WM8960 Breakout Board GNDs - one special GND is AGND?

Sensors Sound & Audio
1

The WM8960 Breakout Board has 6 GNDs, 5 on the left and 1 on the right. There is one unexplained reference to AGND for using headphones in the hookup guide but otherwise, there seems to be no difference between them. I have even studied the schematic and EAGLE board design and they show them all tied together.

So I thought they were all equivalent but I still had some reservations:

  • The WM8960 datasheet describes 5 different grounds (DGND, AGND, SPKGND1, SPKGND2, and GND_PADDLE and say that GND_PADDLE and AGND should be tied together but not necessarily the other ones).
  • The hookup guide always shows microphones being hooked up to the top-left or top-right GNDs, but this could be out of convenience.
  • When looking at the Super Headphones schematic posted on the SparkFun website it shows the breakout board having a separate GND and AGND along with their EAGLE board showing usage of the top-right GND as AGND specifically.

So I wired things up and put all GND together. I was getting large amounts of noise from the microphones. I finally decided to take things apart and discovered that even without a microphone the noise persisted, meaning that hooking common GND to the LIN2 port caused the noise to occur (the microphone was using a pseudo-differential setup with MICBIAS as the positive voltage). Exploring, I discovered there was only minimal noise when LIN2 was connected to only the top-right GND and that GND was not connected to the common GND (but all 5 other GNDs were connected to a common GND).

So ultimately it seems that the top-right GND is in fact “special”? Is it indeed or is something else going on here? Why does the breakout board tie all the different GND together even though the datasheet does not mention doing that?

There is also a little bit of noise left that I will work on eliminating, but it is significantly improved by changing the GND connection. Any other important source of noise to look for? Already going to try reduce exposed wires and move the biasing resistor closer to the microphone.

2

I’d first try to confirm the power supply isn’t contributing to the problem…are you running a mean-well supply? Using a battery power would help isolate whether it is or not

3

I just got a battery today and tested and regardless of if it is on common ground or the “AGND” I get the same clear results. I then tried again with the laptop USB power source I was using before and I also get clean results… I did move everything to a larger breadboard in the meantime… maybe there was noise being introduced by the breadboard itself? This would corroborate the different pins behaving differently as it is possible that only some lines had issues in them.

1 Like