How to measure open loop H(s)?

When designing constant current linear dummy loads, and when I was (attempting) to design a really nice linear power supply with op amps and pass transistors, I consistently run into the devil of the circuits that incorporate feedback. INSTABILITY! Especially if I use fast op amps, or MOSFETs, etc.

My understanding as to why this happens is because I still have some positive gain at the point at the frequencies where signal the total loop has that dreaded -180 degree phase (gain margin) and then where the gain is unity and phase is at -180, again allowing that frequency component to be successively amplified (or never attenuated) as that wave whirls around the loop.

However, with op amp circuits, how am I supposed to get an accurate open loop measurement, when the DC gain is so stupidly high the input to an op amp is essentially a comparator?


http://www.ti.com/lit/an/snva364a/snva364a.pdf

This article suggests whacking in a 20 ohm resistor and injecting a signal into it, and measuring the voltage on either side of it (?) How is this supposed to work? Is there a better method in something like LTspice?

That's a very valuable article. I think it speaks for itself.

-max- (author)  steveastrouk24 days ago

But is that the best way to do it in simulation?

Yes, if you want to avoid unbounded answers ! You are right, its a bitch to do "open-loop" - that's why your link was quite fascinating to me.