EMRFD Message Archive 12744

Message Date From Subject
12744 2016-04-27 14:48:53 jmlynesjr Phase Noise?
Hello all:

Hopefully there aren't any stupid questions on this forum....

I am currently reading EMRFD and following the daily posts. Recently there has been a lot of discussion on Phase Noise.

Can someone take a shot at a layman's definition of phase noise?



12745 2016-04-27 15:30:17 Bill Carver Re: Phase Noise?
12747 2016-04-27 17:42:12 K5ESS Re: Phase Noise?

Well I’ll give it a shot.  Rebuttals/corrections/comments welcome. 


If an oscillator were perfect it would put out a signal of a single frequency with no variation in frequency (and we’ll assume there is no amplitude variation either).  Not even the smallest.  Next let’s say we FM modulate the oscillator with a white noise signal and then reduce the deviation of the FM to much less than one Hz.   The oscillator now outputs a signal with minute random variations in phase.  This phase modulation generates sidebands.  The spectrum of these sidebands is bell shaped with the maximum centered on the oscillator carrier frequency and tapers off as you go away from the carrier.  Phase noise is usually specified by defining the dB ratio of the sideband power in a one Hz slice of the sideband spectrum to the carrier power e.g.  -100 dBc/Hz  at 1 kHz (away from the carrier).   


That’s my cut at it.  I’ll leave it to others to explain what the primary causes of phase noise are.  Probably several.


Mike K5ESS


12749 2016-04-28 17:15:27 jmlynesjr Re: Phase Noise?
Thanks Bill & Mike:

Would oscillator drift be a form of long period phase noise?

12750 2016-04-28 17:50:42 Chris Trask Re: Phase Noise?
Below a certain frequency, such disturbances are known as "short-term drift", but the boundary between that and phase noise isn't exactly a firm number, or so I have understood.

>Thanks Bill & Mike:
>Would oscillator drift be a form of long period phase noise?

Chris Trask
Senior Member IEEE
12751 2016-04-28 19:43:16 kb1gmx Re: Phase Noise?
Would oscillator drift be a form of long period phase noise?

Short answer is no.  Drift is not random nor amplitude variable.

Phase noise, is noise, its both random in frequency and amplitude.
Both of those are the side product of amplifier noise (device) and
the feedback path. 

The example I use is regenerative receiver.  You start with very little
regeneration which is positive feedback and the overall gain is low
as is the noise.  As the feedback is increased to but below oscillation
the noise increases greatly.  Increase the feedback even the smallest 
amount further and the criteria for oscillation is met and the circuit start 
oscillating, also if you are listening to the audio amplifier the noise drops.
At the oscillating point all the noise heard are the result of the active 
device and and losses in the feedback circuit being amplified.
This is all best observed with no antenna or signals to be heard.

That is a simplification but a usable working model.


12752 2016-04-29 06:02:27 Bill Carver Re: Phase Noise?
12758 2016-04-29 15:05:36 jmlynesjr Re: Phase Noise?
Allison, thank you for your response.

While I have an EE from before the days of the HP-35, my career was in realtime software. As such, I didn't do a lot of analog work which I'm now learning about.

I read on a wiki that Phase Noise also equates to clock jitter. I've seen jitter on data comm clock lines, so I suspect that I've got a time domain to frequency domain translation problem in my brain!

Thanks again,
12760 2016-04-29 20:14:45 kb1gmx Re: Phase Noise?
Yep clock jitter is a form of noise though in the frequency domain (time/phase). 

It will contribute noise and if there is fulters in the system then it may also 
acquire amplitude variation as well.

The key is the amount and the desired performace level of the system its in.   

12761 2016-04-29 20:33:34 Dana Myers Re: Phase Noise?
12762 2016-04-29 20:33:43 Graham / KE9H Re: Phase Noise?
In a sampled receiver (SDR), jitter on the sampling clock directly translates into the phase noise performance of the resulting receiver.

If you have enough information, you can translate phase noise to jitter, or jitter to phase noise.

Or, if an approximation is enough, you can make a few assumptions and do the conversion with a lot less information.
Check out this App Note from Analog Devices for a discussion of how to do the conversion or simplified estimations.

--- Graham / KE9H

12763 2016-04-29 21:27:28 Bill Carver Re: Phase Noise?
James, if you look at the data sheet for the Si570 as one example, you
will see both jitter and phase noise numbers. Jitter is what the digital
guys want to see to design their logic and the phase noise helps us
understand how it would perform in an analog system. Just two ways of
looking at the same phenomenon!

The transformation between them does require a clear understanding of
the device being characterized. I think one can to an approximate
translation between them but I believe that's a hip shot without knowing
the internals of the device.