The test circuit itself does not con-tribute much to the total current consumption indi-cated on the meter; some 3 mA is required for eachType 555 timer in the row, while the use of lowpower Type 7555 equivalents should reduce this fig-ure even further. It is, therefore, perfectly feasible tomake the tester into a portable, battery -operatedunit, powered by four penlight type NiCd batteries.

290 SIMPLE SWEEP GENERATOR

The sweep generator is an indispensable piece ofmeasuring equipment for testing the frequencyresponse of AF amplifiers, filters, and loudspeakersystems. At the heart of this design is the well-known Type XR2206 function generator chip fromEXAR. It is seen to the right on the circuitdiagram, in a standard application with 3 capacitorsand a rotary switch for selecting the frequencyrange, and a potentiometer, P5, for adjusting theamplitude of the output signal. The signal fre-quency is a function of the current drawn from pin7 on the XR2206:

I.= 320I/C IFIz1

where I is in milli -amperes, and C is in micro -farads.

It should be noted that pin 7 is internally kept at3 V, which is available at pin 10 also.The left-hand part of the circuit comprises thesawtooth generator, ICI, and a buffer, IC2. Theformer is set up as an integrator, whose sweepperiod depends on the voltage at terminal C. Poten-tiometer P2 enables setting the sweep period be-tween 0.01 and 10 seconds; the maximum durationis adjusted with P4. The sawtooth voltage at pin 6of IC, has an amplitude of 5 Vpp, and can be usedto drive the horizontal deflection 1X1 input of an os-cilloscope via terminal K. The amplitude of thesawtooth voltage is determined by the zener voltageof D1 and the base -emitter voltage of Tz, which isbriefly turned off when the output of IC, exceeds

319

D1

VS

01 s

BS250

G T3_AV

RR

5 V. The collector of this transistor is then pulled toground via R3, so that Ti is switched into conduc-tion. The integrator is reset by making the - inputof IC, positive with respect to the + input with theaid of T3, R5 and Re.Capacitor Ci serves to lengthen the on -time for T,and T3 to ensure that the flyback of the sawtoothis finished.Potentiometer Pi is a voltage divider to define thesawtooth amplitude, and hence the sweep range,while Si makes it possible to turn off the sweepfunction (position F).Opamp IC2 is configured as a buffer stage for in-verting and attenuating the sawtooth voltage, towhich a direct voltage is added also. The output of1C2 carries a sawtooth voltage with an amplitudebetween 0 and 2.85 V, or a direct voltage betweenthe same limits when Si is set to position F. Bearingin mind that the reference voltage of 1C3 is 3 V, thecurrent through R13, and hence the output fre-quency, can be varied by a factor 20, which is themaximum attainable deviation factor in all 3 fre-quency ranges. The frequency scale can be cali-brated with the aid of P3.

Parts list

Resistors (±5%):RI =22KR2;R4;1317=10KR3 = 4K7R5 = 1K2R6 = lOR

Miscellaneous:

= miniature SPDT switch.S2= miniature SPST switch.

131-R12 589

R14

100

R15

6

a = GUSp= 3Vc = 3Vd = 0...2VB5 )S1/F)

ULT.OUT

TIMING

C SYNC. /1OUTP.

Pad 25r,9 12V 1 0 0

CMO

IDP

N

TIMING XR 22°6 OUTPUT

INP.B".ss.61OR VEri ADJUST 2g49 910 13

R121

PS

C9

Tat

S3, T = IkHz...20kHzS = 10011....2kHzR = 10Hz...2001-12

R7 = 1 MOR8 = 68K133;Rio = 820KR1i;R32 = 470KR13 = 2K2RicRis = 33KRi6= 220RP1= 50K linear potentiometerP2 = 100K linear potentiometerP3 = 100K presetP4 =100R presetP5 = 11(0 logarithmic potentiometer

Capacitors:

Ci =3n3C2=12nC3 = 68pCa =1µ; 16 V; radialC6=22nC6 = 220nC7 = 2µ2; 16 V; radialC8 = 1Oµ; 16 V; radialC9 = 2n2Cio=220µ; 16 V; radialCii;Ci2=100n

Semiconductors:

