The Evolution of the IoT

Designing for Rugged

Environments

I.D.E.A. Big

Dual Configurable Logic Design Contest

NXP’s

CLIFF LLOYDBusiness Development Director,Logic Division, NXP

JAN

UAR

Y 2

105

LINK HERECONTENTS

Your Guide to Embedded MCUs and Development Tools.

w w w . e m b e d d e d d e v e l o p e r . c o m

Everything you’re looking for in one place.

22

24

30

38

46

54

60

66

4

12

20

The Evolution of the Internet of Things TI Prepares Customers for the IoT of Tomorrow

Simple Steps to Improving Software QualityFacts of Software Development Life

The Big I.D.E.A. Contest 2015 International Design Engineering Award

Designing for Rugged Environments Embedded Products in Offshore Oil Rigs

Bringing Advanced Sensor Functionality to Consumer, Embedded, and Industrial Applications

Expanding Mobile Application Functionality with Magnetic Gyroscopes

Don’t Lose Your Cool Important Ways to Approach Heat Management

Reduce ESD SusceptibilityWays to Protect Your Electronic Designs

Q & A: The Big I.D.E.A. Cliff Lloyd, NXP Semiconductors

Usinging Shift Registers to Simplify LED Designs

NXP’s Complete Solution for Sensor Processing Applications

to IMPROVING

Mark Moran – Chief Technical Officer, Atollic Inc.

Experience shows us that there are certain attributes of the software development cycle that we must not only live with,

but strive to improve. We may not like it, but we must embrace these realities:

> Code footprints inevitably grow as time goes on

> Factors out of the control of developers and project managers frequently compress project schedules

> Software will always have defects

> The further a project progresses, the more costly it is to find defects and bugs

> From the dawn of human time, good tools are the way to survive and prosper

Immutable Facts of Software Development Life

Simple STEPS

Software QUALITY

GRAND PRIZE

Clemens Valens

JANUARY 1 - ENTRY PHASE

» Fill in and submit questionnaire. (Special bonus drawing for entries

before March 15. Five contestants will win prizes worth $100 each.)

» Advance to the Submission Phase.

Click here to register and start the contest!

MARCH 1 - SUBMISSION PHASE

» Contestants will produce a schematic using NXP’s products.

(Special bonus drawing for schematics before April 15. Ten contestants

win prizes worth $200 each.)

» Selected contestants will advance to the Completion Phase.

Click here to view the Submission Phase rules and instructions.

APRIL 15 - COMPLETION PHASE

» Design kits will be sent to contestants selected from the Submission Phase.

» Entries must be submitted for judging by June 30.

AUGUST 1 - WINNERS ANNOUNCED

» GRAND PRIZE: $3000 Prize Value

» FIRST PLACE: $2000 Prize Value

» SECOND PLACE: $1500 Prize Value

» THIRD PLACE: $1000 Prize Value

» HONORABLE MENTION: $500 Prize Value

NXP Logic Presents

I.D.E.A.2015 International Design Engineering Award

Join NXP and Mouser in the 2015 Big I.D.E.A.,

a design contest featuring NXP’s unique

product line-up across the spectrum,

including Dual Configurable Logic, Smart Analog,

MOSFETs, and Power.

Everyone has a chance to win thousands of

dollars worth of prizes, with awards at every

level of the competition.

The Big

Submission Phase

March 1Entry Phase

January 1

Completion Phase

April 15

Winners Announced

August 1

MANY USES HAVE BEEN DEVELOPED FOR ACCELEROMETERS, INCLUDING INCREASING POWER EFFICIENCY.

Figure 2.

Figure 1.

Many consumer electronics devices can benefit from 9-axis sensor fusion (see Figure 1).

For example, cell phones, gaming systems, health monitors, and wearable electronics could implement high value-added functions such as three-dimension gesture recognition, relative positioning to other users, and motion tracking. To achieve this, devices need 9-axis sensor data and the ability to process it efficiently.

Note that many of these systems do not require the same level of precision that a fighter jet needs to stay in flight. Rather, they need a cost-effective 9-axis sensor fusion implementation that

matches the needs of consumer applications to improve the user experience. With the availability of new MagGyros, such as the KMX62G from Kionix, developers can use 6-axis sensors (accelerometer + magnetometer) to provide the equivalent of 9-axis output (accelerometer + magnetometer + gyroscope) while reducing energy consumption by as much as 90%.

9-AXIS SENSOR FUSION

Using 9-axis sensor fusion—data from an accelerometer, magnetometer, and gyroscope—systems can accurately position themselves in the world, including inclination and orientation as well as changes in position and rotation (see Figure 2). Effectively, sensor fusion uses multiple sensors to fill in the blind spots of individual sensors.

Accelerometers:

An accelerometer measures acceleration in 3 axes. It is important to note that gravity is a form of acceleration, and so an object at rest still has a downward acceleration. This fact enables an accelerometer to determine how a system is oriented relative to “down.”

This information can also be used to determine how level a system is. Many handheld and portable devices use orientation to adjust the display to match how the user is holding the device. Devices commonly using this functionality include cell phones, tablets, and laptops.

Another unique feature of an accelerometer is the ability to tell if a device has been dropped. For example, when a laptop is resting on a table, it detects 1 G. However, if the laptop is knocked off the table, it will detect 0 G as it falls to the floor (similar to how zero-gravity or weightless is simulated by putting a plane into freefall). This enables a system to detect if it is falling and at imminent risk of collision. In the case of a laptop, the system can proactively park

the hard drive head in a safe position to prevent damage to the drive or data.

Many other uses have been developed for accelerometers, including increasing power efficiency. Consider that a fitness tracker that has not moved in a few minutes is likely not being used. The device can then power itself down. This simple use case can substantially improve battery life across a wide range of applications.

Magnetometers:

A magnetometer measures magnetic fields. It can be used in a compass to identify the Earth’s magnetic field. By analyzing the magnitude, direction and rate of change of the detected field, a magnetometer can be used to find the direction of magnetic north.

As a standalone sensor, however, a magnetometer is typically unable to identify the direction of magnetic north. A traditional compass uses a magnetic rod with a bearing that restricts the rod to align in one dimension; proper leveling of the housing allows the user to best align the rod with magnetic north. In an electronic magnetometer, 3-axis of sensing is used to compute the direction and magnitude of the magnetic field. Any tilt between the magnetometer axes and the measured signal will cause error in the reported direction of the field.

A handheld device can use both an accelerometer and magnetometer to determine magnetic north. This is achieved by identifying the tilt of the device using the accelerometer and using this information to supplement the reading from the magnetometer.

EFFECTIVELY, SENSOR FUSION USES MULTIPLE SENSORS TO FILL IN THE BLIND SPOTS OF INDIVIDUAL SENSORS.

QA&

“The idea behind the Big I.D.E.A. Design Contest is to make the engineer aware of the breadth of NXP’s product portfolio.”

CLIFF LLOYD, Business Development Director, Logic Division, NXP Semiconductors, discusses Dual Configurable Logic and this year’s Big I.D.E.A. (International Design Engineering Award).

The Big

I.D.E.A.The following interview was conducted in early January at the NXP offices in San Jose by Glenn ImObersteg, President of Convergence Promotions.

In designs that use LEDs, shifter registers can be very useful. For instance, if the system includes a seven-segment display, a single indicator, or an array of LEDs that form a grid or panel, a standard 8-bit shift register can be used to allow a low pin-count microcontroller to drive multiple LEDs.

In some cases, a 5V, 8-bit register like the 74HC595 can be used to drive LEDs directly. This works best when the LEDs are specified for relatively low voltage and forward current. LEDs that operate with voltages higher than 6V or require forward current that exceeds 20mA will typically require an external driver.

Open-drain outputsAdding open-drain outputs to the shift register creates a single-chip solution that eliminates the need for an external driver. This can yield significant reductions in the bill of materials, since each output of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such device, the NPIC6C596A LED driver from NXP, which combines shift register functions similar to a Z4HC595 with a high-voltage (HV) MOSFET driver.

Using shift registers to reduce size and BOM in LED designsby Michael Lyons, NXP Semiconductor

Summary

Shift registers can help reduce size and BOM in

designs that use LEDs. By providing I/O expansion,

they enable the use of smaller, less expensive

microcontrollers. In some cases, the shift register can

be used to drive the LED directly and thus eliminate

the need for external LED drivers. This adds to

the savings, and makes it possible to drive a wider

variety of LEDs.

Full Article

In designs that use LEDs, shifter registers can be

very useful. For instance, if the system includes a

seven-segment display, a single indicator, or an array

of LEDs that form a grid or panel, a standard 8-bit

shift register can be used to allow a low pin-count

microcontroller to drive multiple LEDs.

Figure 1 gives an example. A single 5V 74HC595

shift register, with serial inputs and serial or

parallel outputs, provides I/O expansion for the

microcontroller. Serial data is applied to the serial

input of the 74HC595 and clocked in via the input

clock. Once the 74HC595 is loaded, the output

clock applies the data to the storage register and

to the parallel and serial outputs. External drivers,

controlled by the 74HC595, then activate the

corresponding LEDs.

Output_CLKI/O

12 V

Microcontroller

74HC595

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

Figure 1. An 8-bit 74HC595 shift register driving multiple LEDs

Using the 74HC595 for I/O expansion means that

it takes only three MCU control pins to drive up to

eight LEDs. Reducing the number of control pins

makes it possible to use an MCU with a lower pin

count, and that can yield a smaller, more cost-

effective design.

Also, because the 74HC595 includes a serial output,

several devices can be cascaded together. Figure 2

gives the layout.

Figure 1. An 8-bit 74HC595 shift register driving multiple LEDs.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 3. Output schematic for shift register with open-drain outputs.

Figure 1 gives an example. A single 5V 74HC595 shift register, with serial inputs and serial or parallel outputs, provides I/O expansion for the microcontroller.

Serial data is applied to the serial input of the 74HC595 and clocked in via the input clock. Once the 74HC595 is loaded, the output clock applies the data to the storage register and to the parallel and serial outputs. External drivers, controlled by the 74HC595, then activate the corresponding LEDs.

Using the 74HC595 for I/O expansion means that it takes only three MCU control pins to drive up to eight LEDs. Reducing the number of control pins makes it possible to use an MCU with a lower pin count, and that can yield a smaller, more cost-effective design. Also, because the 74HC595 includes a serial output, several devices can be cascaded together. Figure 2 gives the layout.

Now, with the cascading, the same three pins on the microcontroller can be used to control up to 16 or 24 LEDs instead of just eight. The ability to cascade shift registers can reduce the total number of microcontrollers needed in the design, and that can help lower costs and reduce size, too.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 4. Output schematic for shift register with open-drain outputs.

Replacing the 74HC595 with the NPIC6C596A eliminates the need for external drivers, creating a design that is more compact and has a lower bill of materials.

NPIC6C devices have open-drain outputs that are tolerant to 33V. Each output is designed to sink 100mA and there is no limit on ground current. All the outputs can actively sink 100mA simultaneously. The outputs include current-limiting circuitry, which sets a 250mA maximum on the sinkable current, and each output also includes thermal protection. Having these protections means the NPIC6C596A device can be used to drive a wider range of LEDs than the 74HC595, including LEDs that operate at higher voltages and with higher forward current.

CONTENTS

CONTENTS

3

44

The Internet of Things (IoT) is rapidly evolving. In order to attain

the expected 50 billion connected devices in 2020, there is a

need to understand the fundamental challenges in obtaining

horizontal and vertical application balance.

With more than 27 years in the high tech industry, Jim Chase

has spent his career working with customers and helping them

get in front of technology trends and challenges. As a trusted

expert, he employs his system solutions approach to business

and consumer cases worldwide. It is that methodology that has

him creating solutions at Texas Instruments (TI) for the IoT and

helping customers connect their products.

of the

of

InternetThings

By Jim Chase — Strategic Marketing, Texas Instruments

The

EVOLUTION

5

TECH REPORT

5

The Internet of Things (IoT) is rapidly evolving. In order to attain

the expected 50 billion connected devices in 2020, there is a

need to understand the fundamental challenges in obtaining

horizontal and vertical application balance.

With more than 27 years in the high tech industry, Jim Chase

has spent his career working with customers and helping them

get in front of technology trends and challenges. As a trusted

expert, he employs his system solutions approach to business

and consumer cases worldwide. It is that methodology that has

him creating solutions at Texas Instruments (TI) for the IoT and

helping customers connect their products.

of the

of

InternetThings

By Jim Chase — Strategic Marketing, Texas Instruments

The

EVOLUTION

66

The Internet of Things (IoT) is generally thought of as connecting things to the Internet

and using that connection to provide some kind of useful remote monitoring and control. This definition of IoT is limited, and references only part of the IoT evolution. It is basically a rebranding of the existing Machine-to-Machine (M2M) market of today.

The Internet of Things is best defined as:

In the 1990s, Internet connectivity began to proliferate in enterprise and consumer markets, but was still limited in its use because of the low performance of the network interconnect. In the 2000s, Internet connectivity became the norm for many applications and today is expected as part of many enterprise, industrial, and consumer products to provide access to information. However, these devices are still primarily things on the Internet that require more human interaction and monitoring through apps and interfaces. The true promise of the IoT is just starting to be realized—when invisible technology operates behind the scenes dynamically responding to how we want “things” to act.

To date, the world has deployed about 5 billion “smart” connected things. Predictions say there will be 50 billion connected devices by 2020 and in our lifetime we will interact in a trillion-node network. Those are really big numbers. How things are fundamentally deployed today is a barrier to realizing those numbers. The industry will only achieve the reality of 50 billion connected devices by simplifying how things connect and communicate today.

THE IOT OF TOMORROW: A SCENARIOThe hotel where I have a reservation knows the approximate time of my arrival because I have allowed Apple and Google to track my location. It also knows that I am hot and sweaty from my trip because of the temperature and moisture sensors that are part of my smartwatch. The hotel room I will stay in is currently dormant—no lights, drapes closed, and the temperature is at optimized dormant levels. The valet knows it is me upon my arrival. He opens my door and the car adjusts the seat because it detects the valet. My preference is to carry my own bag, so I am not accosted by the bell captain. Once in proximity of the hotel lobby, a secure key app is available on my smartphone. By the time I reach the elevator, the room temp has adjusted to coincide with my smartwatch sensors. The light level, music, and privacy settings are to my requirements. As I approach my room, the secure key app unlocks the room door. Once settled for the night, the room detects the lights are turned out, and it changes the temperature setting to my sleep preferences.

In this scenario, I am wearing multiple sensors and actuators, like a watch vibration for alerts. Every room in this particular hotel chain also has multiple sensors and actuators as well as the rental. I am not interacting with my smartphone touchscreen constantly to direct these connected things to take actions even though it is one gateway for my activity. There will be millions of people doing this every day. We will be living in the data.

This vision of IoT will not happen right away. The scale required will only be achieved by creating a lowest common denominator: a simple messaging scheme that everyone on the planet will agree to. It will have to be digitally organic, imitating nature. At present, technology protocols and data structures are limited by their design complexity as well as security, extensibility, and much more. Our connected devices will have to become easier to use even though the complexity of the devices will increase. The line between analog and digital will blur. Every person on the planet will be able to “author” his or her own life environment, even though they know basically nothing about the underlying technology.

THE IOT OF TODAYManufacturers have been connecting things to the Internet before we called it the Internet. By the mid-

1990s, web servers were being added to embedded products. Current M2M manufacturers have been integrating Internet-connected systems into high-value asset tracking, alarm systems, fleet management and the like for more than 15 years. These M2M systems are challenging

to build even though some are based on industry standard protocols. However, it is getting easier to integrate M2M systems as more powerful processors are incorporated into the end nodes. Since these processors support high-level operating systems and languages, the platform can leverage intelligent frameworks. These systems are typically tied into high-end business service layers and are managed by a network operations center (NOC).

Consumers already have connected things like thermostats, energy meters, lighting control systems, music streaming and control systems, remote video streaming boxes, pool systems, and irrigation systems, with more to come. Most of these systems have some connectivity through a website so that a user can manage them through a standard web browser or a smartphone app, which acts as a personal NOC.

An intelligent, invisible network fabric that can be sensed, controlled, and programmed. IoT-enabled products employ embedded technology that allows them to communicate, directly or indirectly, with each other or the Internet.

Every person on the planet will be able to “author” his or her own life environment, even though they know basically nothing about the underlying technology.

7

TECH REPORT

7

The Internet of Things (IoT) is generally thought of as connecting things to the Internet

and using that connection to provide some kind of useful remote monitoring and control. This definition of IoT is limited, and references only part of the IoT evolution. It is basically a rebranding of the existing Machine-to-Machine (M2M) market of today.

The Internet of Things is best defined as:

In the 1990s, Internet connectivity began to proliferate in enterprise and consumer markets, but was still limited in its use because of the low performance of the network interconnect. In the 2000s, Internet connectivity became the norm for many applications and today is expected as part of many enterprise, industrial, and consumer products to provide access to information. However, these devices are still primarily things on the Internet that require more human interaction and monitoring through apps and interfaces. The true promise of the IoT is just starting to be realized—when invisible technology operates behind the scenes dynamically responding to how we want “things” to act.

To date, the world has deployed about 5 billion “smart” connected things. Predictions say there will be 50 billion connected devices by 2020 and in our lifetime we will interact in a trillion-node network. Those are really big numbers. How things are fundamentally deployed today is a barrier to realizing those numbers. The industry will only achieve the reality of 50 billion connected devices by simplifying how things connect and communicate today.

THE IOT OF TOMORROW: A SCENARIOThe hotel where I have a reservation knows the approximate time of my arrival because I have allowed Apple and Google to track my location. It also knows that I am hot and sweaty from my trip because of the temperature and moisture sensors that are part of my smartwatch. The hotel room I will stay in is currently dormant—no lights, drapes closed, and the temperature is at optimized dormant levels. The valet knows it is me upon my arrival. He opens my door and the car adjusts the seat because it detects the valet. My preference is to carry my own bag, so I am not accosted by the bell captain. Once in proximity of the hotel lobby, a secure key app is available on my smartphone. By the time I reach the elevator, the room temp has adjusted to coincide with my smartwatch sensors. The light level, music, and privacy settings are to my requirements. As I approach my room, the secure key app unlocks the room door. Once settled for the night, the room detects the lights are turned out, and it changes the temperature setting to my sleep preferences.

In this scenario, I am wearing multiple sensors and actuators, like a watch vibration for alerts. Every room in this particular hotel chain also has multiple sensors and actuators as well as the rental. I am not interacting with my smartphone touchscreen constantly to direct these connected things to take actions even though it is one gateway for my activity. There will be millions of people doing this every day. We will be living in the data.

This vision of IoT will not happen right away. The scale required will only be achieved by creating a lowest common denominator: a simple messaging scheme that everyone on the planet will agree to. It will have to be digitally organic, imitating nature. At present, technology protocols and data structures are limited by their design complexity as well as security, extensibility, and much more. Our connected devices will have to become easier to use even though the complexity of the devices will increase. The line between analog and digital will blur. Every person on the planet will be able to “author” his or her own life environment, even though they know basically nothing about the underlying technology.

THE IOT OF TODAYManufacturers have been connecting things to the Internet before we called it the Internet. By the mid-

1990s, web servers were being added to embedded products. Current M2M manufacturers have been integrating Internet-connected systems into high-value asset tracking, alarm systems, fleet management and the like for more than 15 years. These M2M systems are challenging

to build even though some are based on industry standard protocols. However, it is getting easier to integrate M2M systems as more powerful processors are incorporated into the end nodes. Since these processors support high-level operating systems and languages, the platform can leverage intelligent frameworks. These systems are typically tied into high-end business service layers and are managed by a network operations center (NOC).

Consumers already have connected things like thermostats, energy meters, lighting control systems, music streaming and control systems, remote video streaming boxes, pool systems, and irrigation systems, with more to come. Most of these systems have some connectivity through a website so that a user can manage them through a standard web browser or a smartphone app, which acts as a personal NOC.

An intelligent, invisible network fabric that can be sensed, controlled, and programmed. IoT-enabled products employ embedded technology that allows them to communicate, directly or indirectly, with each other or the Internet.

Every person on the planet will be able to “author” his or her own life environment, even though they know basically nothing about the underlying technology.

88

The scale required for the IoT will only be achieved by creating a lowest common denominator: a simple messaging scheme that everyone on the planet will agree to.

3Texas Instruments

Consumers already have connected things like thermostats, energy meters, lighting control systems, music

streaming and control systems, remote video streaming boxes, pool systems, and irrigation systems with

more to come. Most of these systems have some connectivity through a Web site so that a user can manage

them through a standard Web browser or a smartphone app, which acts as a personal NOC.

While both the industrial and consumer scenarios are exciting, deployment is not simplified since they are

all disparate vertical systems. The systems may use the exact same protocols and OS underpinnings, but

the communications layers are inconsistent. Each also uses open application programming interfaces (APIs)

without a horizontal connection, which would lead to easier cross-application integration.

Take for example a sprinkler control system. It can have a level of intelligence so it knows when to water

based on sensors and Internet weather data under programmable control. However, it does not know any-

thing about motion sensors around a house that might indicate a reason to delay the zone to avoid drenching

the dog or kids. There are no motion sensor inputs on the sprinkler controller, so other motion control vertical

integration needs to be used to transfer data to another cloud server. Then the two cloud servers need to be

“glued” together somehow. Hopefully, both system integrations allow for some small amount of additional

control. However, hope is never a good word in electronic systems. An additional vertical application written in

Perl, Python, PHP or another programming language on a server can program a connection that allows mo-

tion to delay the sprinkler zone (or other logic the user may want). This is not easy unless you are an expert

and therefore will not lead to rapid deployment.

This need to connect vertical integrations has led to the formation of new web services like IFTTT.com

(If This Then That) and zAPIer.com that allow a user to graphically glue disparate vertical systems together.

However, this requires users to sign up for yet another service to find out if they have an API interface that

meets the specific vertical integration needs. These platforms are set up to provide basic “recipes” such as

“IF I get an email from my wife THEN send a text to my phone.” It is assumed that greater flow control will

The Evolution of the Internet of Things September 2013

Figure 1. IoT-enabled home with connected devices and appliances working invisibly for consumers.

While both the industrial and consumer scenarios are exciting, deployment is not simplified since they are all disparate vertical systems. The systems may use the exact same protocols and OS underpinnings, but the communications layers are inconsistent. Each also uses open application programming interfaces (APIs) without a horizontal connection, which would lead to easier cross-application integration.

Take for example a sprinkler control system. It can have a level of intelligence so it knows when to water based on sensors and Internet weather data under programmable control. However, it does not know anything about motion sensors around a house that might indicate a reason to delay the zone to avoid drenching the dog or kids. There are no

motion sensor inputs on the sprinkler controller, so other motion control vertical integration needs to be used to transfer data to another cloud server. Then the two cloud servers need to be “glued” together somehow. Hopefully, both system integrations allow for some small amount of additional control. However, hope is never a good word in electronic systems. An additional vertical application written in Perl, Python, PHP, or another programming language on a server can program a connection that allows motion to delay the sprinkler zone (or other logic the user may want). This is not easy unless you are an expert and therefore will not lead to rapid deployment.

This need to connect vertical integrations has led to the formation of

new web services like IFTTT.com (If This Then That) and zAPIer.com that allow a user to graphically glue disparate vertical systems together. However, this requires users to sign up for yet another service to find out if they have an API interface that meets the specific vertical integration needs. These platforms are set up to provide basic “recipes” such as “IF I get an email from my wife, THEN send a text to my phone.” It is assumed that greater flow control will come later. Back to the previous example, assuming the sprinkler system has a delay control API, one can glue the recipe into place: IF motion, THEN delay the sprinklers. That is three different services, three sign-ins (which will also have to be managed inside the third service), three different smartphone apps and several points of failure. Now, what if the user wants to integrate this recipe with his or her calendar so the yard is dry for an outdoor family gathering? The scenario becomes more complicated.

While the applications discussed above are interesting, they also do not lead to rapid IoT deployment. Sure, there will be an uptick in the maker communities as well as some new vertical applications and carrier additions, but the IoT is not about simple vertical one-off texts or tweets. That creates interesting demos, but it lacks scalability and integration across vertical systems. The IoT should enable notifications, but it also needs a simple way for devices to run programs and respond to other devices or services to create a sophisticated application without using a complex programming environment.

THE IOT: VERTICAL AND HORIZONTAL BALANCEIt is hard to argue that the Internet as we know it today (technically HTTP) was born of altruistic intent by Tim Berners-Lee to connect everyone around the world with an open platform. Prior, there were only proprietary enterprise networks with little to no sharing of information—the verticals of the pre-Internet days. ARPANET put some basic plumbing and “messaging” protocols in place to get the party started. The architecture was robust and the vertical spin-offs became the Defense Data Network (DDN) and the National Science Foundation Network (NFSNET). Through public and private industry funding, NFSNET eventually became a major part of the Internet backbone. In the Internet of today there are vertical applications on a fundamentally basic platform of connectivity and information passing.

Today, manufacturers have a multitude of vertical application requirements. Some may be altruistic, but most have money behind their requirements. Without that, there would be no next steps. However, there will be an altruistic requirement to gain horizontal balance. The IoT of tomorrow will be the largest horizontal system architecture ever created. Vertical applications will continue to exist; however, the fundamental lowest levels of connectivity and information passing will need to be ubiquitous and invisible in all applications.

Additionally, horizontal balance will require the IoT to look more like an organic system. When cells replicate,

9

TECH REPORT

9

The scale required for the IoT will only be achieved by creating a lowest common denominator: a simple messaging scheme that everyone on the planet will agree to.

3Texas Instruments

Consumers already have connected things like thermostats, energy meters, lighting control systems, music

streaming and control systems, remote video streaming boxes, pool systems, and irrigation systems with

more to come. Most of these systems have some connectivity through a Web site so that a user can manage

them through a standard Web browser or a smartphone app, which acts as a personal NOC.

While both the industrial and consumer scenarios are exciting, deployment is not simplified since they are

all disparate vertical systems. The systems may use the exact same protocols and OS underpinnings, but

the communications layers are inconsistent. Each also uses open application programming interfaces (APIs)

without a horizontal connection, which would lead to easier cross-application integration.

Take for example a sprinkler control system. It can have a level of intelligence so it knows when to water

based on sensors and Internet weather data under programmable control. However, it does not know any-

thing about motion sensors around a house that might indicate a reason to delay the zone to avoid drenching

the dog or kids. There are no motion sensor inputs on the sprinkler controller, so other motion control vertical

integration needs to be used to transfer data to another cloud server. Then the two cloud servers need to be

“glued” together somehow. Hopefully, both system integrations allow for some small amount of additional

control. However, hope is never a good word in electronic systems. An additional vertical application written in

Perl, Python, PHP or another programming language on a server can program a connection that allows mo-

tion to delay the sprinkler zone (or other logic the user may want). This is not easy unless you are an expert

and therefore will not lead to rapid deployment.

This need to connect vertical integrations has led to the formation of new web services like IFTTT.com

(If This Then That) and zAPIer.com that allow a user to graphically glue disparate vertical systems together.

However, this requires users to sign up for yet another service to find out if they have an API interface that

meets the specific vertical integration needs. These platforms are set up to provide basic “recipes” such as

“IF I get an email from my wife THEN send a text to my phone.” It is assumed that greater flow control will

The Evolution of the Internet of Things September 2013

Figure 1. IoT-enabled home with connected devices and appliances working invisibly for consumers.

