Precise Vth extraction of MOS

Hi,

I've found the section MODEL OPTIONS vthmod=vthcc. Using this, one obtains the exact same values provided by the foundry in the related documentation. Anybody has experienced a large mismatch between the two vthmod options? Does this really have a practical consequence or it's just a matter of conventions?

Nicola

Hi Nicola,

Mostly this is a matter of computing the operating point Vth parameter using the constant-current method to get a result that matches the values provided by the foundry. However, it is quite expensive to compute so our recommendation is not to use this in transient simulations - it's mainly for operating point analyses and probably not wise to use when running transient.

Regards,

Andrew.

Hi Andrew,

thank you for the reply. I understand what you mean about the computation, but in my OP I see a difference in Vth higher than 100mV (for all test conditions that I read in the foundry documentation): I wonder if this will really change the silicon results compared to simulations or not.

Moreover, how is it possible to use it for transient simulations as you say? I tried to set vthcc, but this is only used to extract the vth, not in all other parameters of the model. I mean, if I have the same transistor and I choose either standard or vthcc, I will see a different OP( "vth"), but drain current remains the same.

KInd regards,

Nicola

Hi Nicola,

As far as I know, this does not alter the behaviour of the transistor model itself - it's just the computation of the vth operating parameter (this seems to concur with some experiments I just performed - the transient response remains the same, all the currents remain the same, the switching threshold of a transistor remains the same). It's just that the definition of what is the threshold voltage changes.

It impacts transient simulation performance because it still has to do all the extra iterative computations to determine the constant-current vth, just in case you chose to save the vth during the transient (it's possible to save operating point parameters versus time in a transient analysis, e.g. using Outputs->To Be Saved->Select Op Points).

Regards,

Andrew.

Dear Andriew,

Your suggestions indeed match what I also saw in my simulations. 

Thank you for the help!

Kind regards,

Nicola

Dear Andriew,

Your suggestions indeed match what I also saw in my simulations. 

Thank you for the help!

Kind regards,

Nicola