Di =zenerdiode 5V6; 400 mWTi;T2 = BC557T3 = BS2501C3;1C2= CA31401C3= XR2206

S3 = 1 pole, 3 -way rotary switch.PCB Type EPS87419

log

LJY

87419

320

291 SIMPLIFIED WORD COMPARATOR

Primarily intended as a trigger source for an oscillo-scope in the testing of digital circuits, the compara-tor is a derivative of the word recognizer anddelayed trigger published in the July/August 1981issue of Elektor Electronics.When an 8 -bit binary word is recognized during acomparison with a pre -determined value, the pres-ent circuit issues a short trigger pulse. In contrast tothe original circuit, the present one has no provisionfor either a delayed trigger pulse or an external trig-ger input. None the less, the comparator remains analmost indispensable aid in the testing of digital cir-cuits.The unit is based on two four -bit comparators, IC,and IC2. The reference level for them can be set sep-arately with switches Sl . . .S4 and S5 . . .S8 respect-ively. With these switches set as drawn, inputs Aand B are interconnected: this is the don't care pos-ition. With a switch set to its centre position, a highreference level is obtained, while when it is set to the

extreme right position, a low reference level is ob-tained.When all A and B inputs agree, the A = B output ofIC2 goes logic high. Gates N, . .N4 suppress shortspurious pulses that arise during the stabilization ofthe comparator inputs.The size of the binary word can be increased bycascading two or more comparators. Account mustthen be taken of the transit delay which amounts to24 ns per comparator. In some tests this may lead toan unacceptable delay if several comparators areused.The current consumption is about 60 mA per com-parator: 32 mA is drawn by the LS241, and 10 mAby the LS85. This enables the current consumptionof multiple comparators to be calculated quite eas-ily.

Note that each additional IC must be separatelydecoupled by a 100 n capacitor.

321


5V0

2A <8

'Cl

74LS85A>6

AO BO Al 61 A2 B2 A3 B3

5V

16 16

N1 . N4 = IC5 = 74LS00

7 2

6 3

5 4

8 8

5V

R1 R2

47k

R3 R4

47k

R5 R6

47k

R7 R8

47k

OIC2

74LS85A -

80 AO 61 Al B2 A2 B3 A3

9 10 11 12 14 13

85

74LS241

S6 S7

15

-- JT11 13 15 17

Cl

Jam 0

1C:)-85402

100nIC5

5V

292 TWO-TONE RF TEST OSCILLATOR

This test oscillator is useful to ensure optimum op-eration of RF amplifier stages designed to work onthe short-wave bands. Based on two crystaloscilators, it provides considerable output power (10to 100 mW) to enable measuring intermodulationcharacteristics of high level and RF power stages.The quartz crystals used here not only serve as thefrequency determining elements (2...20 MHz), butalso as output filters to prevent one generated signalbeing lost in the other oscillator. With this in mind,tapped inductors Li and L2 ensure freedom ofmutual interference when the oscillator is usedfor frequencies higher than 10 MHz. Both induc-tors are wound as 12 turns of enamelled copper wirewith a centre tap, on either a small balun or asuitably rated core with an air gap. Outputs of equal

20dBm500

322

amplitude can be obtained by adjusting Pi. should be fitted with a heat -sink, and that chokesThe test oscillator consumes about 250 mA from a L3 and L4 should be capable of carrying about60 V supply. This means that both transistors 150 mA.

293 VARIABLE WIEN BRIDGE OSCILLATOR

A Wien bridge oscillator can be made variable byusing two frequency determining parts that arevaried simultaneously at high tracking accuracy.High -quality tracking potentiometers or variablecapacitors are, however, expensive and difficult toobtain. To avoid having to use such a component,this oscillator was designed to operate with a singlepotentiometer. The output frequency, fo, iscalculated from

f0=1/(2TiRCl/a)

where R = R2 = R3 = R4 -= R6, C = C1 = C2 , anda=(Pi+Ri)R. Preset P2 allows adjusting theoverall amplification such that the output signal Parts listhas a reasonably stable amplitude (3.5 Vpp max.)over the entire frequency range.The stated components allow the frequency to be R1=10Kadjusted between 350 Hz and 3.5 kHz. Other fre- R2;R3;R4;R6=100K

quency ranges are readily defined with the aid of R5=2M2

the above formula, although it should be noted that Pi = 1 MO linear potentiometer

the upper frequency limit is determined mainly by P2=5k0 preset

the gain -bandwidth product of the opamps Type Capacitors:OP -221 and TLC272. The current consumption of

Cl;C2=1n5the oscillator depends on the type of opamp used.

