[SOLVED] which one is correct of measuring ldo stability

[ella1923]

which one is the correct way of measuring ldo stability

Hi All,

Why is that the ldo stability is different when measuring it in A & measuring it in B?
Which one is correct?

A.

B.



Thanks,
Ella

 

[goldsmith]

Hey Dude!
Hi
Non of them are not LDO. And about compensation circuit: I think its place is between the out put of op amp and its inverting input.
Best wishes
Goldsmith

 

[ella1923]

Hey Dude!
Hi
Non of them are not LDO. And about compensation circuit: I think its place is between the out put of op amp and its inverting input.
Best wishes
Goldsmith

Sorry. I haven't included the Cout in uF, Resr in mohm & load resistors in the
images. I'm actually using pmos as pass transistor. Currently, I am testing
closed loop stability using the instance iprobe in cadence. Initially, it is
connected like in A & stable. I am expecting the same waveform when
connecting compensation ckt as in B, but it changes phase margin. I wonder
which one is correct...

 

[mtwieg]

What is "IPROBE"? And what is actually in the compensation network? How are you measuring loop response? The feedback path is split, so it really matters. It's hard to tell what's going on what's going on with such a vague diagram.

 

[ella1923]

What is "IPROBE"? And what is actually in the compensation network? How are you measuring loop response? The feedback path is split, so it really matters. It's hard to tell what's going on what's going on with such a vague diagram.

iprobe is an analysis component in cadence when running stb analysis. compensation ckt
consists of RC network connecting error amplifier's output and Vout while another feedback
capacitors connecting Vout to the noninverting input of the error amplifer...


stability using B still shows stable system, but phase margin is not the same in A. Shouldn't
they have the same waveform...

 

[dick_freebird]

Your circuits are topologically different in the compensation so why do
you expect identical response?

I'm not a real big fan of stb. When everything is right it's a handy shortcut.
But determining -whether- things are right, is still easier for me using AC
analysis and transient. Get the circuit working without the iprobe and
then figure out where to insert it.

I much prefer looking at step response for rough stability tweaking.
Once it looks critically damped then I go about proving the phase /
gain margins. LDOs are only small signal when things are sitting still.
The interesting challenge is a near-full-scale load step.

 

[ella1923]

Your circuits are topologically different in the compensation so why do
you expect identical response?

I'm not a real big fan of stb. When everything is right it's a handy shortcut.
But determining -whether- things are right, is still easier for me using AC
analysis and transient. Get the circuit working without the iprobe and
then figure out where to insert it.

I much prefer looking at step response for rough stability tweaking.
Once it looks critically damped then I go about proving the phase /
gain margins. LDOs are only small signal when things are sitting still.
The interesting challenge is a near-full-scale load step.

Circuit works perfectly fine without the iprobe having the same waveform in A.
And it makes more sense to me when checking the transient analysis. It is
easier to check stability when inserting iprobe while modifying or tuning results.
It began to bothers me when I tried to connect it in B. I was thinking it will show
the same waveform. Well, I gain another insight. Thanks.