Beware of Loop Gain - John Dunn, Consultant, Ambertec, P.E., P.C.

It is sometimes suggested that an op-amp can be made less susceptible to electromagnetic interference (EMI) by connecting a capacitance directly between the op-amp's non-inverting and inverting input pins, but doing so is something I very much counsel against.

Here's one example why:  

 

Some years ago, we had a seminar here on Long Island delivered by Bob Pease, then of National Semiconductor, where he suggested using such a capacitor, recommending a value of 1000 pF. I raised the question of such a capacitor possibly causing feedback instability due to loss of phase margin.

"I'm glad you asked that question!", he replied. Then he lowered the 1000 pF suggestion down to 100 pF.

Even then, I would still not like to do this. It is preferable to find ways to keep the EMI away from the op-amp altogether than trying to render that op-amp immune to incoming EMI's effects.

Comments

[John Dunn]

Deliberately putting a capacitance across the two op-amp inputs is not something I would recommend. However, in the case of a photo-diode connected there, that's exactlly what can happpen. In such a case, as you point out, Atul, one would have to take remeidal measures against that diode's capacitance and make a very careful stability analysis.

[Marty]

It is amazing how experiences differ. I was working at Fairchild Camera and they had made thousands of miniiature cockpit cameras for the Airforce. The camera had an automatic electromechanical automatic exposure control which had a most unusual problem. The camera would work perfectly until you tapped it. The exposure control servo would run away and completely shut down the iris.Normal operation could only be resumed by resetting the input power. I was handed the problem with a two foot stack of reports written by engineers who, for many years, had tried numerous solutions. I gathered a couple of expeienced techs and we started to trace and troubleshoot the exposure control's positional and velocity closed loops. In normal operation the voltages in the op-amp's positional loop were precise. However the velocity loop's op-amp's voltages didn't 'compute'. Naturally we changed the op-amp and got the same numbers. Now we really started to look around and discovered some 10 MHz computer clock frequency at the input of that op-amp but its output was perfectly clean. After a little thought it became clear that the 1 MHz op-amp would not pass through the 10MHz and the noise was saturating the amplifier.

Here is where the problem really develops. The assignment was 'fix it but don't touch it' since thousands of these cameras had been delivered to the Military. Well, I'll shorten the story. We found a couple of parts in the area of the op-amps input which could be removed and left connections on the PC board to the inputs of the op-amp. By putting a 100 Pf cap across the inputs we found that we cleaned up the inputs and the DC levels were found to be exactly as predicted. You could hit the camera with a sledge hammer, but the exposure control loop held steady.

Marty Kanner

[Michael Gambuzza]

Really important with photodiode amplifers. Compensation should be worked out for minimum 45 degree phase margin. An excellent book on the subject is entitled "Photodiode Amplifiers" by Jerrold Graeme who was with Burr-Brown for a long time. It has helped me many times on these designs where photodiode capacitance can be quite large.

[Frank Wiedmann]

I present some methods for loop gain simulation on my webpage http://sites.google.com/site/frankwiedmann/loopgain .