How can I use gm/id methodology for sky130 pdk, es...
# analog-design
j
How can I use gm/id methodology for sky130 pdk, especially how to get the necessary plots like ft v/s gm and such?
See my presentation slides on the above page.
j
Thanks.
SIr, I did the technology sweep for nfet. How did you plot the graphs using the data from the sweep?
b
Jupiter Notebooks. See the repo that is given in the slides.
j
Do I need to do something before running the codes? I have copied the codes and tried to plot the graphs but it is showing errors.
b
Nothing else needed. You need to debug on your own.
j
Okay
🙌 1
r
Harald Pretl's course on analog circuit design shows how to do this with the IHP 130nm PDK, you can probably also get some insights and parts of the scripts from there: https://iic-jku.github.io/analog-circuit-design
j
@Roel Jordans Yes I have been using both the courses as a reference. I have been facing some issues with the code for converting the .txt to .mat format. I have been trying to debug it but facing some issues still.
j
https://github.com/chennakeshavadasa/gmid_SKY130 you can check this out, I have uploaded the extracted data along with python scripts to plot them, download it and use the python scripts to generate necessary plots
j
@Boris Murmann Can the tech sweep value obtained be different? Because I got different values for the sweep like I am missing some values of vgs vbs vds. Is this a xschem issue or something else? Any pointers would be helpful
b
This could be anything, different model version, ngspice version, etc. You need to debug on your own. I am also not sure what different means. What kinds of differences are you seeing, quantitatively?
j
Different values for the parameters like vgs vbs vds. The values shown in the txt file does not necessarily match the values for which were swept in the control statement. For example the last value of vgs and vds should be 1.8 (shown in the .txt file in repo you provided), I've got 1.788 and 1.794. These results mess with my plots (notebook for plot from txt is showing errors).