Analog Design Engineer Interview Questions

I drew a resistive voltage divider and was then told, "That won't track over process." I then drew a replica bias structure (diode connected devices that matched the original amplifier) to which I was told, "How are you going to use that if your supply is only 1V?" I then drew a diode connected device with a W/L which was 1/4 the value of the device whose gate needed bias. Was then told, "That's right, but we usually use 1/5 or 1/6 for body effect compensation." Answered the question asked each time but they wanted a very specific answer from the beginning.

I responded with some ideas, and he must have liked them.

I explain using CMOS which i had done. Cudnt explain them properly using BJT (same circuit). PTAT and CTAT, K factor were asked.

Matching in MOSFET transistor inverters is necessary because the electrons that move in the NMOS transistor move at a faster rate than the holes that move in the PMOS transistor. This means when a signal is received, the NMOS will react before the PMOS. To equalize this, they transistors must be matched (by adjusting their widths) so that their transition times are the same.

My answer was we don't need to use it. But we have to change our filter structure. Using UP and DOWN fit well to charge pump and passive RC filter topology. Since DOWN is always 50% duty cycle when there is data transition, it acts as a reference. However, if we took the UP pulse alone, feed it to an active filter with one pole ar zero and one zero at a non- zero frequency we preserve the loop dynamics.

Ideally, I would like THEM to answer this for me!