While both the industrial and consumer scenarios are exciting, deployment is not simplified since they are all disparate vertical systems. The systems may use the exact same protocols and OS underpinnings, but the communications layers are inconsistent. Each also uses open application programming interfaces (APIs) without a horizontal connection, which would lead to easier cross-application integration.

Take for example a sprinkler control system. It can have a level of intelligence so it knows when to water based on sensors and Internet weather data under programmable control. However, it does not know anything about motion sensors around a house that might indicate a reason to delay the zone to avoid drenching the dog or kids. There are no

motion sensor inputs on the sprinkler controller, so other motion control vertical integration needs to be used to transfer data to another cloud server. Then the two cloud servers need to be “glued” together somehow. Hopefully, both system integrations allow for some small amount of additional control. However, hope is never a good word in electronic systems. An additional vertical application written in Perl, Python, PHP, or another programming language on a server can program a connection that allows motion to delay the sprinkler zone (or other logic the user may want). This is not easy unless you are an expert and therefore will not lead to rapid deployment.

This need to connect vertical integrations has led to the formation of

new web services like IFTTT.com (If This Then That) and zAPIer.com that allow a user to graphically glue disparate vertical systems together. However, this requires users to sign up for yet another service to find out if they have an API interface that meets the specific vertical integration needs. These platforms are set up to provide basic “recipes” such as “IF I get an email from my wife, THEN send a text to my phone.” It is assumed that greater flow control will come later. Back to the previous example, assuming the sprinkler system has a delay control API, one can glue the recipe into place: IF motion, THEN delay the sprinklers. That is three different services, three sign-ins (which will also have to be managed inside the third service), three different smartphone apps and several points of failure. Now, what if the user wants to integrate this recipe with his or her calendar so the yard is dry for an outdoor family gathering? The scenario becomes more complicated.

While the applications discussed above are interesting, they also do not lead to rapid IoT deployment. Sure, there will be an uptick in the maker communities as well as some new vertical applications and carrier additions, but the IoT is not about simple vertical one-off texts or tweets. That creates interesting demos, but it lacks scalability and integration across vertical systems. The IoT should enable notifications, but it also needs a simple way for devices to run programs and respond to other devices or services to create a sophisticated application without using a complex programming environment.

THE IOT: VERTICAL AND HORIZONTAL BALANCEIt is hard to argue that the Internet as we know it today (technically HTTP) was born of altruistic intent by Tim Berners-Lee to connect everyone around the world with an open platform. Prior, there were only proprietary enterprise networks with little to no sharing of information—the verticals of the pre-Internet days. ARPANET put some basic plumbing and “messaging” protocols in place to get the party started. The architecture was robust and the vertical spin-offs became the Defense Data Network (DDN) and the National Science Foundation Network (NFSNET). Through public and private industry funding, NFSNET eventually became a major part of the Internet backbone. In the Internet of today there are vertical applications on a fundamentally basic platform of connectivity and information passing.

Today, manufacturers have a multitude of vertical application requirements. Some may be altruistic, but most have money behind their requirements. Without that, there would be no next steps. However, there will be an altruistic requirement to gain horizontal balance. The IoT of tomorrow will be the largest horizontal system architecture ever created. Vertical applications will continue to exist; however, the fundamental lowest levels of connectivity and information passing will need to be ubiquitous and invisible in all applications.

Additionally, horizontal balance will require the IoT to look more like an organic system. When cells replicate,

1010

A human has trillions of cells that are very resilient and can work for 100+ years without a “reboot.” It is no wonder why organic systems are being studied as a basis for fundamental information and device architecture.

they pass fundamental information from one cell to another in the form of DNA. Cells combine to form a hierarchy of automatic mechanisms that use a nervous system to build and protect its cellular architecture – the body’s form of horizontal integration. A human has trillions of cells that are very resilient and can work for 100+ years without a “reboot.” It is no wonder why organic systems are being studied as a basis for fundamental information and device architecture.

One also might conclude that the Internet has the characteristics of an organic system. However, the Internet of today has most of its traffic aggregated into a few very large data pipes. The original Internet was a much “flatter” looking entity and more peer-to-peer in nature. Bandwidth requirements were fairly low with the largest consumer of bandwidth being simple messaging traffic. Media- and time-critical data forced the emergence of big pipes. Client-server architecture is dominant today, primarily driven by content aggregators and big pipe companies. As the industry progresses there will be a gradual shift back to the original flatter architecture. Fat pipes will not go away as heavy bandwidth and time-critical requirements will still exist. However, when trillions of connected devices exist in the IoT, there will be numerous paths for data flow. The aggregate bandwidth of this massive peer-to-peer platform will far exceed the performance of the fat pipes. Since there will be no way to regulate the network, it will become completely neutral and basically invisible. Our great grandchildren will not even know what an “Internet connection” was. That assumes,

of course, that we eventually all agree on the fundamental “currency” of the IoT.

GETTING THE IOT READYPreparing the lowest layers of technology for the horizontal nature of the IoT requires manufacturers to deliver on the most fundamental challenges, including:

• Connectivity: There will not be one connectivity standard that “wins” over the others. There will be a wide variety of wired and wireless standards as well as proprietary implementations used to connect the things in the IoT. The challenge is getting the connectivity standards to talk to one another with one common worldwide data currency.

• Power management: More things within the IoT will be battery powered or use energy harvesting to be more portable and self-sustaining. Line-powered equipment will need to be more energy efficient. The challenge is making it easy to add power management to these devices and equipment. Wireless charging will incorporate connectivity with charge management.

• Security: With the amount of data being sent within the IoT, security is a must. Built-in hardware security and use of existing connectivity security protocols is essential to secure the IoT. Another challenge is simply educating consumers to use the security that is integrated into their devices.

• Complexity: Manufacturers are looking to add connectivity to devices and equipment that has never been

connected before to become part of the IoT. Ease of design and development is essential to get more things connected especially when typical RF programming is complex. Additionally, the average consumer needs to be able to set-up and use their devices without a technical background.

• Rapid evolution: The IoT is constantly changing and evolving. More devices are being added everyday and the industry is still in its naissance. The challenge facing the industry is the unknown. Unknown devices. Unknown applications. Unknown use cases. Given this, there needs to be flexibility in all facets of development. Processors and microcontrollers that range from 16–1500 MHz to address the full spectrum of applications from a microcontroller (MCU) in a small, energy-harvested wireless sensor node to high-performance, multi-core processors for IoT infrastructure. A wide variety of wired and wireless connectivity technologies are needed to meet the various needs of the market. Last, a wide selection of sensors, mixed-signal and power-management technologies are required to provide the user interface to the IoT and energy-friendly designs.

The IoT is expected to transform how we live, work and play. From factory automation and automotive connectivity

to wearable body sensors and home appliances, the IoT is set to touch every facet of our lives. We will “author” our life with networks around us that constantly change and evolve based on our surroundings and inputs from other systems. It will make our lives safer with cars that sense each other to avoid accidents. It will make our lives more green with lighting systems that adjust based on the amount of daylight from windows. It will make our lives healthier with wearables that can detect heart attacks and strokes before they happen. There is a long road ahead to the IoT of 2020. But one thing is for sure, it is going to be amazing.

TEXAS INSTRUMENTS AND THE IOTWith the industry’s broadest IoT-ready portfolio of wired and wireless connectivity technologies, microcontrollers, processors, sensors and analog signal chain and power solutions, TI offers cloud-ready system solutions designed for IoT accessibility. From high-performance home, industrial and automotive applications to battery-powered wearable and portable electronics or energy-harvested wireless sensor nodes, TI makes developing applications easier with hardware, software, tools and support to get anything connected within the IoT. Learn more at www.ti.com/iot.

11

TECH REPORT

11

A human has trillions of cells that are very resilient and can work for 100+ years without a “reboot.” It is no wonder why organic systems are being studied as a basis for fundamental information and device architecture.

they pass fundamental information from one cell to another in the form of DNA. Cells combine to form a hierarchy of automatic mechanisms that use a nervous system to build and protect its cellular architecture – the body’s form of horizontal integration. A human has trillions of cells that are very resilient and can work for 100+ years without a “reboot.” It is no wonder why organic systems are being studied as a basis for fundamental information and device architecture.

One also might conclude that the Internet has the characteristics of an organic system. However, the Internet of today has most of its traffic aggregated into a few very large data pipes. The original Internet was a much “flatter” looking entity and more peer-to-peer in nature. Bandwidth requirements were fairly low with the largest consumer of bandwidth being simple messaging traffic. Media- and time-critical data forced the emergence of big pipes. Client-server architecture is dominant today, primarily driven by content aggregators and big pipe companies. As the industry progresses there will be a gradual shift back to the original flatter architecture. Fat pipes will not go away as heavy bandwidth and time-critical requirements will still exist. However, when trillions of connected devices exist in the IoT, there will be numerous paths for data flow. The aggregate bandwidth of this massive peer-to-peer platform will far exceed the performance of the fat pipes. Since there will be no way to regulate the network, it will become completely neutral and basically invisible. Our great grandchildren will not even know what an “Internet connection” was. That assumes,

of course, that we eventually all agree on the fundamental “currency” of the IoT.

GETTING THE IOT READYPreparing the lowest layers of technology for the horizontal nature of the IoT requires manufacturers to deliver on the most fundamental challenges, including:

• Connectivity: There will not be one connectivity standard that “wins” over the others. There will be a wide variety of wired and wireless standards as well as proprietary implementations used to connect the things in the IoT. The challenge is getting the connectivity standards to talk to one another with one common worldwide data currency.

• Power management: More things within the IoT will be battery powered or use energy harvesting to be more portable and self-sustaining. Line-powered equipment will need to be more energy efficient. The challenge is making it easy to add power management to these devices and equipment. Wireless charging will incorporate connectivity with charge management.

• Security: With the amount of data being sent within the IoT, security is a must. Built-in hardware security and use of existing connectivity security protocols is essential to secure the IoT. Another challenge is simply educating consumers to use the security that is integrated into their devices.

• Complexity: Manufacturers are looking to add connectivity to devices and equipment that has never been

connected before to become part of the IoT. Ease of design and development is essential to get more things connected especially when typical RF programming is complex. Additionally, the average consumer needs to be able to set-up and use their devices without a technical background.

• Rapid evolution: The IoT is constantly changing and evolving. More devices are being added everyday and the industry is still in its naissance. The challenge facing the industry is the unknown. Unknown devices. Unknown applications. Unknown use cases. Given this, there needs to be flexibility in all facets of development. Processors and microcontrollers that range from 16–1500 MHz to address the full spectrum of applications from a microcontroller (MCU) in a small, energy-harvested wireless sensor node to high-performance, multi-core processors for IoT infrastructure. A wide variety of wired and wireless connectivity technologies are needed to meet the various needs of the market. Last, a wide selection of sensors, mixed-signal and power-management technologies are required to provide the user interface to the IoT and energy-friendly designs.

The IoT is expected to transform how we live, work and play. From factory automation and automotive connectivity

to wearable body sensors and home appliances, the IoT is set to touch every facet of our lives. We will “author” our life with networks around us that constantly change and evolve based on our surroundings and inputs from other systems. It will make our lives safer with cars that sense each other to avoid accidents. It will make our lives more green with lighting systems that adjust based on the amount of daylight from windows. It will make our lives healthier with wearables that can detect heart attacks and strokes before they happen. There is a long road ahead to the IoT of 2020. But one thing is for sure, it is going to be amazing.

TEXAS INSTRUMENTS AND THE IOTWith the industry’s broadest IoT-ready portfolio of wired and wireless connectivity technologies, microcontrollers, processors, sensors and analog signal chain and power solutions, TI offers cloud-ready system solutions designed for IoT accessibility. From high-performance home, industrial and automotive applications to battery-powered wearable and portable electronics or energy-harvested wireless sensor nodes, TI makes developing applications easier with hardware, software, tools and support to get anything connected within the IoT. Learn more at www.ti.com/iot.

1212

to IMPROVING

Mark Moran – Chief Technical Officer, Atollic Inc.

Experience shows us that there are certain attributes of the software development cycle that we must not only live with,

but strive to improve. We may not like it, but we must embrace these realities:

> Code footprints inevitably grow as time goes on

> Factors out of the control of developers and project managers frequently compress project schedules

> Software will always have defects

> The further a project progresses, the more costly it is to find defects and bugs

> From the dawn of human time, good tools are the way to survive and prosper

Immutable Facts of Software Development Life

Simple STEPS

Software QUALITY

13

TECH REPORT

13

to IMPROVING

Mark Moran – Chief Technical Officer, Atollic Inc.

Experience shows us that there are certain attributes of the software development cycle that we must not only live with,

but strive to improve. We may not like it, but we must embrace these realities:

> Code footprints inevitably grow as time goes on

> Factors out of the control of developers and project managers frequently compress project schedules

> Software will always have defects

> The further a project progresses, the more costly it is to find defects and bugs

> From the dawn of human time, good tools are the way to survive and prosper

Immutable Facts of Software Development Life

Simple STEPS

Software QUALITY

1414

In almost every application, more lines of code are added every year to support ever-increasing demand for features. Defect rates, measured by defects per line of code, remain approximately (and stubbornly) constant over the years, unless extraordinary measures are taken. Steve McConnell, in his well-regarded tome, Code Complete, estimates the following:

• There is an “industry average” of 15-50 defects per 1000 lines of code (your mileage may vary; embedded apps might have a little less than this rate). Some defects may be benign, and some may turn into those “million-dollar bugs.”

• The post-testing version of Windows code has about 0.5 defects/1000 lines.

• The extraordinary measures used to test and debug Space Shuttle code has shown that there are 0 defects within specific 500K line blocks of code, although this is not uniformly scalable to the millions of lines code in the entire system.

There will be always be bugs in software—be it newly written or well-tested legacy code, Open Source, or brought-in code from reputable vendors. Finding bugs always takes time, and time is always of the essence in every software development project.

From the time a line of code is written, any bug that line may cause gets more expensive to find as every day passes. Finding defects early in development is far cheaper than dealing with them in debug and test, or in the hands of an unhappy customer.

Finding Those Bugs and Find Them EarlyThere is no silver bullet or magic formula in bug eradication; bug elimination is a layered process using a variety of tools. The compiler and linker are traditionally considered the first lines of defense, followed by printf() and JTAG type hardware tools.

By the time you get to the debugging stage with your arsenal of familiar tools like blinking LEDs, printf(), and JTAG debuggers, the bugs are getting expensive to find and the pressure to complete the project is at its highest. What can be done to make this phase of the project as short as possible? The answer: deploy defect-revealing, bug-squashing technology early on.

Code ReviewFormal code review is a well-known bug elimination methodology. This process asks development engineers to study each other’s code, comment on potential problems, decide which to fix or defer, and assign tasks of fixing and testing appropriately. This is one of the best methodologies for finding and eliminating bugs while they are still inexpensive to fix.

Formal code review is somewhat time- and labor-intensive, which is the principle reason why it is not always done by every development team. In an ideal world, code review would be focused on application specific issues, like timing scheme correction, guaranteed packet transfer, and data integrity preservation. Important but easily fixable static construct issues should really be taken care of prior to the point where the team sits down to read and discuss the code and make their decisions.

This is where automatic code inspection can be used very effectively-before formal code review. The Atollic TrueSTUDIO C/C++ IDE includes a MISRA-C checker, which can automatically apply formal coding rules to C code constructs using the MISRA®-C standard. The MISRA-C rules base is the product of hundreds of man-years of experience in writing safety critical C Code. MISRA was originally developed to meet the safety-critical needs of the automotive industry, but is now being applied as the definitive best practice coding standard for C development.

“What if I’m not writing safety critical code?” you ask. Well, you want your code to work as you expect it to work, regardless of whether or not your application requires formal safety critical certification. And following the MISRA rules will help you avoid common defects in software that may turn into bugs down the road.

MISRA-C rules and recommendation have been adopted only after being extensively discussed and voted on by a team of highly experienced experts in the field of embedded software development. The application of MISRA-C rules and recommendations has been shown to dramatically improve code quality and reliability, as well as providing a more consistent “readability” to the code. ii

Using a Code Inspection Tool in Day-to-Day Embedded DevelopmentThe key question at this point is whether or not you can use TrueSTUDIO static analysis tools quickly and reliably to derive useful information without spending a lot of time. The answer: absolutely.

One mouse click kicks off an exhaustive static analysis that applies the entire MISRA-C 2004 rules base and recommendations to the code. Of course, you may configure which code modules in your application that you want checked, which specific rules and recommendations you wish to apply, and other settings based on your needs and the way you work. One of the best things about this inspection utility is that you can do all of this without ever leaving the IDE. Apart from convenience, this offers additional productivity benefits, as we will see in the following section.

Finding defects early in

development is far cheaper than dealing

with them in debug

and test, or in the hands

of an unhappy

customer.

15

TECH REPORT

15

In almost every application, more lines of code are added every year to support ever-increasing demand for features. Defect rates, measured by defects per line of code, remain approximately (and stubbornly) constant over the years, unless extraordinary measures are taken. Steve McConnell, in his well-regarded tome, Code Complete, estimates the following:

• There is an “industry average” of 15-50 defects per 1000 lines of code (your mileage may vary; embedded apps might have a little less than this rate). Some defects may be benign, and some may turn into those “million-dollar bugs.”

• The post-testing version of Windows code has about 0.5 defects/1000 lines.

• The extraordinary measures used to test and debug Space Shuttle code has shown that there are 0 defects within specific 500K line blocks of code, although this is not uniformly scalable to the millions of lines code in the entire system.

There will be always be bugs in software—be it newly written or well-tested legacy code, Open Source, or brought-in code from reputable vendors. Finding bugs always takes time, and time is always of the essence in every software development project.

From the time a line of code is written, any bug that line may cause gets more expensive to find as every day passes. Finding defects early in development is far cheaper than dealing with them in debug and test, or in the hands of an unhappy customer.

Finding Those Bugs and Find Them EarlyThere is no silver bullet or magic formula in bug eradication; bug elimination is a layered process using a variety of tools. The compiler and linker are traditionally considered the first lines of defense, followed by printf() and JTAG type hardware tools.

By the time you get to the debugging stage with your arsenal of familiar tools like blinking LEDs, printf(), and JTAG debuggers, the bugs are getting expensive to find and the pressure to complete the project is at its highest. What can be done to make this phase of the project as short as possible? The answer: deploy defect-revealing, bug-squashing technology early on.

Code ReviewFormal code review is a well-known bug elimination methodology. This process asks development engineers to study each other’s code, comment on potential problems, decide which to fix or defer, and assign tasks of fixing and testing appropriately. This is one of the best methodologies for finding and eliminating bugs while they are still inexpensive to fix.

Formal code review is somewhat time- and labor-intensive, which is the principle reason why it is not always done by every development team. In an ideal world, code review would be focused on application specific issues, like timing scheme correction, guaranteed packet transfer, and data integrity preservation. Important but easily fixable static construct issues should really be taken care of prior to the point where the team sits down to read and discuss the code and make their decisions.

This is where automatic code inspection can be used very effectively-before formal code review. The Atollic TrueSTUDIO C/C++ IDE includes a MISRA-C checker, which can automatically apply formal coding rules to C code constructs using the MISRA®-C standard. The MISRA-C rules base is the product of hundreds of man-years of experience in writing safety critical C Code. MISRA was originally developed to meet the safety-critical needs of the automotive industry, but is now being applied as the definitive best practice coding standard for C development.

“What if I’m not writing safety critical code?” you ask. Well, you want your code to work as you expect it to work, regardless of whether or not your application requires formal safety critical certification. And following the MISRA rules will help you avoid common defects in software that may turn into bugs down the road.

MISRA-C rules and recommendation have been adopted only after being extensively discussed and voted on by a team of highly experienced experts in the field of embedded software development. The application of MISRA-C rules and recommendations has been shown to dramatically improve code quality and reliability, as well as providing a more consistent “readability” to the code. ii

Using a Code Inspection Tool in Day-to-Day Embedded DevelopmentThe key question at this point is whether or not you can use TrueSTUDIO static analysis tools quickly and reliably to derive useful information without spending a lot of time. The answer: absolutely.

One mouse click kicks off an exhaustive static analysis that applies the entire MISRA-C 2004 rules base and recommendations to the code. Of course, you may configure which code modules in your application that you want checked, which specific rules and recommendations you wish to apply, and other settings based on your needs and the way you work. One of the best things about this inspection utility is that you can do all of this without ever leaving the IDE. Apart from convenience, this offers additional productivity benefits, as we will see in the following section.

Finding defects early in

development is far cheaper than dealing

with them in debug

and test, or in the hands

of an unhappy

customer.

1616

Turning Inspection Results into Improved CodeNot all static code-checking tools based on MISRA-C are alike. Many static code-checking tools produce results similar to a list of compiler like warnings (so 1980s). TrueSTUDIO not only produces copious data, but the results are well organized into tables and charts, turning raw data into information that you can use at a glance.

The integration of TrueINSPECTOR into the IDE allows you to do things like go to the line of code in the editor that caused the rules violations by clicking on that

violation entry in the report, making for easy investigation and cross reference. Rules violations are accompanied by a specific example of “good” code and “bad” code in the TrueINSPECTOR output. Warning: Consistent use of TrueINSPECTOR may cause you to become a C programming “language lawyer” and safety critical power user.

Results of the static analysis can be easily exported from the IDE into standard file formats for record keeping, sharing among team members and use in formal code review.

How do you know the rules and recommendations of MISRA-C have really been checked in your code? TrueINSPECTOR itself has been rigorously checked by testing programs designed to specifically determine the efficacy of static code analysis tools. The results speak for themselves.

Benefits to the UserWhat do you, the developer, get out of all of this? After all, features and capabilities are all well and good, but how does this improve your software build process?

“Imagine all the people,” as John Lennon used to say. Using TrueSTUDIO static code analysis is like having the full weight of the MISRA-C committee reviewing your code—with all of their years of experience, the database of all the problems they have encountered with C code constructs, all the cumulative wisdom gained over the years. Your full confidence that static code constructs have been thoroughly examined, giving you an opportunity to address these prior to formal code review. This will speed up the process as discussion of an entire class of potentially dangerous software defects is eliminated prior to full team code reviews. This allows the team more time to be devoted to application-specific issues and design issues knowing full that they have a solid foundation built on sound C code constructs.

From the time a line of code is written, any bug that line may cause gets more expensive to

find as every day passes.

17

TECH REPORT

17

Turning Inspection Results into Improved CodeNot all static code-checking tools based on MISRA-C are alike. Many static code-checking tools produce results similar to a list of compiler like warnings (so 1980s). TrueSTUDIO not only produces copious data, but the results are well organized into tables and charts, turning raw data into information that you can use at a glance.

The integration of TrueINSPECTOR into the IDE allows you to do things like go to the line of code in the editor that caused the rules violations by clicking on that

violation entry in the report, making for easy investigation and cross reference. Rules violations are accompanied by a specific example of “good” code and “bad” code in the TrueINSPECTOR output. Warning: Consistent use of TrueINSPECTOR may cause you to become a C programming “language lawyer” and safety critical power user.

Results of the static analysis can be easily exported from the IDE into standard file formats for record keeping, sharing among team members and use in formal code review.

How do you know the rules and recommendations of MISRA-C have really been checked in your code? TrueINSPECTOR itself has been rigorously checked by testing programs designed to specifically determine the efficacy of static code analysis tools. The results speak for themselves.

Benefits to the UserWhat do you, the developer, get out of all of this? After all, features and capabilities are all well and good, but how does this improve your software build process?

“Imagine all the people,” as John Lennon used to say. Using TrueSTUDIO static code analysis is like having the full weight of the MISRA-C committee reviewing your code—with all of their years of experience, the database of all the problems they have encountered with C code constructs, all the cumulative wisdom gained over the years. Your full confidence that static code constructs have been thoroughly examined, giving you an opportunity to address these prior to formal code review. This will speed up the process as discussion of an entire class of potentially dangerous software defects is eliminated prior to full team code reviews. This allows the team more time to be devoted to application-specific issues and design issues knowing full that they have a solid foundation built on sound C code constructs.

From the time a line of code is written, any bug that line may cause gets more expensive to

find as every day passes.

1818

The probability of introducing a new bug

while attempting to bug fix, modify or extend a complex block of code is considerably

higher than trying to do the same thing

on less complex code blocks.

Don’t Ignore the Code Complexity FactorCode complexity is an important, yet often overlooked code metric. Complexity is something that every developer recognizes, but it is difficult to quantify. TrueINSPECTOR calculates cyclometric complexity metrics for your code, which are presented in easy-to-interpret, at-a-glance tables. Many companies have standards regarding code complexity. Such standards can be monitored and checked quickly and easily by TrueINSPECTOR, making sure that adherence to standards is positive and enforceable.

i Code Complete, June 2014, McConnell

ii MISRA, Development Guidelines for Vehicle

Based Software, November 1994 (PDF version 1.1,

January 2001), Rivett, “Emerging Software Best

Practice and how to be Compliant”, Proceedings

of the 6th International EAEC Congress July 1997

Code is “written one time, and read many times” as the saying goes. Studies have shown that code that is hard to read probably contains more defects that may hide serious bugs. It is certainly more time consuming, frustrating, and expensive to maintain, extend and change as the situation demands. Further, the probability of introducing a new bug while attempting to bug fix, modify or extend a complex block of code is considerably higher than trying to do the same thing on less complex code blocks.

Your Software Is Broken and You Need to Fix ItAll software has defects. Finding those defects before they become important bugs later can save time, frustration, and cost. Atollic TrueSTUDIO gives you another tool in your kit to bring attack the never-ending task of assuring software quality. Many of the perceived barriers to using such and approach are effectively addressed by ease of use and configuration, advanced presentation of results, and productivity benefits of having a tool within the IDE, where the development action takes place. At the end of the day, your code can only benefit by being constructed along the lines of time-tested methods and recommendations that have proved themselves in the most demanding embedded applications.

19

TECH REPORT

19

The probability of introducing a new bug

while attempting to bug fix, modify or extend a complex block of code is considerably

higher than trying to do the same thing

on less complex code blocks.

Don’t Ignore the Code Complexity FactorCode complexity is an important, yet often overlooked code metric. Complexity is something that every developer recognizes, but it is difficult to quantify. TrueINSPECTOR calculates cyclometric complexity metrics for your code, which are presented in easy-to-interpret, at-a-glance tables. Many companies have standards regarding code complexity. Such standards can be monitored and checked quickly and easily by TrueINSPECTOR, making sure that adherence to standards is positive and enforceable.

i Code Complete, June 2014, McConnell

ii MISRA, Development Guidelines for Vehicle

Based Software, November 1994 (PDF version 1.1,

January 2001), Rivett, “Emerging Software Best

Practice and how to be Compliant”, Proceedings

of the 6th International EAEC Congress July 1997

Code is “written one time, and read many times” as the saying goes. Studies have shown that code that is hard to read probably contains more defects that may hide serious bugs. It is certainly more time consuming, frustrating, and expensive to maintain, extend and change as the situation demands. Further, the probability of introducing a new bug while attempting to bug fix, modify or extend a complex block of code is considerably higher than trying to do the same thing on less complex code blocks.

Your Software Is Broken and You Need to Fix ItAll software has defects. Finding those defects before they become important bugs later can save time, frustration, and cost. Atollic TrueSTUDIO gives you another tool in your kit to bring attack the never-ending task of assuring software quality. Many of the perceived barriers to using such and approach are effectively addressed by ease of use and configuration, advanced presentation of results, and productivity benefits of having a tool within the IDE, where the development action takes place. At the end of the day, your code can only benefit by being constructed along the lines of time-tested methods and recommendations that have proved themselves in the most demanding embedded applications.