C3;C4 = 100nThe following values were measured: OP -221:0.5 mA; TLC272: 2 mA; TL072: 2 mA. The con- Semiconductors:struction of the oscillator should present very few Di:D2=1N4148problems since a ready-made circuit board is IC, =TLC272 or TL072 or OP -221available.(PMI Application)

01, D2 =154148

Resistors 1±5%):

Miscellaneous:

PCB Type 87441

4PJ.

a

Al, A2 = ICI = TLC272, TL072, OP -221

323

294 WIEN BRIDGE OSCILLATOR

This AF oscillator can be built with only one activecomponent, and draws so little current that it is

conveniently fed from a 9 V (PP3) battery.The basic circuit of the Wien bridge oscillator isshown in Fig. I. The oscillator consists of two sec-tions, namely the opamp plus R3 -R4 which deter-mine the amplification factor, and positive feedbacknetwork CFR I -C2 -R2 which enables the circuit tooscillate. This network is composed of a low-passsection R2f/C2, and a high pass section RI +CI.The phase difference incurred in these is nulled atthe frequency of oscillation, when the filters form apure ohmic potential divider with an attenuation of3. Therefore, the opamp must have an amplificationof 3, to keep the overall amplification at unity, sothat the oscillation is maintained. The output fre-quency, f0, of the oscillator is

f0= 1/(2nI/R IR2C IC2) [Hz]

but only if RI ,-=--R2 and CI In the practicaldesign shown in Fig. 2, the oscillation frequency isabout 1,000 Hz.Both the inverting and the inverting input of theopamp in Fig. 1 must be held at half the supplyvoltage to ensure minimum current consumption ifthe oscillator is to be fed from a battery. Figure 2shows how this is realized in the practical version ofthe Wien bridge oscillator. Here, resistors R2 andR3 from Fig. 1 are seen as Rea-R2b and R3a-R3b, re-spectively, connected as voltage dividers. This canbe done with impunity, because the voltage sourceis a virtual short circuit for alternating voltages, andthere is also C3 as an effective decoupling device.For an alternating voltage, therefore, the resistors

2

R4

1

are parallel combinations. Evidently, Rea, R2b, R3a,and Rib have two times the calculated resistance ofthe respective components R2 and R3 in Fig. 1.The amplification of the opamp is adjustable withPi, which should be set for reliable oscillation atvirtually no distortion of the output sine wave.When the oscillator is properly aligned, the distor-tion should be less than 0.1%.The use of the Type TIC271 CMOS operationalamplifier results in a current consumption of only0.32 mA at U.=6 V... It is possible to use aspecial low -power opamp such as the type OP -22biased with a resistance of 1 MQ to reduce the cur-rent consumption to 0.1 mA. However, this willcause the oscillation frequency to be limited to1,000 Hz, due to the reduced slew rate at very lowbias settings, which in turn give rise to a strong in-crease in the distortion level.

PMI Application Note ABM.

87415 -2

Al = TLC 271D1, D2 = 1N4148

324

295 RF MODULE FOR INDOOR UNIT

The series of articles on satellite TV reception,started in the September 1986 issue of Elektor Elec-tronics, has met a great deal of interest from ourreaders. Many have successfully ventured out intothe world of centimetre waves and SHF construc-tion methods, and are proud to watch the picturesproduced by a home-made Indoor Unit 0).The construction and alignment of the RF inputstage and the local oscillators is undoubtedly themost difficult phase of the project. However, an at-tractive alternative is now available for those con-structors hesitant about their skills in dealing withvery high frequency components and techniques.The Type HL ECS51 is a ready-made, tunable, 950-1750 MHz to 480 MHz converter of Taiwaneseorigin, and is eminently suitable for taking over thepreviously mentioned functions on the RF board(Type EPS 86082-1). The module is tuned with avoltage between 0 and 20 V, and requires no band(H/L) switch. It has a connection for applying theLNB supply voltage, which is carried to the LNBvia the downlead cable as usual. The module itselfis conveniently fed from 12 VDC, and consumesabout 100 mA. Its RF input is a Type F socket.Figure 1 shows the ready connection of the moduleto the RF board in the IDU. Module pin 3 is the

1

HLECS51

203040-50

AGC input, which is grounded here to achievemaximum gain. A 56 k4 resistor is fitted in serieswith the existing + 33 V tuning voltage rail to en-sure the correct maximum input to pin 5. The IFoutput on the module accepts a common phonoplug, to which a short length of thin (RG174) coaxcable is soldered for connecting to the short trackbetween pin 3 and 4 of MX, on the RF board.Coupling capacitor C7 should be left in place, butMX1, the RF input stage, and both local oscillatorsmay be omitted, since the module takes over theirfunction. In case the RF board is already complete,it is recommended to remove MX1, R3 and bandselector Si. The screen between the RF input stageand the local oscillators may also be omitted, butnot the remaining screens on the board.Before aligning the "modular" Indoor Unit, the in-termediate frequency requires lowering from 610 to480 MHz. The VCO in the SL1451 PLL can be tun-ed to the new centre frequency by increasing the in-ductance of L8. This is easiest to do by making anew inductor as shown in Fig. 2. Use about 5 cm ofsilver plated, 0 1 mm (SWG20) wire to make thesingle turn inductor, ensuring that the underside ofit is just above the PCB surface.The alignment procedure of the module -based In-

RG174

O C7

1 = RF in (from LNB)2 = LNB supply3 = AGC input4 = +12V supply5 = tuning voltage (0...20V)6 = IF output 480 MHz7 = OSC output

86082-1

+12V0

(5)

+33V86082.2

87503-1

325

door Unit is essentially identical to that set forth onpage 54 in Part 2 of (1?. The reference there to TVchannel 36 (600 MHz) should then be read as TVchannel 21/22 (approx. 480 MHz). The IF band -filters can be tuned to the new frequency when therespective trimmers are set for nearly maximumcapacitance.Finally, a note on the results obtained with themodule in combination with the RF board as de-scribed here: the original, completed, and properlyaligned RF board with the BFG65, local oscillators,and MX, fitted gives a slightly better performancewhen relatively weak signals (C/Ns 12 dB) are be-ing received. This is mainly due to its noise figurebeing lower than that of the HL ECS51 module,which is specified for no less than 15 dB in this

2 L8 480 MHz

87503-2

respect. None the less, the module gives goodresults with relatively strong input signals.Literature reference:(1) Indoor Unit for Satellite TV Reception Parts 1-3;Elektor Electronics, October 1986 and followingissues.

296 RGB-TO-MONOCHROME VIDEO COMBINER

This circuit offers impeccable monochrome imageswhen driven by the digital RGB and sync signals ofa high -resolution graphics card such as the onefeatured in Elektor Electronics, issues fromNovember 1985 to March 1986.Transistor T2 in the video combiner/buffer ensuresa short-circuit proof, 75Q impedance monochromeand composite video signal with an rms value ofabout 1V, as usual for connection to a monitor. Thecombination of a PNP and an NPN transis-tor, Ti and T2 respectively, for amplification andcombination of video and sync typically exhibits agood response to the fast rise and fall times of thesesignals, and thus enables sufficient picture defi-nition in the case of, for instance, text presentationat 80 characters per line, or high -resolution graphicsapplications.The input circuit arrangement with Ti and themixer resistors is a D/A converter in its mostrudimentary form; the R, G, B, and I signals are ap-plied to resistors at values that correspond to theluminance percentage of each basic colour, to theeffect that any colour shade is represented as one of16 shades of grey on the monochrome monitor.Where this is desirable, the intensity bit resistor maybe replaced with a 2k5 preset; this enables settingthe intensity ratio. In case the present combiner isused with a video interface that merely supplies avideo and sync signal, or merely RGB signals, theunused inputs may simply be left open.The sync signals are combined with the video signalat the base of T2; depending on the system setup,

I = intensity _IL 1nienun

* 5VR,G,B = video It1-1= hoc sync. LrV = vert. sync.

