corner simulation's query_about the process invariance

hi all,
I am doing the design of constant current LNA independent of the process. Its structure is the common source topology and the drain voltage is fixed to 1 V while gate voltage varies with different process. But the overdrive voltage Vod=Vgs-Vth is nearly fixed for all process. In the simulation,I find the MOS drain current still changes with different corner, ex, ff: 780uA, tt: 662uA and ss: 600uA. According to MOS current equotion of saturation region MOS : I= 1/2*un*W/L*(Vod)^2*(1+lemda*Vds), the current should be nearly constant with constant Vod and Vds. So I guess some main facters related to process result in the drain current increase, could you tell me which parameter leads to the current variation for different corner?
best regards,

shhaha said:

... which parameter leads to the current variation for different corner?

Click to expand...

Different XL, XW variation (due to under-diffusion, lithography conditions) depending on process run, I'd guess. See e.g. these COMMENTS:

Hi Erikl,
Thanks for your reply and the reference. I check the comment and thinks W and L variation with different corner besides short channel effects may be the factor but they are not so crucial for the discrepancy between different corner simulations with same overdrive voltage since their variations are slight with only considering the corner. In my opinions since the ff means the thin film MOS and ss means the thick film mos, the oxide layer will vary more severely than W and L's. The oxide variation will result in the Cox variation due to Cox=E/tox, where tox is the oxide thickness. I guess oxide variation will be the main reason for this case.
Merry Christmas!

shhaha said:

The oxide variation will result in the Cox variation due to Cox=E/tox, where tox is the oxide thickness. I guess oxide variation will be the main reason for this case.

Click to expand...

Of course you are right, Shhaha, but isn't this variation already covered in the model by the gate voltage variation?

shhaha said:

... while gate voltage varies with different process.

Click to expand...

I thought you asked for additional reasons responsible for the process-dependent variation of drain current, which (possibly) aren't covered by parameter changes in the corner sections. Just check those corner-dependent parameters!

Merry Christmas, 2U2!

I get and agree with you! thanks a lot!