Learning Forum

[Rui_Pardal]

Hi all,

I'm modeling capacitors and the dielectric resistance is frequency dependent, which I can easily represent it through Ansys, putting directly the expression where frequency is a variable.

This is good for me, by the way I need to export spice files like '.lib', when we make the equivalent circuit in Q3D, I know that there is a tool that allow us to take spice files directly, but this is not the case. I'm making the circuit directly in simplorer/ circuit environment.

I've already tried to copy and translate the netlist to LTSpice, and there I would generate the lib files, but I've not been successful until now.

Which alternative do we got?

[Dan Dv]

I can speak more to the Circuit tool here than Simplorer. AEDT Circuit cannot export any arbitrary schematic as a .lib file. Nor are the internal netlists generated for that tool guaranteed to work in any other simulator. However the results of an LNA simulation (as well as any set of s-parameters) can be exported as PSPICE .lib file using the Broadband export option in the Network Data Explorer.
Once you have a solved LNA simulation, right mouse click on the Linear Frequency definition in the Project Manager, and select "Network Data Explorer."

Then in the NDE ribbon, select "Broadband."

In the next form, check the "change output file format box" and then select the "PSPICE" radio button.

I would also recommend checking the "enforce model passivity" option to ensure that there are no stability issues with the exported .lib file. Though that is an optional step and the option is not enabled by default. Once you click on the OK button at the, the tool will convert the s-parameter to the SPICE .lib file.
Best regards -Dan

[Rui_Pardal]

At first place, thank you very much -Dan I tried that approach, but that option of Network Data Explorer appears to be "gray" and disabled to me, what is missing ?

[Rui_Pardal]

I know now that we've got to specify ports to serve as source and sink to then access Network Data Explorer and export broadband.
By the way, when we export lib file and then put it in spice , although impedance seems to match perfectly with the impedance of the circuit, we see a difference in ESR curve. What do we need to change in export options to get accurate results?

[Dan Dv]

You should check the reported fitting error in the message manager. That info is also written out in the .fwslog file
genequiv:message Final fitting error: 0.00277931
Note, that is not a percentage being reported but a normalized number. So in the case above the fitting error is 0.277931%. If the reported fitting error is low then your exported spice model is a good match to the s-parameter data. If so, there are not other export options that will help, but something else odd may be going on.
I'm not sure it will make a difference or not here, but I'd also use 50 ohm port impedances on the schematic.