how to measure the loop stability of cmfb loop

[manissri]

how to measure opamp stability

hi all
i am desinging the fully differential opamp with cmfb loop.
how i measure the loop stability of cmfb loop.
and for loop to be stable what should the criteria of gain and phase margin of loop.
my supply voltage is 0-5 volt and common output voltage is 2.5 volt.
regrrrds
manish

 

[manissri]

opamp loop stability

can any body plz tell me how to measure the cmfb loop stability in cmfb loop.
regards
manish

 

[suria3]

differential loop stability

Basically, you need to break the loop of the CMFB. You can break the loop at any point around the loop but the result (Phase Margin) will be the same. The best loop breaking is between the CM dc extraction resistor and opamp that you used for comparing the reference voltage. Then you run AC analysis and proble the gain and phase from the output over input where you break the loop. In this stability measurement, you need to ensure the opamp gain is high enough (60dB and above) for minimizing offset of comparison and the PM is recommended to be 60 degree and above. Then your CMFB will be working fine and your output swing will be symmetric. Thanks

 

[manissri]

thanx suria....

Added after 46 minutes:

hi suria ..
i am sending my circuit which is the fully differential folded cascode with gain boost with cmfb ckt.
plz tell me where i break the loop and apply the ac source .
my suppy is 0-5v and VCM=2.5v
regrds
manish

 

[suria3]

thanx suria....

Added after 46 minutes:

hi suria ..
i am sending my circuit which is the fully differential folded cascode with gain boost with cmfb ckt.
plz tell me where i break the loop and apply the ac source .
my suppy is 0-5v and VCM=2.5v
regrds
manish

Hi Manissri,

I noticed from your schematic that there are 3 possible points you can break to measure the stability of the CMFB loop. You can try one by one breaking the loop and compare the PM and loop gain.

1) 1st break the loop between output of CMFB and that connecting the 2 PMOS of the folded cascode stage. Give AC = 1 at output of CMFB and DC at gate of PMOS. It will be accurate if you use Middle Brook method. Then measure the PM.

2) 2nd connect back the (1) loop and now break the loop between outp from folded cascode and CMFB.

3) 3rd connect back (2) loop and now break the loop between outn from folded cascode and CMFB. Apply AC=1 at output of outp then measure PM.

After measuring all the loop PM, you should be able to get the same loop gain and PM. Pls refer to the attachment of Middle Brook method. The IN pin and OUT suppose to connect at where you break the loop.

Hope this helps you.
Suria

 

[manissri]

Thanx suria ..
one question when i will break my loop then this time my opamp input should be at 2.5 volt ( 0-5 volt suppy).
r i do not apply any input on the opamp input.
what is the correct method.
actually i study somewhere that if u r analysing the loop then all dc input should be at ground (if suppy is 0 -5 volt)
and where is the attachment of middle brook method .
plz send it.
i shall be highly obliged.
regards
manish

 

[suria3]

Hi Manissri,

When you are about to find the PM or the loop gain of a feedback system, you have to run the AC analysis over freq. So, all your input of your opamp should be given dc value. If you ground the input to opamp, then your circuit is not operating and all the dc point of the subsequent stages is not in actual dc operating condition. Thus giving 2.5V at your opamp is fine to measure your loop gain and PM. Pls see the attach file for middle brook measurement.

Thanks,
Suria

 

[manissri]

thanz suria
for ur help..
one question that u have seen me circuit of opamp that there is doublet generation in this circuit which can hamper the settling time.
so the doublet will be a zero and pole which r closer to each other.
so from where this zero will be originated. the pole will be non dominant pole of main folded cascode opamp.
i study the book the zero is due to feed forward path of cascode mos
(cgd) . but i want to know how due to this feed forward path the zero will be produced.

and how this doublet will be removed.
what i have done that i fixed the gain boosted opamp ugb between the dominant pole and nondominant pole of close loop gain of main opamp.
(as the klass bult paper told)
but i want to know by doing this how the doublet r removed.
it would be highly appreciated if u could plz answer my doubts.
thanz and regards
manish


regrds
ma

 

[montage2000]

the phase margin shoule > 60 degree
and gain margin should > 10db