Stability of LDO(difference between AC &transient)?

[Electronicmember]

Hello everyone,

I am designing a LDO right now.

The cirucit works well in the AC simulation, there is sufficient phase margin.

But in the transient simulation, I set the load current as a variable which increases from 0 to 50mA as a pulse signal. When the load current pulse rises from 0 to 50mA, the output of this LDO start to oscillate.

The oscillation happens for all of the transient settings(conservative, moderate and liberal).
But when I selected the option "Euler" in the transient simulation setting, there is only small ringing left at the output load current rising time.

So I am really wondering how to explain the conflict between AC and transient results? Is this circuit really oscillating? which transient result could I trust? and are there some further methods to verify the result?

 

[dick_freebird]

TRAP method leaves an ugly triangular residue at roughly VNTOL
amplitude on a lot of nodes. This can be gained up and look bad.
Or it could be some other reality, like a BJT operating too close to
saturation and turning your marginally-small-signal-stable
amplifier into a large signal relaxation oscillator.

 

[keith1200rs]

Try reducing the maximum time step. You could also use gear integration, but I prefer to avoid that as it can hide instabilities.

Keith

 

[love_analog]

What kind of oscillations are you seeing. If they are triangular wave, then yes, its most likely numerical oscillation caused by simulator (most trapezoidal).

You should always check with transient since its more realistic. Typically I haven't seen LDOs that are "very" stable in ac and oscillate in transient but thats just my observation

 

[LvW]

...........
You should always check with transient since its more realistic. Typically I haven't seen LDOs that are "very" stable in ac and oscillate in transient but thats just my observation

Yes, I can confirm this. But, more than that, that can be be observed not only for LDO's. The reason is simply, that you cannot trust ac simulations if the circuit is unstable. AC analysis are small signal analyses which cannot reveal instabilities.
Simple example: Opamp with positive resistive feedback looks stable in ac analysis.