about subthreshold factor n

[lhlblue]

in some reference book, subthreshold factor n=1+Cd/Cox, and its value is 1.2 ~ 1.5, that is, larger than 1; but in one 0.35um lib, i see that, for nmos n=0.54 and for pmos n=1.32, can n be smaller than 1? is this right? thanks.

 

[erikl]

... in one 0.35um lib, i see that, for nmos n=0.54 and for pmos n=1.32

I think this is a different parameter n , not the subthreshold factor. Where did you find it?

can n be smaller than 1?

The subthreshold factor: no.

 

[an_82]

It should be larger than one absolutely. Can you attach your netlis. file ?

 

[erikl]

... attach your netlis. file ?

Have you ever seen a real-number (float) parameter named n in a netlist file?

 

[an_82]

Have you ever seen a real-number (float) parameter named n in a netlist file?

You can find that in the model of transistor.

 

[erikl]

You can find that in the model of transistor.

Ok, not in a netlist file, but in a transistor model. Could you show an example, please?

BTW: not to be mixed up with the parameter NFACTOR, the subthreshold swing factor, which is always ≧ 1

 

[lhlblue]

the parameter 'n' i said above is NFACTOR, NFACTOR=0.54 for nmos, is that ok?

 

[erikl]

You should have told us this at the beginning. The subthreshold swing factor NFACTOR usually is ≥ 1 , see here: View attachment Physics-Based_Derivation_of_I-V_Model_p2-31.pdf, but for SOI processes I've already seen NFACTOR≈0.8 . As it is an extracted parameter for error compensation of the depletion width capacitance, it apparently may get (calculated) values < 1 .

Not to be mixed up with the subthreshold factor n=1+Cd/Cox .