Power supply Loop evaluation

Switching the load, or squirting in some current from a function generator, is a great way to evaluate loop stability.

One thing that makes sims different from reality is output capacitor ESR. A little ESR in an aluminum or tantalum cap can really stabilize a switcher loop. An ideal L followed by an ideal C is a terrible thing to close a loop around.

Does the ESR make as much of a difference in a current control Loop? From what I have read voltage feedback loop is much more touchy to the output capacitor where is a current Loop is not as touchy?

Depends on the topology. I've found that LT Spice does a pretty accurate job if you add some ESR to the output caps in the sim. It's best to measure the actual cap ESRs.

Any time you take the feedback from across the output cap, ESR adds a zero to the loop. Zeroes are good.

I like the ESR behavior of tantalum caps, but they have to be derated maybe 3:1 on voltage, or they may explode.

Polymer aluminum caps are great parts; esr's may actually be a bit low for best loop stability, but better than zero. I have been known to add a 30 milliohm resistor, or a PCB trace, in series with a polymer or ceramic cap to better define the ESR.

Remember, Bode only works for linear time-invariant circuits. If your loop is non-linear (biased however it is in the simulation), the results are crap. If the Bode plots (either simulated or measured) differ from the step response, the step response is correct.

Only to the extent that the non-linearities matter. _Everything_ is nonlinear, after all. Search on "describing function analysis".

Used with care, you can do more with Bode than might seem apparent -- but at some point, they do become crap.

Yes. Except you can't go assume it's the step response of a linear circuit and analyze from there.