Noise issues and power

Status
Not open for further replies.

flpvsk

New member
Hello,

There were a few people that asked similar question here on the forum and I've tried suggested solutions but failed so far.

I've built a pitch-shiter for Teensy based on this article and code: http://blogs.zynaptiq.com/bernsee/time-pitch-overview/ It works, but depending on how I power the Teensy there's no noise at all or a lot of it.

Here are three cases and results

Case 1 – Powering from an iPad power supply using USB – No noise

IMG_20200810_181427.jpg

Here's the audio: soundcloud 1

Case 2 – Powering from a Mac using USB – Mid level noisy

The layout is the same, but I plug the USB cable into my laptop.

Here's the audio: soundcloud 2

Case 3 – Powering from a Pedal Power Supply through a voltage regulator – Very noisy

IMG_20200810_175608.jpg

Here's the audio: soundcloud 3


The code of the project is here: https://github.com/flpvsk/PitchShifterTeensy. I understand that the problem is a mix of how the algorithm works and the power filtering, because depending on certain parameters in the algorithm the pitch of the noise changes. I do not understand what is the best way to fix it – using software or better filtering on the power side?

My ultimate goal would be to power it up using a 9V power adapter.
 
Here are three cases and results

Case 1 – Powering from an iPad power supply using USB – No noise

Case 2 – Powering from a Mac using USB – Mid level noisy

The layout is the same, but I plug the USB cable into my laptop.
Yes, USB 5V from a computer will be noisy, with components all across the audio band and upwards,
good filtering is advised, but its not the ideal source for an analog circuit.
Case 3 – Powering from a Pedal Power Supply through a voltage regulator – Very noisy

What power supply is this? And regulator? Something very different is happening here, I would
want to investigate the power rail voltage with a 'scope even... Perhaps the "Pedal Power Supply"
is a really bad switch-mode supply that's beating against the sample rate or something like that?
Perhaps the supply isn't regulating properly, is overloaded, or there's a ground loop not present
in the other setups?
 
Thank you for the reply, MarktT!

What power supply is this?

Voodoo Lab Pedal Power 2 Plus It's a standard for pedal power and normally quiet enough even with digital devices.

Regulator: KIA278R05PI There's also a 100uF capacitor between the positive 9V and ground and it helps a tiny bit. Adding more caps before / after the regulator doesn't do anything.

Everything is plugged in into the same wall socket, so I guess there's no possibility of a ground loop?

Could the problem be that I'm connecting all the grounds together? I've seen some people recommending separating digital and analog ground. If yes, what would that mean in the context of that circuit (input buffer -> teensy -> output buffer).
 
If there's a loop, there's a loop, doesn't matter if plugged into same wall socket or not. Current in any part of the loop imposes a voltage across the
rest of the loop... However that power supply is isolated and designed to prevent noise, so it shouldn't be an issue.

Something strange may be happening then - a 'scope might be one way to narrow this down. Perhaps a faulty unit?

Digital and analog ground have to be connected at the ADC or DAC anyway, it wouldn't be different depending on power source either.

How are you recording the output?
 
I might have to borrow a scope in that case.. What would I check? The input and the output of the regulator?

I'm recording the output through an audio interface connected to a laptop. So the full chain is

(Guitar with a Volume knob at 0) -> (Input buffer) -> Teensy -> (Output buffer) -> Soundcard -> Macbook

But I've observed the same problem plugging in into an amp instead of a soundcard. Also, same problem exists if I take the buffers out. I've added them in hoping it would help.
 
Status
Not open for further replies.
Back
Top