phase margin with corner analyses and high temperature

[Junus2012]

Hello Every one

I am simulating the phase margin of an Op-Amp with the corner analyses and with full industrial temperature range
I am getting a typical mean value of 47, and at some corners (WP,WO) and at the elevated temperature the phase margin is dropping down to 40.

I know that the phase margin should be >=45. but still I am wondering if I am deviating from this value at the stated condition is it possible to still consider it stable??

thank you very much

 

[erikl]

Sure, it's still stable, but you'll get some overshoot and ringing (from an input step edge), see e.g. this excerpt from Razavi's book:

 

[Junus2012]

Thank you very much erikl for your reply

the book is shows the P.M of 45 , what about 40 ?? because I am wondering that may be it is standard definition of the P.M from and above 45 while I am getting 40.

if it is still stable as you kindly stated then what the limitation of the non stable P.M??

thank you again

 

[LvW]

Thank you very much erikl for your reply

the book is shows the P.M of 45 , what about 40 ?? because I am wondering that may be it is standard definition of the P.M from and above 45 while I am getting 40.
if it is still stable as you kindly stated then what the limitation of the non stable P.M??
thank you again

There is no "standard definition" regarding minimum PM value.
The excerpt from razavis book clearly shows the impact if the PM is 45 deg or even smaller.
As Erikl has pointed out - the consequence of a PM that is too small is a step response with overshoot and ringing - that`s all.
If you can live with heavy ringing, you can accept a margin of - let`s say - 30 or 25 deg.
The definition of instability is identical to a PM of 0 deg. (or even below).

 

[Junus2012]

thank you friends for your helpful participation
your comments are really clear for me,

 

[LvW]

Yes, that`s no surprise. The step response looks even better for a PM=60 deg.
But - what`s your problem now?

 

[Junus2012]

no nothing just with the stability issue, any way I have updated my last comment you replied after I understood it from you and erikl, you can see I have deleted a part of it.

However, as the fast response behavioral is not a matter in my design, I will still dealing with the circuit with P.M=40 as it still stable as you and erikl cleared to me

thank you

Yes, that`s no surprise. The step response looks even better for a PM=60 deg.
But - what`s your problem now?