The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Analog Verification >> Analog Performance Verification >> Long term multiple cycle jitter simulation
https://designers-guide.org/forum/YaBB.pl?num=1319048545

Message started by pupule on Oct 19^th, 2011, 11:22am

Title: Long term multiple cycle jitter simulation
Post by pupule on Oct 19^th, 2011, 11:22am

Hello,

I would like to know how to simulate long term jitter for a relaxation oscillator.

I have access to spectre/spectreRF simulation suites.

The oscillator is used in a subsystem that is followed by a counter of a high number of counts (>5000) counts and I would like to know how far off the final count value from absolute time ideal time. It is not clear to me how to setup this simulation.

Some values of interest:

Fosc = 50 kHz
# of counts = 5000

I would like to run a simulation that predicts after n# of counts how far off in time does the nth count occur.

i.e. if ideal frequency = 50 kHz and I count 5000 cycles Texpected = 5000*1/(50kHz) = 100mS. I would like to know due to jitter how far off would the final count be from 100 mS.

If I run a PSS/pnoise jitter simulation and consider noise from 10 Hz to 500 kHz. As I increase the number of cycles in the direct plot form the jitter goes up. In the lab as I increase the count number the jitter goes down. The latter seems more intuitive to me as for each cycle the jitter may be positive or negative so the more cycles I consider (higher K) the more averaging occurs and the jitter should go down.

Please let me know if my assumption of jitter should reduce with higher K cycles and any insight on how to setup a simulation to mimic my lab results.