Microphone suddenly provides only noise?

Help !

I’ve been working on a PB V3 + Sensor Board on which i’ve removed the line in module by shorting the 2 contacts under the “cube” pad.

I just finished my wearable, everything was working fine, I closed the Pixelblaze box (it has holes and contains only the Pixelblaze and a few wires for BTN and 2 GPIO buttons) and suddenly the sound reactivity is gone ! Or should I say it’s reacting all the time, even in silence (i’m not talking about the small residual silence noise).

I re-opened the box (maybe it was blocking the sound) and exposed the Pixelblaze to the air and it’s still just reading noise all the time, no sound reaction anymore…

Did I fry the microphone ? (how even !) Is there a way to troubleshoot ?

I tried several patterns, it’s not the code. Microphone is just outputting random noise like full volume, even in silence. Moving wires around did not have any effect.

Thank you…

The sensor board can pick up severe fluctuations in power, does reducing brightness or turning LEDs off completed help the noise at all? Powering with another power supply change anything?

Could try adding bulk capacitors to PB to smooth out the power.

Huge thanks for the (always) fast support.

Wow indeed I was not expecting that AT ALL with lithium batteries.

For the record I’ve been using 18650 and 14500 lithium batteries for a while now with exactly the same configuration and I NEVER encountered this issue. I connect them directly to the Pixelblaze and LED, with a switch to turn ON/OFF.

It seems only one of them is unstable when it gets to low voltage. Weird.
These batteries have an onboard USB to recharge it + integrated BMS, might be the cause.

I will think of a capacitor on the next version. Any capacity you recommend ?

Depends on your current draw, etc., but generally no less than 100uF, probably no need for more than 1000uF, and with a fairly low ESR. Decent aluminum electrolytic caps work well enough. They aren’t the smallest things out there.

The problem could be some resistive component along the way, or lots of little things adding up. Dirty or thin battery contacts, bad cells, bad cells, wires, switches, and a BMS will have some but usually a few 10s of milliohms.

With the system running and under load, measure the voltage between the battery metal itself and the PB VIN, once for GND and once for positive. They should read very close to zero, but will tell you if there is voltage loss outside of the battery somewhere, and which side it’s on. Measure the battery voltage itself under load and idle, and if there’s a difference it means there’s some internal resistance of the battery. If the difference is fairly high, it could mean the cell is bad, or low quality, or just loaded beyond it’s spec.