why the Cgs,Csg,Cgd,Cdg is negative in spectre.How to get ft

[wangkes9]

cgg cggbo spectre

I simulate a common-source curcuit in spectre, and I open calculator,
using op->choose the nmos->choose Cgs.... , and the result show it is a
negative value ,such as -3.7*E-15.

Why, I think it should be positive.

by the way, anyone know how the get the ft of nmos in simulator .Just using the
definition gm/(Cgs+Cgd) in EE204 in berkeley, but the result is something wrong.


Thanks a lot.

 

[drabos]

spectre cgg

It is not so surprising if you get a negative cgs cvalue.
I guess you have realised that there is also Csg value...
By the way,I usually use the cgg value and just simply do not care about cgs, csg...
If you want to calculate this frequency you should calculate the capacitance as the sum of cgg and the overlap capacitances.
If you are interested in the unity gain frequency: one of the oprating point value is this value (i guess fug, but not sure)

 

[wangkes9]

cgg spectre

Thanks a lot.

But the Cgs and Csg are both negtive. In the definition ,Cgs =&Qgate/&Vsourse.
& is derivation symbol. I wonder why it is negtive ?

 

[edge_tv]

cgd cgs ft

Because spectr is using to model this Caps voltage controllt current sources, and the Caps you see are actually Matrix elements. you can look this up at Bsims reference

 

[markty]

cgs negative

Because spectr is using to model this Caps voltage controllt current sources, and the Caps you see are actually Matrix elements. you can look this up at Bsims reference

Can you give more details? thanks.

 

[edge_tv]

ft simulation gm cgs spectre

**broken link removed**
This is the link to the bsims reference. The capacitor model starts at page 69 (4.3), it is a charge model.

 

[frogfat]

cgs csg

Thanks!!edge_tv

I was looking for an answer for that question.

 

[MohEllayali]

Hello everybody,

a question to drabos

I have encountred in forums poeple saying the same thing as you , from your experiece how much accurate is your approximation ?

thanks in advance