Hi all,

I have a question regarding the simulation of flicker noise in SpectreRF.

Consider an NMOS transistor biased in the saturation region. If I cycle the drain-source voltage from on to off (with say a 50% duty, at frequency greater than the flicker corner noise), the flicker noise will be correlated from one cycle to the next. In this case you can model the flicker noise as coloured stationary noise modulated with a square wave, i.e. cyclostationary noise.

However, if I instead cycle the gate-source voltage from on to off, flicker noise will not be correlated from one cycle to the next. As I understand, flicker noise is caused by carriers being trapped in the dielectic, and so cycling the gate-source voltage resets these traps. Right? In this case, modeling the noise current as a stationary source modulated with a square wave is incorrect.

My question is therefore: How does SpectreRF model flicker noise in a switched MOSFET.

Any feedback is greatly appreciated.

Dave