nRF52832 + sgtl5000 wireless audio

[enukic]

Hi, I'm trying to set up wireless audio stream using 2 nRF52832 Dev Kits and SGTL5000 audio boards.
SDK: nRF5 SDK 17.1.0
For the wireless protocol I've used ESB, and the program is based on SDK ESB Tx/Rx examples, alongside with this example from NordicPlayground Github:

https://github.com/NordicPlayground/nRF52-teensy-sgtl5000-audio

Audio stream works, and the latency is decent, although I'd like it to be less (15-20ms).
But the issue for me is when I measure input and output signal with an oscilloscope (where input is a sine signal), the output signal's phase is always slowly shifting.

I am attaching Tx and Rx projects.

Does anyone know what could be the issue here? Is there a way to address it?

Thanks a lot!
Emir

nrf_sgtl500_esb_audio.zip

[Sigurd]

Hi!

But the issue for me is when I measure input and output signal with an oscilloscope (where input is a sine signal), the output signal's phase is always slowly shifting.

Could you explain this a bit more? is it time shifted, or shifted how? Do you have any plots, pictures, etc to share that shows this?

[enukic]

Hi Sigurd,

Thanks for the reply.

It seems like the output signal is slowly shifting in time. For example, if I sit and watch at both signals at the same time, the output signal will slowly slide to the left.
Here is a video of the oscilloscope I made with my phone. I'm sorry, this is the best I could do. The audio is pretty noisy too.

www.youtube.com/watch

First scope (yellow color signal) probe is on the line in pins of the sgtl500, which are connected to my laptop headphones output, and the other scope probe (cyan color signal) is on the 3.5mm jack of the sgtl5000.

If you observe central peak of the sine for both signals, you will notice the cyan signal sliding to the left over the time.

Thank you for looking into this!

Kind regards
Emir

[Sigurd]

enukic said:

First scope (yellow color signal) probe is on the line in pins of the sgtl500, which are connected to my laptop headphones output, and the other scope probe (cyan color signal) is on the 3.5mm jack of the sgtl5000.

Could you make a drawing of this setup?

enukic said:

If you observe central peak of the sine for both signals, you will notice the cyan signal sliding to the left over the time.

Does something at some point crash when enough time has passed?

[enukic]

Sure, here is a drawing

Does something at some point crash when enough time has passed?

I haven't noticed anything crashing. I have observed the system for a couple minutes when testing (multiple times).

[Sigurd]

Difficult to say why the shift is happening. If it's buffering on nRF52 or SGTL5000 side. I don't have a SGTL5000 to test with. But I would expect it to stabilize after a while, if not then, I would expect some buffering warning/errors/overrun/crash/etc to happen, that could point to why it's happening.

[enukic]

Thanks for the support on this. Unfortunately, I wasn't able to solve this issue up to this point.
If you seem appropriate, please close this ticket. I think I have hit a dead end with this one.

Thanks
Emir