How to simulate the stability of a fully differential amplifier using the iprobe?

Use the diffstbprobe from analogLib instead of the iprobe.

Thanks a lot for your great help.

Unfortunately, there is no diffstbprobe in the analogLib. The Cadence version is IC5141 USR6. 

I found a diffprobe constructing by a iprobe and other electronic units in spectreref.pdf of doc directory. It is presented below. Is this diffprobe idential to the diffstbprobe you mentioned? Does it work fine?

Thank you. 

If using and old version like this, you can use cmdmprobe from analogLib instead. This has been replaced by diffstbprobe in later releases, since that is an improved way of measuring the common mode and differential mode stability in a way that can handle "unbalanced" loops - i.e. where  you have significant common mode to differential mode (or vice versa) gain. The cmdmprobe is fine for balanced differential loops, where both halves are the same.

The implementation of cmdmprobe is essentially the same as the netlist  you've provided, but the UI understands how to probe the correct iprobe inside the cmdmprobe.

Regards,

Andrew.

If your circuit is not perfectly symmetric, you can also build your own diffstbprobe from iprobes and ideal baluns as described in http://www.designers-guide.org/Forum/YaBB.pl?num=1190707743.

Thank for all your help.

I've checked the cmdmprobe, which circuit is same as the "diffprobe" subcircuit. Does diffstbprobe has the some circuit inside? Does it contain some iprobe inside?

Thanks a lot. 

We don't reveal the precise details of how diffstbprobe is implemented, but broadly speaking it is along the lines of the balun-based approach that Frank mentioned. 

Regards,

Andrew.

Thank you, Mr. Andrew. 

xxgenerall said:

Thank for all your help.

I've checked the cmdmprobe, which circuit is same as the "diffprobe" subcircuit. Does diffstbprobe has the some circuit inside? Does it contain some iprobe inside?

Thanks a lot. 

In a few experiments I did some time ago, the diffstbprobe behaved exactly like a circuit consisting of two ideal_balun and two iprobe components from analogLib, with one iprobe connected between the d ports of the ideal_balun components, and the other iprobe connected between their c ports. This would be exactly the circuit described in reply #1 of http://www.designers-guide.org/Forum/YaBB.pl?num=1190707743 (3 years before the diffstbprobe was introduced in September 2010 in MMSIM 10.1).

Frank Wiedmann said:

xxgenerall said:

Thank for all your help.

I've checked the cmdmprobe, which circuit is same as the "diffprobe" subcircuit. Does diffstbprobe has the some circuit inside? Does it contain some iprobe inside?

Thanks a lot. 

In a few experiments I did some time ago, the diffstbprobe behaved exactly like a circuit consisting of two ideal_balun and two iprobe components from analogLib, with one iprobe connected between the d ports of the ideal_balun components, and the other iprobe connected between their c ports. This would be exactly the circuit described in reply #1 of http://www.designers-guide.org/Forum/YaBB.pl?num=1190707743 (3 years before the diffstbprobe was introduced in September 2010 in MMSIM 10.1).

To make this a little more clear, here is an image of this configuration: