How to choose the Vds when use the gm/id methodology ?

[jenpeng]

Hello~

When i simulate the gm/Id v.s. I(normalized with W/L) of a single PMOS,

I find that different Vds will induce different curves.

How to choose the Vds when i use the gm/id methodology ?

Thank you !

 

[tompham]

Hi jenpeng

I read in lecture note from professor B. Murmann (uc Stanford), you can choose vds = VDD/2

 

[jenpeng]

Hello tompham :

Thanks for your reply.

I read Professor B. Murmann's lecture before, so i typed this value in my hspice code.

But i just want to know "why Vds=VDD/2 " ?

With different Vds, the curves of gm/Id v.s. I(normalized with W/L) or gm/Id v.s. Vov are different.

How to choose an appropriate Vds ?

 

[Dominik Przyborowski]

It depends to your application. In overall intrinstic gain of mosfet increasing with Vds, but for lateral electric fields in channel higher than 0.8MV/m in nmos, the electrons velocity start to saturate, decreasing a transconductance and other "hot carrier effects" becomes to be significant. For holes the coresponding electric field has value of 1.95MV/m.