Designer’s Guide Community

aaron_do

Senior Fellow

Offline

Posts: 1398

AC simulation and deriv of DC
Oct 22^nd, 2010, 2:12am

Hi all,

I am doing some simple simulations on a transistor, where I the transistor is setup up with a DC voltage across the gate-source and a DC voltage across the drain source. VGS is set to 0.7 V, and I run a DC sweep with VDS going from 0 to 1 V. To find gds, I plot the derivative of the current coming out of the drain. After that, I applied a small signal to the drain on top of VDS, and to find gds, I simply plot the AC (0.1 Hz) current flowing out of the drain. However, I found that the two resutls are not quite equal when VDS is small. What is the reason for this? Shouldn't they be equal?

i.e. gds found from AC analysis is not equal to dIDS/dVDS.

thanks,
Aaron

Back to top

there is no energy in matter other than that received from the environment - Nikola Tesla

IP Logged

Geoffrey_Coram

Senior Fellow

Offline

Posts: 1999
Massachusetts, USA

Re: AC simulation and deriv of DC
Reply #1 - Oct 22^nd, 2010, 5:16am

Probably your ac frequency is small enough that the capacitance isn't affecting you (you could check 0.01 Hz and see if you get the same answer as 0.1 Hz).

I was going to suggest that gds is not equal to dIDS/dVDS because of the resistors inside the MOS model (RD, RG, RS), but that's applies to gds as reported by the model, not the way you compute it.

How small is VDS? How small are your sweep steps? Have you considered how floating-point precision affects your answer?

Back to top

If at first you do succeed, STOP, raise your standards, and stop wasting your time.

IP Logged

aaron_do

Senior Fellow

Offline

Posts: 1398

Re: AC simulation and deriv of DC
Reply #2 - Oct 22^nd, 2010, 7:53am

thanks for the reply. I just used the automatic settings for the sweep steps. I'll try finer steps when I get the chance. VDS is swept from 0 to 1 V. I don't really follow you about the floating-point precision. Could you elaborate? I figure it has something to do with the additional capacitances in the AC model.

thanks,
Aaron

Back to top

there is no energy in matter other than that received from the environment - Nikola Tesla

IP Logged

Geoffrey_Coram Senior Fellow Offline Posts: 1999 Massachusetts, USA	Re: AC simulation and deriv of DC Reply #3 - Oct 25^th, 2010, 11:09am The usual formula for derivative is f(x2)-f(x1)/(x2 - x1) Suppose you took a step of 1pV, so that at the end of the sweep, x2 = 1V and x1 = 1V - 1pV. Now, what do you get when you compute x2-x1? It might not be exactly 1.000000000 pV.
Back to top	If at first you do succeed, STOP, raise your standards, and stop wasting your time. IP Logged

Geoffrey_Coram Senior Fellow Offline Posts: 1999 Massachusetts, USA	Re: AC simulation and deriv of DC Reply #4 - Oct 25^th, 2010, 11:09am Another thing, depending on your simulator: do you have a BYPASS option? Spice3f5 has it, and what happens is the model code says, hey, the voltage you're asking for is "close enough" to the voltage I just evaluated at, so I'll just give you that answer again.
Back to top	If at first you do succeed, STOP, raise your standards, and stop wasting your time. IP Logged

aaron_do Senior Fellow Offline Posts: 1398	Re: AC simulation and deriv of DC Reply #5 - Oct 25^th, 2010, 6:57pm Hi, thanks for all the help. I see what you mean by the precision of the simulation. Finally had the chance to redo the simulations. As you say, it was the sweep step settings. When I increased the number of steps, the results were much closer. thanks, Aaron
Back to top	there is no energy in matter other than that received from the environment - Nikola Tesla IP Logged