The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
https://designers-guide.org/forum/YaBB.pl?num=1355726908

Message started by tulip on Dec 16^th, 2012, 10:48pm

Title: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by tulip on Dec 16^th, 2012, 10:48pm

I need know Gain Margin of a differential VCO, the first figure is the simplified schematic of VCO. In the second figure, I add iprobe to simulate the GM, in the third figure I add cmdmprobe(CMDM=-1), the GM results used by iprobe and cmdmprobe are not the same.

In my opnion, iprobe should be used to simulate the GM, and the result obtained by iprobe is correct.

Is my opnion correct? Please help me, many thinks.

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by raja.cedt on Dec 17^th, 2012, 12:40am

hello,
can you show me how did you use CMDprobe, because cmd probe can be used for fully-diff ckts but here i-probe is suitable as far as i know. Your figs are not clearly visible.

For a +ve feedback ckt, -ve GM margin make sense,but it didn't understand what is the use with finding GM.

Thanks,
Raj.

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by rfidea on Dec 17^th, 2012, 12:11pm

I think that you should use the cmdmprobe since you have a differential circuit.

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by tulip on Dec 17^th, 2012, 4:58pm

I make the figure using cmdmprobe clear.

hello raja.cedit, you said "For a +ve feedback ckt, -ve GM margin make sense,but it didn't understand what is the use with finding GM." what is the meaning +ve feedback ckt? I need finding GM because I want to make sure the circuit can oscillate.

Hello rfidea, this circuit is a differential circuit, but when using iprobe, all the feedback loop is broken, so I think iprobe should be used, but I am not very sure.

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by Frank Wiedmann on Dec 18^th, 2012, 5:03am

If you have a recent version, you should use the diffstbprobe instead of the cmdmprobe. As we have a differential mode of oscillation here, I would not use a single-ended probe.

Of course, an oscillator does not really have a dc operating point, this is just a simulation artifact. You also probably have nonlinear effects in your oscillator that the linear stb analysis does not take into account. You might get a more accurate result with a pstb analysis as shown in http://support.cadence.com/wps/mypoc/cos?uri=deeplinkmin:DocumentViewer;src=wp;q=ApplicationNotes/Custom_IC_Design/Pstbv1.pdf.

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by rfidea on Dec 18^th, 2012, 1:29pm

Sure Frank, the probe to use today is the diffstbprobe, cmdmprobe is the old one.

Tulpin, the idea with the stb analysis in that you DO NOT need to open the loop. The analysis can measure loop gain and phase anyway. It injecting signals in both directions and by measuring the signal levels around the probe and compensating for impedances it can measure loop gain and phase

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by tulip on Dec 24^th, 2012, 7:59pm

Thank you very much.
Hello Frank, according to your suggestion, I used pstb, but the GM is only several udB, I think because at this time the oscillator is already oscillate, so the GM near zero dB is a reasonable result, but this do not give me more information on gain margin.
Thanks a lot.

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by vp1953 on Dec 25^th, 2012, 12:27pm

Hi Tulip,

For steady state oscillation, GM=0db, no? So the results look good

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by tulip on Dec 25^th, 2012, 6:34pm

Yes, Pstb give the result of GM close to 0dB.

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by raja.cedt on Dec 27^th, 2012, 11:34am

Dear vp1953,
I didn't understand why 0dB GM will be reasonable? I guess -ve gain margin will be reasonable because PSTB will give starting information, there pole are on the RH side. In steady state 0 GM make scene (but not 0dB, which is having lot of damping and hence osc may die),

Is PSTB gives steady state info or starting info??

Thanks,
Raj.

Title: Re: Should I use cmdmprobe or iprobe in GM simulation of differential VCO?
Post by rfidea on Dec 27^th, 2012, 1:59pm

pstb is the stability analysis in periodic steady state by CDS Virtuoso. stb is the small signal variant.

So, for an oscillating VCO the GM in pstb should be =1 (=0dB) and for a non-oscillating VCO the GM in stb should be >1 (>0dB) to secure oscillation start.