TPA3118D2: 1MHz Sine on Input pins and a lot of noise on speaker output

Florian Born

Part Number: TPA3118D2
Other Parts Discussed in Thread: PCM5100

Tool/software:

I designed a Digital Audio circuit with the TPA3118D2 as the audio amplifier.

It's using a PCM5100 as a I2S DAC which is connected to the TPA3118. I've attached the parts of the schematics to see the design.

Generally I can hear sound output but it has a lot of noise. When checking the audio inputs with a scope I can see a ~1MHz sine wave modulated on the input channel.

Scope picture shows measurement directly at the TPA3118 pin.

I tried disconnecting the DAC output and the output signal is very nice following the I2S data, meaning the 1MHz is coming from the TPA3118.

Interestingly this wave is only present to the inputs connecting to the DAC. The negative inputs RINN/LINN don't show the sine wave.

Anyone has an idea what this is and how to fix it?

The PCB was soldered in an oven and not by hand, so gnd connections should be solid.

PCM5100 SCH

TPA3118 SCH

Scope RINP

1 month ago

0 Sydney Northcutt 1 month ago

TI__Genius 17580 points

Hello,

Thanks for reaching out. I am reviewing your data and will follow up by the end of the day.

Regards,
Sydney Northcutt

0 Sydney Northcutt 1 month ago in reply to Sydney Northcutt

TI__Genius 17580 points

Hello,

With the DAC connected to the amplifier, can you probe the amplifier inputs close to the TPA3118D2 device?

Can you also probe the power supply?

At first observation I think the issue is somewhere on the routing from the DAC to the amp. Are there any other connections or probes / test points in this path?

If you remove the DAC and apply an external signal close to the TPA3118D2 pins is the same observation observed on the output?

Is there any other external circuitry on the board that you can turn off while testing? Do yo have anything operating around 1MHz?

Regards,
Sydney Northcutt

0 Florian Born 1 month ago in reply to Sydney Northcutt

Prodigy 10 points

Hi Sydney!

This is the PCB Layout of the DAC->Amp Path.

The colored nets are the RINP and LINP traces. Underneath is a solid GND plane.

Power Supply looks like that, coming directly from a bench supply.

Internal voltages are generated by LDOs only.

These are the measurements from one channel. Yellow being directly at DAC side, turquoise being directly on TPA side.

On the other channel I removed the connection and the DAC output looks like that

0 Florian Born 1 month ago in reply to Florian Born

Prodigy 10 points

Just tried to inject a waveform, it's the 1kHz test waveform from the scope.
I connected it to the lightest green trace and measured at the TPA Pin.
The resistor between the light green net and the PCM5100 is not populated.

This time the signal is clean

0 Sydney Northcutt 1 month ago in reply to Florian Born

TI__Genius 17580 points

Hello,

I see. So to confirm, the amp is outputting correctly when the DAC is disconnected, but the DAC when the amp is disconnected is outputting with this 1 MHz signal?

Regards,
Sydney Northcutt

0 Florian Born 1 month ago in reply to Sydney Northcutt

Prodigy 10 points

Not quite.

When DAC and AMP are disconnected they both work as expected. An injected signal also works as expected.
As soon as DAC and AMP are connected, I can see the 1MHz Sine.

+1 Sydney Northcutt 1 month ago in reply to Florian Born

TI__Genius 17580 points

Hello,

I see. Can you remove the cap to ground and resistors in the input path and replace the resistors with shorts and observe the output?

Regards,
Sydney Northcutt

0 Florian Born 20 days ago in reply to Sydney Northcutt

Prodigy 10 points

Sorry for the late reply.

This seems to do the trick. Would you care to explain why?

Audio

Audio forum

TPA3118D2: 1MHz Sine on Input pins and a lot of noise on speaker output