Diagnosing a dull pop, followed by dull/dark and quiet output

[andicniko]

Thank you for the fantastic design and instructions. I'm hoping your experience might help diagnose a problem I'm having.

Issue:
After a few minutes of use (varies, anywhere from 1 to 10 minutes), my microphone outputs a dull pop(s), after which the output goes very dull/dark and quiet.

I have attached:

• audio and an image illustrating the pop and change in output.
• an image of my board (in case there is an obvious error in the assembly?).

Audio.zip

Things I have tried so far (in case it helps troubleshoot):

• Unplugging and leaving overnight resolves the issue temporarily. The output is back to normal, until a few minutes of use reproduces the same issue.
• The issue occurs even if I don't interact with the mic (i.e. it's not triggered by input or humidity from my breath, it happens even if recording silence from a distance).
• I don't think it is a faulty capsule, board, connector or cable. I have tried multiple of each (I was building two mics so had spares).
• It don't think the issue is my audio interface. Mine has two inputs/preamps (Arturia Minifuse 2), and the issue happens with both. Switching between inputs doesn't resolve the issue once it starts. The issue doesn't persist if you switch to a second identical mic (again, I was building two mics so was able to test this).

Do you have any ideas what the issue could be? Is it possibly a phantom power issue? Thanks in advance.

[Spirit532]

Thanks for building the design!

I encountered this exact issue in a previous design revision, where I didn't have the R101-R110 chain discharging the diaphragm, which was simply slamming the input of the opamp(U1A pin 3) beyond its rail, causing it to go silent.
The reason it's detected as a pop is because mics are AC coupled(C2/C1).
You can verify that this is the case by measuring the output of U1A(pin 1). When it goes quiet, you should see the chip saturate and sit high. If it doesn't sit high, then it's either your capsules or potentially 1/5(see below).

There may be multiple causes, any one of which would lead to this, in the order of easiest to test(but not necessarily the most likely):
1 - Your capsules are just built differently to mine and can't handle the high of a bias. You may attempt to lower the bias voltage by sequentially removing C16&D7 and shorting the nodes labelled 72V and 60V, then doing the same for C15&D6, 72 -> 48V if it's not enough.

2 - The 1G bleeder at the input needs to be smaller. You may short up to... I don't know, let's say 5 resistors in the R101-R110 chain, halving the total resistance.

3 - The board isn't clean enough(and it won't be if you do №4). It must be perfectly clean. Spotless. No finger grease or dust. I use brake cleaner(the non-chlorinated kind), followed by a distilled water wash, followed by drying with isopropyl alcohol.

4 - Your board's solder mask is black, which may be slightly conductive(even though it's not carbon, obviously). At these currents, dealing essentially entirely with electrostatics, that may be enough to cause problems. Not much you can do, but maybe scraping it away around the different areas highlighted in blue, all the way to the substrate.

5 - Your capsules are already wet, not from your breath but improper storage(including at the factory or in the mail). Unfortunately it's really hard to say how to dry a capsule correctly, but what seems to work for others is either placing it in a sealed container with a pile of fresh desiccant next to it for a few days, or by baking it in a (clean!) oven at no more than 40C for 5-10 hours. The latter may cause permanent damage to the capsule, especially if the oven overshoots to 50-60C.

[andicniko]

Great answer thank you! Problem solved.

Yes it was the chip saturating and sitting high. I can also confirm it was because my capsules seem to be built differently.

Solution: I ordered new capsules that matched your pictures in appearance (yellow tape, blue and white lead wires) and can't reproduce the issue anymore. It was cheap enough to try before doing any surgery on the PCB, and luckily worked :).

I guess my problematic capsules (green tape, yellow lead wires) came from a different factory and some unwritten specification was different.

It is a fantastic mic by the way. Thanks again.

[Spirit532]

You're welcome :)

The other capsules probably had too high of a leakage current. Might be solvable by reducing the 1G chain's resistance.
They could also have been collapsing(diaphragm sticks to the frame), which would also lead to this behavior. That would've been the bias voltage being too high.