1830 Hz noise when there is no input

[frsaba]

I am building a guitar effects processor using the Teensy 4.1 with the audio board. The boards are stacked and directly connected with pin headers and sockets. I'm using the PassThroughStereo example to send out the input from the line in pins. I hear an unpleasant high pitched ringing noise which most curiously goes away when there is some input signal. I've attached a sample wav file that demonstrates this. I have determined the noise to be at a constant 1830 Hz. The noise is the same on the line out, headphone out and USB.
This happens even when the line in pins are directly connected to ground. Powering from a 5V wall adapter or my laptop on battery has no effect.
The mic input is also noisy even when shorted to ground, although it is not high pitched rather normal white noise.

Is my board faulty?

 

[Pio]

I have encountered this issue, it's not really Teensy related, but the SGTL5000 codec chip.
I was able to narrow it down to a resonance happening due to impedance of the capacitor combo and the SGTL5000's analog power supply pin resulting in a high pitched (around 1.8-2kHz) tone polluting all the audio. It did show up with the recent SGTL5000 batches. Maybe something has changed in the manufacturing process. I did not have that problem ~2 years ago with exactly the same circuit.

The solution was to add more capacitance to the analog VDD supply (pin5 in the QFN32 package) which killed the resonant tone and made it silent.
For the Teensy Audio board you could try the following:
1. Simpler method, might work - add more capacitance to the digital 3V3 bus. Analog supply is derived from it using an LC filter.
2. If it doesn't help, try adding more capacitance directly to the analog supply pin, ie, 0603 10uF capacitor placed ontop of the existing 2.2µF one.

 

[MarkT]

Maybe it is fussy about the ESR of the decoupling capacitor?

 

[frsaba]

Update: adding a capacitor to the 3V3 line did not have an effect, however: I have found the adcHighPassFilterDisable function which helped immensely: the 1830 Hz ringing is gone, now there is only conventional noise on line out! I would still like to reduce it though (I plan on building a LM4562 preamp stage which I reckon should help significantly). The headphone out, though also improved is still not acceptable, but I'm not sure of the best way to record and analyze it.

 

[PaulStoffregen]

If I were to grab a brand new audio shield, how would I reproduce this problem? Do I run one of the audio library examples? Is listening with headphones enough?

 

[frsaba]

If I were to grab a brand new audio shield, how would I reproduce this problem? Do I run one of the audio library examples? Is listening with headphones enough?

For me this happens with the PassThroughStereo example with line in selected and the ADC high pass filter enabled. Ringing noise is equally apparent in line out, headphone out and USB. Disabling the ADCHPF makes it go away.