Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Help me with loop stability check

Status
Not open for further replies.
M

mike_bihan

Full Member level 3
Joined
Mar 21, 2002
Messages
182
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,296
Activity points
1,531
Loop stability check

Hi, gurus on control theory:

I need help on loop stability check.

I did small signal analysis and is going to do more transient simulation by injecting a current pulse into the loop node. I observe the response to judge the phase margin. It is somewhat like step response analysis.

Can anyone introduce any document on it? Is this method valid as step response analysis? How to judge the phase margin from the waveform overshoot/undershoot?

Many thanks,
Han
 

Re: Loop stability check

Dear Mike,

You can do the small signal analysis(which is ac analysis) and check the loop phase margin. If you want to check with the step response.there is no point in cutting the loop. Remember that a transient is a real time analysis and in most cases, you put the same set up as reality in transient analysis. Of course in the step response, you should see atleast a second order settling in the output to your step input.(what i mean by this is that, your output should settle after 2 peaks max). This will ensure that your circuit is even stable in the closed loop conditions as well.

Added after 9 minutes:

Dear Mike,

I am sure that this document helps you...........It talks of Peak Overshoot and the phase margin. This is suitable for 2nd order systems, which most of our circuits fall in.
 

    M

    mike_bihan

    Points: 2
    Helpful Answer Positive Rating
Re: Loop stability check

mike, yes u could do that.

but what i normally do is i put a series RC with an impulse voltage with steep rise/fall time and observe all the node voltages to ensure they settle in a satisfactory manner. Do not rely entirely on AC, u could be in for a surprise.
Inject the signal into different nodes at the same time or once at a time, try many different combinations.

And most importantly, predict your parasitics(inductance, ccpacitance, resistance), and put it in ur sim...
absolutely crucial.
 

Re: Loop stability check

Run ac analysis,see amplitude and phase response.find the phase when the gain equate 0dB,and calculate the phase margin is larger than 45° or not.
 

Re: Loop stability check

I have developed a method to check phase margin by pulse injection. Pls advice if you have different idea.

I assume the disturb by pulse injection is same as step response.

Thanks,
Han

Added after 2 minutes:

doc
 

Re: Loop stability check

Vamsi Mocherla said:
Dear Mike,

You can do the small signal analysis(which is ac analysis) and check the loop phase margin. If you want to check with the step response.there is no point in cutting the loop. Remember that a transient is a real time analysis and in most cases, you put the same set up as reality in transient analysis. Of course in the step response, you should see atleast a second order settling in the output to your step input.(what i mean by this is that, your output should settle after 2 peaks max). This will ensure that your circuit is even stable in the closed loop conditions as well.

Added after 9 minutes:

Dear Mike,

I am sure that this document helps you...........It talks of Peak Overshoot and the phase margin. This is suitable for 2nd order systems, which most of our circuits fall in.
Click to expand...

Vamsi Mocherla:

Both your attached doc and inputs are helpful. thanks.
I attached one pdf on this method. Still, i have no idea to have the step response of the close-loop. I can only do some pulse injection and observe its behavior there. I assume it is close to step response settling.

Thanks,
Han
 

Re: Loop stability check

So does the method of injecting current work in estimating phase margin? You can probably compare your result with one using AC analysis. In the new cadence there is actually a stability analysis where it will do automatically for you without having to break the loop yourself.

Please let us know if the method of current injection is comparable to the AC analysis. I am curious

Thanks
 

Re: Loop stability check

ccw27 said:
So does the method of injecting current work in estimating phase margin? You can probably compare your result with one using AC analysis. In the new cadence there is actually a stability analysis where it will do automatically for you without having to break the loop yourself.

Please let us know if the method of current injection is comparable to the AC analysis. I am curious

Thanks
Click to expand...

I am curious about the new function on cadence loop analysis. Is there any doc or manual on it? Can you upload it?
I confirmed my method get similar results as AC analysis. However, there is one mistake in my document, i will update it later.

thanks,
Han

Added after 1 hours 57 minutes:

document updated.
 

Loop stability check

The loop stability is OK when all the
phase margin > 0,of course we should
use the worst phase margin.
 

Re: Loop stability check

hey mike_bihan,

Insert iprobe into the loop where you usually break and in ADE select stb analysis and choose the probe you just inserted. Run the analysis as you will would for ac. If not clear let me know.
 

Re: Loop stability check

phase margin is great than 60 you loop is stable
 

Re: Loop stability check

but what i normally do is i put a series RC with an impulse voltage with steep rise/fall time and observe all the node voltages to ensure they settle in a satisfactory manner. Do not rely entirely on AC, u could be in for a surprise.
Inject the signal into different nodes at the same time or once at a time, try many different combinations.
Click to expand...


Hi,

What typical values do u use for R and C? Also how much impulse voltage do u inject? 100mV?

Thanks
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top