Hi,
I have built some I2S stuff, DSP circuits, DAC out etc. over the years, but I have never had any equipment to measure quality details such as phase jitter.
I would like to ask for suggestions of reasonably priced equipment but good enough quality for measuring phase jitter.
Thank you in advance.
BN
I have built some I2S stuff, DSP circuits, DAC out etc. over the years, but I have never had any equipment to measure quality details such as phase jitter.
I would like to ask for suggestions of reasonably priced equipment but good enough quality for measuring phase jitter.
Thank you in advance.
BN
Ok, which jitter have you heard about in digital audio discussions?
I probably know so little about it that I don't even got the terminology right...
I probably know so little about it that I don't even got the terminology right...
Clock jitter. Jitter is defined as 1/10UI - i.e. maximum deviation from nominal time/frequencyif it is < 1/10 of one full clock cycle (if it is larger, its called wander). So, OK - it will be a phase change but we never name it like that really.
It's quite complicated in some sense - as everything really when you start to dig deeper.
Do you want to measure actual jitter on clocks or its consequence on the analog side?
//
It's quite complicated in some sense - as everything really when you start to dig deeper.
Do you want to measure actual jitter on clocks or its consequence on the analog side?
//
I am probably more interested in the "consequences on the analog side".
Unless I really need to know the actual clock jitter to fix any problems I might find.
Unless I really need to know the actual clock jitter to fix any problems I might find.
BN, do you want to DIY some measuring tool, or just buy a commercial measuring equipment?
There is also the J-Test signal for home testing:
https://www.stereophile.com/content/case-jitters
There is also the J-Test signal for home testing:
https://www.stereophile.com/content/case-jitters
Then you shall measure distortion and noise because that is what is created by jitter. But you will need some really good equipment to do it. Try the Cosmos ADC - there is a thread about it here on the forum.
//
If there is a DIY tool available, I would be interested.
For the "J-Test" signal you would need some kind of spectrum analyser, right? I do not have that.
For the "J-Test" signal you would need some kind of spectrum analyser, right? I do not have that.
Phase jitter: Phase jitter is the term used for looking at the phase fluctuations themselves, i.e. the deviations in the position of the phase against what would be expected from a pure signal at any given time. Accordingly phase jitter is measured in radians.
https://www.google.com/search?q=pha...0i390i650l4.5838j0j1&sourceid=chrome&ie=UTF-8
Phase jitter versus period jitter animation attached (hint: open attachment in new browser tab).
https://www.google.com/search?q=pha...0i390i650l4.5838j0j1&sourceid=chrome&ie=UTF-8
Phase jitter versus period jitter animation attached (hint: open attachment in new browser tab).
Attachments
If you get a Cosmos and use the free REW measurement application - you do have a spectrum analyser - a quite good one at that 😉
//
//
Also, there are multiple jitter type definitions:
Cycle−to−cycle jitter, period jitter, and time interval error (TIE) jitter are measured in time domain...
https://www.mouser.com/pdfdocs/src-tutorials/Basics-of-Clock-Jitter.pdf
In addition: J-test is mostly useful for evaluating deterministic jitter. Random jitter can also be visualized to some extent, but its hard to quantify using J-test.
Cycle−to−cycle jitter, period jitter, and time interval error (TIE) jitter are measured in time domain...
https://www.mouser.com/pdfdocs/src-tutorials/Basics-of-Clock-Jitter.pdf
In addition: J-test is mostly useful for evaluating deterministic jitter. Random jitter can also be visualized to some extent, but its hard to quantify using J-test.
Now I have the COSMOS ADC hardware.
I will need some advice how to set up the REW RTA to see what I am looking for.
If I am measuring a single 1kHz signal, I expect that the clock jitter will manifest itself as a widening of the 1kHz spectrum peak.
What quantitative information about any clock jitter and its effects can I get from the spectrum?
Can someone recommend any specific forum thread that discusses this?
I will need some advice how to set up the REW RTA to see what I am looking for.
If I am measuring a single 1kHz signal, I expect that the clock jitter will manifest itself as a widening of the 1kHz spectrum peak.
What quantitative information about any clock jitter and its effects can I get from the spectrum?
Can someone recommend any specific forum thread that discusses this?
The problem with Cosmos ADC is that it cannot use external clock so it is not possible to sync Cosmos ADC's clock with DAC's clock. This widens the noise skirts around the fundamental and no-window FFT (rectangular FFT in REW) is not possible. But you can also use the J-test signal in REW. This may show some jitter artefacts of clocks.
I used to use an Agilent 5052B to measure phase noise & jitter, but those start at about $130k...
For a quick and dirty test, feed the DUT a low distortion 1K sine and compare before & after spectrums to see how much the skirt of the 1KHz spike widens at low amplitudes (before it disappears into the noise floor). But there's not much one can do about it short of redesigning the clock oscillator & its power supply.
The widening of the skirt isn't useful for much except comparative tests. The frequency resolution is too low to determine pSec of jitter or RMS jitter.
For a quick and dirty test, feed the DUT a low distortion 1K sine and compare before & after spectrums to see how much the skirt of the 1KHz spike widens at low amplitudes (before it disappears into the noise floor). But there's not much one can do about it short of redesigning the clock oscillator & its power supply.
The widening of the skirt isn't useful for much except comparative tests. The frequency resolution is too low to determine pSec of jitter or RMS jitter.
Last edited:
What is the recommended setting of the REW RTA in this case?
Window type, FFT length, etc.?
This thread has some examples of phase noise measurements:
https://www.diyaudio.com/community/threads/phase-noise-in-ds-dacs.387862/
As I said Cosmos ADC is not really suited for this task as you are not measuring DAC clock alone since Cosmos ADC uses its own clock. Another thing is that the noise skirt is more affected by amplitude modulated noise from DAC/ADC Vref.
https://www.diyaudio.com/community/threads/phase-noise-in-ds-dacs.387862/
As I said Cosmos ADC is not really suited for this task as you are not measuring DAC clock alone since Cosmos ADC uses its own clock. Another thing is that the noise skirt is more affected by amplitude modulated noise from DAC/ADC Vref.