The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Measurements >> Phase Noise and Jitter Measurements >> Phase noise bump
https://designers-guide.org/forum/YaBB.pl?num=1174556880

Message started by Visjnoe on Mar 22^nd, 2007, 2:47am

Title: Phase noise bump
Post by Visjnoe on Mar 22^nd, 2007, 2:47am

Dear all,

please find enclosed a phase noise profile of a VCO: I'm seeing a 'bump' in its profile at the oscillation frequency.

What would be the physical explanation for this bump?

Thanks in advance!

Peter

Title: Re: Phase noise bump
Post by mg777 on Mar 22^nd, 2007, 9:04am

The 'bump' is called a spur. I see a 32 MHz fund, a weak second harmonic, and a stronger third. The center freq is not specified.

Is the plot measured data or sim? If its measured then the 32 MHz could be a digital clock from an adjacent block, with duty cycle distortion. It's like how the old synthesized signal generators would exhibit spurs corresponding to the rpm of the equipment's cooling fan. If it's sim then your oscillator may have a weak second resonance riding on top of the fundamental mode - except that given the small signal level for this second mode we'd expect it to be highly sinusoidal.

M.G.Rajan
www.eecalc.com

Title: Re: Phase noise bump
Post by Ken Kundert on Mar 22^nd, 2007, 9:43am

You are seeing the harmonics of the VCO's fundamental frequency. Each harmonic is also contaminated with phase noise, that is why you are seeing them in the PNoise results.

-Ken

Title: Re: Phase noise bump
Post by Visjnoe on Mar 22^nd, 2007, 11:12am

Dear,

okay, thanks for the answer!

So, if I get this right, the 'first bump' is actually the second harmonic? Correct?

The plot is indeed obtained from simulation, not measurement and plots the phase noise relative to the fundamental carrier frequency (~32MHz).

Kind Regards

Peter

Title: Re: Phase noise bump
Post by Ken Kundert on Mar 22^nd, 2007, 12:53pm

You are doing a relative frequency sweep, so 0 Hz is the fundamental frequency. Thus, 32MHz would be the second harmonic. This would more obvious if you did an absolute frequency sweep from 0 to say 100MHz. They you would see the phase noise "tents" at 0, 32MHz, 64MHz, and 96MHz, which represent harmonics 0, 1, 2, and 3.

-Ken

Title: Re: Phase noise bump
Post by jeffyan on Sep 19^th, 2007, 7:54pm

Ken Kundert wrote on Mar 22^nd, 2007, 12:53pm:

hi Ken,
if we do pss+pnoise analysis to PFD&CP, i think, we should set a absolute frequency sweep, right?
thanks.
jeff

Title: Re: Phase noise bump
Post by Ken Kundert on Sep 20^th, 2007, 12:30am

It's your choice. Generally if you want to see close in noise, particularly on a log scale, you use relative. If you want to see the noise over a wide bandwidth, absolute is preferred.

-Ken

Title: Re: Phase noise bump
Post by jeffyan on Sep 20^th, 2007, 2:17am

Ken Kundert wrote on Sep 20^th, 2007, 12:30am:

It's your choice. Generally if you want to see close in noise, particularly on a log scale, you use relative. If you want to see the noise over a wide bandwidth, absolute is preferred.

-Ken

thanks Ken.
but the nosie plot is different near the dc and fref.
to PLL, the noise of cp is lowe-pass, and i think the noise near the fref can be filtered out greatly for narrow bandwidth PLL.So i think, using sweep type=absolute makes more sense in noise analysis of PLL. is it correct?

thanks.

Title: Re: Phase noise bump
Post by Ken Kundert on Sep 20^th, 2007, 7:43am

Oh yeah. Of course. Sorry, I wasn't thinking.

-Ken