Designer’s Guide Community

newic

Senior Member

Offline

Posts: 138

stb phase margin
Dec 12^th, 2014, 5:57am

Hi

I use spectre stb analysis with the same setup as in
http://www.lumerink.com/courses/ece5411/Handouts/Loop%20Stability%20Analysis.pdf...

The phase will start at +180o degree right. As long as the phase is >0 degree below unity gain freq (ugf), the design is consider stable right.

What if the phase starts at 0 degree? will it consider stable? The stb still reported +ve phase margin. Is stb reported phase margin reliable?

Back to top

IP Logged

raja.cedt Senior Fellow Offline Posts: 1516 Germany	Re: stb phase margin Reply #1 - Dec 12^th, 2014, 6:02am check your circuit, some where you might have wrongly connected such that it will be +ve feedback. Check your dc operating point once before you start stb. Thanks, Raj.
Back to top	IP Logged

sheldon

Community Fellow

Offline

Posts: 751

Re: stb phase margin
Reply #2 - Dec 12^th, 2014, 9:13am

Remember that the Spectre stability analysis is reporting the loop
gain. If the phase starts at 0 degrees then it is likely that either
the inputs are reversed or the circuit is not biased properly.

If the phase is 0 degrees at zero then the circuit has positive feedback.
If the gain is low then the circuit may behave. That is if the gain is
low, then it won't latch up.

Since phase margin measured at the unity gain frequency, the phase
at dc is not relevant. That even loops with multiple poles and zeroes
can be stable and have positive margin. If the effect of the poles
and zeroes can be neglected at the cross over frequency.

No, if you are using the Bode plot then
1) if the gain is 1, then phase margin needs to be positive,
typically greater than 45 degrees
2) if the phase is 0, then the gain needs to be less than zero

Back to top

IP Logged

newic Senior Member Offline Posts: 138	Re: stb phase margin Reply #3 - Dec 12^th, 2014, 11:16pm Hi Sheldon, Does it mean that I cannot rely on the stb phase margin report right? the STB result a +ve phase margin but when i look at the phase response, it starts at 0degree at 1Hz. it should be unstable right. In most of the corners, it starts at +180o at 1Hz, but in some cases, it starts in 0 degree.
Back to top	IP Logged

newic Senior Member Offline Posts: 138	Re: stb phase margin Reply #4 - Dec 12^th, 2014, 11:18pm hi Raj, the DC value is correct. i have confirmed that
Back to top	IP Logged

buddypoor

Community Fellow

Offline

Posts: 529
Bremen, Germany

Re: stb phase margin
Reply #5 - Dec 13^th, 2014, 2:15am

newic wrote on Dec 12^th, 2014, 11:16pm:

Hi Sheldon,

Does it mean that I cannot rely on the stb phase margin report right?
the STB result a +ve phase margin but when i look at the phase response, it starts at 0degree at 1Hz. it should be unstable right.
In most of the corners, it starts at +180o at 1Hz, but in some cases, it starts in 0 degree.

Let me state the following:

1.) For stable feedback circuits the phase of the loop gain always must start at -180deg . This is necessary for negative dc feedback (stable operational point).

2.) In this case, the gain margin is to be determined at a frequency where the loop phase reaches -360deg (0 deg).

3.) Sometimes, not the loop gain is analyzed but the product of all units within the loop only (without the minus sign at the summing node).
This product is, of course, identical to the negative of the loop gain. In this case, the phase starts at 0 deg and stability is checked at the -180deg point.

Back to top

LvW (buddypoor: In memory of the great late Buddy Rich)

IP Logged

newic Senior Member Offline Posts: 138	Re: stb phase margin Reply #6 - Dec 14^th, 2014, 5:28pm hi buddypoor based on the setup in http://www.lumerink.com/courses/ece5411/Handouts/Loop%20Stability%20Analysis.pdf.... where the ac source /iprobe at the feedback path. the phase should start at +180o from stb plot right. as long as the phase > 0o before UGF. it is stable.
Back to top	IP Logged

Frank Wiedmann Community Fellow Offline Posts: 677 Munich, Germany	Re: stb phase margin Reply #7 - Dec 16^th, 2014, 12:58am I suggest that you take a look at http://www.designers-guide.org/Forum/YaBB.pl?num=1294178255.
Back to top	IP Logged

buddypoor

Community Fellow

Offline

Posts: 529
Bremen, Germany

Re: stb phase margin
Reply #8 - Dec 16^th, 2014, 3:55am

newic wrote on Dec 14^th, 2014, 5:28pm:

hi buddypoor
based on the setup in http://www.lumerink.com/courses/ece5411/Handouts/Loop%20Stability%20Analysis.pdf....
where the ac source /iprobe �at the feedback path.
the phase should start at +180o from stb plot right. as long as the phase > 0o before UGF. it is stable.

I think both cases are identical:
* Phase from +180deg to zero deg;
* Phase from -180deg to -360deg.

By the way: I couldn`t open the given link.

Back to top

LvW (buddypoor: In memory of the great late Buddy Rich)

IP Logged

newic Senior Member Offline Posts: 138	Re: stb phase margin Reply #9 - Dec 16^th, 2014, 7:28pm the link is broken due to the "." in the end of the link pls use the http://www.lumerink.com/courses/ece5411/Handouts/Loop%20Stability%20Analysis.pdf...
Back to top	IP Logged

newic Senior Member Offline Posts: 138	Re: stb phase margin Reply #10 - Dec 16^th, 2014, 7:31pm To add on, my circuit has 2 loops: main loop and a fast loop. most of the corners start at 180o but in few cases, it starts at 0o. Can STB analyse 2 loops correctly? as i only break in the main loop (feedback path).
Back to top	IP Logged

buddypoor Community Fellow Offline Posts: 529 Bremen, Germany	Re: stb phase margin Reply #11 - Dec 17^th, 2014, 3:30am newic wrote on Dec 16^th, 2014, 7:31pm: Can STB analyse 2 loops correctly? as i only break in the main loop (feedback path). More than that, how to proceed if no main loop can be identified because there are two nested loops? (I think, STB analysis requires that both loops are to be opened at the same time)
Back to top	LvW (buddypoor: In memory of the great late Buddy Rich) IP Logged

newic Senior Member Offline Posts: 138	Re: stb phase margin Reply #12 - Dec 17^th, 2014, 7:09pm Hi buddypoor "(I think, STB analysis requires that both loops are to be opened at the same time) " how to achieve that in STB if having 2 loops? In the STB setup, only one probe is allowed.
Back to top	IP Logged

buddypoor Community Fellow Offline Posts: 529 Bremen, Germany	Re: stb phase margin Reply #13 - Dec 18^th, 2014, 12:53am newic wrote on Dec 17^th, 2014, 7:09pm: In the STB setup, only one probe is allowed. Yes - and that is the problem (in some cases) because we have to decide which of the loops is dominating and mainly determines the circuits behaviour? Up to now - I didn`t find any answer to this general question.
Back to top	LvW (buddypoor: In memory of the great late Buddy Rich) IP Logged

raja.cedt Senior Fellow Offline Posts: 1516 Germany	Re: stb phase margin Reply #14 - Dec 18^th, 2014, 2:24am Hi Newic, When you have several loops you should find a common point where you can break all the loop. For example band gap, opamp o/p is prefect point to keep stb-probe. I also don't know how to proceed if I can't find a common break point. Thanks, Raj.
Back to top	IP Logged