Hello everyone I m trying to make a plot using the internal open-source-silicon.dev #xschem

Hello everyone. I'm trying to make a plot using ...

Emmanuel Innocent

04/23/2024, 3:11 PM

Hello everyone. I'm trying to make a plot using the internal xschem plotter. I have attached the intended relationship to be plotted and the xschem postfix representation of the ( relationship) expression. I just want to confirm if the postfix representation of the expression is correct because I'm not getting the expected plot. The xschem plot and the expected(reference) plot are also attached here. @Stefan Schippers @Tim Edwards

Luke Harold Miles

04/23/2024, 8:55 PM

looks equivalent to deriv(id) / id to me. Maybe • use a smaller timestep and • scroll your graph to the right one pixel (or set the graph's xmin=2*stepsize) to exclude the values near zero? (deriv is weird on first value)

Luke Harold Miles

04/23/2024, 9:00 PM

• Could also use an ammeter.sym to make sure you're grabbing the right value, since the symbol referencing is weird. • Putting this at top of your ngspice control block should improve accuracy for small currents:

Copy code

.option GMIN=1e-18
.option method=gear

• maybe also add epsilon in your plot to avoid dividing by zero (or start plot with larger current). If your ammeter is named `vm`:

Copy code

i(vm) 1e-15 + deriv() i(vm) 1e-15 + /

Stefan Schippers

04/23/2024, 9:21 PM

The expression for

λ

divides

gds

Idsat

not by

Id

Stefan Schippers

04/23/2024, 9:35 PM

If output resistance is needed I would simply calculate 1/(d(Id)/d(Vds)):

Emmanuel Innocent

04/24/2024, 4:05 PM

Thank you very much @Luke Harold Miles. I'd try that out.

Emmanuel Innocent

04/24/2024, 6:35 PM

Yes @Stefan Schippers thank you. I wanted to estimate the value for \lambda in the gf180 SPICE file. Since I need Idsat, please how do I get the value for Idsat?

5 Views

Open in Slack

Previous Next