BF 451

G

R

B

Cfl VBE

CSYNC 8Fs

9 in

di 9,60 connector

4

video

combiner

05V / 35mA

1N41413

N4148

V

H

2

Ground

R6

GroundRedGreen 4Blue 5

Intensity 6

Reserved 7

Horizontal DriveVertical Drive

Color / GraphicsDirecliDriveAdapter

+50

8646

the sync input may have to be slightly modified.Where an inverted CSYNC (composite sync) signalis available, as in the case of the Elektor Electronics

326

graphics card, all parts relevant to the sync inputsmay be dispensed with, except for Di. Resistors R5and R6 determine the black level of the compositevideo signal, to which the sync component is addedby Di, which is capable of lowering the basevoltage of T2 to a value, lower than the black level.Diode D2 allows separate, inverted sync signals tobe applied to the combiner, and T3, Rio, Rig andRii may be added for video systems that output apositive -going, separate sync.The combiner may also be used with the video -

interface incorporated in the IBM PC (see ElektorElectronics, May 1985 issue); if this computer isequipped with a monochrome video board, the syncmixer only requires D2, R11, R12, and T3. The IBMcolour board requires the addition of Rio, whereasD2 may be left out. Figure 2 shows how the videocombiner is typically connected to the IBM colourboard. Note the use of the 'reserved' line (7) for the+ 5V supply voltage of the combiner circuit.

297 SCART SWITCH

This circuit is not really a technical novelty, but ithas its practical uses. If, for instance, it is desired toconnect a video recorder AND a computer perma-nently to the SCARF socket at the back of amodern television set, it will be found that that isimpossible. All that can be done is to connect eitherthe video recorder or the computer. But the propos-ed SCART switch offers a solution to this problem.The switch is constructed from a small(110 x 60 x 30 mm) metal case, a six -pole change-

over switch, two SCART sockets, one SCART plug,and a length of screened coaxial cable. Suitableholes should be provided in the case to receive thetwo sockets, a cable outlet, and the switch. Thevarious components are connected together asshown in the accompanying diagram. The SCARTplug is connected at the free end of the cable, whichshould not be longer than 1 metre. The connectionsto the sockets and plug are also identified in thetable.

7-t

20

-0-

18 16 18 12 10

21 19 1 15 ta 1 9 5 a 1

01 ::

r

1 0 I= =I 0 =I i=1 CI Ci

RA

09 :

...40

I' 1;

05 ',

LA

6

I

Yn, I B

U 1 I

8 II,I0° 1

10t 1, : 0 G

1120 I

II

114

0 le') n)16 I

116 I

0'9 I V120

.77.71: Vi--LI

LA = left-hand channelRA= right-hand channel

I

,),321

1I'Ll

is 1L

lie=

15

't,13

',E1

11

)''Ll

9

1, i=i

LL

LL

i=R = red MN

t 1G=arB = blue

86418 -1 1 V = video signal in -0

327

The completed switch should find a home beside,under, or on top of the TV set: the SCARF plug isinserted into the SCARF socket at the back of theset. The two SCA' sockets on the case are then

used to receive the computer and video recorder re-spectively. From then on, it is a simple matter ofswitching between recorder and computer!

SCART connector

Pin Function Level

1 Audio output (right-hand)or channel 2

2 Audio output (right-hand)or channel 2

3 Audio output (left-hand)or channel 1 or mono

4 Audio earth5 Blue earth6 Audio input Ileft-hand)

or channel 1 or mono7 Blue component

8 Switching voltage:0 = TV reception1 = operation of

associated units9 Green earth

10 Not used11 Green component12 Not used13 Red earth14 Not used15 Red component16 Blanking signal

1 = blanking

17 Video earth18 Blanking signal earth19 Video output

20 Video input21 Housing screen and/or

earth

0.5 V for outputimpedances ?_1 k52

0.5 V for inputimpedances 5.10 k00.5 V for outputimpedances 5_1 kg

0.5 V for inputimpedances ?...10 k52Difference betweenpeak value and blankingsignal level = 0.7 V;load impedance = 75 0;superimposed directvoltage = 0...2 V0 = 0...2 V1 = 9.5...12 V

Identical to 7

Identical to 70 = 0...0.4 V1 = 1...3 VLoad resistance = 75 Q

Difference betweenpeak white level andsync signal = 1 V;Output resistance = 75 0;Superimposed directvoltage = 0...2 VSynchronization signalonly = 0.3 VppIdentical to 19Connected to chassis

