Audio noise problem.

[Braeden Hamson]

So I'm effectively building my own bluetooth speaker. I have an RN52 bluetooth module mounted to a PCB I made.

https://www.microchip.com/wwwproducts/en/RN52

I then output the audio from the bluetooth module to an audio amp

https://www.digikey.com/en/products/detail/pui-audio-inc/AMP2X15/8567350

I've configured the amp to run in mono differential mode. I'm sending an audio + and - signal to the amp. Not sure if those are the correct terms but I hope you can understand what I mean. I shorted the out1+ and out1- together as instructed (same with out2)

Now, I can get audio to the speaker so the system is working. However there is a nasty high frequency whine when I do so. 12V is going to the PCB and then it heads straight to the amp from there. On the PCB itself is a linear regulator, that in all honesty is stressed out and probably overloaded.

Interestingly when I power the audio amp with my power supply and then power the board itself with 5V USB through the Arduino on the board the high frequency noise goes away. However I pick up what sounds to me like a ground loop, a 60Hz drone. Which makes sense because they don't share a ground. When I'm powering the amp through the PCB they do share a ground.

If it will help I can record the high frequency noise.



Here's the schematic for what it's worth. I'll attach the files as well.

https://drive.google.com/file/d/1FHvSMnOOzbkGnSzdkvgecrVPBWnQCM2W/view?usp=sharing

Any help is greatly appreciated!

 

[Harald Kapp]

I shorted the out1+ and out1- together as instructed (same with out2)

I hope you did so using some resistors. Otherwise you can have nasty current spikes when the outputs have different voltages, as would be expected from a stereo signal.

However there is a nasty high frequency whine

Since this whine goes away when you power the board from a lab supply, it is probably noise from the switching regulator.

12V is going to the PCB and then it heads straight to the amp from there.

The least you can do is add an LC filter to reduce noise on the power supply line.

Also ensure that the GND on your PCB is solid, A ground plane would be best. If you can't use a plane, use solid wire, avoid ground loops. This article explains it in more detail. Eliminating ground loops will also help reduce the 60 Hz hum picked up from mains.

Which makes sense because they don't share a ground.

A small wonder this works at all. You need at least one common reference, usually this is ground.

 

[Braeden Hamson]

I hope you did so using some resistors. Otherwise you can have nasty current spikes when the outputs have different voltages, as would be expected from a stereo signal.

The datasheet said to short them, there are dip switches that configure the board to run in this mode.

Since this whine goes away when you power the board from a lab supply, it is probably noise from the switching regulator.

I'll remove the linear regulator to see if it's the problem.


Also ensure that the GND on your PCB is solid, A ground plane would be best. If you can't use a plane, use solid wire, avoid ground loops. This article explains it in more detail. Eliminating ground loops will also help reduce the 60 Hz hum picked up from mains.

I have full board ground planes that are via stitched. AFAIK all the devices are well grounded. But I'll check again.

A small wonder this works at all. You need at least one common reference, usually this is ground.[


In normal operation I have common grounds. However when I eliminate the common ground the high frequency noise goes away, and I just get the expected ground loop 60 Hz noise. This was just for testing as I noticed a lot of ground noise on the pcb.

Thanks for your help as always!

 

[Harald Kapp]

The datasheet said to short them, there are dip switches that configure the board to run in this mode.

o.k.

I'll remove the linear regulator to see if it's the problem.

The switching regulator is the problem, not the linear regulator.

I have full board ground planes that are via stitched.

Very good.