TPA3122D2 oscillating output

Hi, Martin,

Martin,

Hmmmmmm, sounds like a ground loop issue.

The output is still switching, but there's no sound?

Another place ot look is the bypass cap. It has a comparator on it for pop avoidance, so if the voltage on it isn't up to snuff due to leakage, it may go into the mode you describe.

You may want to buy an EVM to play around with.

As always, I encourage customers to copy our EVM designs. I just noticed we missed putting the gerbers on the web. Here they are. I'm going to post them to the web as well.

-d2

Hello Don,

thank you for your fast reply.
I did some testing with a 1kHz input signal. Sometimes i can hear the 1kHz at the output and the signal looks like picture "1" im not sure if the overlain higher frequency has to be there. Usually I cant hear the 1kHz at the output and the signal looks like "2". "3" shows the same as"2" but with another timestep, my picoscope measures 81kHz at this point.
All three measurements are made with the same setting, the change comes for example just by touching the iPod.

I have also measured the current through my bypass cap and my meter says 0.0uA (after charging up in under 1 sec), i guess thats ok.

I'd like a layout with only THT components, the recomended EVM design uses many SMD components. Correct me if I am wrong, but i think it should be possible to get this running with THT components.

I will look into the Ground loop suggestion in the next days and do some more testing. Feel free to bring up more weak points that could cause the problem.

Best Regards

Martin

Hi, Martin,

I looked at your schematic until I was cross-eyed, and I honestly am not sure what you are trying to do.

Are you trying to feed the stereo signal from the iPod to the L/R inputs of our chip, and then connect the outputs BTL across your speaker?

I also didn't fully understand your plots.Why is there so much hash on them?

-d2

Hi Don,

my schematic should be closely to figure_29 from the TI Datasheet (see attachment).
Here is what i want to do:

-mix the L+R channel to a mono signal (by using the C2+C3) this works with the TPA3106d2
-decouple DC from the iPod C1 (C1 may also be needed for BTL configuration?)
-drive a speaker in BTL

The plots show the output measured while a speaker is connected. The hash is the malfunction I try to find?

I'm sorry if your eyes suffer:)

Best Regards

Martin

Hi, Martin,

If you feed L into one input, and R into the other input, the BTL output will be L-R, not L+R as I suspect you want. Music might sound sort of ok, but you will be missing nearly all the bass.

You need to invert one of the channels before applying it to our device.

Now, for the hash, I'm not sure what's going on there, still... Have you made any progress?

-d2

Hello Don,

is there a way to invert the signal with only passive components?

I will do some testing with the right signal and give a feedback. Perhaps the hash is created by this faulty signal circuit.

Best Regards

Martin

Hi, Martin,

Since we also sell op-amps, we have only ever tried it using them. :)

-d2