June 04, 2001
Question:Why compensate in the feedback loop for output capacitor ESR?
Original Question: Why has it been frequently recommended to compensate the ESR Zero by a pole in the compensator. What happens if the ESR zero is kept uncompensated? N. S. 05/17/01
Answer: The ESR (Equivalent Series Resistance) of the output capacitor provides a zero that can provide phase margin to the control loop. Sometimes output capacitors are fused in high reliability application to prevent a failure of the output bus if a capacitor shorts. This added resistance has the same effect. Also, in the early days, it was sometimes suggested that additional resistance be added in series with the capacitor for compensating the control loop -- a poor approach because of the degradation of load transient response.
There are two reasons I can think of why one might compensate for the ESR zero and there may be more. One is purely esthetic. Some designers like the overall control loop to go down at a pure slope of one (there is no reason to do this and it may not give the best response) and to do this they need to compensate for the ESR zero.
The other is more important. After you go through unity gain at your desired cross-over frequency, you do not want the gain to come back up at a high frequency, so it is always good design practice (mandatory for me) to make sure the gain is solidly killed at high frequency. This prevents surprises by noise, etc., causing problems. However, the pole that kills the high frequency gain need not necessarily compensate for the ESR.
There may be other reasons I am unaware of.
Posted by Jerrold Foutz at June 4, 2001 10:40 AM