Class D Guitar Amp Noise Issues

[Ben7128]

I have a Vox Mini3 G2 3W guitar modelling amp. It has two noise issues. With no inputs plugged in, running on batteries there is considerable varying white noise in the speaker, and headphones even with gain & volume at 0, no FX and cleanest amp model. Note that unused inputs short correctly. With various guitars or mic plugged in, when I increase the volume there is a loud buzzing until I touch metal on the guitar body. The guitars are noise free using other amps. There is no real difference if i switch between mains plugpack and batteries.

See following pics. The driver amp is TPA3130 Class D in mono (full bridge) configuration.  There is approx 20V 400kHz pcm signal at the speaker (wrt BATT GND) - is this normal or should the output filters have removed the carrier altogether?   There is a varying 40 mVpp noise at the amplifier input, f = approx 200kHz.

I welcome suggestions on how to fault find this, as it a great sounding amp which I would like to use for recording. Thx, Ben

[nanodocl]

Hi, do you have the schematics for this amp?

[DrGeoff]

First check that it is the amp itself responsible for the noise, and not something else like the stage before it or the power supply.
The output filter and supply decoupling are important to get correct in class D design, along with the layout.
Is the switching waveform you measure at the load (speaker) terminals? If it is in bridge mode I assume you are only measuring one side wrt ground? If so then try measuring the other side with the second channel and view them together to see the phase relationship.
You are not going to hear 200kHz so that isn't the source of your noise, however a decent output filter will attenuate this considerably. Is it hiss you are hearing or specific digital tones?
It should be possible to get very quiet output with no signal for the TI class D amps, I have designed several using other TI chips and they are dead quiet and very efficient.

[Ben7128]

Hi, do you have the schematics for this amp?

Wish I did. Does anyone have a link to it or something like it?

[2N3055]

TPA3130 has analog inputs.

Which means you need to check if signal coming in is good or noise is already there.

Many small power class D amps don't have  or have very simple output filters and you will see PWM on speaker out.
Download datasheet, it has many details about TPA3130  and it's application..
At least you have that part of schematic.

[Ben7128]

Thank you for your replies. I didn't say, but other than this noise, the amp and speaker are working fine. The noise is like wind noise on a mic - I guess you could call that hiss - and not a specific tone. Maybe resistor shot noise. 

I have TPA3130 datasheet and I think I have a fair idea of how it works. But it will help to know if the carrier wave should appear at the speaker or not. At the moment I feel like a dog chasing its tail.

The first pic below shows noise at TPA3130 input (pin 4).

The second pic below shows the output of inductors L13, L14 wrt BATT GND. The carrier waveforms appear to be in phase.  I tried adding extra capacitors (10n, 1u, 10u) to ground at the measuring point with no real change. 

[Ben7128]

Pics across the speaker terminals and L & R inductor output showing slight diffs but basically in phase.

[2N3055]

Like I said, PWM on speaker is OK. As you can see signals are in phase, meaning speaker sees no current between them.

Noise might mean bad contact.. Did you try gently tapping on different parts of PCB to see if you can hear difference..
If there is an analog potentiometer in a path they can be culprit..

[Ben7128]

Like I said, PWM on speaker is OK. As you can see signals are in phase, meaning speaker sees no current between them.

Noise might mean bad contact.. Did you try gently tapping on different parts of PCB to see if you can hear difference..
If there is an analog potentiometer in a path they can be culprit..

Could not find any mechanical issues or microphonic parts other than the input connector itself. It's hard to access parts of the signal path. Being a digital amp, the pots are all controllers so audio doesn't pass through them. The noise appears present on both sides of the main DSP. I think I will try to nut out a schematic for the main signal path. One thing I forgot to mention, the noise does not get louder or softer vs volume and gain changes. I.e. it is always present at a fairly constant level and gets masked when the volume is cranked out. Which is OK for loud playing but very annoying for quieter playing and between songs.

[Ben7128]

I fixed the second issue (buzzing when guitar plugged in). It was due to a semi-dry joint on the input jack. I believe this is a common issue with small amps as they are not constructed as solidly as larger ones. The bursts of white noise is still a mystery as I am not able to trace it back to source due to the multi-layer board construction.