Designer’s Guide Community

JC

Junior Member

Offline

Posts: 23

How to measure node charge using Spectre
May 31^st, 2007, 3:52am

I just wonder if it is possible to measure it directly or not? I use the CAPTAB option that allows to print the total capacitance on a node during Transient simualtion.

However, I found the value of capacitance is derived by C=dQ/dV in Spectre, so I think Spectre should know the node charge during simulation. And in my case the capacitance is due to parasitic effect and will change during operating. So I cann't use Q=C*V to calculate the node charge.

Is there a way to measure charge node directly? or How can I calculate node charge by knowing node capactiance?

JC

Back to top

IP Logged

Geoffrey_Coram Senior Fellow Offline Posts: 1999 Massachusetts, USA	Re: How to measure node charge using Spectre Reply #1 - Jun 1^st, 2007, 9:17am Please read this thread: http://www.designers-guide.org/Forum/YaBB.pl?num=1172655940/0#1 Next time, use the search function first!
Back to top	If at first you do succeed, STOP, raise your standards, and stop wasting your time. IP Logged

JC Junior Member Offline Posts: 23	Re: How to measure node charge using Spectre Reply #2 - Jun 1^st, 2007, 9:34am I had read that thread yet, but it doesn't answer my question. So I post again. And I know how to print the node capacitance using CAPTAB option, However my question is "How to measure the node charge while the capacitance is caused by parasitic effect and is nonliear?" JC
Back to top	IP Logged

Geoffrey_Coram

Senior Fellow

Offline

Posts: 1999
Massachusetts, USA

Re: How to measure node charge using Spectre
Reply #3 - Jun 4^th, 2007, 4:21am

I'm not aware of any simulator that gives you direct access to the node charge.

You can't get Q from C because C is the partial derivative of Q and there's no (easy) way to integrate back. If you had 2-terminal nonlinear capacitors, you could integrate from 0 to v, but MOSFETs are 4-terminal nonlinear capacitors, and in many MOS models, the capacitance is non-reciprocal, so the integral would be path-dependent: the value of the integral depends on how you get from (0,0,0) to (vgs,vds,vbs)!

Back to top

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

IP Logged