18
HW5

HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

Page 1: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

HW5

Page 2: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

1. (60 points; OTA integrator design)

a) For the above connection give the nonlinear differential equation for the voltage output versus the voltage input using an MOS OTA.

Page 3: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through
Page 4: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

Design an OTAC circuit to give this differential equation using a MOS OTA made of 4007 transistors (with one tail current setting resistor). Use two 6V power supply batteries, set the tail current at 5mA, and use a 10nFd capacitor.

Page 5: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

Rb

1.0 3.0 10 30 100 300 1.0K 3.0K 10K

I(Rbias)

-25mA

-20mA

-15mA

-10mA

-5mA

0mA

I ran a sweep for {Rb}, the value of the bias resistor. The sweep intersects 5mA (the desired current) at about 1.142 kΩ.

Page 6: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

b) Sketch (by hand) the low frequency output current versus input differential voltage that you expect for this integrator with small signal sine wave inputs.

Page 7: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

Correction: 𝑣𝑜 would be a negative cosine wave, because it is the integration of 𝑖𝑐, and 𝑖𝑐 = −𝑖𝑜

Page 8: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

Determine the small signal gm and run Spice to check the integrator operation.

Page 9: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

Do a frequency response to determine over what frequency range it will work.

Page 10: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

Frequency plot of the current amplitude through the output capacitor. No significant current goes through the transistor until around 1GHz, this seems really high…

Page 11: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

2. (40 points; current sources, mirrors)

a) Compare NMOS current sources made with one input/output pair of transistors versus a circuit using a cascode of two input/output pairs.

Page 12: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

One-pair:

Voltage sweep on {Vds1}. Measure the current into the drain of OP2.

Page 13: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

V_Vds1

0V 2V 4V 6V 8V 10V 12V 14V 16V 18V 20V

I(OP2:d)

0A

1.0mA

2.0mA

3.0mA

4.0mA

One-pair plot of 𝑉𝐷𝑆 across the mirror NMOS, vs the current through its drain. The saturation region still has a rising slope, which is undesirable.

Page 14: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

Two-pair (cascode):

Voltage sweep on {Vo}, measure current through the transistors (same through both) on the mirrored side).

Page 15: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

V_Vo

0V 2V 4V 6V 8V 10V 12V 14V 16V 18V 20V

I(Ca4:d)

0A

50uA

100uA

150uA

200uA

250uA

300uA

Two-pair cascode plot of 𝑉𝑜𝑢𝑡 across the both mirror transistors, vs the current through those mirror transistors. The saturation is now flat with a constant current, but that current is much lower than the one-pair.

Page 16: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

b) Design a 4 transistor CMOS bidirectional current mirror (that is, one which will work with input currents of any sign). Note that this will require a bias offset current. Check your design in Spice.

Page 17: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

For the bidirectional mirror, I could not figure out how to get the proper behavior in pspice with just 4 transistors. I used an NMOS cascode sink and a PMOS cascodesource in parallel, both of which connect to the load resistor (in the middle).

Page 18: HW5 - University Of Marylandnewcomb/courses/fall2015/303H/HW5Solutions.pdf · HW5 I did a bias-point simulation with a positive 20V source, which should have let current run through

HW5

I did a bias-point simulation with a positive 20V source, which should have let current run through the NMOS cascode, but not through the PMOS cascode. But the bulk terminal on P_cas1 was drawing -18.34mA. I don’t know why this happened.