328

298 SYNCHRONIZATION SEPARATOR I

Many monitors have separate inputs for the line(horizontal) and field synchronization pulses. Ifyour computer only provides composite sync pulses,the circuit described here makes it possible to splitthe composite sync signal, CSYNC, into properline, HS, and field, VS, pulses.It is possible to feed the CSYNC as line sync pulsesdirect to the monitor, which is the reason that theCSYNC input is connected direct to the HS outputterminal.To derive the field sync pulses from the compositesignal, a dual retriggerable monostable type74LS123 is needed. The first mono period is slightlylonger than the distance between two line syncpulses. As the monostable is retriggered by each linesync pulse, it only resets in the absence of a linesync pulse, that is, at the onset of a blanking inter-val. The first monostable then triggers the second,which generates a VS pulse at its Q output. Whenthe second mono period has lapsed, the firstmonostable has already been provided with moreline sync pulses, so that monostable 2 is not trig-gered again until the next blanked interval. The

CSYNC = HS

vs

85414

overall result is that all line sync pulses are sup-pressed, while monostable 2 provides field syncpulses.

299 SYNCHRONIZATION SEPARATOR II

Many monitor chassis currently offered by com-puter surplus stores have separate inputs forhorizontal and vertical synchronization signals.Most home micros, however, have a compositevideo output, so that some form of interfacing is re-quired to drive these bargain monitors.The Type TBA950-2 is a sync separator chip whichis frequently encountered on TV chassis. In its stan-dard application circuit, it requires to be driven bya flyback signal derived from the output of the linefrequency oscillator. Without this signal, which isapplied to pin 10, the sync pulse would end upsomewhere among the picture lines. To be able touse the TBA950-2 in the present application, thehorizontal pulse is slightly shifted with the aid of adouble monostable multivibrator, IC2.The operation of the circuit should be clear fromthe accompanying timing diagram. The outputpulse from the TBA950 is fairly wide (26 pis), and itspositive edge triggers the first MMV (Q1), whosenegative output pulse transition in turn triggers the

second MMV in the 4538 package. The line syncpulse for the monitor is available positive and nega-tive at IC2 outputs Q2 and Q2, respectively.Adjust the circuit as follows: set P2 to the centre ofits travel, and adjust the frequency control, 131 , suchthat the image is stable. Next, position the image byadjusting P3. If the correct position can not be ob-tained, the phase control, P2, must be carefullyreadjusted, followed by P3. The vertical sync pulseis available at pin 7 of the TBA950-2. Finally, thedashed resistors and diodes are required if the moni-tor inputs are designed to accept signals with apeak -to -peak amplitude of 5 V.

329

12V

01

VIDEO ci

0C) ID10p 16V

R1

* see text

2VIDEO

IC1-2

R7

IC1TBA 950.2

0

R10

R11

C3

70

c.9

10p 16V

R8

C5

105MKP

C5mlm 1467

470n

P1

14

R9

12

16

11 10

P21136 CS

50p16V

4TR1

01 -TR

R12

16

IC24538

C10

330P14 15 3

02

02

Litt 112 1

MMV-01

MMV-02

.ops

9

1k

R15 *L 1k

87414-1

5V0

0

300 VIDEO AMPLIFIER FOR B/W TELEVISIONSETS

It appears that the use of portable, mains operated receiver as a monitor (Elektor Electronics,television receivers as monitor in a computer system November 1984) described an all -embracing ampli-has become very popular. The article use your TV fier, but here we propose a much simpler one.

i 12

330

To raise the standard video signal of 1 Vpp to thelevel required by the television receiver, a preampli-fier with a bandwidth of not less than 10 MHz is re-quired. With careful construction of the presentamplifier, this bandwidth is guaranteed, and shouldactually be of the order of close to 20 MHz. Witha supply voltage of 12 V, the direct -voltage outputis 4 V. If different supply voltages are used, the DCoutput is retained at that level by suitably alteringthe values of R, and R2 (which form a voltage div-ider). However, the supply voltage should not belower than 10 V, nor higher than 15 V. The amplifi-cation depends on the ratio R7 : R8; if higher ampli-fication is needed, the value of R7 should be in-creased.The respectable bandwidth is achieved by low valuebase and collector resistors: with this arrangement,even audio transistors may be used in this, essential-ly HF, circuit. In any case, the cut-off frequency ofa BC 547 is 300 MHz, and that of a BC 557 is150 MHz.The input impedance is strictly determined by R3;

750

15 mA 12V

85439

its value of 82 Q is near enough the required im-pedance, but if you really want to be a purist, thereare 75 Q resistors available at some stockists, or youcan connect a 100 Q resistor in parallel with a330 Q one.

