The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> phase noise for oscillator
https://designers-guide.org/forum/YaBB.pl?num=1053630141

Message started by Steven on May 22^nd, 2003, 12:02pm

Title: phase noise for oscillator
Post by Steven on May 22^nd, 2003, 12:02pm

Hi I have designed an oscillator with 2.4G oscillation frequency. I am new to the phase noise simulation in SPectre. I followed the pss and pnoise set to simulate phase noise, it gives me error when running pss, complaining something like the oscillation node is too small relative to the period of oscillator. Can i know why?? I have tried using traponly method but still has this error.

Please advise. Thanks.

Title: Re: phase noise for oscillator
Post by Jitter Man on May 27^th, 2003, 9:44am

I suspect that either you did not start the oscillator, or the oscillation is quenched before the steady-state portion of the PSS analysis begins because you did not give it a large enough kick to get it started properly. You can use either initial conditions or a transient source to start the oscillator, as described in section 4.4.2 in The Designer's Guide to SPICE and Spectre by Kundert (http://www.designers-guide.com/books.html).

[glb]The Jitter Man[/glb]

Title: Re: phase noise for oscillator
Post by junkwar2003 on Jan 16^th, 2004, 5:59pm

I ran into similar problem. Basically I placed a psin source driving a dummy ckt. and then it was fine. Also you must enable the ossillator option and select the driven and output node. When there are no periodic source in the schematic spectre gives that error for some reason [i used ipulse to kick start the ossc]. I am still novice in using spectre but someone else maybe able to better explain. But method i mentioned would get you around your problem.

Parag

Title: Re: phase noise for oscillator
Post by Jitter Man on Jan 16^th, 2004, 9:22pm

Junkwar,
You are only partially correct. You do need to let the PSS analysis know that the circuit is autonomous (an oscillator) by specifying a pair of node to the PSS analysis. The voltage difference on the nodes should have a significant component at the fundamental frequency, so they could be connected to the output or perhaps to the resonator (the nodes do not load the circuit in anyway).

The warning about the period being too small is an indication that the voltage difference on the nodes is constant valued. This could be because you connected the PSS analysis to the wrong nodes, but more often it is because the oscillator is not oscillating.

In an autonomous analysis, there should be no sources that produce a time-varying steady state signal. So you should get rid or the psin source. If the oscillator really does start, the simulator will not converge because there will be two different periods present.

[glb]Jitter Man[/glb]

Title: Re: phase noise for oscillator
Post by Alez on Sep 6^th, 2004, 4:36am

Hi Steven,

This might help:

In case that your oscillator really does oscillate, you

might forget to set some options for PSS analysis.

Goto PSS-> Options -> Time Step Parameter

Then set "time step" and "max. time step" as you set for Transient analysis (normally one-tenth of the oscillation period)

Hope this help, Good luck

Alez

Title: Re: phase noise for oscillator
Post by shanks on Dec 15^th, 2004, 9:55pm

Is there any procedure to accurately measure jittre in PLL, or can we measure phase noise for each block individually..?

Title: Re: phase noise for oscillator
Post by Jitter Man on Dec 15^th, 2004, 10:50pm

Check out http://www.designers-guide.com/Analysis/PLLnoise+jitter.pdf

[glb]The Jitter Man[/glb]