1.#QNAN message when I run Monte Carlo

[Yasutaka]

I am a student. I recently installed the evaluation version of Multisim 10.1.1 for trial.

I edited my circuit, which so far works great when I run DC operating point, transient and AC analyses, etc.

I now need to do Monte Carlo analysis with my design, but face a problem which I cannot solve.

I edited the tolerance of some components and specified for few runs (e.g. I randomly chose 30 runs) of DC operating point analysis, but I get the "1.#QNAN" message for any available node I select.

In precisely, the message says:

"1.#QNAN (1.#QNAN% lower than nominal, equal to mean)"

at the output value column.

Can anyone advise me where I should look at to solve the problem?

[ENIA]

Hi Yasutaka,

Could you please add your simulation file, so I could have a look on your settings and parametritzation. Without this is very difficult to say, where the problem is.

Thanks.

Best regards,

ENIA

NI Germany

[ENIA]

Hi Yasutaka,

here a good tutorial for you. Maybe this will help you a bit for a start. Last section is with Monte Carlo Analysis. There is solution also available.

http://cnx.org/content/m13737/latest/

Best regards,

ENIA

NI Germany

[Yasutaka]

Hi ENIA,

Thank you for introducing a tutorial. I will take a look.

Maybe I can solve the problem by myself.

I also attached the ms10 file.

Is this file enough for you to investigate?

In this message, I couldn't attach the generated netlist or the simulation setting file (*.dat)

For the Monte Carlo, I set 30 runs of DC operating point analysis with the resistor having 5% tolerance.

I hope that you can spot out my mistakes... 

Thanking you in advance.

Yasutaka 

[OStep]

Hi ENIA,

I believe the root problem is that the MOSFET model you are trying to use is not supported. The Level parameter is set to 8, telling Multisim to use the BSIM3.3 moset device. However, the version parameter in the model is set to "3.2" and many other parameters are unrecognized (as shown in the simulation tab of the spreadsheet and in the simulation error log). So, it is quite possible that the parameters belong to BSIM3.2, which we do not support.

In Multisim V11, because of several imporvements to the overall simulator, even though many parameters are still unrecognized (and thus ignored), I do get numeric values as opposed to #Qnan for the outputs. Of course, they probably have little significance.

To remedy the problem, I would recommended using a different mosfet model, perhaps even the basic Schichman-Hodges at Level 1. 

Hope that helps.

[ENIA]

Max thanks for answering. The question is comming originally from Yasutaka.

Hi Yasutaka,

summarizing, your problem is the MOSFET transistor model which you are using, which is not supported by Multisim. At leve 8 we support model version 3.3 and not 3.2. You have probably imported the model from other software, which is using different equations than the one which we support. That is why you are getting #QNAN as the result. As Max proposed you could use other model, important is, that the model is supported by Multisim. Than your simulation will be giving correct results.

I hope we could help you.

Best regards,

ENIA

NI Germany 

[Yasutaka]

Dear Enia, dear Max,

Thank you for taking a detail look at my problem.

I now understand.

I will look into alternative MOSFET models to utilize Multisim.

I truly appreciate your great technical support.

Thank you very much indeed.

Best regards,

Yasutaka