GRAND PRIZEClemens Valens

JANUARY 1 - ENTRY PHASE » Fill in and submit questionnaire. (Special bonus drawing for entries before March 15. Five contestants will win prizes worth $100 each.)

» Advance to the Submission Phase.

Click here to register and start the contest!

MARCH 1 - SUBMISSION PHASE » Contestants will produce a schematic using NXP’s products. (Special bonus drawing for schematics before April 15. Ten contestants win prizes worth $200 each.)

» Selected contestants will advance to the Completion Phase.

Click here to view the Submission Phase rules and instructions.

APRIL 15 - COMPLETION PHASE » Design kits will be sent to contestants selected from the Submission Phase.

» Entries must be submitted for judging by June 30.

AUGUST 1 - WINNERS ANNOUNCED » GRAND PRIZE: $3000 Prize Value

» FIRST PLACE: $2000 Prize Value

» SECOND PLACE: $1500 Prize Value

» THIRD PLACE: $1000 Prize Value

» HONORABLE MENTION: $500 Prize Value

NXP Logic Presents

I.D.E.A.2015 International Design Engineering Award

Join NXP and Mouser in the 2015 Big I.D.E.A.,

a design contest featuring NXP’s unique

product line-up across the spectrum,

including Dual Configurable Logic, Smart Analog,

MOSFETs, and Power.

Everyone has a chance to win thousands of

dollars worth of prizes, with awards at every

level of the competition.

The Big

Submission Phase

March 1Entry Phase

January 1Completion Phase

April 15Winners Announced

August 1

Click here to register and start the contest!

Click here to view the Submission Phase rules and instructions.

GRAND PRIZEClemens Valens

JANUARY 1 - ENTRY PHASE » Fill in and submit questionnaire. (Special bonus drawing for entries before March 15. Five contestants will win prizes worth $100 each.)

» Advance to the Submission Phase.

Click here to register and start the contest!

MARCH 1 - SUBMISSION PHASE » Contestants will produce a schematic using NXP’s products. (Special bonus drawing for schematics before April 15. Ten contestants win prizes worth $200 each.)

» Selected contestants will advance to the Completion Phase.

Click here to view the Submission Phase rules and instructions.

APRIL 15 - COMPLETION PHASE » Design kits will be sent to contestants selected from the Submission Phase.

» Entries must be submitted for judging by June 30.

AUGUST 1 - WINNERS ANNOUNCED » GRAND PRIZE: $3000 Prize Value

» FIRST PLACE: $2000 Prize Value

» SECOND PLACE: $1500 Prize Value

» THIRD PLACE: $1000 Prize Value

» HONORABLE MENTION: $500 Prize Value

NXP Logic Presents

I.D.E.A.2015 International Design Engineering Award

Join NXP and Mouser in the 2015 Big I.D.E.A.,

a design contest featuring NXP’s unique

product line-up across the spectrum,

including Dual Configurable Logic, Smart Analog,

MOSFETs, and Power.

Everyone has a chance to win thousands of

dollars worth of prizes, with awards at every

level of the competition.

The Big

Submission Phase

March 1Entry Phase

January 1Completion Phase

April 15Winners Announced

August 1

2222

Design ChallengesWhile drilling for oil and gas on oil rigs, the content of the extracted media needs to be analyzed in”rea1-time” to control pressure and drill pipe speed, as well as eliminate dangerous situations. The equipment also needs to operate both above and below water at varying temperatures and great depths.

Other requirements include:

» Maximum reliability, highest mean time to failure (MTTF)

» Highest possible operating temperature

» Withstand vibrations up to 30g

» Fast (low-latency) connection to a FPGA for serial interfaces

» Smallest possible size, because of very limited available space

» Storage of data on rugged media

» Second NOR-Flash device for maximum data retention and firmware update

Design Solutions » Data storage on mSATA-module

» TQ Systems QoriQ embedded module (Freescale P2020 QoriQ™P2 platform MCU) provides:

» A reliable NOR-Flash to boot from

» PCIexpress line for the connection to the mSATA device

» SPI-Bus for second NOR-Flash

» FPGA-interconnection via Local Bus for lowest latency

» 3x Ethernet interfaces (1x to measuring board, 1x for SW-update, 1x for data upload)

» ECC for DDR3 to detect and correct bit-failures

» Double-precision FPU for calculation of ADC-Data

» Industrial grade temperature range: -40°C to +85°C

» Ruggedized connector system for highest reliability

Save Design Time and MoneyTQMP2020 embedded modules with Freescale QorIQ MCU:

» Are the smallest in the industry without compromising quality and reliability

» Bring out all the processor signals to the TE connectors

» Can reduce development time by as much as 12 months

» The TQMP2020 module comes with a Freescale QoriQ™ Power Architecture® MCU and supports Linux and QNX operating systems

» The full-function STKP2020 Starter Kit is an easy and inexpensive platform to test and evaluate the TQMP2020 module

FOR MORE INFORMATION on the TQMP2020 and

other modules, or to purchase modules or starter kits:

www.convergencepromotions.com/TQ-USA.

TQ-USA is a brand of modules distributed in

North America by Convergence Promotions

RUGGED ENVIRONMENTS

Designing for

Embedded Products in Offshore Oil Rigs

www.convergencepromotions.com/TQ-USA.

23

TECH REPORT

23

Design ChallengesWhile drilling for oil and gas on oil rigs, the content of the extracted media needs to be analyzed in”rea1-time” to control pressure and drill pipe speed, as well as eliminate dangerous situations. The equipment also needs to operate both above and below water at varying temperatures and great depths.

Other requirements include:

» Maximum reliability, highest mean time to failure (MTTF)

» Highest possible operating temperature

» Withstand vibrations up to 30g

» Fast (low-latency) connection to a FPGA for serial interfaces

» Smallest possible size, because of very limited available space

» Storage of data on rugged media

» Second NOR-Flash device for maximum data retention and firmware update

Design Solutions » Data storage on mSATA-module

» TQ Systems QoriQ embedded module (Freescale P2020 QoriQ™P2 platform MCU) provides:

» A reliable NOR-Flash to boot from

» PCIexpress line for the connection to the mSATA device

» SPI-Bus for second NOR-Flash

» FPGA-interconnection via Local Bus for lowest latency

» 3x Ethernet interfaces (1x to measuring board, 1x for SW-update, 1x for data upload)

» ECC for DDR3 to detect and correct bit-failures

» Double-precision FPU for calculation of ADC-Data

» Industrial grade temperature range: -40°C to +85°C

» Ruggedized connector system for highest reliability

Save Design Time and MoneyTQMP2020 embedded modules with Freescale QorIQ MCU:

» Are the smallest in the industry without compromising quality and reliability

» Bring out all the processor signals to the TE connectors

» Can reduce development time by as much as 12 months

» The TQMP2020 module comes with a Freescale QoriQ™ Power Architecture® MCU and supports Linux and QNX operating systems

» The full-function STKP2020 Starter Kit is an easy and inexpensive platform to test and evaluate the TQMP2020 module

FOR MORE INFORMATION on the TQMP2020 and

other modules, or to purchase modules or starter kits:

www.convergencepromotions.com/TQ-USA.

TQ-USA is a brand of modules distributed in

North America by Convergence Promotions

RUGGED ENVIRONMENTS

Designing for

Embedded Products in Offshore Oil Rigs

http://www.kionix.comjchong@kionix.com

2424

SENSORS PLAY A KEY ROLE IN ENABLING MODERN ELECTRONICS

to more intelligently act within their environment. A cell phone, for

example, uses sensors to detect when a user is holding the device

to his or her ear so it can save power by turning off the display.

Through sensor fusion technology, where data from multiple

sensors is used in combination, even more advanced features

can be enabled.

Bringing Advanced Sensor Functionality to Consumer, Embedded, and Industrial Applications

Kionix, Inc. Worldwide Headquarters36 Thornwood DriveIthaca, NY 14850(607) 257-1080 (voice)(607) 257-1146 (fax)info@kionix.com

Kionix, Inc. is a ROHM Group Company

Sensors are popping up everywhere. A wide range of consumer electronics devices make extensive

use of sensors, including smartphones, tablets, and gaming equipment. To capitalize upon the growing gaming market, set-top box vendors are beginning to integrate sensor-based gaming capabilities into their systems. Perhaps the largest growth sector for intelligent sensing is the Internet of Things. This market is focused on extending awareness and connectivity everywhere and includes everything from wearable devices such as pedometers and portable medical equipment to complex industrial systems that can self-monitor themselves and their environment.

The benefits of embedding sensors are apparent, but designing in sensors presents multiple challenges to OEMs. Some sensors must be continuously monitored, putting a load on the system’s main applications processor. Data processing algorithms are often compute-intensive as well. Developers must minimize device size and power consumption. Finally, sensor design and algorithm development requires expertise an OEM may not have readily available in-house.

SENSOR HUBS

For many applications, a dedicated sensor hub provides a compact way to integrate sensor monitoring and analysis capabilities in a cost-effective manner that minimizes power consumption. As its name states, a sensor hub has the ability to accept inputs from multiple sensors such as accelerometers, gyroscopes, magnetometers, and pressure sensors. It also has an integrated MCU to handle real-time algorithm processing.

For existing systems, baseband and applications processors are power hungry components often called upon to do many tasks. Consider the use case of a cell phone running a video call with full video to the display while managing a cellular connection, streaming audio to a headset via Bluetooth, providing Wi-Fi hotspot capabilities to a tablet, and tracking the device’s GPS location. In many instances, it is advantageous to offload the real-time monitoring of sensors for features like orientation positioning and drop detection away from the applications processor to a more efficient sensor hub.

SENSOR FUNCTIONALITY

to Consumer, Embedded, and Industrial Applications

Bringing Advanced

A SENSOR HUB HAS THE ABILITY TO ACCEPT INPUTS FROM MULTIPLE SENSORS SUCH AS ACCELEROMETERS, GYROSCOPES, MAGNETOMETERS, AND PRESSURE SENSORS. IT ALSO HAS AN INTEGRATED MCU TO HANDLE REAL-TIME ALGORITHM PROCESSING.

Kionix is a ROHM Group Company http://www.kionix.com jchong@kionix.com

25

TECH REPORT

25

SENSORS PLAY A KEY ROLE IN ENABLING MODERN ELECTRONICS

to more intelligently act within their environment. A cell phone, for

example, uses sensors to detect when a user is holding the device

to his or her ear so it can save power by turning off the display.

Through sensor fusion technology, where data from multiple

sensors is used in combination, even more advanced features

can be enabled.

Bringing Advanced Sensor Functionality to Consumer, Embedded, and Industrial Applications

Kionix, Inc. Worldwide Headquarters36 Thornwood DriveIthaca, NY 14850(607) 257-1080 (voice)(607) 257-1146 (fax)info@kionix.com

Kionix, Inc. is a ROHM Group Company

Sensors are popping up everywhere. A wide range of consumer electronics devices make extensive

use of sensors, including smartphones, tablets, and gaming equipment. To capitalize upon the growing gaming market, set-top box vendors are beginning to integrate sensor-based gaming capabilities into their systems. Perhaps the largest growth sector for intelligent sensing is the Internet of Things. This market is focused on extending awareness and connectivity everywhere and includes everything from wearable devices such as pedometers and portable medical equipment to complex industrial systems that can self-monitor themselves and their environment.

The benefits of embedding sensors are apparent, but designing in sensors presents multiple challenges to OEMs. Some sensors must be continuously monitored, putting a load on the system’s main applications processor. Data processing algorithms are often compute-intensive as well. Developers must minimize device size and power consumption. Finally, sensor design and algorithm development requires expertise an OEM may not have readily available in-house.

SENSOR HUBS

For many applications, a dedicated sensor hub provides a compact way to integrate sensor monitoring and analysis capabilities in a cost-effective manner that minimizes power consumption. As its name states, a sensor hub has the ability to accept inputs from multiple sensors such as accelerometers, gyroscopes, magnetometers, and pressure sensors. It also has an integrated MCU to handle real-time algorithm processing.

For existing systems, baseband and applications processors are power hungry components often called upon to do many tasks. Consider the use case of a cell phone running a video call with full video to the display while managing a cellular connection, streaming audio to a headset via Bluetooth, providing Wi-Fi hotspot capabilities to a tablet, and tracking the device’s GPS location. In many instances, it is advantageous to offload the real-time monitoring of sensors for features like orientation positioning and drop detection away from the applications processor to a more efficient sensor hub.

SENSOR FUNCTIONALITY

to Consumer, Embedded, and Industrial Applications

Bringing Advanced

A SENSOR HUB HAS THE ABILITY TO ACCEPT INPUTS FROM MULTIPLE SENSORS SUCH AS ACCELEROMETERS, GYROSCOPES, MAGNETOMETERS, AND PRESSURE SENSORS. IT ALSO HAS AN INTEGRATED MCU TO HANDLE REAL-TIME ALGORITHM PROCESSING.

Kionix is a ROHM Group Company http://www.kionix.com jchong@kionix.com

2626Kionix, Inc. | 36 Thornwood Drive, Ithaca, NY 14850 | +1 (607) 257-1080 | info@kionix.com 3

When the device isn’t being actively used, it is valuable to power down the applications processor for intervals to save power. However, in order to detect use or measure movement (e.g., when someone picks up the device from a table or during use as a pedometer), the accelerom-eter needs to be continuously monitored. Here the sensor hub can manage the sensor data at an optimized low power rate while letting the main applications processor sleep. The impact of this power efficiency is especially noticeable for features like wake-on-approach where the device turns on when a person brings his or her hand close.

Additionally, in specialized applications that don’t require the sophistica-tion of a smartphone, the CPU integrated into the sensor hub may provide all the processing capabilities required. For example, a sensor hub like the KX23H from Kionix is built on the advanced computing capabilities of the ARM Cor-tex-M0 architecture. It offers the right level of processing performance to effi-ciently monitor multiple sensors and implement complex analysis algorithms. It also has enough headroom to perform all of the functionality required for a targeted application such as a pedometer or heart monitor without the need for an additional applications processor.

SENSOR FUSION

One of the primary benefits of using a sensor hub is the ability to perform what is known as sensor fusion. In its generic sense, sensor fusion refers to combining data from multiple sensors to yield information that cannot be captured by each sensor separately. A good example of sensor fusion is the use of two cameras to calculate the three-dimensional position of an object.

One way of looking at the “1+1>2” result of sensor fusion is to realize that every sensor has its blind spots. Consider the challenge of measuring motion or change in orientation. When an accelerometer is laying flat on a table at rest, the effect of gravity causes an acceleration to be detected in the Z direction with no force on the X or Y axes. If the accelerometer is tipped, the acceleration due to gravity is distributed on the X and Y axes as well, allowing one to measure motion and orientation. However, if the accelerometer is placed back on the table and then rotated while still remaining flat on the table, the acceleration due to gravi-ty always acts fully on the Z axis while the acceleration on the X and Y axes remain at zero. Thus the outputs remain the same and there is no indication of

Figure 1

Figure 2

Kionix, Inc. | 36 Thornwood Drive, Ithaca, NY 14850 | +1 (607) 257-1080 | info@kionix.com 3

When the device isn’t being actively used, it is valuable to power down the applications processor for intervals to save power. However, in order to detect use or measure movement (e.g., when someone picks up the device from a table or during use as a pedometer), the accelerom-eter needs to be continuously monitored. Here the sensor hub can manage the sensor data at an optimized low power rate while letting the main applications processor sleep. The impact of this power efficiency is especially noticeable for features like wake-on-approach where the device turns on when a person brings his or her hand close.

Additionally, in specialized applications that don’t require the sophistica-tion of a smartphone, the CPU integrated into the sensor hub may provide all the processing capabilities required. For example, a sensor hub like the KX23H from Kionix is built on the advanced computing capabilities of the ARM Cor-tex-M0 architecture. It offers the right level of processing performance to effi-ciently monitor multiple sensors and implement complex analysis algorithms. It also has enough headroom to perform all of the functionality required for a targeted application such as a pedometer or heart monitor without the need for an additional applications processor.

SENSOR FUSION

One of the primary benefits of using a sensor hub is the ability to perform what is known as sensor fusion. In its generic sense, sensor fusion refers to combining data from multiple sensors to yield information that cannot be captured by each sensor separately. A good example of sensor fusion is the use of two cameras to calculate the three-dimensional position of an object.

One way of looking at the “1+1>2” result of sensor fusion is to realize that every sensor has its blind spots. Consider the challenge of measuring motion or change in orientation. When an accelerometer is laying flat on a table at rest, the effect of gravity causes an acceleration to be detected in the Z direction with no force on the X or Y axes. If the accelerometer is tipped, the acceleration due to gravity is distributed on the X and Y axes as well, allowing one to measure motion and orientation. However, if the accelerometer is placed back on the table and then rotated while still remaining flat on the table, the acceleration due to gravi-ty always acts fully on the Z axis while the acceleration on the X and Y axes remain at zero. Thus the outputs remain the same and there is no indication of

Figure 1

Figure 2

When the device isn’t being actively used, it is valuable to power down the applications processor for intervals to save power. However, in order to detect use or measure movement (e.g., when someone picks up the device from a table or during use as a pedometer), the accelerometer needs to be continuously monitored. Here the sensor hub can manage the sensor data at an optimized low power rate while letting the main applications processor sleep. The impact of this power efficiency is especially noticeable for features like wake-on-approach where the device turns on when a person brings his or her hand close.

Additionally, in specialized applications that don’t require the sophistication of a smartphone, the CPU integrated into the sensor hub may provide all the processing capabilities required. For example, a sensor hub like the KX23H from Kionix is built on the advanced computing capabilities of the ARM Cortex-M0 architecture. It offers the right level of processing performance to efficiently monitor multiple sensors and implement complex analysis algorithms. It also has enough headroom to perform all of the functionality required for a targeted application such as a pedometer or heart monitor without the need for an additional applications processor.

SENSOR FUSION

One of the primary benefits of using a sensor hub is the ability to perform what is known as sensor fusion. In its generic sense, sensor fusion refers to combining data from multiple sensors to yield information that cannot be captured by each sensor separately. A good example of sensor fusion is the use of two cameras to calculate the three-dimensional position of an object.

control pointing devices and augmented reality displays.

THE KX23H SENSOR HUB

To help OEMs introduce advanced sensor functionality to consumer, embedded, and industrial designs, Kionix offers the KX23H sensor hub (Figure 3). Key features of the KX23H include:

HIGH PERFORMANCE—Built on a 32-bit ARM Cortex-M0 core running at 32MHz, the KX23H offloads real-time sensor fusion algorithms from a system’s main applications processor.

ON-CHIP ACCELEROMETER—Based on Kionix’s highest performance design, the accelerator integrated into the KX23H provides 16-bit resolution with a 256-byte FIFO/FILO buffer and very low power consumption down to 1uA. It also features FlexSet Performance Optimization, allowing dynamic adjustment of power and noise parameters to match the state and activity of the device.

POWER EFFICIENCY—In sleep mode, the ARM core consumes only 2.5uA. It also offers an efficient operational power of 1.5mA @ 32kHz and just 6mA @ 32MHz.

OPTIMAL LEVEL OF INTEGRATION—Designed for sensor monitoring and processing, the KX23H doesn’t integrate the extraneous capabilities of a general-purpose processor, like a USB port or large memory banks that add unnecessary cost and reduce power efficiency. In addition, because the accelerometer and Cortex-M0 are paired, they can work more efficiently than a CPU with an external accelerometer.

Kionix, Inc. | 36 Thornwood Drive, Ithaca, NY 14850 | +1 (607) 257-1080 | info@kionix.com 4

motion or change in orientation. In this situation, the accelerometer has a blind spot. The addition of data from gyroscope or magnetometer would allow one to compensate for this blind spot and identify the device’s motion and orientation.

Another example of how sensors can complement each other is an eCom-pass application. Many people think that the magnetic field points due north, but in reality the direction of the magnetic field varies across the globe and one has to hold a traditional compass level to indicate true north. A magnetometer alone can’t be used to compute true north unless it is held perfectly level. An ac-celerometer is typically used to measure tilt with respect to level and so paired with a magnetometer the two can accurately compute magnetic north.

One form of sensor fusion that is commonly used in portable devices is known as nine-axis sensor fusion. Nine-axis sensor fusion combines the tri-axis data from three different sensors—accelerometer, magnetometer, and gyro-scope—to accurately determine a device’s orientation and motion in 3-D space. In this way, the blind spots of each sensor are filled in using information from the other sensors. Nine-axis sensor fusion brings value to consumers by provid-ing a level of functionality that enables devices to accurately track themselves in 3-D space. With this accuracy, many new features can be introduced such remote control pointing devices and augmented reality displays.

THE KX23H SENSOR HUB

To help OEMs introduce advanced sensor functionality to consumer, embed-ded, and industrial designs, Kionix offers the KX23H sensor hub (Figure 3). Key features of the KX23H include:

• HIGH PERFORMANCE—Built on a 32-bit ARM Cortex-M0 core running at 32 MHz, the KX23H offloads real-time sensor fusion algorithms from a system’s main applications processor.

• ON-CHIP ACCELEROMETER—Based on Kionix’s highest performance design, the accelerator integrated into the KX23H provides 16-bit resolution with a 256-byte FIFO/FILO buffer and very low power consumption down to 1uA. It also features FlexSet Performance Optimization, allowing dynam-ic adjustment of power and noise parameters to match the state and activity of the device.

• POWER EFFICIENCY —In sleep mode, the ARM core consumes only 2.5 uA. It also offers an efficient operational power of 1.5 mA @ 32 kHz and just 6 mA @ 32 MHz.

• OPTIMAL LEVEL OF INTEGRATION—Designed for sensor monitoring and processing, the KX23H doesn’t integrate the extraneous capabilities of a general-purpose processor, like a USB port or large memory banks, that add unnecessary cost and reduce power efficiency. In addition, because the ac-

Figure 3

One way of looking at the “1+1>2” result of sensor fusion is to realize that every sensor has its blind spots. Consider the challenge of measuring motion or change in orientation. When an accelerometer is laying flat on a table at rest, the effect of gravity causes an acceleration to be detected in the Z direction with no force on the X- or Y-axes. If the accelerometer is tipped, the acceleration due to gravity is distributed on the X- and Y-axes as well, allowing one to measure motion and orientation. However, if the accelerometer is placed back on the table and then rotated while still remaining flat on the table, the acceleration due to gravity always acts fully on the Z-axis, while the acceleration on the X- and Y-axes remain at zero. Thus the outputs remain the same and there is no indication of motion or

change in orientation. In this situation, the accelerometer has a blind spot. The addition of data from gyroscope or magnetometer would allow one to compensate for this blind spot and identify the device’s motion and orientation.

Another example of how sensors can complement each other is an eCompass application. For a traditional compass, the magnetic rod is aligned with the earth’s magnetic field, and the user inherently levels the housing to obtain a stable reading for magnetic north. In an electronic compass, the signal is typically detected on three orthogonal magnetic sensing elements, and slight variations in tilt or level significantly alter the readings on the three axes and hence the interpretation of direction. An accelerometer is used to measure the tilt relative to gravity and so paired with a magnetometer the two can accurately compute magnetic north.

One form of sensor fusion that is commonly used in portable devices is known as nine-axis sensor fusion. Nine-axis sensor fusion combines the tri-axis data from three different sensors—accelerometer, magnetometer, and gyro- scope—to accurately determine a device’s orientation and motion in 3-D space. In this way, the blind spots of each sensor are filled in using information from the other sensors. Nine-axis sensor fusion brings value to consumers by providing a level of functionality that enables devices to accurately track themselves in 3-D space. With this accuracy, many new features can be introduced such remote

Figure 3.

Figure 2.

Figure 1.

27

TECH REPORT

27Kionix, Inc. | 36 Thornwood Drive, Ithaca, NY 14850 | +1 (607) 257-1080 | info@kionix.com 3

When the device isn’t being actively used, it is valuable to power down the applications processor for intervals to save power. However, in order to detect use or measure movement (e.g., when someone picks up the device from a table or during use as a pedometer), the accelerom-eter needs to be continuously monitored. Here the sensor hub can manage the sensor data at an optimized low power rate while letting the main applications processor sleep. The impact of this power efficiency is especially noticeable for features like wake-on-approach where the device turns on when a person brings his or her hand close.

Additionally, in specialized applications that don’t require the sophistica-tion of a smartphone, the CPU integrated into the sensor hub may provide all the processing capabilities required. For example, a sensor hub like the KX23H from Kionix is built on the advanced computing capabilities of the ARM Cor-tex-M0 architecture. It offers the right level of processing performance to effi-ciently monitor multiple sensors and implement complex analysis algorithms. It also has enough headroom to perform all of the functionality required for a targeted application such as a pedometer or heart monitor without the need for an additional applications processor.

SENSOR FUSION

One of the primary benefits of using a sensor hub is the ability to perform what is known as sensor fusion. In its generic sense, sensor fusion refers to combining data from multiple sensors to yield information that cannot be captured by each sensor separately. A good example of sensor fusion is the use of two cameras to calculate the three-dimensional position of an object.

One way of looking at the “1+1>2” result of sensor fusion is to realize that every sensor has its blind spots. Consider the challenge of measuring motion or change in orientation. When an accelerometer is laying flat on a table at rest, the effect of gravity causes an acceleration to be detected in the Z direction with no force on the X or Y axes. If the accelerometer is tipped, the acceleration due to gravity is distributed on the X and Y axes as well, allowing one to measure motion and orientation. However, if the accelerometer is placed back on the table and then rotated while still remaining flat on the table, the acceleration due to gravi-ty always acts fully on the Z axis while the acceleration on the X and Y axes remain at zero. Thus the outputs remain the same and there is no indication of

Figure 1

Figure 2

Kionix, Inc. | 36 Thornwood Drive, Ithaca, NY 14850 | +1 (607) 257-1080 | info@kionix.com 3

When the device isn’t being actively used, it is valuable to power down the applications processor for intervals to save power. However, in order to detect use or measure movement (e.g., when someone picks up the device from a table or during use as a pedometer), the accelerom-eter needs to be continuously monitored. Here the sensor hub can manage the sensor data at an optimized low power rate while letting the main applications processor sleep. The impact of this power efficiency is especially noticeable for features like wake-on-approach where the device turns on when a person brings his or her hand close.

Additionally, in specialized applications that don’t require the sophistica-tion of a smartphone, the CPU integrated into the sensor hub may provide all the processing capabilities required. For example, a sensor hub like the KX23H from Kionix is built on the advanced computing capabilities of the ARM Cor-tex-M0 architecture. It offers the right level of processing performance to effi-ciently monitor multiple sensors and implement complex analysis algorithms. It also has enough headroom to perform all of the functionality required for a targeted application such as a pedometer or heart monitor without the need for an additional applications processor.

SENSOR FUSION

One of the primary benefits of using a sensor hub is the ability to perform what is known as sensor fusion. In its generic sense, sensor fusion refers to combining data from multiple sensors to yield information that cannot be captured by each sensor separately. A good example of sensor fusion is the use of two cameras to calculate the three-dimensional position of an object.

