https://open-source-silicon.dev logo
Join Slack
Powered by
At the risk that I'm not the first asking here: do...
# xschem
c
At the risk that I'm not the first asking here: do graphs in xschem with a log x scale work? I seem to be able to plot g_m/I_d vs. I_d on a linear scale, but not on a log scale.
I can work around the buggy log x scale, but is that really necessary, or am I just missing something basic?
s
Usually when you switch to x log scale do a 'full zoom' (f in the middle of the graph area) to rescale.
c
Wors sometimes … but as I'm trying to plot noise spectra and dc sweeps, xschem graph plotting seems to get confused … or I am confused about the finer points of scaling. log log output noise with 
set sqrnoise
results in confusion. With gaw, I get meaningful graphs for both dc sweep and noise, although not simultaneously.
s
@Christoph Maier can you please share the raw file and the variables you need to plot? So I can do a test. There are gazillion plot types ngspice sets in the 
Plotname:
line and may be this is yet a new plot type I need to add.
c
s
Thank you will look asap
I have tested your example with set sqrnoise enabled. I have added some options to increase the precision since you are dividing by derivatives of almost constant signals (vg) I have added in the schematic the same graphs plotted in ngspice plus the spectrum graph. Added an image in the schematic of the ngspice plots, they seem to match. This example was useful to do a small change in xschem since due to random rounding errors in the ngspice generated data it did not correctly recognize wrap arounds (see following image) DC plot is still visible though.
pfet_g5v0d10v5_noise.sch
2.png
5 Views
Open in Slack
PreviousNext
Powered by