Slew-rate amplifier - overshoot

Hi guys,

Does anyone has experienced a high overshoot with you amplifier, say 2 stage amplifier, slews? What might cause this effect? It's because the opamp goes out of its small signal mode and it starts to work as an open loop system? If yes, how it manages to get back, or in other words, recover from the overshoot?

Regards

Do you mean like the purple or green lines here ?

Hi, thanks for you reply.

What I mean is something like this:



My PM is around 55 deg. and this happens when I apply an input square voltage of (aroundW) 100mV.

What might happening?

Regards

Can be large signal behavior of the compensation circuit, e.g. if the first stage goes into saturation during slew rate limiting. Most commercial OPs are designed in a way that these unwanted effects are avoided.

What you mean by first stage goes into saturation during slew rate limiting? Can you exemplify? I thought it could be related to something regarding the fact of the opamp not being operating in the small signal regime, that is, the compensation is not able to handle to catch that behaviour and so it simple lets it go to whichever place it wants and then return.

CAMALEAO said:

My PM is around 55 deg. and this happens when I apply an input square voltage of (aroundW) 100mV.

Click to expand...

It is because of the PM. Design for PM ~ 100 and you will see no overshoot.

- - - Updated - - -

But then, your slew rate will decrease.

Only PM is necessary but insufficient condition for the stability of a system.Root Locus and Bode Plots should also be examined.Overshoot is a kind of stability fact and generally not desired.

Even slewing? Slew is large signal and stability make sense in small signal analysis.

Compensation networks for small signal stability will
impose a slew rate limitation, and uncompensated
(or marginally stable) amps will overshoot (if not
oscillate).

Particularly in bipolar amps, if you "wind up" any of
the stages into saturation or cutoff, the overshoot
will be worse than when the loop remains linear.
MOS amps can also show this if devices "go dark"
and require internal slew time to get back to linear
operation.

The design of piece-part op amps is often a battle
between desired slew rate and stability; you see
the result as (for example) families of 1X, 5X, 10X
gain-stable op amps based around a single design
with varying compensation capacitor values. The
BW / slew desired for an A=10 loop could not be
had from the A=1 stable amplifier, and the needed
stability could not be had from the A=10 stable amp
in a lower gain configuration. This is a fundamental
technology bind (fT and parasitics, @ voltage).

It can be calculated that PM of 59 degree is the condition for no overshoot of an OP in feedback with two poles (if I remember right). PM of 100 degree is no useful aim and would require additional lead-lag compensation even in an originally first order system.

I agree that you have to look at pole zero locations for an exact analysis. But the problem of this thread is clearly non-linear behavior, a problem beyond linear analysis, PM or pole-zero pattern. The red waveform in post #3 will be hardly achieved by a linear system.

"Wind-up", as brought in by dick_freebird is a descriptive term for what I mentioned in post #4, similar to integrator wind-up of PI controller in saturation. It's a non-linear circuit property which isn't accessed by pole-zero and gain parameters. You have to perform large signal circuit analysis to see it.