During characterization of a 90nm lib gm = f(VGS) decrease

[jamzallag]

gm = f(VGS) decrease

During the characterization of a 90nm library, I'm plotting gm = f(vgs) and gm reaches to a max and then decrease. Why gm reaches a max and then decrease?

Thanks,
Joel.

 

[dick_freebird]

gm = f(VGS) decrease

That's natural, your peak is usually near VT because
you see the strongest rate-of-change where you are
still exponential (subthreshold, highest range) and lose
slope (gm) once you transition to square-law.