In the square-law equations you use (V_DS - V_DS,sat) while other books use

only V_DS. Would you comment on the difference?

Yes, in Eq. (6.42) we derived, for saturation operation,

I_D = [(KP/2)•(W/L)•(V_GS - V_THN)²]•(1 + l•(V_DS - V_DS,sat))

and for the triode region, Eq. (6.36)

I_D = (KP)•(W/L)•[(V_GS - V_THN)•V_DS - V ²_DS/2]

It's important to note that at the border between the triode and saturation

regions, where V_DS = V_DS,sat = V_GS - V_THN, these equations are equal and

given by

I_D = I_D,sat = [(KP/2)•(W/L)•(V_GS - V_THN)²]

Okay, now as you have indicated others remove the V_DS,sat and use simply

I_D = [(KP/2)•(W/L)•(V_GS - V_THN)²]•(1 + l•V_DS)

which is the implementation also used in the Level 1 SPICE model.

So here's the rub, if you don't use V_DS,sat then there is a discontinuity

at the border of the triode and saturation regions when V_DS = V_DS,sat. To

avoid this discontinuity you have to also multiply the equation for

operation in the triode region by (1 + l•V_DS) or

I_D = (KP•W/L)•[(V_GS - V_THN)• V_DS - V ²_DS/2]•(1 + l•V_DS)

This is often not mentioned or indicated by other authors, but, again it is what

is used in the level 1 MOSFET spice model.

See, also, the bottom of page 146 in the third edition.

Return