The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> HSPICE RF versus SpecteRF benchmarking
https://designers-guide.org/forum/YaBB.pl?num=1183733585

Message started by Visjnoe on Jul 6^th, 2007, 7:53am

Title: HSPICE RF versus SpecteRF benchmarking
Post by Visjnoe on Jul 6^th, 2007, 7:53am

Dear all,

does anyone have some experience on HSPICE RF versus SpectreRF benchmarking?

I'm currently benchmarking both simulators on a 0.18um CMOS crystal oscillator and both phase noise profiles are in line (well, let's say within 5dB) for most part of the frequency range, except for low frequency offsets.

It appears HSPICE RF does not show a 1/f phase noise region (-30dB/decade): the phase noise lowers @-20dB/decade before reaching a noise floor. SpectreRF does show 1/f noise at low frequency offsets...

I examined the device models for Spectre and HSPICE and both list the same noise parameters (NOIA,NOIB etc.)

Does anyone have some experience/insights on this?

Regards

Peter

Title: Re: HSPICE RF versus SpecteRF benchmarking
Post by Visjnoe on Jul 10^th, 2007, 5:13am

Dear all,

I have benchmarked SpectreRF versus HSPICE RF on a 0.18um CMOS crystal oscillator. The absent -30dB/decade region in the HSPICE RF simulation (previous posting) has been corrected and was related to a testbench setup issue.

Running both simulators in accurate enough settings, I end up with a phase noise difference of about 2dB-3dB over the entire frequency offset range.
This kind of deviation between both simulators was also reported by Li Yi in @SNUG Singapore (2005).

The RMS jitter differs by about 20%-25%. Unfortunately, I cannot provide silicon correlation results.

Regards

Peter