Clock "noise" in audio signal

[telengard]

I built a DIY sampler and it is working well except one thing.  I hear a high pitched whine when using it.  The pitch of the whine changes when I move the pot that controls clock speed.  With no audio going into the device I look at the output and see a periodic signal which changes frequency as I change the clock pot.  It's ~60mv p2p.

Also, if I measure with my scope with the probe ground attached to the input power ground terminal and then my signal probe at the ground pin of the clock chip, I see that periodic signal.  I guess that surprised me, but I'm also a newbie.

Just looking for hints on where to possibly look, or if this is just normal behavior that maybe can be tempered a bit, but not eliminated?  Other folks have built this and don't seem to be having this issue, so I figure I did something wrong.  I've gone over the solder joints many times and it looks OK to my eyes.  I've also eliminated the 9v wall wart I'm using having anything to do with it by using a 9v battery and my Rigol bench supply.  Same issue in all cases.

If this is too broad of a question, I can give more details.

Any guidance greatly appreciated!

[TimFox]

Can you show a scope trace of the periodic waveform?  What is its frequency with respect to the clock?

[telengard]

Can you show a scope trace of the periodic waveform?  What is its frequency with respect to the clock?

Hi, yup.

They seem to match.  Here's a pic (sorry didn't have a pen drive handy to do real screenshots) of the audio output



and a picture of the generated clock (from a 4046n which seems to be a phased lock loop chip)



what was confusing to me is I checked the ground signal with respect to the - of the input power terminal and I see a stronger spike than on the audio output

[Bassman59]

Is it your PCB layout, or someone else's? How many layers?

I suspect the layout is, well, suboptimal. It's easy enough to do mixed-signal layout, but you need to make sure you follow the rules. Route digital stuff away from analog to the extent possible. Don't use split ground planes. Make sure digital traces always go over a ground plane for the return current. And so forth.

[telengard]

Is it your PCB layout, or someone else's? How many layers?

I suspect the layout is, well, suboptimal. It's easy enough to do mixed-signal layout, but you need to make sure you follow the rules. Route digital stuff away from analog to the extent possible. Don't use split ground planes. Make sure digital traces always go over a ground plane for the return current. And so forth.

I've been reading and have seen a lot of these recommendations.  This is not mine at all, it's just a kit I purchased and soldered up.

Others who have purchased the kit haven't complained of this so my guess is I may have done something wrong (or their hearing or tolerance of the noise is better than mine  )

[telengard]

Are there any basics or suggestions I can look into to possibly find the culprit?  Is there a process for trying to narrow this down by process of elimination?

[Vovk_Z]

Try add small inductors (about 3-10uH) somewere in power lines, to shorten an radiating antenna.
LC filters always help, (more or less.)

[telengard]

Try add small inductors (about 3-10uH) somewere in power lines, to shorten an radiating antenna.
LC filters always help, (more or less.)

Thanks for the tips, this is something I will try.  This is sort of a learning experience for me, is there a way to diagnose or troubleshoot the root cause of the noise w/ an oscope?  As I mentioned, other folks have built this kit and aren't having these issues, so I either did something wrong, or a component has failed.  I'd like to try and figure that out.

[Vovk_Z]

  As I mentioned, other folks have built this kit and aren't having these issues, so I either did something wrong, or a component has failed.  I'd like to try and figure that out.

- or they are just not very curious and don't mind to how good it works.

[AndyC_772]

Also, if I measure with my scope with the probe ground attached to the input power ground terminal and then my signal probe at the ground pin of the clock chip, I see that periodic signal.  I guess that surprised me, but I'm also a newbie.

This is an absolutely classic grounding issue, which may be only partially to do with your design. If two points are supposed to be connected together, but you somehow measure a voltage between them, then there must be some resistance - or in this case, inductance - between them.

Firstly, remove the croc clip GND lead from your scope probe and, instead, use the very short GND spike that came in your probe's accessory kit. Put that spike on the GND pin of the audio output, and measure the audio signal. Does it look better now? If so then the spikes are a result of the inductance of the croc clip lead, and don't actually exist on the audio signal at all.

If there's little or no difference, then you do have a layout issue which could be difficult to fix after the fact.

[eblc1388]

Also, if I measure with my scope with the probe ground attached to the input power ground terminal and then my signal probe at the ground pin of the clock chip, I see that periodic signal.

Disconnect the 0V wire of the input power terminal and connect it directly to the ground pin of the clock chip. Doing this ensure that the switching current of the clock chip does not flow via other paths before it returns to the 0V terminal.

 

[hamster_nz]

Do you have a schematic? Or even the part numbers of the major parts?

Sigma Delta  DACs can have 'idle tomes' that are a fraction of the sample clock, which can be more apparent if the sample is a fixed value... See https://www.analog.com/en/analog-dialogue/articles/using-sigma-delta-converters-1.html

[telengard]

Do you have a schematic? Or even the part numbers of the major parts?

Sigma Delta  DACs can have 'idle tomes' that are a fraction of the sample clock, which can be more apparent if the sample is a fixed value... See https://www.analog.com/en/analog-dialogue/articles/using-sigma-delta-converters-1.html

Yup, the schematics are here in 4 parts.  https://github.com/ToddBailey/WTPA2/tree/master/hardware

While poking around I think I managed to fry a quad op-amp chip, it's now really quiet as far as output.  I had checked across some of the pots and one of them made a slight hum when I did it and then the audio out was super quiet.  I've purchased a replacement.  I did read up on why that could happen, and if I'm understanding right I may have caused a feedback loop?

I also determined one source of noise I was seeing on my scope (but was inaudible).  It was the LED lighting over my bench, it was around 170kHz.  Having it on or off had no effect on the clock noise (which is audible, and changes pitch when I use the pots to control that).  But it was making it harder to see stuff on the scope.

Once I have the op-amp chip replaced I'll resume the clock noise issue.  Hopefully not blowing something else poking around.   

Thanks everyone for the suggestions, I'm learning a LOT!