Designer’s Guide Community

‹ Previous topic | Next topic ›

Pages: 1

Zero Diagonal Found in Jacobian! (Read 5226 times)

schehrazi

Community Member

Offline

Posts: 45
University of CA, LA

Zero Diagonal Found in Jacobian!
Nov 26^th, 2006, 10:06am

Hi,

When does Spectre prompt the following convergence issue?

Zero diagonal found in Jacobian at 'nodeX' and 'nodeX'

Does a certain circuit condition lead to this error? I should say that I am trying to simulate a mixer with VerilogA square law transistors using PSS. I am My VerilogA model covers both saturation and triode operation regions.

Thank you so much.

IP Logged

Geoffrey_Coram

Senior Fellow

Offline

Posts: 1999
Massachusetts, USA

Re: Zero Diagonal Found in Jacobian!
Reply #1 - Nov 27^th, 2006, 5:16am

You've probably got a poorly-defined model. What happens in subthreshold? If you don't have a GMIN or subthreshold current equation, then you'd likely get a zero diagonal message. The Jacobian (matrix) is loaded with partial derivatives of all the currents with respect to the node voltages. If there's a zero diagonal at nodeX, then it means that, when linearized about the operating point, changing the voltage on nodeX will not affect the current into nodeX.

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

IP Logged

schehrazi Community Member Offline Posts: 45 University of CA, LA	Re: Zero Diagonal Found in Jacobian! Reply #2 - Dec 12^th, 2006, 11:12am Thanks, I fixed it by changing the initial condition parameters of PSS.
Back to top	IP Logged

Pages: 1

‹ Previous topic | Next topic ›