LM48100Q-Q1: Why am I getting noise on the LM48100Q audio amplifier, even from a flashing LED ?

Hi John

Sorry that I can't open you link, IT blocked them.

But from your schematic and layout, indeed possible coupling noise. Single end input using, the audio quality would be lower, and your input trace is indeed too long.

And another thing possible could have a try. The R4/R3 resistor makes IN1 high impedance. This 2 resistor seems quite close to LM48100. Have a try to remove this 2 resistor, and short the resistor place to GND. Possibly would make the results better.

Thanks Shadow He,

The noise is on channel IN2, not channel IN1.

Channel IN2 is connected to the DAC output of the STM32

Channel IN1 is connected to the DAC output (left and right) of a bluetooth audio receiver MH-M18 but this is temporarily disabled.

See block diagram below (bottom right)...

The links that you are unable to view show videos of the audible noise on channel IN2 (from the STM32 DAC) whilst channel IN1 is disabled.

Strangely, I don't get noise on channel IN1 when it is enabled, even though the bluetooth receiver DAC channels tracks are routed similar to the STM32 DAC for channel IN2.

Below are the scope captures of amp channel IN2, but I can't see any noise for the hiss or interference from the LED....

The PCB track routing is shown below:

Green is the STM32 DAC output

Red is the bluetooth receiver DAC right output

Blue is the bluetooth receiver DAC left output

The analogue ground (GNDA) for the STM32 DAC is shown in yellow below, it goes back to the battery negative terminal...

I have uploaded the videos (from the links) - see below...

Hi John

A little strange, that the waveform you captured at input line, seems not match the noise.

In your video, the LED is switching at around 10Hz? Or maybe a little higher. 

But the input waveform, the noise frequency is more likely to be 25KHz. It is quite similar to the noise on the 5V power supply, maybe you could double check on this point. And 25KHz itself is not able to be heard.

And your output waveform shows, not a very clear fixed frequency on it, more seems to be white noise. You could try to do FFT analysis about it to see more clearly, any fundamental frequency could be find out from this waveform. And is below waveform captured with led working?

My guessing for now, is the noise mainly conducted through GND plane, as both your power supply and input has similar frequency. And led circuit seems raised the background noise level, make it more clearly heard from speaker.

Thanks Shadow,

The waveforms captured were with the LED flashing.

I finally got round to measuring the FFT of the STM32 DAC output, but it does not mean much to me...

The measurement was taken at the STM32 DAC output pin and a local ground....

Hi John

FFT looks quite clean, not having a special noise peak, just a DC. 

But the back ground noise level seems too high, in your FFT analysis, around -50dBV, which is 3mV. For the audio side, 50uV~100uV white noise, already possibly been heard from some sensitive speakers. 3mV is enough to generate clear sound. 

MCU's internal DAC usually not specially designed for audio, and possibly have this background noise level. You could check the spec of STM32. 

Thanks Shadow

Which parameters are required for the STM32 DAC ?  Below is the SNR and THD...

Strangely, when I built the prototype using development boards, I don't get the same hiss that I do with my custom PCB which is surprising given the rats nest of wires...