Cannot plot transient response (no graph/empty window)

Dear hikotoshi,

hikotoshi said:

Whenever I try to do a transient simulation of any circuit, I only get a black window, as seen below. DC and other simulations work fine. I've made a very simple inverter schematic as shown in the picture and I don't see any errors in the Spectre log. This problem occurs on any circuit, I've tried various. What could be the reason for this?

Did you check the CIW window for any issues? It sounds as if a netlist cannot be created -  possibly due to an invalid parameter in your transient input pulse. This will prevent spectre from simulating the circuit. In my limited experience, if a netlist cannot be generated, the response you describe is common. If there is no information in the CIW, You might try to enable the  +diagnose option to provide more information. If you have not done this before, some instructions are at Cadence On0line support at URL:

support.cadence.com/.../ArticleAttachmentPortal

It is a bit difficult for me to be certain on the issue as you have not included the netlist or input.scs file or the tools you are using...

Shawn

I have noticed the following in the CDS.log file:

net /INPUT selected but not highlighted

net /OUTPUT selected but not highlighted

No valid waveforms for plotting

No image saved: Possible reasons could be that none of the windows

or subwindows are visible or are non-empty.

Could this indicate the source of the problem? Both signals are set to plot & save.

I would advise you to contact customer support - I don't think it's a netlisting issue (since you wouldn't the dc sweep results either in that case). It does seem a bit strange, although you didn't mention anything about the versions used. It would be interesting to see if you can see the results in the results browser, and whether you can plot from there. However, I think the best thing would be to work with colleagues in support - a screen sharing session is likely to get the bottom of this quickly.

Andrew.

Dear hikotoshi,

hikotoshi said:

I've attached the input.scs. Tools are GF 22FDX-EXT.

input.scs:

Thank you for providing your input.scs file! This is quite helpful - to me anyway!

If you have not followed through with Andrew's suggestion to contact customer support, I think I may have found an issue that will preclude spectre from performing the transient simulation.

Examining your input.scs file, the input pulse statement to your inverter is defined as follows:

hikotoshi said:

V1 (INPUT 0) vsource type=pulse val0=0 val1=vdd period=1/F rise=0.1/F \
fall=0.1/F

However, it is missing the "width" attribute. Without the "width" atttribute, it is not clear how spectre can establish the pulse width. From the UNIX spectre -h vsource command, an example vpulse statement is provided below. Note the parameter "width" which is missing from the vpulse statement in your input.scs

Sample Instance Statement:
vpulse1 (1 0) vsource type=pulse val0=0 val1=5 period=100n rise=10n fall=10n width=40n

Did you specificy a width in the ADE setup GUI highlighted in Figure 1 hikotoshi?

Shawn

Figure 1

That still wouldn't cause only the dc sweep results to appear - it would most likely cause either a non-pulse waveform or an error. In practice though, if the width is omitted, it is inferred from the period and the rise/fall times (i.e. I think it's (period-rise-fall)/2). I just tried the netlist (with some made up models rather than looking for the right model files) and it runs OK, and I can plot INPUT and OUTPUT (from command line).

So that's not it...

The issue is very unlikely to be related to an error in the netlist - I'm struggling to think of what it might be though, since it doesn't sound like symptoms I've seen before...

Andrew

That still wouldn't cause only the dc sweep results to appear - it would most likely cause either a non-pulse waveform or an error. In practice though, if the width is omitted, it is inferred from the period and the rise/fall times (i.e. I think it's (period-rise-fall)/2). I just tried the netlist (with some made up models rather than looking for the right model files) and it runs OK, and I can plot INPUT and OUTPUT (from command line).

So that's not it...

The issue is very unlikely to be related to an error in the netlist - I'm struggling to think of what it might be though, since it doesn't sound like symptoms I've seen before...

Andrew