One way of looking at the “1+1>2” result of sensor fusion is to realize that every sensor has its blind spots. Consider the challenge of measuring motion or change in orientation. When an accelerometer is laying flat on a table at rest, the effect of gravity causes an acceleration to be detected in the Z direction with no force on the X or Y axes. If the accelerometer is tipped, the acceleration due to gravity is distributed on the X and Y axes as well, allowing one to measure motion and orientation. However, if the accelerometer is placed back on the table and then rotated while still remaining flat on the table, the acceleration due to gravi-ty always acts fully on the Z axis while the acceleration on the X and Y axes remain at zero. Thus the outputs remain the same and there is no indication of

Figure 1

Figure 2

When the device isn’t being actively used, it is valuable to power down the applications processor for intervals to save power. However, in order to detect use or measure movement (e.g., when someone picks up the device from a table or during use as a pedometer), the accelerometer needs to be continuously monitored. Here the sensor hub can manage the sensor data at an optimized low power rate while letting the main applications processor sleep. The impact of this power efficiency is especially noticeable for features like wake-on-approach where the device turns on when a person brings his or her hand close.

Additionally, in specialized applications that don’t require the sophistication of a smartphone, the CPU integrated into the sensor hub may provide all the processing capabilities required. For example, a sensor hub like the KX23H from Kionix is built on the advanced computing capabilities of the ARM Cortex-M0 architecture. It offers the right level of processing performance to efficiently monitor multiple sensors and implement complex analysis algorithms. It also has enough headroom to perform all of the functionality required for a targeted application such as a pedometer or heart monitor without the need for an additional applications processor.

SENSOR FUSION

One of the primary benefits of using a sensor hub is the ability to perform what is known as sensor fusion. In its generic sense, sensor fusion refers to combining data from multiple sensors to yield information that cannot be captured by each sensor separately. A good example of sensor fusion is the use of two cameras to calculate the three-dimensional position of an object.

control pointing devices and augmented reality displays.

THE KX23H SENSOR HUB

To help OEMs introduce advanced sensor functionality to consumer, embedded, and industrial designs, Kionix offers the KX23H sensor hub (Figure 3). Key features of the KX23H include:

HIGH PERFORMANCE—Built on a 32-bit ARM Cortex-M0 core running at 32MHz, the KX23H offloads real-time sensor fusion algorithms from a system’s main applications processor.

ON-CHIP ACCELEROMETER—Based on Kionix’s highest performance design, the accelerator integrated into the KX23H provides 16-bit resolution with a 256-byte FIFO/FILO buffer and very low power consumption down to 1uA. It also features FlexSet Performance Optimization, allowing dynamic adjustment of power and noise parameters to match the state and activity of the device.

POWER EFFICIENCY—In sleep mode, the ARM core consumes only 2.5uA. It also offers an efficient operational power of 1.5mA @ 32kHz and just 6mA @ 32MHz.

OPTIMAL LEVEL OF INTEGRATION—Designed for sensor monitoring and processing, the KX23H doesn’t integrate the extraneous capabilities of a general-purpose processor, like a USB port or large memory banks that add unnecessary cost and reduce power efficiency. In addition, because the accelerometer and Cortex-M0 are paired, they can work more efficiently than a CPU with an external accelerometer.

Kionix, Inc. | 36 Thornwood Drive, Ithaca, NY 14850 | +1 (607) 257-1080 | info@kionix.com 4

motion or change in orientation. In this situation, the accelerometer has a blind spot. The addition of data from gyroscope or magnetometer would allow one to compensate for this blind spot and identify the device’s motion and orientation.

Another example of how sensors can complement each other is an eCom-pass application. Many people think that the magnetic field points due north, but in reality the direction of the magnetic field varies across the globe and one has to hold a traditional compass level to indicate true north. A magnetometer alone can’t be used to compute true north unless it is held perfectly level. An ac-celerometer is typically used to measure tilt with respect to level and so paired with a magnetometer the two can accurately compute magnetic north.

One form of sensor fusion that is commonly used in portable devices is known as nine-axis sensor fusion. Nine-axis sensor fusion combines the tri-axis data from three different sensors—accelerometer, magnetometer, and gyro-scope—to accurately determine a device’s orientation and motion in 3-D space. In this way, the blind spots of each sensor are filled in using information from the other sensors. Nine-axis sensor fusion brings value to consumers by provid-ing a level of functionality that enables devices to accurately track themselves in 3-D space. With this accuracy, many new features can be introduced such remote control pointing devices and augmented reality displays.

THE KX23H SENSOR HUB

To help OEMs introduce advanced sensor functionality to consumer, embed-ded, and industrial designs, Kionix offers the KX23H sensor hub (Figure 3). Key features of the KX23H include:

• HIGH PERFORMANCE—Built on a 32-bit ARM Cortex-M0 core running at 32 MHz, the KX23H offloads real-time sensor fusion algorithms from a system’s main applications processor.

• ON-CHIP ACCELEROMETER—Based on Kionix’s highest performance design, the accelerator integrated into the KX23H provides 16-bit resolution with a 256-byte FIFO/FILO buffer and very low power consumption down to 1uA. It also features FlexSet Performance Optimization, allowing dynam-ic adjustment of power and noise parameters to match the state and activity of the device.

• POWER EFFICIENCY —In sleep mode, the ARM core consumes only 2.5 uA. It also offers an efficient operational power of 1.5 mA @ 32 kHz and just 6 mA @ 32 MHz.

• OPTIMAL LEVEL OF INTEGRATION—Designed for sensor monitoring and processing, the KX23H doesn’t integrate the extraneous capabilities of a general-purpose processor, like a USB port or large memory banks, that add unnecessary cost and reduce power efficiency. In addition, because the ac-

Figure 3

One way of looking at the “1+1>2” result of sensor fusion is to realize that every sensor has its blind spots. Consider the challenge of measuring motion or change in orientation. When an accelerometer is laying flat on a table at rest, the effect of gravity causes an acceleration to be detected in the Z direction with no force on the X- or Y-axes. If the accelerometer is tipped, the acceleration due to gravity is distributed on the X- and Y-axes as well, allowing one to measure motion and orientation. However, if the accelerometer is placed back on the table and then rotated while still remaining flat on the table, the acceleration due to gravity always acts fully on the Z-axis, while the acceleration on the X- and Y-axes remain at zero. Thus the outputs remain the same and there is no indication of motion or

change in orientation. In this situation, the accelerometer has a blind spot. The addition of data from gyroscope or magnetometer would allow one to compensate for this blind spot and identify the device’s motion and orientation.

Another example of how sensors can complement each other is an eCompass application. For a traditional compass, the magnetic rod is aligned with the earth’s magnetic field, and the user inherently levels the housing to obtain a stable reading for magnetic north. In an electronic compass, the signal is typically detected on three orthogonal magnetic sensing elements, and slight variations in tilt or level significantly alter the readings on the three axes and hence the interpretation of direction. An accelerometer is used to measure the tilt relative to gravity and so paired with a magnetometer the two can accurately compute magnetic north.

One form of sensor fusion that is commonly used in portable devices is known as nine-axis sensor fusion. Nine-axis sensor fusion combines the tri-axis data from three different sensors—accelerometer, magnetometer, and gyro- scope—to accurately determine a device’s orientation and motion in 3-D space. In this way, the blind spots of each sensor are filled in using information from the other sensors. Nine-axis sensor fusion brings value to consumers by providing a level of functionality that enables devices to accurately track themselves in 3-D space. With this accuracy, many new features can be introduced such remote

Figure 3.

Figure 2.

Figure 1.

2828

SMALLER FOOTPRINT—The KX23H offers a great deal of functionality in a small 3mm x 3mm x 0.9mm package.

OVERCOMING THE CHALLENGE OF SOFTWARE

Traditionally, OEMs have had to design sensor algorithms from scratch or port generic code to their chosen applications processor. This means OEMs are responsible for developing, testing, and verifying their own monitoring and analysis algorithms. This presents a challenge for many OEMs that don’t have their own design resources or in-house expertise. This is typical for many companies in the IoT space that are still emerging and don’t have the volumes to justify internal algorithm development. The ability to partner with an expert and offload sensor design can substantially reduce system architecture and design complexity.

To help speed time-to-market, Kionix offers a full suite of sensor software, ranging from device drivers for Windows, Android, and Open Source systems to its extensive Sensor Fusion Library that will be optimized for the ARM Cortex-M0 architecture of the KX23H. Kionix is an expert in motion sensing for a wide range of application nodes and has developed a comprehensive library of application-specific algorithms based on its extensive in-field experience, including:

SCREEN ORIENTATION—Signals when the user has rotated the device and the screen orientation should be changed.

FREE FALL—Determines when a device is falling (Figure4). This is important for devices like laptops that have a hard disk drive. If falling can be identified in time, the hard drive can be parked to minimize damage. It can also be used to assist in detection and trigger an alert when a person falls. Lastly, it can be used to monitor for drops for purposes such as managing warranty claims.

PEDOMETRY—Accurately measures steps taken by a user.

MOTION WAKE UP—Many devices can be put to sleep when they are not in motion. With this feature, systems can quickly wake once a user picks up the device.

TAP/DOUBLE TAP DETECTION—Detect the direction of a user tapping on a device and identify whether it is a single or double tap.

While many of these algorithms may seem straightforward to implement, there are subtle issues that can significantly impact the customer experience. For example, the difference between whether the device is flat or vertical can affect the screen orientation algorithm. In such cases, special care must be taken to improve user experience and responsiveness while preventing jitter or false identification of orientation. In other cases, some device functions might interfere with motion detection. For example, the vibration from certain

ring tones from the speaker can, to an accelerometer, be quite similar to a finger tapping, thus triggering false tap events. Only through careful algorithm design can these situations be minimized and the user experience maintained.

THE PATH TO ASIC

One of the key advantages Kionix brings to high-volume OEMs is a clear path to cost down the sensor hub to a custom device. An ASIC is often the lowest cost and most power-efficient route to implementing sensor-based functionality. However, ASIC design typically takes between 12 and 24 months and assumes that an OEM already has most of the IP in place to create the ASIC. This adds a delay that can be detrimental in the fast-changing consumer electronics industry by making OEMs late to market when introducing innovative new features or optimizations.

Instead, OEMs can lead the market by quickly implementing new features using the flexibility and programmability of the KX23H. This has the advantage of enabling OEMs to test features in the market and optimize them. Furthermore, any unexpected issues that arise can be addressed before the design has been committed to silicon.

Once a design is stable, Kionix is able to cost-down an OEM’s design to create a custom device that achieves power and cost savings not possible with a software-based MCU design. Thus, OEM’s are able to leverage the benefits of both low-risk design, a quick-to-

market approach, as well as high-volume economies of scale.

OEMs can also leverage Kionix’s comprehensive portfolio of sensors that share a common sensor fusion library. Kionix sensors offer excellent noise levels, accuracy, and robustness to ensure the best reliability and performance for today’s demanding applications.

To further simplify design, Kionix’s roadmap for its sensor hub products includes integrating more of the sensor solution for its customers over time. With its strong customer support team, Kionix is able to deliver turnkey solutions, both for new designs and existing applications. This includes working with OEMs to develop custom algorithms that can be preloaded onto the sensor hub during manufacturing calibration. In this way, OEMs can focus on their value-add rather than having to reinvent their own sensor IP.

With the KX23H sensor hub, OEMs can leverage the flexibility and programmability of the ARM Cortex-M0 paired with the high performance of its embedded accelerometer to bring new features and functionality to market quickly and at a lower risk. The KX23H is ideal for augmenting the existing capabilities of a system at a lower cost and power footprint via the sensor hub approach, as well as providing main processor functionality to those targeted applications requiring motion processing and computation capabilities in a small, low power package.

Kionix, Inc. | 36 Thornwood Drive, Ithaca, NY 14850 | +1 (607) 257-1080 | info@kionix.com 5

celerometer and Cortex-M0 are paired, they can work more efficiently than a CPU with an external accelerometer.

• SMALLER FOOTPRINT—The KX23H offers a great deal of functionality in a small 3mm x 3mm x 0.9mm package.

OVERCOMING THE CHALLENGE OF SOFTWARE

Traditionally, OEMs have had to design sensor algorithms from scratch or port generic code to their chosen applications processor. This means OEMs are responsible for developing, testing, and verifying their own monitoring and analysis algorithms. This presents a challenge for many OEMs that don’t have their own design resources or in-house expertise. This is typical for many companies in the IoT space that are still emerging and don’t have the volumes to justify internal algorithm development. The ability to partner with an expert and offload sensor design can substantially reduce system architecture and design complexity.

To help speed time-to-market, Kionix offers a full suite of sensor software, ranging from device drivers for Windows, Android, and Open Source sys-tems to its extensive Sensor Fusion Library that will be optimized for the ARM Cortex-M0 architecture of the KX23H. Kionix is an expert in motion sensing for a wide range of application nodes and has developed a comprehensive library of application-specific algorithms based on its extensive in-field experience, including:

• SCREEN ORIENTATION—Signals when the user has rotated the device and the screen orientation should be changed.

• FREE FALL—Determines when a device is falling (Figure 4). This is im-portant for devices like laptops that have a hard disk drive. If falling can be identified in time, the hard drive can be parked to minimize damage. It can also be used to detect and trigger an alert when a person falls.

• PEDOMETRY—Accurately measures steps taken by a user.

• MOTION WAKE UP—Many devices can be put to sleep when they are not in motion. With this feature, systems can quickly wake once a user picks up the device.

• TAP/DOUBLE TAP DETECTION—Detect the direction of a user tapping on a device and identify whether it is a single or double tap.

Figure 4

ONCE A DESIGN IS STABLE, KIONIX IS ABLE TO COST-DOWN AN OEM’S DESIGN TO CREATE A CUSTOM DEVICE THAT ACHIEVES POWER AND COST SAVINGS NOT POSSIBLE WITH A SOFTWARE-BASED MCU DESIGN.

Figure 4.

29

TECH REPORT

29

SMALLER FOOTPRINT—The KX23H offers a great deal of functionality in a small 3mm x 3mm x 0.9mm package.

OVERCOMING THE CHALLENGE OF SOFTWARE

Traditionally, OEMs have had to design sensor algorithms from scratch or port generic code to their chosen applications processor. This means OEMs are responsible for developing, testing, and verifying their own monitoring and analysis algorithms. This presents a challenge for many OEMs that don’t have their own design resources or in-house expertise. This is typical for many companies in the IoT space that are still emerging and don’t have the volumes to justify internal algorithm development. The ability to partner with an expert and offload sensor design can substantially reduce system architecture and design complexity.

To help speed time-to-market, Kionix offers a full suite of sensor software, ranging from device drivers for Windows, Android, and Open Source systems to its extensive Sensor Fusion Library that will be optimized for the ARM Cortex-M0 architecture of the KX23H. Kionix is an expert in motion sensing for a wide range of application nodes and has developed a comprehensive library of application-specific algorithms based on its extensive in-field experience, including:

SCREEN ORIENTATION—Signals when the user has rotated the device and the screen orientation should be changed.

FREE FALL—Determines when a device is falling (Figure4). This is important for devices like laptops that have a hard disk drive. If falling can be identified in time, the hard drive can be parked to minimize damage. It can also be used to assist in detection and trigger an alert when a person falls. Lastly, it can be used to monitor for drops for purposes such as managing warranty claims.

PEDOMETRY—Accurately measures steps taken by a user.

MOTION WAKE UP—Many devices can be put to sleep when they are not in motion. With this feature, systems can quickly wake once a user picks up the device.

TAP/DOUBLE TAP DETECTION—Detect the direction of a user tapping on a device and identify whether it is a single or double tap.

While many of these algorithms may seem straightforward to implement, there are subtle issues that can significantly impact the customer experience. For example, the difference between whether the device is flat or vertical can affect the screen orientation algorithm. In such cases, special care must be taken to improve user experience and responsiveness while preventing jitter or false identification of orientation. In other cases, some device functions might interfere with motion detection. For example, the vibration from certain

ring tones from the speaker can, to an accelerometer, be quite similar to a finger tapping, thus triggering false tap events. Only through careful algorithm design can these situations be minimized and the user experience maintained.

THE PATH TO ASIC

One of the key advantages Kionix brings to high-volume OEMs is a clear path to cost down the sensor hub to a custom device. An ASIC is often the lowest cost and most power-efficient route to implementing sensor-based functionality. However, ASIC design typically takes between 12 and 24 months and assumes that an OEM already has most of the IP in place to create the ASIC. This adds a delay that can be detrimental in the fast-changing consumer electronics industry by making OEMs late to market when introducing innovative new features or optimizations.

Instead, OEMs can lead the market by quickly implementing new features using the flexibility and programmability of the KX23H. This has the advantage of enabling OEMs to test features in the market and optimize them. Furthermore, any unexpected issues that arise can be addressed before the design has been committed to silicon.

Once a design is stable, Kionix is able to cost-down an OEM’s design to create a custom device that achieves power and cost savings not possible with a software-based MCU design. Thus, OEM’s are able to leverage the benefits of both low-risk design, a quick-to-

market approach, as well as high-volume economies of scale.

OEMs can also leverage Kionix’s comprehensive portfolio of sensors that share a common sensor fusion library. Kionix sensors offer excellent noise levels, accuracy, and robustness to ensure the best reliability and performance for today’s demanding applications.

To further simplify design, Kionix’s roadmap for its sensor hub products includes integrating more of the sensor solution for its customers over time. With its strong customer support team, Kionix is able to deliver turnkey solutions, both for new designs and existing applications. This includes working with OEMs to develop custom algorithms that can be preloaded onto the sensor hub during manufacturing calibration. In this way, OEMs can focus on their value-add rather than having to reinvent their own sensor IP.

With the KX23H sensor hub, OEMs can leverage the flexibility and programmability of the ARM Cortex-M0 paired with the high performance of its embedded accelerometer to bring new features and functionality to market quickly and at a lower risk. The KX23H is ideal for augmenting the existing capabilities of a system at a lower cost and power footprint via the sensor hub approach, as well as providing main processor functionality to those targeted applications requiring motion processing and computation capabilities in a small, low power package.

Kionix, Inc. | 36 Thornwood Drive, Ithaca, NY 14850 | +1 (607) 257-1080 | info@kionix.com 5

celerometer and Cortex-M0 are paired, they can work more efficiently than a CPU with an external accelerometer.

• SMALLER FOOTPRINT—The KX23H offers a great deal of functionality in a small 3mm x 3mm x 0.9mm package.

OVERCOMING THE CHALLENGE OF SOFTWARE

Traditionally, OEMs have had to design sensor algorithms from scratch or port generic code to their chosen applications processor. This means OEMs are responsible for developing, testing, and verifying their own monitoring and analysis algorithms. This presents a challenge for many OEMs that don’t have their own design resources or in-house expertise. This is typical for many companies in the IoT space that are still emerging and don’t have the volumes to justify internal algorithm development. The ability to partner with an expert and offload sensor design can substantially reduce system architecture and design complexity.

To help speed time-to-market, Kionix offers a full suite of sensor software, ranging from device drivers for Windows, Android, and Open Source sys-tems to its extensive Sensor Fusion Library that will be optimized for the ARM Cortex-M0 architecture of the KX23H. Kionix is an expert in motion sensing for a wide range of application nodes and has developed a comprehensive library of application-specific algorithms based on its extensive in-field experience, including:

• SCREEN ORIENTATION—Signals when the user has rotated the device and the screen orientation should be changed.

• FREE FALL—Determines when a device is falling (Figure 4). This is im-portant for devices like laptops that have a hard disk drive. If falling can be identified in time, the hard drive can be parked to minimize damage. It can also be used to detect and trigger an alert when a person falls.

• PEDOMETRY—Accurately measures steps taken by a user.

• MOTION WAKE UP—Many devices can be put to sleep when they are not in motion. With this feature, systems can quickly wake once a user picks up the device.

• TAP/DOUBLE TAP DETECTION—Detect the direction of a user tapping on a device and identify whether it is a single or double tap.

Figure 4

ONCE A DESIGN IS STABLE, KIONIX IS ABLE TO COST-DOWN AN OEM’S DESIGN TO CREATE A CUSTOM DEVICE THAT ACHIEVES POWER AND COST SAVINGS NOT POSSIBLE WITH A SOFTWARE-BASED MCU DESIGN.

Figure 4.

jchong@kionix.comhttp://www.kionix.com

3030

MOBILE APPLICATIONS

BEING ABLE TO DETERMINE A DEVICE’S

position and movement is becoming a standard

feature in many portable systems. Systems such

as cell phones and tablets use the 6-axis data from

accelerometers and magnetometers to enable

key functions that make consumer electronics

interfaces easier and more intuitive to use.

Next-generation devices are moving to 9-axis

sensor fusion employing gyroscopic capabilities to

further improve the user experience. For example,

a health monitor can track users more accurately

when it can differentiate between activities like

walking, swimming, and running. Gyroscopic data

also enables new interface capabilities, such as

gesture recognition, where a user can flick his or

her wrist to bring up the display.

Bringing Advanced Sensor Functionality to Consumer, Embedded, and Industrial Applications

Kionix, Inc. Worldwide Headquarters36 Thornwood DriveIthaca, NY 14850(607) 257-1080 (voice)(607) 257-1146 (fax)info@kionix.com

Kionix, Inc. is a ROHM Group Company

with Magnetic Gyroscopes

Expanding the Functionality of

By John Chong, Vice President of Product and Business Development, Kionix

Kionix is a ROHM Group Company http://www.kionix.com jchong@kionix.com

31

TECH REPORT

31

MOBILE APPLICATIONS

BEING ABLE TO DETERMINE A DEVICE’S

position and movement is becoming a standard

feature in many portable systems. Systems such

as cell phones and tablets use the 6-axis data from

accelerometers and magnetometers to enable

key functions that make consumer electronics

interfaces easier and more intuitive to use.

Next-generation devices are moving to 9-axis

sensor fusion employing gyroscopic capabilities to

further improve the user experience. For example,

a health monitor can track users more accurately

when it can differentiate between activities like

walking, swimming, and running. Gyroscopic data

also enables new interface capabilities, such as

gesture recognition, where a user can flick his or

her wrist to bring up the display.

Bringing Advanced Sensor Functionality to Consumer, Embedded, and Industrial Applications

Kionix, Inc. Worldwide Headquarters36 Thornwood DriveIthaca, NY 14850(607) 257-1080 (voice)(607) 257-1146 (fax)info@kionix.com

Kionix, Inc. is a ROHM Group Company

with Magnetic Gyroscopes

Expanding the Functionality of

By John Chong, Vice President of Product and Business Development, Kionix

Kionix is a ROHM Group Company http://www.kionix.com jchong@kionix.com

3232

MANY USES HAVE BEEN DEVELOPED FOR ACCELEROMETERS, INCLUDING INCREASING POWER EFFICIENCY.

Figure 2.

Figure 1.

Many consumer electronics devices can benefit from 9-axis sensor fusion (see Figure 1).

For example, cell phones, gaming systems, health monitors, and wearable electronics could implement high value-added functions such as three-dimension gesture recognition, relative positioning to other users, and motion tracking. To achieve this, devices need 9-axis sensor data and the ability to process it efficiently.

Note that many of these systems do not require the same level of precision that a fighter jet needs to stay in flight. Rather, they need a cost-effective 9-axis sensor fusion implementation that

matches the needs of consumer applications to improve the user experience. With the availability of new MagGyros, such as the KMX62G from Kionix, developers can use 6-axis sensors (accelerometer + magnetometer) to provide the equivalent of 9-axis output (accelerometer + magnetometer + gyroscope) while reducing energy consumption by as much as 90%.

9-AXIS SENSOR FUSION

Using 9-axis sensor fusion—data from an accelerometer, magnetometer, and gyroscope—systems can accurately position themselves in the world, including inclination and orientation as well as changes in position and rotation (see Figure 2). Effectively, sensor fusion uses multiple sensors to fill in the blind spots of individual sensors.

Accelerometers:

An accelerometer measures acceleration in 3 axes. It is important to note that gravity is a form of acceleration, and so an object at rest still has a downward acceleration. This fact enables an accelerometer to determine how a system is oriented relative to “down.”

This information can also be used to determine how level a system is. Many handheld and portable devices use orientation to adjust the display to match how the user is holding the device. Devices commonly using this functionality include cell phones, tablets, and laptops.

Another unique feature of an accelerometer is the ability to tell if a device has been dropped. For example, when a laptop is resting on a table, it detects 1 G. However, if the laptop is knocked off the table, it will detect 0 G as it falls to the floor (similar to how zero-gravity or weightless is simulated by putting a plane into freefall). This enables a system to detect if it is falling and at imminent risk of collision. In the case of a laptop, the system can proactively park

the hard drive head in a safe position to prevent damage to the drive or data.

Many other uses have been developed for accelerometers, including increasing power efficiency. Consider that a fitness tracker that has not moved in a few minutes is likely not being used. The device can then power itself down. This simple use case can substantially improve battery life across a wide range of applications.

Magnetometers:

A magnetometer measures magnetic fields. It can be used in a compass to identify the Earth’s magnetic field. By analyzing the magnitude, direction and rate of change of the detected field, a magnetometer can be used to find the direction of magnetic north.

As a standalone sensor, however, a magnetometer is typically unable to identify the direction of magnetic north. A traditional compass uses a magnetic rod with a bearing that restricts the rod to align in one dimension; proper leveling of the housing allows the user to best align the rod with magnetic north. In an electronic magnetometer, 3-axis of sensing is used to compute the direction and magnitude of the magnetic field. Any tilt between the magnetometer axes and the measured signal will cause error in the reported direction of the field.

A handheld device can use both an accelerometer and magnetometer to determine magnetic north. This is achieved by identifying the tilt of the device using the accelerometer and using this information to supplement the reading from the magnetometer.

EFFECTIVELY, SENSOR FUSION USES MULTIPLE SENSORS TO FILL IN THE BLIND SPOTS OF INDIVIDUAL SENSORS.

33

TECH REPORT

33

MANY USES HAVE BEEN DEVELOPED FOR ACCELEROMETERS, INCLUDING INCREASING POWER EFFICIENCY.

Figure 2.

Figure 1.

Many consumer electronics devices can benefit from 9-axis sensor fusion (see Figure 1).

For example, cell phones, gaming systems, health monitors, and wearable electronics could implement high value-added functions such as three-dimension gesture recognition, relative positioning to other users, and motion tracking. To achieve this, devices need 9-axis sensor data and the ability to process it efficiently.

Note that many of these systems do not require the same level of precision that a fighter jet needs to stay in flight. Rather, they need a cost-effective 9-axis sensor fusion implementation that

matches the needs of consumer applications to improve the user experience. With the availability of new MagGyros, such as the KMX62G from Kionix, developers can use 6-axis sensors (accelerometer + magnetometer) to provide the equivalent of 9-axis output (accelerometer + magnetometer + gyroscope) while reducing energy consumption by as much as 90%.

9-AXIS SENSOR FUSION

Using 9-axis sensor fusion—data from an accelerometer, magnetometer, and gyroscope—systems can accurately position themselves in the world, including inclination and orientation as well as changes in position and rotation (see Figure 2). Effectively, sensor fusion uses multiple sensors to fill in the blind spots of individual sensors.

Accelerometers:

An accelerometer measures acceleration in 3 axes. It is important to note that gravity is a form of acceleration, and so an object at rest still has a downward acceleration. This fact enables an accelerometer to determine how a system is oriented relative to “down.”

This information can also be used to determine how level a system is. Many handheld and portable devices use orientation to adjust the display to match how the user is holding the device. Devices commonly using this functionality include cell phones, tablets, and laptops.

Another unique feature of an accelerometer is the ability to tell if a device has been dropped. For example, when a laptop is resting on a table, it detects 1 G. However, if the laptop is knocked off the table, it will detect 0 G as it falls to the floor (similar to how zero-gravity or weightless is simulated by putting a plane into freefall). This enables a system to detect if it is falling and at imminent risk of collision. In the case of a laptop, the system can proactively park

the hard drive head in a safe position to prevent damage to the drive or data.

