Noise Figure of foundry MOSFET model.

Hi, in the current RF CMOS process, foundries would provide a complete RF MOSFET model with all necessary "external" components such as substrate resistor, gate resister.... By including this, they will model the noise from these portions. They will also provide all characterization reports for their models to show how accurate they are. However, in these reports, I've never seen the simulation vs. measurement report about noise. I have never seen any thing about how accurate their model can predict the NFmin, Γopt...

One process I have used reported their measurement NFmin, Γopt without de-embeded! nothing about the simulation accuracy about that.

Another process do not do this because they even don't have this measurement equiments!

I am quite curious, how people design RFIC using these processes if they need to simulate the NF? Are these RF MOSFET models able to predict the noise performance? How accurate are they?

Thanks.

The models of deep submicron CMOS are acceptably good (really).
However, it is the real noise performance that counts (no help with good device models if you gat a lot of noise pick up elsewhere :) )

A.T.B