How to verify a wordclock signal with a scope .

[Tubetec]

So I hooked up the scope to the clock signal on the Alesis BRC remote ,
first I was able to look at the BNC output ,then the Adat sync clocksource ,
results with the Tek 2215 were good , a nice clear trace ,
BNC out shows some overshoot and resonance on the leading edge positive peak ,
Adat clock looked a little better , but still had the little squiggle on the waveform ,

Of course I forgot terminate the outputs , that might explain the slightly poor looking waveforms .
I'll have another go tomorrow in the light of day .

 

[PermO]

Here's what mine looks like (RME FF800) on a cheap China portable scope;

 

[Tubetec]

Similar shape to what I see , but more wiggles on the horizontal axis
I best use a T junction and 75 ohm terminator otherwise its unlikely to give the view under realistic working conditions .

Im wondering would there be any point in trying REW at 192 KHZ sample rate to look at the 48khz clock ?
Im guessing the components of a 48khz square wave are way beyond the nyquist at 192khz , so probably mangle the waveform beyond recognition ,
Maybe the frequency counter on the DDS Gen could give me a better measure of the timing .

How would I go about measureing the variance of a clock signal , say from power up until normal opperating conditions are reached ?

 

[PermO]

I tried with a T juction and 75Ohm terminator, and BNC direct... gave the same picture.

It's time for the great "Show us your clockpulse" topic

 

[JohnRoberts]

the best verification is does the circuit work.... budget scopes and probes can give dodgy trace images.

JR

 

[Tubetec]

Im getting a very nice image off the phosphor on the TEK ,

Everything is working as it should with this piece of gear and what it connects to ,
Ill try and hijack a smart phone and take a few hi res pics later .

 

[sodderboy]

You are also affecting sync circuit impedance by paralleling a second load. Perhaps measure from a distribution amp?
I never got that deep with the clocks- if it locked, we rocked.
Mike

 

[Newmarket]

You are also affecting sync circuit impedance by paralleling a second load. Perhaps measure from a distribution amp?
I never got that deep with the clocks- if it locked, we rocked.
Mike

Yes but the 10M resistance you get with a typical x10 Probe setup is fairly minimal in terms of affecting the result ( + some pF to Reference/Ground).
Could use x100 probes (typically used for HV work up to some kV) if willing to take the noise penalty from the extra attenuation.

 

[Newmarket]

Im wondering would there be any point in trying REW at 192 KHZ sample rate to look at the 48khz clock ?
Im guessing the components of a 48khz square wave are way beyond the nyquist at 192khz , so probably mangle the waveform beyond recognition ,

Definitely - you'd get the harmonic at 144kHz and everything else out of band.

Maybe the frequency counter on the DDS Gen could give me a better measure of the timing .

How would I go about measureing the variance of a clock signal , say from power up until normal operating conditions are reached ?

One thing might be to try averaging on the scope. Depending on timebase this should indicate timing tightness over a period of time - although I guess it might be restricted to a relatively short time ?
Then FFT the resulting waveform data ?

 

[arthurdb]

What are you looking for? Overshoot or ringing, under a certain threshold at least, won't affect your clock or sync stability which is really all that matters. You don't need a perfectly square wave to have a perfectly stable clock. Clock inputs would have some sort of buffer like a schmitt trigger to clean up any noise, ringing, or overshoot and, as already mentioned, your probing might deform your square wave anyways.

What really matters is anything related to timing. Is the frequency correct? Is it stable? How much jitter do you have? The rest... no so much and you I reckon it is safe to assume it is ok if the timing looks good and your gear works as it should.

 

[Newmarket]

Jitter may be in the order of ps. To measure to that level you need either some serious (£££) general test kit or a dedicated audio test setup - AP2; dScope etc.

 

[JohnRoberts]

An old friend, Ethan Winer, just wrote an article about clock jitter.
https://audioxpress.com/article/aud...eJODoSDyeyfGZJ5Hyy8YVaYfyOCZIZVBOzpQyq9trmfc8

JR

 

[terry setter]

JR: I tried to download Ethan Winer's files (mentioned in the article you linked). They were not available. Too bad, I'd love to hear them. Perhaps you could let Ethan know.

 

[Newmarket]

Definitely - you'd get the harmonic at 144kHz and everything else out of band.

No Newmarket - you're wrong
Sample rate is 192kHz so by Nyquist you can't properly measure any frequency that is not <96kHz.

 

[JohnRoberts]

JR: I tried to download Ethan Winer's files (mentioned in the article you linked). They were not available. Too bad, I'd love to hear them. Perhaps you could let Ethan know.

I contacted Ethan and he suggested that it may be an SSL issue (address should be https... with the extra 's'). It still didn't play from my apple Safari browser.

I went to my downloads folder and advised it to play with the quicktime player and I hear the music file...

Try looking at your downloads file and maybe use a different player.

Note: I am not the resident IT expert here....

JR

 

[cyrano]

The two video files downloaded fine for me. So, either it's fixed, or there's a problem on your side.

Could be a too old browser, as Ethan suggested, or a routing problem with your provider. You can't do anything about that routing problem. Your provider will probably fix that.

There is something funny with those files. I can't preview them on a Mac. Wouldn't surprise me for the Windows (.avi) version, but even the Quicktime movie doesn't play in preview. Both play fine in VLC however...

 

[terry setter]

I opened QuickTime (on my Mac) and used the "Open Location" feature under "File" in the menu. I put in the address (https://ethanwiner.com/jitter.mov) and I can hear it fine now. However, my computer showed a message saying "this file contains information that is not compatible with QuickTime". I though Ethan mentioned that there would also be visual content, so I'm assuming that it is the visual material that was not retrievable by QuickTime. Thanks for the suggestion.

What I notice as much as the increase in graininess and spitting sounds is the reduction in the "space" around the sound sources. It's narrower and shallower. I point this out because it might be the basis of comments about how jitter can negatively effect stereo imaging. For me, the relative position of sound sources remained pretty stable, but with the addition of a massive amount of jitter noise, they became far more two-dimensional sounding and far less engaging.

Thanks to Ethan for the work and JR for the link.

 

[kags]

An old friend, Ethan Winer, just wrote an article about clock jitter.
https://audioxpress.com/article/aud...eJODoSDyeyfGZJ5Hyy8YVaYfyOCZIZVBOzpQyq9trmfc8

JR

Excellent article from Ethan. Avid (then Digidesign) issued a white paper around 2008 that came to similar conclusions. That is, external word clock will not improve the jitter performance of a device, although it is sometimes necessary to have an external word clock in complex setups. This is, obviously, contrary to what marketers of word clock generators would have us believe. One company even claims to add noise to their jitter ("jither") to make the pico-second jitter sound even more musical (!?).

In my experience in a mastering situation, clocks and converters may sound slightly different, but there is no way to know which is "correct" or "more better." These days every ADC and DAC has an incredible clock and converter set and as long as the rest of the design is decent, we get insane performance even from inexpensive interfaces and DACs.

FWIW, I often point people to this site, which explores jitter in a very exaggerated, but effective way. https://www.sereneaudio.com/blog/what-does-jitter-sound-like

 

[Tubetec]

thanks,
I'll check out those links tomorrow ,