Many other uses have been developed for accelerometers, including increasing power efficiency. Consider that a fitness tracker that has not moved in a few minutes is likely not being used. The device can then power itself down. This simple use case can substantially improve battery life across a wide range of applications.

Magnetometers:

A magnetometer measures magnetic fields. It can be used in a compass to identify the Earth’s magnetic field. By analyzing the magnitude, direction and rate of change of the detected field, a magnetometer can be used to find the direction of magnetic north.

As a standalone sensor, however, a magnetometer is typically unable to identify the direction of magnetic north. A traditional compass uses a magnetic rod with a bearing that restricts the rod to align in one dimension; proper leveling of the housing allows the user to best align the rod with magnetic north. In an electronic magnetometer, 3-axis of sensing is used to compute the direction and magnitude of the magnetic field. Any tilt between the magnetometer axes and the measured signal will cause error in the reported direction of the field.

A handheld device can use both an accelerometer and magnetometer to determine magnetic north. This is achieved by identifying the tilt of the device using the accelerometer and using this information to supplement the reading from the magnetometer.

EFFECTIVELY, SENSOR FUSION USES MULTIPLE SENSORS TO FILL IN THE BLIND SPOTS OF INDIVIDUAL SENSORS.

3434

Figure 4.

Gyroscopes:

An accelerometer at rest on a desk can sense that the device is at rest. However, if you rotate the system with the accelerometer at the center of the rotation, the accelerometer will not be able to detect the movement. This is effectively a blind spot for the accelerometer. If the system has access to a gyroscope, the gyroscope can detect the rotation.

Unfortunately, implementing gyroscopic technology is difficult to justify for many applications. Of the three types of sensors, gyroscopes are the largest and most expensive. Although the sensor provides useful information, too often the cost, size, and/or power consumption

of a physical gyroscope exceeds the value it represents. For these reasons, developers have often had to limit position and orientation functionality to the 6-axis data provided by an accelerometer plus magnetometer.

THE KMX62G MAGGYRO

Today, advances in software algorithms and low-noise, low-latency sensors make it possible to simulate gyroscopic output using data from only an accelerometer and magnetometer. This is the approach used in MagGyros, where the system computes a device’s rotational direction and speed based on knowledge of its previous position. In other words, when a system can track orientation over time, it can extrapolate rotational data.

To enable developers to bring gyroscopic capabilities to a wide range of new applications, Kionix offers the KMX62G MagGyro. The KMX62G takes Kionix’s KMX62 acceleromator/magnetometer and enhances it with industry-leading sensor fusion software and auto-calibration algorithms (see Figure 3). This makes the KMX62G more than just an acceleromator/magnetometer. It is the industry’s first highly accurate gyroscopic emulator, providing 9-axis positioning capabilities.

THE KMX62G OFFERS SEVERAL KEY BENEFITS TO OEMS:

The Right Level of Accuracy: Because it estimates rather than directly measures rotational speed, the accuracy of a MagGyro is not as high as that of a physical gyroscope. However, gyroscopic accuracy comes into play most often when distance is involved, such as when flying a plane: being off by one degree over 100 miles can put you far off course. Close activities such as gesture recognition, on the other hand, can tolerate relatively large errors. Thus, consumer electronics applications focused on the user experience do not require the scientific levels of accuracy of a physical gyroscope.

Power: A physical gyroscope is always resonating, which means it continuously draws power. Typically, a physical gyroscope operates at 2.5 V or greater and consumes 4000 to 7000 μA (see Figure 4). Note that this is for the gyroscope alone.

With the KMX62G MagGyro, the accelerometer + magnetometer sensors consume 450 μA. Even when the power required for the microcontroller executing the software algorithms is taken into account (~500 μA), the entire MagGyro operation is less than 1000 μA. Thus, the KMX62G achieves 5X to 10X better energy efficiency compared to a physical gyroscope.

Start Up Time: Because they consume so much power, systems often turn off the gyroscope as often as they can. However, a physical gyroscope can take between 50 and 100ms for the output to stabilize. Thus, a power-efficient implementation with a physical gyroscope results in slow responsiveness that can negatively impact the user experience.

Figure 3.

Kionix, Inc. ©2013 All Rights Reserved

Power  

#1  Reason  •  1/5th  to  1/10th  of  

power  consump,on  

of  physical  gyro    

Opera'on  Voltage  •  Gyro  (typ)    >  2.5V  •  MagGyro  (typ)    >1.8V  

µA·∙MagGyro    Sensors  450µA    MicroController    500µA  

Physical  Gyro  4000-­‐7000µA  

Figure  4:  Original  Slide  14  from  uAMagGyro  

KMX62G Micro-Amp Magnetic Gyro

1 Includes power consumed by Atmel ATUC128L4U operating at 48 MHz executing at 3MIPS to produce gyro output 2 Via sensor fusion software

OF THE THREE TYPES OF SENSORS, GYROSCOPES ARE THE LARGEST AND MOST EXPENSIVE.

35

TECH REPORT

35

Figure 4.

Gyroscopes:

An accelerometer at rest on a desk can sense that the device is at rest. However, if you rotate the system with the accelerometer at the center of the rotation, the accelerometer will not be able to detect the movement. This is effectively a blind spot for the accelerometer. If the system has access to a gyroscope, the gyroscope can detect the rotation.

Unfortunately, implementing gyroscopic technology is difficult to justify for many applications. Of the three types of sensors, gyroscopes are the largest and most expensive. Although the sensor provides useful information, too often the cost, size, and/or power consumption

of a physical gyroscope exceeds the value it represents. For these reasons, developers have often had to limit position and orientation functionality to the 6-axis data provided by an accelerometer plus magnetometer.

THE KMX62G MAGGYRO

Today, advances in software algorithms and low-noise, low-latency sensors make it possible to simulate gyroscopic output using data from only an accelerometer and magnetometer. This is the approach used in MagGyros, where the system computes a device’s rotational direction and speed based on knowledge of its previous position. In other words, when a system can track orientation over time, it can extrapolate rotational data.

To enable developers to bring gyroscopic capabilities to a wide range of new applications, Kionix offers the KMX62G MagGyro. The KMX62G takes Kionix’s KMX62 acceleromator/magnetometer and enhances it with industry-leading sensor fusion software and auto-calibration algorithms (see Figure 3). This makes the KMX62G more than just an acceleromator/magnetometer. It is the industry’s first highly accurate gyroscopic emulator, providing 9-axis positioning capabilities.

THE KMX62G OFFERS SEVERAL KEY BENEFITS TO OEMS:

The Right Level of Accuracy: Because it estimates rather than directly measures rotational speed, the accuracy of a MagGyro is not as high as that of a physical gyroscope. However, gyroscopic accuracy comes into play most often when distance is involved, such as when flying a plane: being off by one degree over 100 miles can put you far off course. Close activities such as gesture recognition, on the other hand, can tolerate relatively large errors. Thus, consumer electronics applications focused on the user experience do not require the scientific levels of accuracy of a physical gyroscope.

Power: A physical gyroscope is always resonating, which means it continuously draws power. Typically, a physical gyroscope operates at 2.5 V or greater and consumes 4000 to 7000 μA (see Figure 4). Note that this is for the gyroscope alone.

With the KMX62G MagGyro, the accelerometer + magnetometer sensors consume 450 μA. Even when the power required for the microcontroller executing the software algorithms is taken into account (~500 μA), the entire MagGyro operation is less than 1000 μA. Thus, the KMX62G achieves 5X to 10X better energy efficiency compared to a physical gyroscope.

Start Up Time: Because they consume so much power, systems often turn off the gyroscope as often as they can. However, a physical gyroscope can take between 50 and 100ms for the output to stabilize. Thus, a power-efficient implementation with a physical gyroscope results in slow responsiveness that can negatively impact the user experience.

Figure 3.

Kionix, Inc. ©2013 All Rights Reserved

Power  

#1  Reason  •  1/5th  to  1/10th  of  

power  consump,on  

of  physical  gyro    

Opera'on  Voltage  •  Gyro  (typ)    >  2.5V  •  MagGyro  (typ)    >1.8V  

µA·∙MagGyro    Sensors  450µA    MicroController    500µA  

Physical  Gyro  4000-­‐7000µA  

Figure  4:  Original  Slide  14  from  uAMagGyro  

KMX62G Micro-Amp Magnetic Gyro

1 Includes power consumed by Atmel ATUC128L4U operating at 48 MHz executing at 3MIPS to produce gyro output 2 Via sensor fusion software

OF THE THREE TYPES OF SENSORS, GYROSCOPES ARE THE LARGEST AND MOST EXPENSIVE.

3636

From startup, the KMX62G takes approximately 15-20ms to start outputting 9-axis data. This arises from the need to collect first samples and process them. Once the data pipeline has been established, this delay effectively goes away. However, for many applications, the low energy draw of the KMX62G enables the device to be running most of the time, eliminating any start up delay for those applications or use cases where this matters.

Cost: The KMX62G can be implemented for substantially less cost than an equivalent 9-axis solution with a physical gyroscope. It provides an excellent intermediate solution between not having a physical gyroscope and offering features based on 9-axis positioning.

PERFORMANCE AND QUALITY

The performance of a MagGyro is highly dependent upon the quality of its components. Consider that in order to simulate a gyroscope, the MagGyro algorithms use the sensor data from both an accelerometer and magnetometer. Noise in either sensor will quickly

erode the accuracy of the gyroscopic calculations. In addition, the same sensor data is then used in sensor fusion algorithms to compute 9-axis positioning. This means the error from these sensors may be compounded. Thus, low noise performance in both the accelerometer and magnetometer is essential to achieving accuracy in 9-axis system.

Synchronization between the sensors is also critical. If the reading for the accelerometer is associated with a reading from the magnetometer captured at a different time, an error can result in the MagGyro output. Designing a tight circuit to ensure accurate synchronization of these sensors can be difficult when using external components. To simplify design, many designers prefer to eliminate synchronization issues by using a single component like the KMX62G that combines the accelerometer and magnetometer together. In addition, the latency of magnetometers can vary depending on their underlying technology. The magnetometer in the KMX62 has one of the lowest latencies on the market. This enables the accurate synchronization of data between the accelerometer and magnetometer and results in high performance of Kionix’s MagGyro solution.

The final piece of the KMX62G MagGyro is the software implementation of the simulated gyroscopic algorithms. There are numerous functions that make up

these algorithms, and the quality of the implementation determines the system’s overall performance and, in consequence, the user experience. These algorithms are implemented on a host processor or sensor hub. An Application Programming Interface (API) is provided to simplify product design. Supported platforms include Qualcomm’s Snapdragon and Atmel’s AVR UC3 and ARM-based SAM D20. The KMX62G is also certified for Windows 8 and 8.1.

Accurate MagGyro algorithms are fairly complex in their implementation. For example, averaging of signals reduces noise. However, averaging takes time, impacting latency and responsiveness. To maximize accuracy, Kionix utilizes adaptive software that dynamically adjusts averaging. When the system is moving quickly and small errors aren’t as noticeable, averaging is reduced to improve responsiveness. Similarly, when movements are smaller and responsiveness is less important, averaging is increased to improve accuracy. This provides an optimal user experience based on how the system is currently being used.

This focus on human use cases is part of the value Kionix offers with its KMX62G MagGyro. It excels in human perception and user experience applications. It does this by trading off a little of its performance and in return enhances the user’s experience. By designing its

algorithms for a high quality experience, Kionix enables the KMX62G to bring new functionality cost-effectively to a whole new realm of applications.

One other factor to consider when evaluating sensor fusion algorithms is the data processing requirements. Key metrics to consider are MIPS, code space, and RAM space. Algorithms that consume too many system resources can actually add cost to the system. If they place too much of a load on the host processor or sensor hub, they may even impact the primary function of the device and degrade the overall user experience. In terms of software loading, the MagGyro algorithm requires less than 3 MIPS. It addition, it can run on systems with as little as 128 KBytes of Flash and 32 KBytes of RAM. This is all that is needed for sensor calibration, magnetic anomaly rejection, sensor fusion and synthetic gyroscope calculations. This is a reasonable load for the value of being able to provide 9-axis data including gyroscopic functionality.

With the availability of the KMX62G MagGyro, there is now a cost-effective option for next-generation consumer electronics systems that can benefit from gyroscopic capabilities. By trading off an appropriate level of performance, the KMX62G offers lower cost, higher power efficiency, and a smaller footprint compared to systems based on physical gyroscopes.

BY DESIGNING ITS ALGORITHMS FOR A HIGH QUALITY EXPERIENCE, KIONIX ENABLES THE KMX62G TO BRING NEW FUNCTIONALITY COST-EFFECTIVELY TO A WHOLE NEW REALM OF APPLICATIONS.

37

TECH REPORT

37

From startup, the KMX62G takes approximately 15-20ms to start outputting 9-axis data. This arises from the need to collect first samples and process them. Once the data pipeline has been established, this delay effectively goes away. However, for many applications, the low energy draw of the KMX62G enables the device to be running most of the time, eliminating any start up delay for those applications or use cases where this matters.

Cost: The KMX62G can be implemented for substantially less cost than an equivalent 9-axis solution with a physical gyroscope. It provides an excellent intermediate solution between not having a physical gyroscope and offering features based on 9-axis positioning.

PERFORMANCE AND QUALITY

The performance of a MagGyro is highly dependent upon the quality of its components. Consider that in order to simulate a gyroscope, the MagGyro algorithms use the sensor data from both an accelerometer and magnetometer. Noise in either sensor will quickly

erode the accuracy of the gyroscopic calculations. In addition, the same sensor data is then used in sensor fusion algorithms to compute 9-axis positioning. This means the error from these sensors may be compounded. Thus, low noise performance in both the accelerometer and magnetometer is essential to achieving accuracy in 9-axis system.

Synchronization between the sensors is also critical. If the reading for the accelerometer is associated with a reading from the magnetometer captured at a different time, an error can result in the MagGyro output. Designing a tight circuit to ensure accurate synchronization of these sensors can be difficult when using external components. To simplify design, many designers prefer to eliminate synchronization issues by using a single component like the KMX62G that combines the accelerometer and magnetometer together. In addition, the latency of magnetometers can vary depending on their underlying technology. The magnetometer in the KMX62 has one of the lowest latencies on the market. This enables the accurate synchronization of data between the accelerometer and magnetometer and results in high performance of Kionix’s MagGyro solution.

The final piece of the KMX62G MagGyro is the software implementation of the simulated gyroscopic algorithms. There are numerous functions that make up

these algorithms, and the quality of the implementation determines the system’s overall performance and, in consequence, the user experience. These algorithms are implemented on a host processor or sensor hub. An Application Programming Interface (API) is provided to simplify product design. Supported platforms include Qualcomm’s Snapdragon and Atmel’s AVR UC3 and ARM-based SAM D20. The KMX62G is also certified for Windows 8 and 8.1.

Accurate MagGyro algorithms are fairly complex in their implementation. For example, averaging of signals reduces noise. However, averaging takes time, impacting latency and responsiveness. To maximize accuracy, Kionix utilizes adaptive software that dynamically adjusts averaging. When the system is moving quickly and small errors aren’t as noticeable, averaging is reduced to improve responsiveness. Similarly, when movements are smaller and responsiveness is less important, averaging is increased to improve accuracy. This provides an optimal user experience based on how the system is currently being used.

This focus on human use cases is part of the value Kionix offers with its KMX62G MagGyro. It excels in human perception and user experience applications. It does this by trading off a little of its performance and in return enhances the user’s experience. By designing its

algorithms for a high quality experience, Kionix enables the KMX62G to bring new functionality cost-effectively to a whole new realm of applications.

One other factor to consider when evaluating sensor fusion algorithms is the data processing requirements. Key metrics to consider are MIPS, code space, and RAM space. Algorithms that consume too many system resources can actually add cost to the system. If they place too much of a load on the host processor or sensor hub, they may even impact the primary function of the device and degrade the overall user experience. In terms of software loading, the MagGyro algorithm requires less than 3 MIPS. It addition, it can run on systems with as little as 128 KBytes of Flash and 32 KBytes of RAM. This is all that is needed for sensor calibration, magnetic anomaly rejection, sensor fusion and synthetic gyroscope calculations. This is a reasonable load for the value of being able to provide 9-axis data including gyroscopic functionality.

With the availability of the KMX62G MagGyro, there is now a cost-effective option for next-generation consumer electronics systems that can benefit from gyroscopic capabilities. By trading off an appropriate level of performance, the KMX62G offers lower cost, higher power efficiency, and a smaller footprint compared to systems based on physical gyroscopes.

BY DESIGNING ITS ALGORITHMS FOR A HIGH QUALITY EXPERIENCE, KIONIX ENABLES THE KMX62G TO BRING NEW FUNCTIONALITY COST-EFFECTIVELY TO A WHOLE NEW REALM OF APPLICATIONS.

3838

Important ways to approach heat management

When you watch a movie on your laptop, the fan

almost immediately starts to sound like a jet engine, and feels warm enough to start a fire. Similarly, if you leave your smartphone in the glove box in the summer, it won’t work until it cools down. Keeping electronics a reasonable temperature is

not magic, although it may feel like it at times as the rules and situations are not very familiar to electrical engineers. The effects of heat on electronics are also somewhat difficult to predict and often can be better understood using statistics, such as mean time between failures to show the lifetime averages.

Don’t Lose Your

COOL

39

TECH REPORT

39

Important ways to approach heat management

When you watch a movie on your laptop, the fan

almost immediately starts to sound like a jet engine, and feels warm enough to start a fire. Similarly, if you leave your smartphone in the glove box in the summer, it won’t work until it cools down. Keeping electronics a reasonable temperature is

not magic, although it may feel like it at times as the rules and situations are not very familiar to electrical engineers. The effects of heat on electronics are also somewhat difficult to predict and often can be better understood using statistics, such as mean time between failures to show the lifetime averages.

Don’t Lose Your

COOL

4040

When current flows through a resistor, power is dissipated in the form of heat. Resistors are rated for the amount of power they can dissipate before they burn out or burn up. Electronics are filled with transistors that turn on and off at literally unimaginable speeds but in some ways act as resistors. Transistors are typically understood to be in two states: open, wherein no current flows, or closed, wherein the resistance is nominally zero and current flows unimpeded. However, sometimes reality does not exactly match the concept. For example, when a transistor is open, there is a small leakage of current through what is considered a very high resistance burning a small amount of power. When the transistor is closed, the resistance is not truly zero, and that minor resistance also adds to the total used power. However, the greatest amount of power used is during the transition between the two states.

Although assumed to be a perfect square wave for most considerations, clock

signals have finite edge widths where a transistor is partially on and partially off. During this time, a current is flowing yet the resistance is much higher than when the transistor is in the closed state. In the large view, the amount of power in one clock cycle for that transistor is miniscule. However, when you multiply that one transistor by the hundreds of millions, or even billions of transistors in a processor, it jumps phenomenally. Multiply that by the billions of cycles per second and suddenly the trivial amount of power consumed has become a serious concern. Designers have tried many different methods to reduce the power consumed by making the clock edges sharper, at the expense of increased cost and electromagnetic emissions, or by decreasing the voltage level, which increases the susceptibility to spurious signals or digital errors. Many processors now are able to shut down portions of the chip when not in use to decrease power usage. This is very helpful, yet does not decrease the peak power usage.

In addition to avoid burning holes in desks, there are many reasons why we care about how much heat is being generated by electronics. The effects of heat on electronics vary greatly, and they are not always negative. For example, the chemical reactions within batteries tend to work better at room temperature. Yet batteries store better at cooler temperatures and becoming too hot while running can lead to catastrophic and explosive results.

TEMPERATURE RATINGS

Most electronics have a minimum and maximum temperature rating, typically 0C to 70C for commercial applications. The minimum temperature rating of electronics is usually not an issue except during the winter when electronics are left in the car or attic. Most of the concern related to cold is either the condensation that could be formed when bringing the electronics into the warmth, much like glasses fogging when coming inside, or the rapid expansion of traces and connections when electricity flows reheat them. Fortunately, these problems are easily avoided by either keeping electronics at a reasonable temperature or letting them come to room temperature before using them.

The damage caused by heat has some similarities with those with cold. Much as a cold part coming to room temperature causes unequal expansion of the different materials, changing from room temperature to levels better suited for baking cakes also causes different

expansions. As different materials have different coefficients of thermal expansion, some materials will expand much more than others and break connections. Another issue created by excessive heat is the increased resistance of copper, which creates a spiral effect and can contribute to runaway conditions. Even if the circuit remains functional and the temperature levels off at acceptable levels, the increased temperature and resistance means that more power is being used by the circuit to accomplish the same things, thus decreasing the overall efficiency. While copper increases its resistance with increasing temperatures, the silicon substrate of most integrated circuits is also susceptible to heat, causing it to lose its semiconductor properties over a certain threshold. This leads to completely unpredictable behavior changes in the circuit. Finally, excessive heat can also encourage whiskering—a concern that has already increased due to the transition back to lead-free solder.

HEAT MANAGEMENT

There are a myriad of ways to approach heat management. The simplest could be to reduce the amount of heat created by your product. Microprocessor manufacturers over the last two decades have struggled with keeping the temperature of their products at a reasonable level as they produce a large amount of heat in a very small area. As mentioned, decreasing the voltage, and therefore power consumption, has been very successful in the past. There

Power consumption versus resistance Affect of heat expansion

The effects of heat on electronics vary greatly, and they are not always negative.

41

TECH REPORT

41

When current flows through a resistor, power is dissipated in the form of heat. Resistors are rated for the amount of power they can dissipate before they burn out or burn up. Electronics are filled with transistors that turn on and off at literally unimaginable speeds but in some ways act as resistors. Transistors are typically understood to be in two states: open, wherein no current flows, or closed, wherein the resistance is nominally zero and current flows unimpeded. However, sometimes reality does not exactly match the concept. For example, when a transistor is open, there is a small leakage of current through what is considered a very high resistance burning a small amount of power. When the transistor is closed, the resistance is not truly zero, and that minor resistance also adds to the total used power. However, the greatest amount of power used is during the transition between the two states.

Although assumed to be a perfect square wave for most considerations, clock

signals have finite edge widths where a transistor is partially on and partially off. During this time, a current is flowing yet the resistance is much higher than when the transistor is in the closed state. In the large view, the amount of power in one clock cycle for that transistor is miniscule. However, when you multiply that one transistor by the hundreds of millions, or even billions of transistors in a processor, it jumps phenomenally. Multiply that by the billions of cycles per second and suddenly the trivial amount of power consumed has become a serious concern. Designers have tried many different methods to reduce the power consumed by making the clock edges sharper, at the expense of increased cost and electromagnetic emissions, or by decreasing the voltage level, which increases the susceptibility to spurious signals or digital errors. Many processors now are able to shut down portions of the chip when not in use to decrease power usage. This is very helpful, yet does not decrease the peak power usage.

In addition to avoid burning holes in desks, there are many reasons why we care about how much heat is being generated by electronics. The effects of heat on electronics vary greatly, and they are not always negative. For example, the chemical reactions within batteries tend to work better at room temperature. Yet batteries store better at cooler temperatures and becoming too hot while running can lead to catastrophic and explosive results.

TEMPERATURE RATINGS

Most electronics have a minimum and maximum temperature rating, typically 0C to 70C for commercial applications. The minimum temperature rating of electronics is usually not an issue except during the winter when electronics are left in the car or attic. Most of the concern related to cold is either the condensation that could be formed when bringing the electronics into the warmth, much like glasses fogging when coming inside, or the rapid expansion of traces and connections when electricity flows reheat them. Fortunately, these problems are easily avoided by either keeping electronics at a reasonable temperature or letting them come to room temperature before using them.

The damage caused by heat has some similarities with those with cold. Much as a cold part coming to room temperature causes unequal expansion of the different materials, changing from room temperature to levels better suited for baking cakes also causes different

expansions. As different materials have different coefficients of thermal expansion, some materials will expand much more than others and break connections. Another issue created by excessive heat is the increased resistance of copper, which creates a spiral effect and can contribute to runaway conditions. Even if the circuit remains functional and the temperature levels off at acceptable levels, the increased temperature and resistance means that more power is being used by the circuit to accomplish the same things, thus decreasing the overall efficiency. While copper increases its resistance with increasing temperatures, the silicon substrate of most integrated circuits is also susceptible to heat, causing it to lose its semiconductor properties over a certain threshold. This leads to completely unpredictable behavior changes in the circuit. Finally, excessive heat can also encourage whiskering—a concern that has already increased due to the transition back to lead-free solder.

HEAT MANAGEMENT

There are a myriad of ways to approach heat management. The simplest could be to reduce the amount of heat created by your product. Microprocessor manufacturers over the last two decades have struggled with keeping the temperature of their products at a reasonable level as they produce a large amount of heat in a very small area. As mentioned, decreasing the voltage, and therefore power consumption, has been very successful in the past. There

Power consumption versus resistance Affect of heat expansion

The effects of heat on electronics vary greatly, and they are not always negative.

4242

are other ways to decrease the power output at more of a board versus IC level. Any power conversion is a great opportunity to find savings in heat and efficiency, such as replacing a low drop-out voltage regulator with a switching regulator. This has its own trade-offs like efficiency for cost and electromagnetic noise, however, it is a viable option. Other ways to increase efficiency are to reduce any bit-banging that may be required on any embedded microcontrollers. By choosing systems that are able to handle complex or difficult data management in hardware, the amount of required instruction cycles is reduced, allowing more time to go into sleep mode, wherein the chip has time to cool.

• Heat MediationThere are times that the amount of heat generated cannot be reduced, at which point the heat must be removed. The first step of heat remediation is to set goals. Establish what levels are acceptable, then identify which devices are most susceptible to heat, as well as what devices create the most heat. If these items are not documented, it is difficult to balance the different trade-offs that will be necessary in the design process. Also, without a set goal, there is no way to know when it has been achieved.

• Environmental ConditionsDirectly after setting those goals, establish your constraints. While all electronics ideally would be placed directly in the flow of dry, cold, clean air, the reality is that most electronics are in very tough environments – e.g. cellphones reside next to warm bodies, computers are placed in dusty office corners with restricted airflow, and embedded systems are placed in airplane locations with wildly oscillating pressure, temperature, and humidity conditions. Defining these constraints requires flexibility. Make sure you know the space availability, any I/O requirements, whether or not these items will be in contact with air, and whether there will be a place to mechanically and thermally connect to the enclosure. Also, determine if there will be encapsulating material such as potting on any of the ICs. Write these items down along with the items that you can change, so that all your variables are clearly defined. While this document should be well-organized and clear, it is also a working document that will certainly change as requirements are refined.

• Heat Simulation ToolsSimulation tools are more prevalent today, and there are many heat

simulation tools available to get general ideas of how devices will heat up and where that heat will go. These are powerful but limited by the same issue plaguing any computer program, that of garbage in, garbage out. When properly used, these can provide a general concept of where to start and what needs changing; however, they are only as accurate as the information provided.

• Natural ConvectionNatural convection is an inexpensive solution that allows your electronics to cool, and it should be the first choice. If the device is not producing enormous amounts of heat and there is flexibility in the spacing of the enclosure, you may be able to passively cool it. This is a fantastic option as there is no fan to power or to get clogged. However, it requires even more exactness in other portions of the heat management technique. Placement and orientation of the heat-producing and heat-sensitive components needs to be managed, taking into account that heat rises. To oversimplify the matter, do not put all of the heat-producing components directly below all of the heat-sensitive components. Also, if in an enclosed space, determine if the passive method will continue to work after the ambient temperature has increased due to the heat producing ICs.

• Heat SinksTo increase the effectiveness of passive or active heat removal, heat sinks are sometimes used. A heat sink can be the oddly shaped metal piece attached to a computer processor, but it can also

be the circuit board attached to the IC. A properly laid out and drilled PCB can quickly move heat away from the IC and spread it throughout the board. It is important to remember that heat sinks do not make heat disappear; all they do is expedite the movement of heat from one point to another. In a tight enclosure, where the ambient temperature will rise with the heat put out by an IC, a heat sink will not change the overarching issue of the ambient temperature getting too warm. The idea is that a heat sink that is touching a very large heat mass—for example, the atmosphere—will cause such an inconsequential increase of temperature to that mass that it can be assumed to have no effect whatsoever on the temperature.

• FansIf the heat needs to be actively drawn away from something, fans are tried-and-true solutions, but carry their own concerns. They draw their own power, which may negatively affect other aspects of the design, plus they can be

Heat Dissipation

Microprocessor manufacturers over the last two decades have struggled with keeping the temperature of their products at a reasonable level as they produce a large amount of heat in a very small area.

43

TECH REPORT

43

are other ways to decrease the power output at more of a board versus IC level. Any power conversion is a great opportunity to find savings in heat and efficiency, such as replacing a low drop-out voltage regulator with a switching regulator. This has its own trade-offs like efficiency for cost and electromagnetic noise, however, it is a viable option. Other ways to increase efficiency are to reduce any bit-banging that may be required on any embedded microcontrollers. By choosing systems that are able to handle complex or difficult data management in hardware, the amount of required instruction cycles is reduced, allowing more time to go into sleep mode, wherein the chip has time to cool.

• Heat MediationThere are times that the amount of heat generated cannot be reduced, at which point the heat must be removed. The first step of heat remediation is to set goals. Establish what levels are acceptable, then identify which devices are most susceptible to heat, as well as what devices create the most heat. If these items are not documented, it is difficult to balance the different trade-offs that will be necessary in the design process. Also, without a set goal, there is no way to know when it has been achieved.

• Environmental ConditionsDirectly after setting those goals, establish your constraints. While all electronics ideally would be placed directly in the flow of dry, cold, clean air, the reality is that most electronics are in very tough environments – e.g. cellphones reside next to warm bodies, computers are placed in dusty office corners with restricted airflow, and embedded systems are placed in airplane locations with wildly oscillating pressure, temperature, and humidity conditions. Defining these constraints requires flexibility. Make sure you know the space availability, any I/O requirements, whether or not these items will be in contact with air, and whether there will be a place to mechanically and thermally connect to the enclosure. Also, determine if there will be encapsulating material such as potting on any of the ICs. Write these items down along with the items that you can change, so that all your variables are clearly defined. While this document should be well-organized and clear, it is also a working document that will certainly change as requirements are refined.

• Heat Simulation ToolsSimulation tools are more prevalent today, and there are many heat

simulation tools available to get general ideas of how devices will heat up and where that heat will go. These are powerful but limited by the same issue plaguing any computer program, that of garbage in, garbage out. When properly used, these can provide a general concept of where to start and what needs changing; however, they are only as accurate as the information provided.

• Natural ConvectionNatural convection is an inexpensive solution that allows your electronics to cool, and it should be the first choice. If the device is not producing enormous amounts of heat and there is flexibility in the spacing of the enclosure, you may be able to passively cool it. This is a fantastic option as there is no fan to power or to get clogged. However, it requires even more exactness in other portions of the heat management technique. Placement and orientation of the heat-producing and heat-sensitive components needs to be managed, taking into account that heat rises. To oversimplify the matter, do not put all of the heat-producing components directly below all of the heat-sensitive components. Also, if in an enclosed space, determine if the passive method will continue to work after the ambient temperature has increased due to the heat producing ICs.

• Heat SinksTo increase the effectiveness of passive or active heat removal, heat sinks are sometimes used. A heat sink can be the oddly shaped metal piece attached to a computer processor, but it can also

be the circuit board attached to the IC. A properly laid out and drilled PCB can quickly move heat away from the IC and spread it throughout the board. It is important to remember that heat sinks do not make heat disappear; all they do is expedite the movement of heat from one point to another. In a tight enclosure, where the ambient temperature will rise with the heat put out by an IC, a heat sink will not change the overarching issue of the ambient temperature getting too warm. The idea is that a heat sink that is touching a very large heat mass—for example, the atmosphere—will cause such an inconsequential increase of temperature to that mass that it can be assumed to have no effect whatsoever on the temperature.

• FansIf the heat needs to be actively drawn away from something, fans are tried-and-true solutions, but carry their own concerns. They draw their own power, which may negatively affect other aspects of the design, plus they can be

Heat Dissipation

Microprocessor manufacturers over the last two decades have struggled with keeping the temperature of their products at a reasonable level as they produce a large amount of heat in a very small area.

www.aapcb.com

4444

Advanced Assembly was founded to help engineers assemble their prototype and low-volume PCB orders. Based on years of experience within the printed circuit board industry, Advanced Assembly developed a proprietary system to deliver consistent, machine surface mount technology (SMT) assembly in 1-5 days. It’s our only focus. We take the hassle out of PCB assembly and make it easy, so you can spend time on other aspects of your design.

20100 E. 32nd Pkwy #225 | Aurora, CO 80011 | www.aapcb.com | 1-800-838-5650

When testing for temperature reliability, try to emulate real-world conditions as much as possible.

noisy, need to be cleaned on occasion, and will leave your product susceptible to damage if they fail. If possible, avoid fans, but if not possible, use them with caution.

When testing for temperature reliability, try to emulate real-world conditions as much as possible. Temperature testing is not purely about surviving certain temperatures, it includes other variables such as air flow, humidity levels, and concurrent physical strains. A product that has been tested up to one hundred degrees Celsius with forced air does not indicate whether or not that product will, in an enclosed environment without forced air, heat itself up to even

higher temperatures. Actual tests, like computer simulations, are subject to the same rule of garbage in, garbage out. If the parameters are not correctly setup, then the test results will not be helpful.

Yet another aspect of engineering that must be taken into consideration when designing is that temperature control and testing are not the be-all and end-all of any product design. A well-designed product will balance all of the needs of the project in an efficient, cost-effective solution to best provide a solution for your customer. Creating a device that always runs well below the temperature threshold but is too bulky to be conveniently used is a failure compared to a device that runs barely within the temperature threshold and has a convenient form factor. However, creating a device that has a great interface but survives less than three days under normal operating conditions is an absolute failure. Keep heat management in your personal toolbox and pull it out every once in a while during the design to make sure that you are on the right track, not just designing another fancy looking brick.

Garbage in – garbage out

CLICK HERE

Your Circuit Starts Here.Sign up to design, share, and collaborate

on your next project—big or small.

Click Here to Sign Up

4646

ESD Susceptibility

Every person who has purchased computer components or any bare electronics is familiar with

the silvery sheen of the antistatic bag. While it is common knowledge that electrostatic discharge (ESD) is bad for electronics, fewer are those who understand why. That nearly instantaneous transfer of electrons from one point to another, whether eliciting yelps of pain, or causing a temporary blanking of a TV monitor, or resulting in a cell phone that either restarts or decides to never start again, ESD is a part of everybody’s life. And it’s not very pleasant for people or electronics. While these discharges cannot be eliminated, there are steps that can be taken to protect any electronics that you design.

REDUCE

While these discharges cannot be eliminated, there are steps that can be taken to protect any electronics that you design.

Static electricity can come from a myriad of sources, but most people are familiar with the static electricity generated by their own bodies. When walking across the carpet, or shifting in a chair, a person’s body will physically pick up or lose electrons via the triboelectric effect. As the carpet and chair are not conductive, the electrons will not flow back to where they came from, but will build up. This will put the body at a different voltage potential from those objects around it. Static discharge is the equalization of charges on two entities that come together. If there is a greater electron density on one object than another, then when they touch, the electrons will try to equalize so there is the same density on both objects. If the

two objects are conductive, the equalization process happens quickly. Technically, that equalization is an electrical current because electrons are flowing. Enough electrons moving quickly enough will produce a current that is noticeable and sometimes painful or destructive. For this reason, discharges are most frequent and noticeable when touching metal, as metal is very conductive and allows for a rapid transfer, in other words, a large current.

A larger build-up in potential difference will also lead to larger currents. In dry weather, it is very common for there to be significantly more electric shocks than other times. This is due to the dryness increasing the resistance between objects, creating a greater insulated barrier.

47

TECH REPORT

47

ESD Susceptibility

Every person who has purchased computer components or any bare electronics is familiar with

the silvery sheen of the antistatic bag. While it is common knowledge that electrostatic discharge (ESD) is bad for electronics, fewer are those who understand why. That nearly instantaneous transfer of electrons from one point to another, whether eliciting yelps of pain, or causing a temporary blanking of a TV monitor, or resulting in a cell phone that either restarts or decides to never start again, ESD is a part of everybody’s life. And it’s not very pleasant for people or electronics. While these discharges cannot be eliminated, there are steps that can be taken to protect any electronics that you design.

REDUCE

While these discharges cannot be eliminated, there are steps that can be taken to protect any electronics that you design.

Static electricity can come from a myriad of sources, but most people are familiar with the static electricity generated by their own bodies. When walking across the carpet, or shifting in a chair, a person’s body will physically pick up or lose electrons via the triboelectric effect. As the carpet and chair are not conductive, the electrons will not flow back to where they came from, but will build up. This will put the body at a different voltage potential from those objects around it. Static discharge is the equalization of charges on two entities that come together. If there is a greater electron density on one object than another, then when they touch, the electrons will try to equalize so there is the same density on both objects. If the

two objects are conductive, the equalization process happens quickly. Technically, that equalization is an electrical current because electrons are flowing. Enough electrons moving quickly enough will produce a current that is noticeable and sometimes painful or destructive. For this reason, discharges are most frequent and noticeable when touching metal, as metal is very conductive and allows for a rapid transfer, in other words, a large current.

A larger build-up in potential difference will also lead to larger currents. In dry weather, it is very common for there to be significantly more electric shocks than other times. This is due to the dryness increasing the resistance between objects, creating a greater insulated barrier.

4848

The ideal level of conductance is a balance between the two extremes. While there needs to be a low enough resistance to allow the electrons to flow and reach that equilibrium, it has to be a high enough resistance to make certain that this flow is too low to cause any damage. While certain ESD bags for components are only designed to resist the accumulation of static electricity, other ESD bags are dissipative in nature so they will constantly and slowly make sure that the bag is at the same voltage potential as its surroundings. On the other end, conductive ESD bags are also used. These will change voltage potentials more quickly but protect their contents by routing the current through the bag itself and not through any of its contents.

While the main focus has thus far been about current due to direct contact, ESD can be induced by electromagnetic waves or pulses, an indirect source. If a device is going to be in an electromagnetically active area, then you need to avoid or

protect inadvertent antenna traces. A strong enough electromagnetic pulse can create a voltage differential across a long trace that will induce a damaging current. It’s reasonable to ask at what level of discharge is damaging. This is a difficult question to answer because there are many factors, which depends on the setup of the circuit. However, if there is an audible snap sound when there’s a discharge, then the voltage difference was at least approximately 5,000 volts. If you can feel it, it’s above approximately 2,000 volts. Unprotected integrated circuits can be destroyed at potentials less than 2,000 volts. It should be noted that the voltage potential is not itself what typically causes the problem but the current that melts or destroys portions of the circuit. That being said, there would be no current without the voltage, and this topic causes a fair amount of confusion.

The damage that is caused by ESD varies depending on the circuit and where the discharge occurs on the circuit board or integrated circuit. The most common damage that occurs to practically all types of electronics is simply the damage caused by overcurrent. A very fast, yet high amount of current can melt silicon and literally blow traces off of circuit boards, basically creating shorts or opens where they should not be. In today’s prominently complementary metal-oxide-semiconductor (CMOS) based technology, silicon-controlled rectifier (SCR) latchups are a major concern.

Due to the intrinsic design of CMOS circuits, without proper safeguards, a high voltage can activate a normally benign parasitic structure within the silicon substrate. This parasitic structure is, in essence, a thyristor between power and ground that will conduct high amounts of current through the substrate until all power is removed or the channel has burned itself open. This connection does not necessarily have to be between power and ground, however, those have the greatest potential of causing irreversible damage.

There are several different models that are used in testing the susceptibility of electronics to ESD. These models are for simulating a discharge from a human body to a device, a conductive object to a device, and a charged device discharging to another object. The model for transference from a human body to a device is called the Human Body Model and is simulated by discharging

a 100pF capacitor through a 1.5kOhm resistor to the device being tested. It is easiest to remember this model to think that the 1.5kOhm resistor is to simulate the resistance level of skin. For a conductive object, the model is called the Machine Model and instead of a 1.5kOhm resistor, a .5uH inductor is used and the capacitor is 200pF. As this lets a significantly greater current flow, the Machine Model is generally considered to be more stringent. The final model, in which a device is charged, most likely from the triboelectric effect while sliding through packaging on it’s way to be placed, is called the Charged Device Model. This test is performed by electrically isolating the device and then connecting it to a high voltage source, bringing the overall potential up. Once the device has been charged, it is mechanically touched by a grounded probe. Like the Machine Model, this is considered a more stringent test as it does not have a limiting resistor.

If there is an audible snap sound when there’s a discharge, then the voltage difference was at least approximately 5,000 volts.

Human Body Model Machine Model

49

TECH REPORT

49

The ideal level of conductance is a balance between the two extremes. While there needs to be a low enough resistance to allow the electrons to flow and reach that equilibrium, it has to be a high enough resistance to make certain that this flow is too low to cause any damage. While certain ESD bags for components are only designed to resist the accumulation of static electricity, other ESD bags are dissipative in nature so they will constantly and slowly make sure that the bag is at the same voltage potential as its surroundings. On the other end, conductive ESD bags are also used. These will change voltage potentials more quickly but protect their contents by routing the current through the bag itself and not through any of its contents.

While the main focus has thus far been about current due to direct contact, ESD can be induced by electromagnetic waves or pulses, an indirect source. If a device is going to be in an electromagnetically active area, then you need to avoid or

protect inadvertent antenna traces. A strong enough electromagnetic pulse can create a voltage differential across a long trace that will induce a damaging current. It’s reasonable to ask at what level of discharge is damaging. This is a difficult question to answer because there are many factors, which depends on the setup of the circuit. However, if there is an audible snap sound when there’s a discharge, then the voltage difference was at least approximately 5,000 volts. If you can feel it, it’s above approximately 2,000 volts. Unprotected integrated circuits can be destroyed at potentials less than 2,000 volts. It should be noted that the voltage potential is not itself what typically causes the problem but the current that melts or destroys portions of the circuit. That being said, there would be no current without the voltage, and this topic causes a fair amount of confusion.

The damage that is caused by ESD varies depending on the circuit and where the discharge occurs on the circuit board or integrated circuit. The most common damage that occurs to practically all types of electronics is simply the damage caused by overcurrent. A very fast, yet high amount of current can melt silicon and literally blow traces off of circuit boards, basically creating shorts or opens where they should not be. In today’s prominently complementary metal-oxide-semiconductor (CMOS) based technology, silicon-controlled rectifier (SCR) latchups are a major concern.

Due to the intrinsic design of CMOS circuits, without proper safeguards, a high voltage can activate a normally benign parasitic structure within the silicon substrate. This parasitic structure is, in essence, a thyristor between power and ground that will conduct high amounts of current through the substrate until all power is removed or the channel has burned itself open. This connection does not necessarily have to be between power and ground, however, those have the greatest potential of causing irreversible damage.

There are several different models that are used in testing the susceptibility of electronics to ESD. These models are for simulating a discharge from a human body to a device, a conductive object to a device, and a charged device discharging to another object. The model for transference from a human body to a device is called the Human Body Model and is simulated by discharging

a 100pF capacitor through a 1.5kOhm resistor to the device being tested. It is easiest to remember this model to think that the 1.5kOhm resistor is to simulate the resistance level of skin. For a conductive object, the model is called the Machine Model and instead of a 1.5kOhm resistor, a .5uH inductor is used and the capacitor is 200pF. As this lets a significantly greater current flow, the Machine Model is generally considered to be more stringent. The final model, in which a device is charged, most likely from the triboelectric effect while sliding through packaging on it’s way to be placed, is called the Charged Device Model. This test is performed by electrically isolating the device and then connecting it to a high voltage source, bringing the overall potential up. Once the device has been charged, it is mechanically touched by a grounded probe. Like the Machine Model, this is considered a more stringent test as it does not have a limiting resistor.

If there is an audible snap sound when there’s a discharge, then the voltage difference was at least approximately 5,000 volts.

Human Body Model Machine Model

5050

To make a working environment ESD safe, there are a multitude of approaches. At the workbench, having a properly grounded mat and strap can make a significant difference. If this is not sufficient, there are commercially available ionizers that neutralize the charges on nonconductive objects. Take care in the implementation of ionizers, though, as they can actually compound the problem if setup incorrectly. There are also special ESD garments that can be worn by workers to also reduce the build up on their bodies as they move about the workspace or shift on their chairs and desks. Further methods to reduce static build up as workers move around the production floor is a specialized floor coating. Unsurprisingly, carpet is a non-ideal floor surface if dealing with electrostatically sensitive devices.

To ESD harden individual devices is a bit more difficult as the methods have to be more carefully tailored to the device, its performance requirements, and its environments. Despite this, there are general tips that can be used, even if the implementation varies. Before pursuing these tips, however, first view the data sheets of the more sensitive integrated circuits to see if there are counsels specific to those devices. They may have recommendations that emphasize or contradict the general tips.

The biggest step for the device itself is prevention, which is done through proper grounding. If the electronics are kept at the same voltage level as its environment, there won’t be

the opportunity to develop a voltage differential. If the electronics are embedded in a larger system, then providing adequate grounding between the different aspects of the system will help significantly. For smaller, embedded devices that reside in user’s pockets, it is much more difficult to ground them properly as they’re floating and usually encased in nonconductive enclosures. These enclosures could technically be made of a dissipative or ESD-resistant material, but this will create trade-offs with the other aspects of the materials, such as durability. For those devices that do have a physical connection to power and ground, properly grounding the electronics and the chassis will cause any discharge to be routed directly to the power ground, not the electronics.

With proper grounding in place, then methods for decoupling the circuit from potential surges is the next step. The simplest method, which is also rather effective, is to implement decoupling capacitors at places where a discharge

is most expected. These capacitors are usually placed between power and ground near the integrated circuits, which serves the dual purpose of compensating for any dips in the voltage as the circuit pulls more current while also shunting any high-frequency noise directly to ground. These capacitors are also typically placed on any communication lines that leave the device itself and are accessible to the outside world. As people are plugging and unplugging their connections, such as USB or audio jacks, there can be discharges down those lines. The biggest problem with capacitors in this situation, though, is how they can affect high-speed communication signals. In the case of the USB, a capacitor on the signal lines could significantly slew the signal, rendering the USB useless. Any high-frequency line will be adversely affected by a capacitor, so other methods must be considered.

Aside from capacitors, there are many other devices that can be used to protect electronics, varistors and transient-voltage-suppression (TVS) diodes being prominent on this list. Varistors, which conduct only at high-voltage levels, can be placed between any line of concern and ground. If there is a spike of high voltage that exceeds the clamping voltage, the varistor will switch on, shunting that surge directly to ground. TVS diodes are functionally similar to varistors in that they shunt any high-voltage surges while acting like an open at lower voltages. TVS diodes typically have a faster response time than varistors but in general are not capable

Reducing issues with ESD should involve a two-fold approach. The environment in which a product is made, from manufacturer to distributor, should be designed specifically to reduce the build up of static. Also, products need to be able to survive the world of the consumer, so these products must be hardened against the static that they will confront once they are with the end user. There are certain tried and true methods that will help with either aspect but their application depends on the situation, budget, and standards. A two-dollar toy will not have seventy cents worth of ESD hardening added to the circuitry. However, a control board for a fighter plane that will be in an incredibly noisy environment while needing to maintain a zero percent failure rate requires that can a significant amount of time and money be invested to ensure the board has been appropriately hardened.

Charged Device Model

For small, embedded devices that reside in user’s pockets, it is much more difficult to ground them properly as they’re floating and usually encased in nonconductive enclosures.

51

TECH REPORT

51

To make a working environment ESD safe, there are a multitude of approaches. At the workbench, having a properly grounded mat and strap can make a significant difference. If this is not sufficient, there are commercially available ionizers that neutralize the charges on nonconductive objects. Take care in the implementation of ionizers, though, as they can actually compound the problem if setup incorrectly. There are also special ESD garments that can be worn by workers to also reduce the build up on their bodies as they move about the workspace or shift on their chairs and desks. Further methods to reduce static build up as workers move around the production floor is a specialized floor coating. Unsurprisingly, carpet is a non-ideal floor surface if dealing with electrostatically sensitive devices.

To ESD harden individual devices is a bit more difficult as the methods have to be more carefully tailored to the device, its performance requirements, and its environments. Despite this, there are general tips that can be used, even if the implementation varies. Before pursuing these tips, however, first view the data sheets of the more sensitive integrated circuits to see if there are counsels specific to those devices. They may have recommendations that emphasize or contradict the general tips.

The biggest step for the device itself is prevention, which is done through proper grounding. If the electronics are kept at the same voltage level as its environment, there won’t be

the opportunity to develop a voltage differential. If the electronics are embedded in a larger system, then providing adequate grounding between the different aspects of the system will help significantly. For smaller, embedded devices that reside in user’s pockets, it is much more difficult to ground them properly as they’re floating and usually encased in nonconductive enclosures. These enclosures could technically be made of a dissipative or ESD-resistant material, but this will create trade-offs with the other aspects of the materials, such as durability. For those devices that do have a physical connection to power and ground, properly grounding the electronics and the chassis will cause any discharge to be routed directly to the power ground, not the electronics.

With proper grounding in place, then methods for decoupling the circuit from potential surges is the next step. The simplest method, which is also rather effective, is to implement decoupling capacitors at places where a discharge

is most expected. These capacitors are usually placed between power and ground near the integrated circuits, which serves the dual purpose of compensating for any dips in the voltage as the circuit pulls more current while also shunting any high-frequency noise directly to ground. These capacitors are also typically placed on any communication lines that leave the device itself and are accessible to the outside world. As people are plugging and unplugging their connections, such as USB or audio jacks, there can be discharges down those lines. The biggest problem with capacitors in this situation, though, is how they can affect high-speed communication signals. In the case of the USB, a capacitor on the signal lines could significantly slew the signal, rendering the USB useless. Any high-frequency line will be adversely affected by a capacitor, so other methods must be considered.

Aside from capacitors, there are many other devices that can be used to protect electronics, varistors and transient-voltage-suppression (TVS) diodes being prominent on this list. Varistors, which conduct only at high-voltage levels, can be placed between any line of concern and ground. If there is a spike of high voltage that exceeds the clamping voltage, the varistor will switch on, shunting that surge directly to ground. TVS diodes are functionally similar to varistors in that they shunt any high-voltage surges while acting like an open at lower voltages. TVS diodes typically have a faster response time than varistors but in general are not capable

Reducing issues with ESD should involve a two-fold approach. The environment in which a product is made, from manufacturer to distributor, should be designed specifically to reduce the build up of static. Also, products need to be able to survive the world of the consumer, so these products must be hardened against the static that they will confront once they are with the end user. There are certain tried and true methods that will help with either aspect but their application depends on the situation, budget, and standards. A two-dollar toy will not have seventy cents worth of ESD hardening added to the circuitry. However, a control board for a fighter plane that will be in an incredibly noisy environment while needing to maintain a zero percent failure rate requires that can a significant amount of time and money be invested to ensure the board has been appropriately hardened.

Charged Device Model

For small, embedded devices that reside in user’s pockets, it is much more difficult to ground them properly as they’re floating and usually encased in nonconductive enclosures.

www.aapcb.com

5252

of handling such high power spikes. Both of these, while effective, still have their own drawbacks. Despite not being designed as capacitors, these both have intrinsic capacitance and small, yet finite amounts of leakage current. While these are both designed to be minimal, it will require custom calculations to determine if they will cause adverse affects on your circuits. While they may provide a shunting effect, they’re also not immune to being destroyed in discharges, so they must be sized to accommodate the current that they’re expecting.

Another method, besides shunting to ground, is to simply reduce the amount of electricity that can flow down a trace or lead. This is very simply done by putting a resistor in series of whichever line needs protection. This is a highly discouraged method for any lines that are anticipated to have non-negligible amounts of current. Otherwise the

resistors will simply be turning power into heat, which is rarely a good idea. However, if the line is a simple communications line with no possibility of current flow, then it is a simple way to make sure that if a voltage is induced on the line, it will be absorbed more by the resistor than the circuit behind it.

Electrostatic discharge and its effects are an interesting topic with many potential problems and an equal number of potential solutions. Knowing the advantages and disadvantages of each solution ensure sound engineering decisions that protect products from the wildly varying voltages found in the real world. While a one-size-fits-all solution would be ideal, that is rarely found in the engineering. Running basic calculations, implementing solutions, and thoroughly testing using different models will help in the never-ending goal of making products more robust.

Advanced Assembly was founded to help engineers assemble their prototype and low-volume PCB orders. Based on years of experience within the printed circuit board industry, Advanced Assembly developed a proprietary system to deliver consistent, machine surface mount technology (SMT) assembly in 1-5 days. It’s our only focus. We take the hassle out of PCB assembly and make it easy, so you can spend time on other aspects of your design.

20100 E. 32nd Pkwy #225 | Aurora, CO 80011 | www.aapcb.com | 1-800-838-5650

CLICK HERE

eeweb.com/register

Join Today

READY TO LAUNCH

For the launch of the Tiva C Series Connected LaunchPad, TI has partnered with Exosite, mentioned briefly above, to provide easy access to the LaunchPad from the Internet. The LaunchPad takes about 10 minutes to set up and you can immediately interact with it across the Internet and do things like turn an LED on and off remotely from the website and see the reported temperature as well. It can also display approximate geographic location based on the assigned IP address and display a map of all other connected LaunchPad owners if they are active and plugged-in to Exosite. “In addition, it supports a basic game by enabling someone to interface to the Connected LaunchPad through a serial port from a terminal while someone else is playing with them through their browser. It is basically showing how you can interact remotely with this product and a user even if you are across the globe,” Folkens explained.

START DEVELOPING

The Tiva C Series Connected LaunchPad is shipping now and the price is right; at $19.99 USD, it is less than half the price of other Ethernet-ready kits. The LaunchPad comes complete with quick start and user guides, and ample online support to ensure developers of all backgrounds are well equipped to begin creating cloud-based applications. “We have assembled an online support team to monitor the Engineering-to-Engineering (or E2E) Community,” Folkens said. “Along with this, you also got a free Code Composer Studio Integrated Development Environment, which allows developers to use the full capability. We also support other tool chains like Keil, IAR and Mentor Embedded.

Affordable, versatile, and easy to use, the Tiva Series Connected LaunchPad is well suited for a broad audience and promises to facilitate the expansion of ingenious IoT applications in the cloud. As Folkens concluded, “The target audiences actually are the hobbyists, students and professional engineers. A better way of looking at it is that we are targeting people with innovative ideas and trying to help them get those ideas launched into the cloud.”

