No, the lambda value is the one fixed dimension. Everything else is scalable.
v
Varun Majji
01/12/2021, 2:49 PM
so sir if I want to design an inverter using this skywater 130nm tech file, I should make a polysilicon width of 130nm, it's not like in magic sky130nm is scaled to some other measurement like osu18 right!
t
tnt
01/12/2021, 3:50 PM
Actually you need to make it 150nm
tnt
01/12/2021, 3:51 PM
(poly.1a)
rule states minimum width of poly is 150 nm.
j
Jecel Assumpção Jr
01/12/2021, 5:54 PM
You draw it 150nm wide, but when the source and drain diffusion step drive the impurities deeper into the silicon they will also spread sideways a bit under the polysilicon. That means that the actual channel length will be 130nm which is why the technology is called sky130
v
Varun Majji
01/12/2021, 9:25 PM
oh got it thankyou sir
t
Tim Edwards
01/12/2021, 9:32 PM
@Varun Majji: OSU018 (which is a standard cell library, not a technology) is based on one of the SCMOS tech files, which means it has DRC based on lambda units. It is not "scaled to some other measurement", though. There is a mapping between lambda and physical units, and if there is only one output style, then the mapping is fixed. If you type "box" in magic, you get the output in microns (as well as lambda and internal units) based on the output style.
Linen is a search-engine friendly community platform. We offer integrations with existing Slack/Discord communities and make those conversations Google-searchable.