Is the CM3 the same as the Pi 3B?

[BudO]

We have been working on building a Custom Board to process Audio recording and playback using the WM8731 Codec. We tested several Audio Boards using this chip and came up with a Prototype Board that worked as a Hat on the Pi 3B. The Codec Operated "as advertised" although I'm not sure if we have exactly the right driver (Audioinjector-Pi-Soundcard). The board is completely operational so we took it off from the Pi and and connected it to a CM3L on the Pi IO Board. We are using the same SD Card that we booted and operated the Pi with and used the same GPIOs on the CM3L that we used on the Pi, but it doesn't operate properly. We are able to record and use the audio pass through function, but the Codec will not playback. We can see activity on the data out GPIO from the IO Board (GPIO 21 ALT 0), but nothing from the Codec on GPIO 20. we are controling the Codec with GPIOs 2 and 3, ALT 0 and followed the connections: Bitclock GPIO 18 and LR Clock GPIO 19.

The only noticeable difference is the GPIOs are slightly more noisy on the CM3 Board and we think this might be the extra leads from the IO Board (about 6").

Any help would be greatly apreciated.

[PhilE]

Have you got pull-ups on GPIOs 2 & 3? They aren't there as standard on a Compute Module.

I'd also like to see "raspi-gpio get" output from both platforms.

[BudO]

PhilE

Thanks for the quick reply!

We have tried the board with and without the pull ups and it didn't seem to make a difference.

I have attached the GPIO Settings

Thanks

[PhilE]

I'm not a hardware engineer, but the fact that you are running the same software does suggest a hardware problem. Is the codec the master or the slave (or one of each) when it comes to the bit and frame clocks? Can you try reversing the relationship to see if it works better?

[BudO]

PhilE

Thanks for all of your help, we did find it was a hardware problem. We reworked our prototype board and got rid of all of the long interconnects and it started working properly.