54

QA&

“The idea behind the Big I.D.E.A. Design Contest is to make the engineer aware of the breadth of NXP’s product portfolio.”

CLIFF LLOYD, Business Development Director, Logic Division, NXP Semiconductors, discusses Dual Configurable Logic and this year’s Big I.D.E.A. (International Design Engineering Award).

The Big

I.D.E.A.The following interview was conducted in early January at the NXP offices in San Jose by Glenn ImObersteg, President of Convergence Promotions.

INDUSTRY INTERVIEW

55

QA&

“The idea behind the Big I.D.E.A. Design Contest is to make the engineer aware of the breadth of NXP’s product portfolio.”

CLIFF LLOYD, Business Development Director, Logic Division, NXP Semiconductors, discusses Dual Configurable Logic and this year’s Big I.D.E.A. (International Design Engineering Award).

The Big

I.D.E.A.The following interview was conducted in early January at the NXP offices in San Jose by Glenn ImObersteg, President of Convergence Promotions.

CLICK HERE

56

C L I C K H E R E

To watch Tom Wolf’s Dual Configurable Logic video, click the image below:

Tom Wolf in his video series on logic says that the configurable logic has been around since the ‘60s—is that right?

Yes, in one form or another. They were originally programmable where you blew fuses—as such, they were dedicated to a particular application. Then they had configurable logic in the form of PLAs, where they were mass-programmed in the factory for a particular application. But those types again were dedicated to a particular application. Later on, they came out with the FPGA (Field-Programmable

Gate Arrays) so the engineer or the manufacturing site could program them when they needed them. And again, once they were programmed, they were set for that particular application.

We’ve come a long way from the early programmable devices to today’s configurable logic though, haven’t we?

Our configurable logic devices are not programmed by anybody—they are configured by the layout of the PCB so you can put the same device down in different locations on the PCB and depending on how the PCB is laid out, the part would perform the function it is configured to perform in that location.

What part did NXP play in the configurable logic idea?

We developed the idea about 6 years ago, around 2009.

Last year, you introduced the AXP Logic family, which was a breakthrough product. Can you tell me a little bit about AXP?

Yes, we saw a downward migration in voltage from 3.3V to 1.8V to 1.2V to 0.8V. There were no logic devices on the market that were specified down below 1.2V. Yet, we saw on new ASIC product road maps that the new processes are going down to 0.8V and that in the future, they would not be able to support devices operating at 3.3V. So there will be a need for a lower operating voltage logic device. Now you don’t see many products on the market today, but there will be in the years to come. In a couple of years, there will be a greater need for the lower voltage AXP family.

A year after the introduction of the AXP family, you introduced dual configurable logic. Can you tell us a little bit about the need for that and why you introduced a dual configurable product on the heels of the AXP?

That’s a very good question. When we introduced the AXP family, the first part we introduced was a configurable part. When we ran a design contest to encourage engineers to become familiar with using the part, we realized from their submissions that there was a need for multiple parts that were different on the PCB. So in the past, you could buy an AND gate and if you needed two AND gates, you would buy two in the same package so you could save space; however, if you needed an AND gate and an OR gate, you would have to buy two different devices—each packaged

separately in their own package—and put them on the PCB, which took up a lot of space. So we thought, why don’t we offer a combination of devices in a single package? So we started doing that, but in logic, there are about 50 or 60 different functions and you can’t possibly make a combination of all those 50 or 60 functions. We did an analysis and it turns out there would be over 400 different combinations, and there would be no way we could offer all of these different combinations of logic functions. So we decided to take a look at doing a family of dual configurable logic devices, so that the engineer can configure each of the parts individually. That way, they can not only get the 400-plus different functions, but they can get even more. That’s how we came up with the idea.

How do the multiple functions in the package reduce system costs, inventory control, and operations like qualification?

The benefits of the dual configurable logic device are significant in many areas. For example, for the design engineer, he can save both space and power because you only have to put one part on the board instead of two. In addition to the physical space the package itself takes up, you will also save the clearance space that is required around each package since they can’t be placed physically touching each other. Also, the power consumption of a dual configurable logic device is less than the total power consumed by two individual gates. Other benefits for the design engineer are that he does not have to

order individual, specific functions and wait for them to arrive, he can take a configurable device and configure it in the function that he needs at any time.

Apart from the space and power savings of the device, what are some of the other benefits of the dual configurable logic device?

From a purchasing and logistics point of view, both the purchasing manager and the logistics manager need to manage only one device. Also, from a cost perspective, better pricing can sometimes be obtained from purchasing higher volumes of one specific device rather than having the volume spread over a number of different devices.

Are there any other benefits to the customer when they use a dual configurable logic device?

Yes. From a production and assembly point of view, a pick-n-place machine needs a dedicated magazine for each device type or component that is being placed on the PCB. For example, if an application requires 5 different logic devices, then you need 5 magazines to be reserved and loaded with the 5 different logic devices. However, by using the dual configurable logic device, only one magazine needs to be reserved for the logic component as its function is determined by the PCB layout—thus, freeing up the other 4 magazines for other components.

Do you know of any company that is using configurable logic for increased assembly efficiency?

Our configurable logic devices

are configured by the layout

of the PCB so you can

put the same device down in different

locations on the PCB.

INDUSTRY INTERVIEW

57

C L I C K H E R E

To watch Tom Wolf’s Dual Configurable Logic video, click the image below:

Tom Wolf in his video series on logic says that the configurable logic has been around since the ‘60s—is that right?

Yes, in one form or another. They were originally programmable where you blew fuses—as such, they were dedicated to a particular application. Then they had configurable logic in the form of PLAs, where they were mass-programmed in the factory for a particular application. But those types again were dedicated to a particular application. Later on, they came out with the FPGA (Field-Programmable

Gate Arrays) so the engineer or the manufacturing site could program them when they needed them. And again, once they were programmed, they were set for that particular application.

We’ve come a long way from the early programmable devices to today’s configurable logic though, haven’t we?

Our configurable logic devices are not programmed by anybody—they are configured by the layout of the PCB so you can put the same device down in different locations on the PCB and depending on how the PCB is laid out, the part would perform the function it is configured to perform in that location.

What part did NXP play in the configurable logic idea?

We developed the idea about 6 years ago, around 2009.

Last year, you introduced the AXP Logic family, which was a breakthrough product. Can you tell me a little bit about AXP?

Yes, we saw a downward migration in voltage from 3.3V to 1.8V to 1.2V to 0.8V. There were no logic devices on the market that were specified down below 1.2V. Yet, we saw on new ASIC product road maps that the new processes are going down to 0.8V and that in the future, they would not be able to support devices operating at 3.3V. So there will be a need for a lower operating voltage logic device. Now you don’t see many products on the market today, but there will be in the years to come. In a couple of years, there will be a greater need for the lower voltage AXP family.

A year after the introduction of the AXP family, you introduced dual configurable logic. Can you tell us a little bit about the need for that and why you introduced a dual configurable product on the heels of the AXP?

That’s a very good question. When we introduced the AXP family, the first part we introduced was a configurable part. When we ran a design contest to encourage engineers to become familiar with using the part, we realized from their submissions that there was a need for multiple parts that were different on the PCB. So in the past, you could buy an AND gate and if you needed two AND gates, you would buy two in the same package so you could save space; however, if you needed an AND gate and an OR gate, you would have to buy two different devices—each packaged

separately in their own package—and put them on the PCB, which took up a lot of space. So we thought, why don’t we offer a combination of devices in a single package? So we started doing that, but in logic, there are about 50 or 60 different functions and you can’t possibly make a combination of all those 50 or 60 functions. We did an analysis and it turns out there would be over 400 different combinations, and there would be no way we could offer all of these different combinations of logic functions. So we decided to take a look at doing a family of dual configurable logic devices, so that the engineer can configure each of the parts individually. That way, they can not only get the 400-plus different functions, but they can get even more. That’s how we came up with the idea.

How do the multiple functions in the package reduce system costs, inventory control, and operations like qualification?

The benefits of the dual configurable logic device are significant in many areas. For example, for the design engineer, he can save both space and power because you only have to put one part on the board instead of two. In addition to the physical space the package itself takes up, you will also save the clearance space that is required around each package since they can’t be placed physically touching each other. Also, the power consumption of a dual configurable logic device is less than the total power consumed by two individual gates. Other benefits for the design engineer are that he does not have to

order individual, specific functions and wait for them to arrive, he can take a configurable device and configure it in the function that he needs at any time.

Apart from the space and power savings of the device, what are some of the other benefits of the dual configurable logic device?

From a purchasing and logistics point of view, both the purchasing manager and the logistics manager need to manage only one device. Also, from a cost perspective, better pricing can sometimes be obtained from purchasing higher volumes of one specific device rather than having the volume spread over a number of different devices.

Are there any other benefits to the customer when they use a dual configurable logic device?

Yes. From a production and assembly point of view, a pick-n-place machine needs a dedicated magazine for each device type or component that is being placed on the PCB. For example, if an application requires 5 different logic devices, then you need 5 magazines to be reserved and loaded with the 5 different logic devices. However, by using the dual configurable logic device, only one magazine needs to be reserved for the logic component as its function is determined by the PCB layout—thus, freeing up the other 4 magazines for other components.

Do you know of any company that is using configurable logic for increased assembly efficiency?

Our configurable logic devices

are configured by the layout

of the PCB so you can

put the same device down in different

locations on the PCB.

www.thebigidea2015.com

58

Yes and, as a matter of fact, when I visited them this past fall, they said they now encourage their design engineers use configurable logic devices as much as possible in place of individual logic gates.

The 74AUP2G57 is the centerpiece of the new “Big I.D.E.A” Design Contest that just launched. The idea behind this design contest is to pull together logic with other standard products from NXP—load switches, ESD protection and filtering components, small signal MOSFETs, and small signal diodes. What was the rationale behind this?

The idea behind this design contest is to make the engineer aware of the breadth of NXP’s product portfolio. Whenever you have an application, the heart of the application, the chip or ASIC—no matter what it is—it doesn’t stand alone. It has components connected to both its inputs and outputs. For example, when you have a device with a push button or connector, you have to have some way of protecting the chip from any possible damage from ESD. To do so, you need some ESD protection devices to protect the chip/ASIC’s inputs. On the back end you will most likely have to drive something like a light bulb, a relay, motor, display, etc. Thus you might need a power MOS device or a load switch. The dual configurable logic design contest is intended to not only make the design engineer familiar with the dual configurable logic device but also to make him aware of the other components and

to encourage him to use them. Thus, we have included some power MOS devices, load switches, and ESD and filtering devices in our design contest. Of course, the center of the design contest is the dual configurable logic.

How will the submitted designs be judged and are there any ways the engineers can increase the number of points they can receive?

The engineers that submit their designs will get a certain amount of points for creativity, and will get additional points for documentation and demonstration. The demonstration can be performed by either submitting a video, showing curve traces, or sending in the modeling files. If they use some of the other components we have suggested, like ESD protection diodes, they will receive a 20% multiplier. The same goes for the use of power MOS, load switches, and some of the protection diodes. Finally, because this contest is being co-sponsored by Mouser, if they use the MultiSIM Blue program to do the schematic capture and simulation, they will receive an additional 20% multiplier. The engineers can actually double their score by using all of these other components in the contest.

Thank you, Cliff. We’re looking forward to seeing some exciting and creative applications from this contest later this year.

When it comes to

design, the engineer can save

both space and power

because you only have to put one part on the

board instead of two.

To view the Big I.D.E.A Contest page, click the link below: www.thebigidea2015.com

CLICK HERE

Evolve to app-based controlwith AIR for Wiced Smart!

Get “mobile smart”in 3 easy steps:

Get your AIR for Wiced Smart dev kit at your distributor of choice. (See our website for a current list.)

Develop your wireless link and basic app using our exclusive Atmosphere development tool.

With our AIR for Wiced Smart module on board, proceed in record time to a prototype and final, mobile-app development!

1905

TodayJOIN THEEVOLUTION.

Learn more

1945

2005

If you’re ready to evolve from fixed control panels populated with dials, buttons, keypads, and LCD displays to mobile-app based control of your embedded product – check out Anaren’s AIR for Wiced Smart module, featuring Broadcom’s Wiced Smart Bluetooth® chip (BCM20737). Not only does our small-footprint, SMT, and pre-certified all-in-one module save you the time, effort, and trouble of designing your own radio... It’s supported by our industry-exclusive Atmosphere development ecosystem that lets you develop your basic embedded code and app code in one, easy-to-use development tool – for a far speedier product development cycle and time-to-market. Follow the steps at left to join the evolution, right now!

www.anaren.com/AIRforWiced800-411-6596In Europe: 44-2392-232392

6060

By Michael Lyons, NXP Semiconductor

Shift registers can help

reduce size and BOM in

designs that use LEDs.

By providing I/O expansion, they

enable the use of smaller, less

expensive microcontrollers. In

some cases, the shift register can

be used to drive the LED directly

and thus eliminate the need for

external LED drivers. This adds to

the savings, and makes it possible

to drive a wider variety of LEDs.

Shift Registers

LED Designs

Using

to SIMPLIFY

61

TECH REPORT

61

By Michael Lyons, NXP Semiconductor

Shift registers can help

reduce size and BOM in

designs that use LEDs.

By providing I/O expansion, they

enable the use of smaller, less

expensive microcontrollers. In

some cases, the shift register can

be used to drive the LED directly

and thus eliminate the need for

external LED drivers. This adds to

the savings, and makes it possible

to drive a wider variety of LEDs.

Shift Registers

LED Designs

Using

to SIMPLIFY

6262

In designs that use LEDs, shifter registers can be very useful. For instance, if the system includes a seven-segment display, a single indicator, or an array of LEDs that form a grid or panel, a standard 8-bit shift register can be used to allow a low pin-count microcontroller to drive multiple LEDs.

In some cases, a 5V, 8-bit register like the 74HC595 can be used to drive LEDs directly. This works best when the LEDs are specified for relatively low voltage and forward current. LEDs that operate with voltages higher than 6V or require forward current that exceeds 20mA will typically require an external driver.

Open-drain outputsAdding open-drain outputs to the shift register creates a single-chip solution that eliminates the need for an external driver. This can yield significant reductions in the bill of materials, since each output of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such device, the NPIC6C596A LED driver from NXP, which combines shift register functions similar to a Z4HC595 with a high-voltage (HV) MOSFET driver.

Using shift registers to reduce size and BOM in LED designsby Michael Lyons, NXP Semiconductor

Summary

Shift registers can help reduce size and BOM in

designs that use LEDs. By providing I/O expansion,

they enable the use of smaller, less expensive

microcontrollers. In some cases, the shift register can

be used to drive the LED directly and thus eliminate

the need for external LED drivers. This adds to

the savings, and makes it possible to drive a wider

variety of LEDs.

Full Article

In designs that use LEDs, shifter registers can be

very useful. For instance, if the system includes a

seven-segment display, a single indicator, or an array

of LEDs that form a grid or panel, a standard 8-bit

shift register can be used to allow a low pin-count

microcontroller to drive multiple LEDs.

Figure 1 gives an example. A single 5V 74HC595

shift register, with serial inputs and serial or

parallel outputs, provides I/O expansion for the

microcontroller. Serial data is applied to the serial

input of the 74HC595 and clocked in via the input

clock. Once the 74HC595 is loaded, the output

clock applies the data to the storage register and

to the parallel and serial outputs. External drivers,

controlled by the 74HC595, then activate the

corresponding LEDs.

Output_CLKI/O

12 V

Microcontroller

74HC595

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

Figure 1. An 8-bit 74HC595 shift register driving multiple LEDs

Using the 74HC595 for I/O expansion means that

it takes only three MCU control pins to drive up to

eight LEDs. Reducing the number of control pins

makes it possible to use an MCU with a lower pin

count, and that can yield a smaller, more cost-

effective design.

Also, because the 74HC595 includes a serial output,

several devices can be cascaded together. Figure 2

gives the layout.

Figure 1. An 8-bit 74HC595 shift register driving multiple LEDs.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 3. Output schematic for shift register with open-drain outputs.

Figure 1 gives an example. A single 5V 74HC595 shift register, with serial inputs and serial or parallel outputs, provides I/O expansion for the microcontroller.

Serial data is applied to the serial input of the 74HC595 and clocked in via the input clock. Once the 74HC595 is loaded, the output clock applies the data to the storage register and to the parallel and serial outputs. External drivers, controlled by the 74HC595, then activate the corresponding LEDs.

Using the 74HC595 for I/O expansion means that it takes only three MCU control pins to drive up to eight LEDs. Reducing the number of control pins makes it possible to use an MCU with a lower pin count, and that can yield a smaller, more cost-effective design. Also, because the 74HC595 includes a serial output, several devices can be cascaded together. Figure 2 gives the layout.

Now, with the cascading, the same three pins on the microcontroller can be used to control up to 16 or 24 LEDs instead of just eight. The ability to cascade shift registers can reduce the total number of microcontrollers needed in the design, and that can help lower costs and reduce size, too.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 4. Output schematic for shift register with open-drain outputs.

Replacing the 74HC595 with the NPIC6C596A eliminates the need for external drivers, creating a design that is more compact and has a lower bill of materials.

NPIC6C devices have open-drain outputs that are tolerant to 33V. Each output is designed to sink 100mA and there is no limit on ground current. All the outputs can actively sink 100mA simultaneously. The outputs include current-limiting circuitry, which sets a 250mA maximum on the sinkable current, and each output also includes thermal protection. Having these protections means the NPIC6C596A device can be used to drive a wider range of LEDs than the 74HC595, including LEDs that operate at higher voltages and with higher forward current.

63

TECH REPORT

63

In designs that use LEDs, shifter registers can be very useful. For instance, if the system includes a seven-segment display, a single indicator, or an array of LEDs that form a grid or panel, a standard 8-bit shift register can be used to allow a low pin-count microcontroller to drive multiple LEDs.

In some cases, a 5V, 8-bit register like the 74HC595 can be used to drive LEDs directly. This works best when the LEDs are specified for relatively low voltage and forward current. LEDs that operate with voltages higher than 6V or require forward current that exceeds 20mA will typically require an external driver.

Open-drain outputsAdding open-drain outputs to the shift register creates a single-chip solution that eliminates the need for an external driver. This can yield significant reductions in the bill of materials, since each output of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such device, the NPIC6C596A LED driver from NXP, which combines shift register functions similar to a Z4HC595 with a high-voltage (HV) MOSFET driver.

Using shift registers to reduce size and BOM in LED designsby Michael Lyons, NXP Semiconductor

Summary

Shift registers can help reduce size and BOM in

designs that use LEDs. By providing I/O expansion,

they enable the use of smaller, less expensive

microcontrollers. In some cases, the shift register can

be used to drive the LED directly and thus eliminate

the need for external LED drivers. This adds to

the savings, and makes it possible to drive a wider

variety of LEDs.

Full Article

In designs that use LEDs, shifter registers can be

very useful. For instance, if the system includes a

seven-segment display, a single indicator, or an array

of LEDs that form a grid or panel, a standard 8-bit

shift register can be used to allow a low pin-count

microcontroller to drive multiple LEDs.

Figure 1 gives an example. A single 5V 74HC595

shift register, with serial inputs and serial or

parallel outputs, provides I/O expansion for the

microcontroller. Serial data is applied to the serial

input of the 74HC595 and clocked in via the input

clock. Once the 74HC595 is loaded, the output

clock applies the data to the storage register and

to the parallel and serial outputs. External drivers,

controlled by the 74HC595, then activate the

corresponding LEDs.

Output_CLKI/O

12 V

Microcontroller

74HC595

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

Figure 1. An 8-bit 74HC595 shift register driving multiple LEDs

Using the 74HC595 for I/O expansion means that

it takes only three MCU control pins to drive up to

eight LEDs. Reducing the number of control pins

makes it possible to use an MCU with a lower pin

count, and that can yield a smaller, more cost-

effective design.

Also, because the 74HC595 includes a serial output,

several devices can be cascaded together. Figure 2

gives the layout.

Figure 1. An 8-bit 74HC595 shift register driving multiple LEDs.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 3. Output schematic for shift register with open-drain outputs.

Figure 1 gives an example. A single 5V 74HC595 shift register, with serial inputs and serial or parallel outputs, provides I/O expansion for the microcontroller.

Serial data is applied to the serial input of the 74HC595 and clocked in via the input clock. Once the 74HC595 is loaded, the output clock applies the data to the storage register and to the parallel and serial outputs. External drivers, controlled by the 74HC595, then activate the corresponding LEDs.

Using the 74HC595 for I/O expansion means that it takes only three MCU control pins to drive up to eight LEDs. Reducing the number of control pins makes it possible to use an MCU with a lower pin count, and that can yield a smaller, more cost-effective design. Also, because the 74HC595 includes a serial output, several devices can be cascaded together. Figure 2 gives the layout.

Now, with the cascading, the same three pins on the microcontroller can be used to control up to 16 or 24 LEDs instead of just eight. The ability to cascade shift registers can reduce the total number of microcontrollers needed in the design, and that can help lower costs and reduce size, too.

Using shift registers to reduce size and BOM in LED designs 2

74HC595 74HC595

Output_CLKI/O

Microcontroller

External drivers

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 2. Cascading 74HC595 devices to drive more LEDs

Now, with the cascading, the same three pins on

the microcontroller can be used to control up to

16 or 24 LEDs instead of just eight. The ability to

cascade shift registers can reduce the total number

of microcontrollers needed in the design, and that

can help lower costs and reduce size, too.

In some cases, a 5 V, 8-bit register like the 74HC595

can be used to drive LEDs directly. This works

best when the LEDs are specified for relatively low

voltage and forward current. LEDs that operate with

voltages higher than 6 V or require forward current

that exceeds 20 mA will typically require an external

driver.

Open-drain outputs

Adding open-drain outputs to the shift register

creates a single-chip solution that eliminates the

need for an external driver. This can yield significant

reductions in the bill of materials, since each output

of the shift register can drive the LEDs directly.

Figure 3 gives the output schematic for one such

device, the NPIC6C596A LED driver from NXP,

which combines shift register functions similar to a

74HC595 with a high-voltage (HV) MOSFET driver.

Output schematic

33 V

GND

Qn

Figure 3. Output schematic for shift register with open-drain

outputs

Figure 4 shows the NPIC6C596A used in place of

the 74HC595.

Figure 4. Output schematic for shift register with open-drain outputs

NPIC6C596A NPIC6C596A

Output_CLKI/O

Microcontroller

Open-drain outputsdrive LEDS directly

Storage register

Shift registerI/O

I/O

Serial input

Input_CLK

12 V

Storage register

Shift register

Serial output

Figure 4. Output schematic for shift register with open-drain outputs.

Replacing the 74HC595 with the NPIC6C596A eliminates the need for external drivers, creating a design that is more compact and has a lower bill of materials.

NPIC6C devices have open-drain outputs that are tolerant to 33V. Each output is designed to sink 100mA and there is no limit on ground current. All the outputs can actively sink 100mA simultaneously. The outputs include current-limiting circuitry, which sets a 250mA maximum on the sinkable current, and each output also includes thermal protection. Having these protections means the NPIC6C596A device can be used to drive a wider range of LEDs than the 74HC595, including LEDs that operate at higher voltages and with higher forward current.

6464

Using shift registers to reduce size and BOM in LED designs 3

Replacing the 74HC595 with the NPIC6C596A

eliminates the need for external drivers, creating a

design that is more compact and has a lower bill of

materials.

NPIC6C devices have open-drain outputs that are

tolerant to 33 V. Each output is designed to sink 100

mA and there is no limit on ground current. All the

outputs can actively sink 100 mA simultaneously. The

outputs include current-limiting circuitry, which sets

a 250 mA maximum on the sinkable current, and

each output also includes thermal protection. Having

these protections means the NPIC6C596A device

can be used to drive a wider range of LEDs than

the 74HC595, including LEDs that operate at higher

voltages and with higher forward current.

Protection features

Figure 5 shows the behavior of the current-

limiting circuitry on the open-drain outputs of the

NPIC6C596A. The circuitry limits the maximum

current each output can sink. As the drain voltage

increases, the drain source current decreases. This

protects the outputs and the components they

are driving. At 25 °C, the output clamp is typically

activated when the drain source current is 250 mA.

-40º C

-10º C

25º C

80º C

125º C

0.4

0.3

0.2

0.1

00 10 20 30

VDS(V)

IDS(A)

Figure 5. Current-limiting behavior in NPIC6C596A

Figure 6 shows how the open-drain outputs of the

NPIC6C596A provide thermal protection. The clamp

current is inversely proportional to temperature. As

the temperature increases, the output resistance

increases, thus limiting the drain source current

and preventing damage to the output and the

components it drives. At 25 °C, the output typically

limits the drain source current to 120 mA.

5 V

10 V

15 V

20 V

25 V

30 V

0.3

0.2

0.1

0-25 25 75 125

TAMB(ºC)

IDS(A)

Figure 6. Thermal protection in NPIC6C596A

Multiple options

Table 1 shows the NPIC6C LED drivers available

from NXP. The NPIC6C596 and the NPIC6C596A

are 8-bit solutions, while the NPIC6C4894 is a 12-bit

solution. All include a serial output for cascading.

Data is propagated through the shift register

on the rising edge of the input clock. With the

NPIC6C595 and the NPIC6C4894, the same rising

edge is used to clock data to the serial output QS.

The NPIC6C596 and NPIC6C596A delay the serial

output to the next falling edge of the input clock.

The delay provides a longer data hold time, which

improves timing margin and makes it easier to

cascade many shift registers.

The NPIC6C596 and NPIC6C4894 can be used

between 4.5 and 5.5 V, making them suitable for 5.0

V control logic interfaces. The NPIC6C596A can be

used from 2.3 to 5.5 V, so it can be used with 5.0,

3.3, and 2.5 V control logic interfaces. All NPIC6C

devices operate from -40 to +125 °C and with an

input clock frequency of at least 10 MHz.

Using shift registers to reduce size and BOM in LED designs 3

Replacing the 74HC595 with the NPIC6C596A

eliminates the need for external drivers, creating a

design that is more compact and has a lower bill of

materials.

NPIC6C devices have open-drain outputs that are

tolerant to 33 V. Each output is designed to sink 100

mA and there is no limit on ground current. All the

outputs can actively sink 100 mA simultaneously. The

outputs include current-limiting circuitry, which sets

a 250 mA maximum on the sinkable current, and

each output also includes thermal protection. Having

these protections means the NPIC6C596A device

can be used to drive a wider range of LEDs than

the 74HC595, including LEDs that operate at higher

voltages and with higher forward current.

Protection features

Figure 5 shows the behavior of the current-

limiting circuitry on the open-drain outputs of the

NPIC6C596A. The circuitry limits the maximum

current each output can sink. As the drain voltage

increases, the drain source current decreases. This

protects the outputs and the components they

are driving. At 25 °C, the output clamp is typically

activated when the drain source current is 250 mA.

-40º C

-10º C

25º C

80º C

125º C

0.4

0.3

0.2

0.1

00 10 20 30

VDS(V)

IDS(A)

Figure 5. Current-limiting behavior in NPIC6C596A

Figure 6 shows how the open-drain outputs of the

NPIC6C596A provide thermal protection. The clamp

current is inversely proportional to temperature. As

the temperature increases, the output resistance

increases, thus limiting the drain source current

and preventing damage to the output and the

components it drives. At 25 °C, the output typically

limits the drain source current to 120 mA.