301 VIDEO BUFFER/REPEATER

This universal video amplifier is intended as abuffer/repeater in a long coaxial cable to keep thesignal at a reasonable level. Its gain is about 6 dB.The circuit is built from readily available compo-nents: some transistors and a few others.The circuit consists of a two -stage amplifier, Ti andT2, and an emitter follower that functions as im-pedance converter. The bandwidth at -3 dB is notless than 20 MHz. Current consumption at asupply voltage of 12 V amounts to about 20 mA.The power supply needs to be regulated to preventlines and other noise on the screen.The buffer/repeater is very suitable for being com-bined with the video selector featured elsewhere in

12 V

this issue. The present circuit, with R, omitted, isthen used as a buffer for the output of the inverter.Its input impedance is then around 4 k52.

302 VIDEO DISTRIBUTION AMPLIFIER

The Type TEA5114 from Thomson-CSF comprisesthree electronic switches followed by a buffer/ampli-fier. Normally the voltage amplification is 2 (6 dB).When the input voltage exceeds 1.2 Vpp, or when

the output voltage exceeds 1.5 Vpp, an internalselector reduces the amplification to unity (0 dB).The threshold of 1.2 Vpp is created with the aid ofvoltage divider It4-R5, which also forms the input

331

termination of 75 Q. Series resistors R1 -R3 ensure75 Q output impedance for driving video equip-ment via standard coax cable. The TEA5114 can beused as a video source selector also, provided eachinput has its own 75 Q termination network. Thenon -connected inputs should then be fitted with acoupling capacitor. Channel selection is effected bycontrolling the logic level at pins 10, 12 and 15.Note that the logic 1 (high) level corresponds to+ 2.5 V here.

IC 1

TEA 5114

750videoinput

750video outputs

C2

05

1000

70...140InA

87466

0

12V

0

0

303 VIDEO SELECTOR

It is sometimes useful, or even necessary, to use thesame screen for more than one video source. Somesimple video selectors used for this purpose sufferrather badly from crosstalk. The present circuit doesnot have this drawback: the unused channel(s) isshorted out with a switch.When CH(annel) 1 is switched in, electronicswitches ES, and ES2 are closed and ES4 is open.The other channel(s) is effectively choked becauseES5 and ES6 are both open and any residualcrosstalk is shorted to earth by ES5. Each channeluses its own IC so that there is no risk of cross chan-nel interference via the chips.As the switches have a certain impedance in the onstate, there will be some losses when the output isterminated into 75 Q. It is, therefore, best to bufferthe output; for instance, with the videobuffer/repeater described elsewhere in this issue.The input of the video selector must be terminatedinto 75 Q. The -3 dB bandwidth is about 8 MHz.Current consumption amounts to 1. . . 2 mA de-pending on the supply voltage. A high supplyvoltage is preferable, because the electronic switchesthen have the lowest impedance in the on state.

CH1

CH2

5 . . 15 VES1 ES4 = 1C1 = 4066ES6 ES8 = 1C2 = 4066

ES1

0

52

ES2

5

ES4

ES5

I et'

ES3

5

ES6

ES8

1" = CH10

5...15 V

EST

5

2.29

332

303 CIRCUITS is the latest in Bektor Electronics' well -liked series or books for theelectronic enthusiast. professionol or amateur dike Uke Its predecessors. it

offers a comprehensive coection of practical ideas. concept, and day.,mans in the gamut of electronics.

Unlike its predecessor, 303 CIRCUITS is arranged In 11 subject sections to makeit easier for the reader to find that long -sought circuit.

In wee over 300 page, 303 CIRCUITS offers 32 Audio and HI-li project& 14 cir-cuits for Cars & Bicycles: 43 Computer & Microprocessor circuits: 11 Electraphonics projects: 24 HF & VHF circuit, 16 circuits fora number of hobbies &pastimes: 54 projects for Home & Garden: 29 Power Supply circuits; 29 circuitsfor Test & Measurement equipment; 9 IV & Video projects: as well as 42 DesignIdeas

Ilex 0- OS

111111.11161 111'

Documents

CIRCUITS - americanradiohistory.com · 1984 (p. 28) issue of Elektor Electronics. In the design stages, stability problems were en-countered when opamps with JFET inputs (TL074;