Hi guys, does anyone have a rational reason as to ...
# sky130
t
Hi guys, does anyone have a rational reason as to why the default area parameter of diode models in ngspice are set to 1e12 to represent a 1 micron^2 diode? It seems a bit odd to measure diode area in picometers squared.
b
1e-6*1e-6 = 1e-12. Diode area has been specified like this in Spice since its inception.
t
Thanks for the reply! So, a 100 micron^2 device would be represented by 1e14 like in this message? https://open-source-silicon.slack.com/archives/C016UL7AQ73/p1658700853488009 Shouldn't it be 1e-10 - to stay in SI units?
b
Ok, I now understand what you mean. It's important to distinguish simulator and pdk. Based on what @Stefan Schippers wrote, there is an odd scale factor baked into the Is of the diode model in the SkyWater pdk.
So yes, it seems that a 100um^2 device in the SkyWater pdk needs an area parameter of 1e14. This doesn't have much to do with ngspice.
t
Ah, sorry, I should've been clearer. So, its just a quirk of the PDK.