5 V

10 V

15 V

20 V

25 V

30 V

0.3

0.2

0.1

0-25 25 75 125

TAMB(ºC)

IDS(A)

Figure 6. Thermal protection in NPIC6C596A

Multiple options

Table 1 shows the NPIC6C LED drivers available

from NXP. The NPIC6C596 and the NPIC6C596A

are 8-bit solutions, while the NPIC6C4894 is a 12-bit

solution. All include a serial output for cascading.

Data is propagated through the shift register

on the rising edge of the input clock. With the

NPIC6C595 and the NPIC6C4894, the same rising

edge is used to clock data to the serial output QS.

The NPIC6C596 and NPIC6C596A delay the serial

output to the next falling edge of the input clock.

The delay provides a longer data hold time, which

improves timing margin and makes it easier to

cascade many shift registers.

The NPIC6C596 and NPIC6C4894 can be used

between 4.5 and 5.5 V, making them suitable for 5.0

V control logic interfaces. The NPIC6C596A can be

used from 2.3 to 5.5 V, so it can be used with 5.0,

3.3, and 2.5 V control logic interfaces. All NPIC6C

devices operate from -40 to +125 °C and with an

input clock frequency of at least 10 MHz.

Figure 5. Current-limiting behavior in NPIC6C596A. Figure 6. Thermal protection in NPIC6C596A.

Protection FeaturesFigure 5 shows the behavior of the current-limiting circuitry on the open-drain outputs of the NPIC6C596A. The circuitry limits the maximum current each output can sink. As the drain voltage increases, the drain source current decreases. This protects the outputs and the components they are driving. At 25 °C, the output clamp is typically activated when the drain source current is 250 mA.

Figure 6 shows how the open-drain outputs of the NPIC6C596A provide thermal protection. The clamp current is inversely proportional to temperature. As the temperature increases, the output resistance increases, thus limiting the drain source current and preventing damage to the output and the components it drives. At 25°C, the output typically limits the drain source current to 120mA.

The NPIC6C596 and NPIC6C596A delay the serial output to the next falling

edge of the input clock. The delay provides a longer data hold time, which improves timing margin and makes it easier to cascade many shift registers.

The NPIC6C596 and NPIC6C4894 can be used between 4.5V and 5.5V, making them suitable for 5.0 V control logic interfaces. The NPIC6C596A can be used from 2.3V to 5.5V, so it can be used with 5.0V, 3.3V, and 2.5V control logic interfaces. All NPIC6C devices operate from -40 to +125°C and with an input clock frequency of at least 10MHz.

NPIC6C LED drivers are available in industry-standard SO and TSSOP packages, as well as the space-saving DQFN leadless package, which is up to 76 percent smaller than a TSSOP and 40 percent smaller than a QFN. DQFN packages also include a heat sink and are the packages of choice for space-constrained applications that use higher currents. Automotive variants are also available.

Table 1. NPIC6C LED drivers from NXP

NPIC6C LED drivers are available in industry-standard SO and TSSOP packages, as well as the space-

saving DQFN leadless package, which is up to 76 percent smaller than a TSSOP and 40 percent smaller

than a QFN. DQFN packages also include a heat sink and are the packages of choice for space-constrained

applications that use higher currents. Automotive variants are also available.

Table 2. Package options for NPIC6C LED drivers

Conclusion

When LEDs are part of the design, shift registers make it possible to use a smaller, less expensive

microcontroller. Standard 8-bit shift registers like the 74HC595 are available from a number of suppliers,

including NXP. Shift registers that are equipped with open-drain outputs, like the NPIC6C series from NXP,

go a step further, because they eliminate the need for external LED drivers.

More about the NPIC6C series can be found at http://www.nxp.com/products/logic/family/NPIC/#overview

Type number Format Supply voltage (V) fmax (MHz) Tamb (°C) QS clock Packages

NPIC6C595 8-bit 4.5 to 5.5 10 -40 to +125 RiseSO16, TSSOP16, DQFN16

NPIC6C596 8-bit 4.5 to 5.5 10 -40 to +125 FallSO16, TSSOP16, DQFN16

NPIC6C596A 8-bit 2.3 to 5.5 10 -40 to +125 RiseSO16, TSSOP16, DQFN16

NPIC6C4894 12-bit 4.5 to 5.5 10 -40 to +125 RiseSO20, TSSOP20, DQFN20

Package suffix D PW BQ D PW

16-pin 16-pin 16-pin 20-pin 20-pin

Package SOT109-1 SOT403-1 SOT763-1 SOT163-1 SOT360-1

Width (mm) 6.00 6.40 2.50 10.30 6.40

Length (mm) 9.90 5.00 3.50 12.80 6.50

Height (mm) 1.75 1.10 1.00 2.65 1.10

Pitch (mm) 1.27 0.65 0.50 1.27 0.65

www.nxp.com

© 2014 NXP B.V.

All rights reserved. Reproduction in whole or in part is prohibited without the prior written consent of the copyright owner. The

information presented in this document does not form part of any quotation or contract, is believed to be accurate and reliable and

may be changed without notice. No liability will be accepted by the publisher for any consequence of its use. Publication thereof

does not convey nor imply any license under patent or other industrial or intellectual property rights.

Date of release: May 2014

Published in the USA

Table 1. NPIC6C LED Drivers from NXP.

Table 1. NPIC6C LED drivers from NXP

NPIC6C LED drivers are available in industry-standard SO and TSSOP packages, as well as the space-

saving DQFN leadless package, which is up to 76 percent smaller than a TSSOP and 40 percent smaller

than a QFN. DQFN packages also include a heat sink and are the packages of choice for space-constrained

applications that use higher currents. Automotive variants are also available.

Table 2. Package options for NPIC6C LED drivers

Conclusion

When LEDs are part of the design, shift registers make it possible to use a smaller, less expensive

microcontroller. Standard 8-bit shift registers like the 74HC595 are available from a number of suppliers,

including NXP. Shift registers that are equipped with open-drain outputs, like the NPIC6C series from NXP,

go a step further, because they eliminate the need for external LED drivers.

More about the NPIC6C series can be found at http://www.nxp.com/products/logic/family/NPIC/#overview

Type number Format Supply voltage (V) fmax (MHz) Tamb (°C) QS clock Packages

NPIC6C595 8-bit 4.5 to 5.5 10 -40 to +125 RiseSO16, TSSOP16, DQFN16

NPIC6C596 8-bit 4.5 to 5.5 10 -40 to +125 FallSO16, TSSOP16, DQFN16

NPIC6C596A 8-bit 2.3 to 5.5 10 -40 to +125 RiseSO16, TSSOP16, DQFN16

NPIC6C4894 12-bit 4.5 to 5.5 10 -40 to +125 RiseSO20, TSSOP20, DQFN20

Package suffix D PW BQ D PW

16-pin 16-pin 16-pin 20-pin 20-pin

Package SOT109-1 SOT403-1 SOT763-1 SOT163-1 SOT360-1

Width (mm) 6.00 6.40 2.50 10.30 6.40

Length (mm) 9.90 5.00 3.50 12.80 6.50

Height (mm) 1.75 1.10 1.00 2.65 1.10

Pitch (mm) 1.27 0.65 0.50 1.27 0.65

www.nxp.com

© 2014 NXP B.V.

All rights reserved. Reproduction in whole or in part is prohibited without the prior written consent of the copyright owner. The

information presented in this document does not form part of any quotation or contract, is believed to be accurate and reliable and

may be changed without notice. No liability will be accepted by the publisher for any consequence of its use. Publication thereof

does not convey nor imply any license under patent or other industrial or intellectual property rights.

Date of release: May 2014

Published in the USA

Table 2. Package options for NPCIC6C LED drivers.

ConclusionWhen LEDs are part of the design, shift registers make it possible to use a smaller, less expensive microcontroller. Standard 8-bit shift registers like the 74HC595 are available from a

More about the NPIC6C series can be found at: http://www.nxp.com/products/logic/family/NPIC/#overview

number of suppliers, including NXP. Shift registers that are equipped with open-drain outputs, like the NPIC6C series from NXP, go a step further, because they eliminate the need for external LED drivers.

http://www.nxp.com/products/logic/family/NPIC/#overview

65

TECH REPORT

65

Using shift registers to reduce size and BOM in LED designs 3

Replacing the 74HC595 with the NPIC6C596A

eliminates the need for external drivers, creating a

design that is more compact and has a lower bill of

materials.

NPIC6C devices have open-drain outputs that are

tolerant to 33 V. Each output is designed to sink 100

mA and there is no limit on ground current. All the

outputs can actively sink 100 mA simultaneously. The

outputs include current-limiting circuitry, which sets

a 250 mA maximum on the sinkable current, and

each output also includes thermal protection. Having

these protections means the NPIC6C596A device

can be used to drive a wider range of LEDs than

the 74HC595, including LEDs that operate at higher

voltages and with higher forward current.

Protection features

Figure 5 shows the behavior of the current-

limiting circuitry on the open-drain outputs of the

NPIC6C596A. The circuitry limits the maximum

current each output can sink. As the drain voltage

increases, the drain source current decreases. This

protects the outputs and the components they

are driving. At 25 °C, the output clamp is typically

activated when the drain source current is 250 mA.

-40º C

-10º C

25º C

80º C

125º C

0.4

0.3

0.2

0.1

00 10 20 30

VDS(V)

IDS(A)

Figure 5. Current-limiting behavior in NPIC6C596A

Figure 6 shows how the open-drain outputs of the

NPIC6C596A provide thermal protection. The clamp

current is inversely proportional to temperature. As

the temperature increases, the output resistance

increases, thus limiting the drain source current

and preventing damage to the output and the

components it drives. At 25 °C, the output typically

limits the drain source current to 120 mA.

5 V

10 V

15 V

20 V

25 V

30 V

0.3

0.2

0.1

0-25 25 75 125

TAMB(ºC)

IDS(A)

Figure 6. Thermal protection in NPIC6C596A

Multiple options

Table 1 shows the NPIC6C LED drivers available

from NXP. The NPIC6C596 and the NPIC6C596A

are 8-bit solutions, while the NPIC6C4894 is a 12-bit

solution. All include a serial output for cascading.

Data is propagated through the shift register

on the rising edge of the input clock. With the

NPIC6C595 and the NPIC6C4894, the same rising

edge is used to clock data to the serial output QS.

The NPIC6C596 and NPIC6C596A delay the serial

output to the next falling edge of the input clock.

The delay provides a longer data hold time, which

improves timing margin and makes it easier to

cascade many shift registers.

The NPIC6C596 and NPIC6C4894 can be used

between 4.5 and 5.5 V, making them suitable for 5.0

V control logic interfaces. The NPIC6C596A can be

used from 2.3 to 5.5 V, so it can be used with 5.0,

3.3, and 2.5 V control logic interfaces. All NPIC6C

devices operate from -40 to +125 °C and with an

input clock frequency of at least 10 MHz.

Using shift registers to reduce size and BOM in LED designs 3

Replacing the 74HC595 with the NPIC6C596A

eliminates the need for external drivers, creating a

design that is more compact and has a lower bill of

materials.

NPIC6C devices have open-drain outputs that are

tolerant to 33 V. Each output is designed to sink 100

mA and there is no limit on ground current. All the

outputs can actively sink 100 mA simultaneously. The

outputs include current-limiting circuitry, which sets

a 250 mA maximum on the sinkable current, and

each output also includes thermal protection. Having

these protections means the NPIC6C596A device

can be used to drive a wider range of LEDs than

the 74HC595, including LEDs that operate at higher

voltages and with higher forward current.

Protection features

Figure 5 shows the behavior of the current-

limiting circuitry on the open-drain outputs of the

NPIC6C596A. The circuitry limits the maximum

current each output can sink. As the drain voltage

increases, the drain source current decreases. This

protects the outputs and the components they

are driving. At 25 °C, the output clamp is typically

activated when the drain source current is 250 mA.

-40º C

-10º C

25º C

80º C

125º C

0.4

0.3

0.2

0.1

00 10 20 30

VDS(V)

IDS(A)

Figure 5. Current-limiting behavior in NPIC6C596A

Figure 6 shows how the open-drain outputs of the

NPIC6C596A provide thermal protection. The clamp

current is inversely proportional to temperature. As

the temperature increases, the output resistance

increases, thus limiting the drain source current

and preventing damage to the output and the

components it drives. At 25 °C, the output typically

limits the drain source current to 120 mA.

5 V

10 V

15 V

20 V

25 V

30 V

0.3

0.2

0.1

0-25 25 75 125

TAMB(ºC)

IDS(A)

Figure 6. Thermal protection in NPIC6C596A

Multiple options

Table 1 shows the NPIC6C LED drivers available

from NXP. The NPIC6C596 and the NPIC6C596A

are 8-bit solutions, while the NPIC6C4894 is a 12-bit

solution. All include a serial output for cascading.

Data is propagated through the shift register

on the rising edge of the input clock. With the

NPIC6C595 and the NPIC6C4894, the same rising

edge is used to clock data to the serial output QS.

The NPIC6C596 and NPIC6C596A delay the serial

output to the next falling edge of the input clock.

The delay provides a longer data hold time, which

improves timing margin and makes it easier to

cascade many shift registers.

The NPIC6C596 and NPIC6C4894 can be used

between 4.5 and 5.5 V, making them suitable for 5.0

V control logic interfaces. The NPIC6C596A can be

used from 2.3 to 5.5 V, so it can be used with 5.0,

3.3, and 2.5 V control logic interfaces. All NPIC6C

devices operate from -40 to +125 °C and with an

input clock frequency of at least 10 MHz.

Figure 5. Current-limiting behavior in NPIC6C596A. Figure 6. Thermal protection in NPIC6C596A.

Protection FeaturesFigure 5 shows the behavior of the current-limiting circuitry on the open-drain outputs of the NPIC6C596A. The circuitry limits the maximum current each output can sink. As the drain voltage increases, the drain source current decreases. This protects the outputs and the components they are driving. At 25 °C, the output clamp is typically activated when the drain source current is 250 mA.

Figure 6 shows how the open-drain outputs of the NPIC6C596A provide thermal protection. The clamp current is inversely proportional to temperature. As the temperature increases, the output resistance increases, thus limiting the drain source current and preventing damage to the output and the components it drives. At 25°C, the output typically limits the drain source current to 120mA.

The NPIC6C596 and NPIC6C596A delay the serial output to the next falling

edge of the input clock. The delay provides a longer data hold time, which improves timing margin and makes it easier to cascade many shift registers.

The NPIC6C596 and NPIC6C4894 can be used between 4.5V and 5.5V, making them suitable for 5.0 V control logic interfaces. The NPIC6C596A can be used from 2.3V to 5.5V, so it can be used with 5.0V, 3.3V, and 2.5V control logic interfaces. All NPIC6C devices operate from -40 to +125°C and with an input clock frequency of at least 10MHz.

NPIC6C LED drivers are available in industry-standard SO and TSSOP packages, as well as the space-saving DQFN leadless package, which is up to 76 percent smaller than a TSSOP and 40 percent smaller than a QFN. DQFN packages also include a heat sink and are the packages of choice for space-constrained applications that use higher currents. Automotive variants are also available.

Table 1. NPIC6C LED drivers from NXP

NPIC6C LED drivers are available in industry-standard SO and TSSOP packages, as well as the space-

saving DQFN leadless package, which is up to 76 percent smaller than a TSSOP and 40 percent smaller

than a QFN. DQFN packages also include a heat sink and are the packages of choice for space-constrained

applications that use higher currents. Automotive variants are also available.

Table 2. Package options for NPIC6C LED drivers

Conclusion

When LEDs are part of the design, shift registers make it possible to use a smaller, less expensive

microcontroller. Standard 8-bit shift registers like the 74HC595 are available from a number of suppliers,

including NXP. Shift registers that are equipped with open-drain outputs, like the NPIC6C series from NXP,

go a step further, because they eliminate the need for external LED drivers.

More about the NPIC6C series can be found at http://www.nxp.com/products/logic/family/NPIC/#overview

Type number Format Supply voltage (V) fmax (MHz) Tamb (°C) QS clock Packages

NPIC6C595 8-bit 4.5 to 5.5 10 -40 to +125 RiseSO16, TSSOP16, DQFN16

NPIC6C596 8-bit 4.5 to 5.5 10 -40 to +125 FallSO16, TSSOP16, DQFN16

NPIC6C596A 8-bit 2.3 to 5.5 10 -40 to +125 RiseSO16, TSSOP16, DQFN16

NPIC6C4894 12-bit 4.5 to 5.5 10 -40 to +125 RiseSO20, TSSOP20, DQFN20

Package suffix D PW BQ D PW

16-pin 16-pin 16-pin 20-pin 20-pin

Package SOT109-1 SOT403-1 SOT763-1 SOT163-1 SOT360-1

Width (mm) 6.00 6.40 2.50 10.30 6.40

Length (mm) 9.90 5.00 3.50 12.80 6.50

Height (mm) 1.75 1.10 1.00 2.65 1.10

Pitch (mm) 1.27 0.65 0.50 1.27 0.65

www.nxp.com

© 2014 NXP B.V.

All rights reserved. Reproduction in whole or in part is prohibited without the prior written consent of the copyright owner. The

information presented in this document does not form part of any quotation or contract, is believed to be accurate and reliable and

may be changed without notice. No liability will be accepted by the publisher for any consequence of its use. Publication thereof

does not convey nor imply any license under patent or other industrial or intellectual property rights.

Date of release: May 2014

Published in the USA

Table 1. NPIC6C LED Drivers from NXP.

Table 1. NPIC6C LED drivers from NXP

NPIC6C LED drivers are available in industry-standard SO and TSSOP packages, as well as the space-

saving DQFN leadless package, which is up to 76 percent smaller than a TSSOP and 40 percent smaller

than a QFN. DQFN packages also include a heat sink and are the packages of choice for space-constrained

applications that use higher currents. Automotive variants are also available.

Table 2. Package options for NPIC6C LED drivers

Conclusion

When LEDs are part of the design, shift registers make it possible to use a smaller, less expensive

microcontroller. Standard 8-bit shift registers like the 74HC595 are available from a number of suppliers,

including NXP. Shift registers that are equipped with open-drain outputs, like the NPIC6C series from NXP,

go a step further, because they eliminate the need for external LED drivers.

More about the NPIC6C series can be found at http://www.nxp.com/products/logic/family/NPIC/#overview

Type number Format Supply voltage (V) fmax (MHz) Tamb (°C) QS clock Packages

NPIC6C595 8-bit 4.5 to 5.5 10 -40 to +125 RiseSO16, TSSOP16, DQFN16

NPIC6C596 8-bit 4.5 to 5.5 10 -40 to +125 FallSO16, TSSOP16, DQFN16

NPIC6C596A 8-bit 2.3 to 5.5 10 -40 to +125 RiseSO16, TSSOP16, DQFN16

NPIC6C4894 12-bit 4.5 to 5.5 10 -40 to +125 RiseSO20, TSSOP20, DQFN20

Package suffix D PW BQ D PW

16-pin 16-pin 16-pin 20-pin 20-pin

Package SOT109-1 SOT403-1 SOT763-1 SOT163-1 SOT360-1

Width (mm) 6.00 6.40 2.50 10.30 6.40

Length (mm) 9.90 5.00 3.50 12.80 6.50

Height (mm) 1.75 1.10 1.00 2.65 1.10

Pitch (mm) 1.27 0.65 0.50 1.27 0.65

www.nxp.com

© 2014 NXP B.V.

All rights reserved. Reproduction in whole or in part is prohibited without the prior written consent of the copyright owner. The

information presented in this document does not form part of any quotation or contract, is believed to be accurate and reliable and

may be changed without notice. No liability will be accepted by the publisher for any consequence of its use. Publication thereof

does not convey nor imply any license under patent or other industrial or intellectual property rights.

Date of release: May 2014

Published in the USA

Table 2. Package options for NPCIC6C LED drivers.

ConclusionWhen LEDs are part of the design, shift registers make it possible to use a smaller, less expensive microcontroller. Standard 8-bit shift registers like the 74HC595 are available from a

More about the NPIC6C series can be found at: http://www.nxp.com/products/logic/family/NPIC/#overview

number of suppliers, including NXP. Shift registers that are equipped with open-drain outputs, like the NPIC6C series from NXP, go a step further, because they eliminate the need for external LED drivers.

6666

Sensor Processing ApplicationsThis all-in-one solution, based on the ultra-low power LPC54102

microcontroller, provides everything needed to bring sensor-based

motion and other sensor-processing applications to market quickly. NXP

has partnered with Bosch Sensortec to offer an integrated solution that

makes it easy to incorporate motion/inertia and other sensor data into

a variety of end applications. The solution includes sensors for

6- and 9-axis motion vectors as well as

temperature, proximity

and ambient light.

NXP’s Complete Solution for

MOTION-BASED, ALWAYS-ON

SENSOR PROCESSING/MOTION

The LPC54102 Sensor Processing/Motion Solution includes everything you need to add motion and other sensor-based features to your applications. Using the ultra-low-power, dual-core LPC54102 microcontroller, it provides a scalable solution for a family of connected sensor-based products.

Solution Includes:

• Reference hardware including the LPCXpresso54102 board and a sensor shield with nine sensors and Bluetooth LE wireless connectivity

• Bosch BSX Lite sensor fusion middleware

• Bosch sensor drivers

• LPC Sensor Framework to simplify sensor-processing and support additional sensors

• Complete reference documentation

FLEXIBLE, EXPENDABLE REFERENCE HARDWARE

The Solution’s LPCXpresso54102 development board includes:

• Large set of pinouts for measurement and prototyping

• Built-in Hi-Speed USB debug probe (Link2) and support for external debug probes

• On-board 1.8/3.3V or external power supply options for LPC54102

• Built-in MCU power consumption and supply voltage measurement for LPC54102 MCU and sensor shield board

• UART, I2C and SPI port bridging form LPC54102 target to USB via Link2 probe

The sensor shield board includes:

• Bosch Sensortec BMI0155 intertial measurement unit, BMC150 digital compass, BMM150 magnetometer, and BMP280 pressure/temp sensors

• MAX44000 ambient light/proximity sensor

• Vishay VSOP98260 IR remote control sensor/driver

• ACKme AMS002 Bluetooth LE module

• Headers for easy prototyping of additional SPI and I2C sensors

• Sensor Interface & Data Processing

The Solution includes NXP’s Sensor Framework and the Bosch BSX Lite sensor fusion middleware needed to compute Euler orientation data (heading, roll, pitch and yaw), linear acceleration, gravity, and quaternion using 6- and 9-axis accelerometer, magnetometer and gyroscope data. 

The LPC Sensor Framework software provides an easy to use, modular software approach that handles sensor initialization, calibration, sensor interfacing for getting raw data, timing/bundling/batching of raw and processed sensor data, interfacing with the Sensor Fusion Library, power management, and providing a clear command/processed-data interface to either the user application or a host processor over I2C or SPI.

WEBINAR

USB Type-C is a new Connector Standard billed as the ‘Last Connector’, that will allow multi-function signaling (e.g. USB, Display Port, etc.) to pass over a small form factor, compact, high reliability connector. NXP’s offers a broad portfolio of best-in-class solutions for this emerging market, including USB-Power Delivery, X-Bar Switches, Microcontrollers, Authentication, AC/DC adapter components and Load Switches and ESD/EMI solutions. This webinar will give an overview of the Type-C interface and NXP solutions.

To register, click the link below:

https://webinar.techonline com/19368?keycode=xxxxxx

https://webinar.techonline.com/19368?keycode=xxxxxx

67

TECH REPORT

67

Sensor Processing ApplicationsThis all-in-one solution, based on the ultra-low power LPC54102

microcontroller, provides everything needed to bring sensor-based

motion and other sensor-processing applications to market quickly. NXP

has partnered with Bosch Sensortec to offer an integrated solution that

makes it easy to incorporate motion/inertia and other sensor data into

a variety of end applications. The solution includes sensors for

6- and 9-axis motion vectors as well as

temperature, proximity

and ambient light.

NXP’s Complete Solution for

MOTION-BASED, ALWAYS-ON

SENSOR PROCESSING/MOTION

The LPC54102 Sensor Processing/Motion Solution includes everything you need to add motion and other sensor-based features to your applications. Using the ultra-low-power, dual-core LPC54102 microcontroller, it provides a scalable solution for a family of connected sensor-based products.

Solution Includes:

• Reference hardware including the LPCXpresso54102 board and a sensor shield with nine sensors and Bluetooth LE wireless connectivity

• Bosch BSX Lite sensor fusion middleware

• Bosch sensor drivers

• LPC Sensor Framework to simplify sensor-processing and support additional sensors

• Complete reference documentation

FLEXIBLE, EXPENDABLE REFERENCE HARDWARE

The Solution’s LPCXpresso54102 development board includes:

• Large set of pinouts for measurement and prototyping

• Built-in Hi-Speed USB debug probe (Link2) and support for external debug probes

• On-board 1.8/3.3V or external power supply options for LPC54102

• Built-in MCU power consumption and supply voltage measurement for LPC54102 MCU and sensor shield board

• UART, I2C and SPI port bridging form LPC54102 target to USB via Link2 probe

The sensor shield board includes:

• Bosch Sensortec BMI0155 intertial measurement unit, BMC150 digital compass, BMM150 magnetometer, and BMP280 pressure/temp sensors

• MAX44000 ambient light/proximity sensor

• Vishay VSOP98260 IR remote control sensor/driver

• ACKme AMS002 Bluetooth LE module

• Headers for easy prototyping of additional SPI and I2C sensors

• Sensor Interface & Data Processing

The Solution includes NXP’s Sensor Framework and the Bosch BSX Lite sensor fusion middleware needed to compute Euler orientation data (heading, roll, pitch and yaw), linear acceleration, gravity, and quaternion using 6- and 9-axis accelerometer, magnetometer and gyroscope data. 

The LPC Sensor Framework software provides an easy to use, modular software approach that handles sensor initialization, calibration, sensor interfacing for getting raw data, timing/bundling/batching of raw and processed sensor data, interfacing with the Sensor Fusion Library, power management, and providing a clear command/processed-data interface to either the user application or a host processor over I2C or SPI.

WEBINAR

USB Type-C is a new Connector Standard billed as the ‘Last Connector’, that will allow multi-function signaling (e.g. USB, Display Port, etc.) to pass over a small form factor, compact, high reliability connector. NXP’s offers a broad portfolio of best-in-class solutions for this emerging market, including USB-Power Delivery, X-Bar Switches, Microcontrollers, Authentication, AC/DC adapter components and Load Switches and ESD/EMI solutions. This webinar will give an overview of the Type-C interface and NXP solutions.

To register, click the link below:

https://webinar.techonline com/19368?keycode=xxxxxx

Click Here

Click here

CLICK HERE

Sierra Circuits:A Complete PCB Resource

PLUS: The Ground ” Myth in PrintedCircuits

“

PCB Resin Reactor+

Ken BahlCEO of Sierra Circuits

Let There Be

How Cree reinvented the light bulb

LIGHT

David ElienVP of Marketing & Business

Development, Cree, Inc.

New LED Filament Tower

Cutting Edge Flatscreen Technologies

+

+

M o v i n g T o w a r d s

a Clean Energy

FUTURE— Hugo van Nispen, COO of DNV KEMA

MCU Wars 32-bit MCU Comparison

Cutting Edge

SPICEModeling

Freescale and TI Embedded

Modules

ARMCortex

Programming

From Concept to

Reality Wolfgang Heinz-Fischer

Head of Marketing & PR, TQ-Group

Low-Power Design Techniques

TQ-Group’s Comprehensive Design Process

+

+

PowerDeveloper

Octobe r 20 13

Designing forDurability

View more EEWeb magazines— Click Here