Designer’s Guide Community

leander

New Member

Offline

Posts: 3
Beijing

#nodes and #equations in Spectre
Apr 13^th, 2011, 6:56pm

I am doing a postlayout simulation for a PA. Their are a lot of parasitic Rs and Cs. The simulator reports insufficient memory when doing PSS analysis. I then notice from the inventory that the equations is much more than the nodes (>30X). Why this is the case? I thought the number of equations should equal the number of nodes plus independent sources and inductors.

Circuit inventory:
nodes 9929
equations 347049
iprobe 166629
ahdl simulator 1
bsim4 1400
bsource_1 313
capacitor 75946
diode 960
inductor 334
isource 2
port 2
resistor 83473
vsource 2

Back to top

IP Logged

boe Community Fellow Offline Posts: 615	Re: #nodes and #equations in Spectre Reply #1 - Apr 14^th, 2011, 3:58am leander, You have ~166 000 iprobes. I expect that the simulator solves those too. Also, you may have equations for derivatives of signals. B O E
Back to top	IP Logged

Geoffrey_Coram

Senior Fellow

Offline

Posts: 1999
Massachusetts, USA

Re: #nodes and #equations in Spectre
Reply #2 - Apr 14^th, 2011, 8:22am

boe is correct; if iprobes are implemented as 0V voltage sources, then that would add a row for the branch current and an extra node, which gets you quite close to the number of equations.

BSIM4 can also add several internal nodes per device (up to 7: internal drain, internal source, two internal gate nodes, and four internal body nodes).

Back to top

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

IP Logged

boe Community Fellow Offline Posts: 615	Re: #nodes and #equations in Spectre Reply #3 - Apr 18^th, 2011, 10:43am Geoffrey, additional question: I'd expect the #nodes in the inventory to include the internal nodes of any (BSIM or other) models. Is that correct? B O E
Back to top	IP Logged

Geoffrey_Coram Senior Fellow Offline Posts: 1999 Massachusetts, USA	Re: #nodes and #equations in Spectre Reply #4 - Apr 25^th, 2011, 6:01am BOE - I think #nodes generally does not include internal nodes of models.
Back to top	If at first you do succeed, STOP, raise your standards, and stop wasting your time. IP Logged

John O Donovan Junior Member Offline Posts: 29 San Jose, CA	Re: #nodes and #equations in Spectre Reply #5 - Apr 25^th, 2011, 10:46am Geoffrey is correct. Nodes are circuit nodes, not including any internal nodes of devices. Regards, John
Back to top	IP Logged

boe Community Fellow Offline Posts: 615	Re: #nodes and #equations in Spectre Reply #6 - May 6^th, 2011, 11:32am Geoffrey and John, Thanks! B O E
Back to top	IP Logged