The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> Does Spectre RF (PSS+PNOISE) work for low beat frequnecy?
https://designers-guide.org/forum/YaBB.pl?num=1416344472

Message started by Homer on Nov 18^th, 2014, 1:01pm

Title: Does Spectre RF (PSS+PNOISE) work for low beat frequnecy?
Post by Homer on Nov 18^th, 2014, 1:01pm

I'm designing a switch controlled circuit. It contains one opamp, several switches and some resistors. All the switches are controlled by one clock. The clock frequency is 40KHz.

Now I'm trying to run pss+pnoise simulation to check the total noise from 30Hz to 4MHz. Because of the clock, the beat frequency is set to 40KHz.

I studied some posts in this forum and some papers from google. It seems the noise integration range should be from 0~(beat frequency)/2. For all the samples I found, the beat frequencies are very high. What can I do for such low beat frequency?

Title: Re: Does Spectre RF (PSS+PNOISE) work for low beat frequnecy?
Post by Ken Kundert on Nov 18^th, 2014, 6:13pm

That advice stems from the type of noise they are observing, which is sampled noise. When you sample a signal, the spectrum repeats as the sample frequency, and since the signals are real, they are symmetric about DC, so you only need to measure from 0 to half the sample frequency.

It sounds as if you are not sampling your signals, so you would not follow this advice.

I would expect SpectreRF to work fine for you as long as you choose a large enough value for maxsidebands and maxacfreq.

-Ken

Title: Re: Does Spectre RF (PSS+PNOISE) work for low beat frequnecy?
Post by Homer on Dec 15^th, 2014, 6:58am

Hi Ken,

Thanks for the reply. I tried to set he max sidebands to high (100) and it took quite long time for the simulation. And then I set it to 7. The simulation was much faster and the result was similar, only 1% difference. I think 7 should be enough for me.

Now I have another question. I think the integration from 0 to beatfrequecy/2 is for the total noise across the whole noise frequency range. Is there any methods to divide it into different frequency range? For example, what's the noise from 0 to 30Hz? And what is noise higher than 1MHz? The project is an imaging related design. We need to know what's the noise for different frequency range. We can do it with traditional noise simulation. But I can't find any solution with pss+pnoise simulation.