My First Audio PCB Layout - An OPA Alice with Hex inverter

[Pariah Zero]

Disaster! Sortof.

Good news: The hex booster checks out so far. (It boots the voltage from 12 to 80V as I tended)

And when I plug in the OPA Alice (without a capsule - using just a 68 pF ceramic cap as a simulated capsule), the circuit seems pretty quiet by itself. (If I have my DAW measuring and the Zoom U-22 cranked up to max gain, it shows -72 dbFS for the noise level, and I'm reasonably sure a fair amount of that is the audio interface -- I have to leave for an appointment before I could test further.)

Next up: testing against my Scarlett 2i2 gen3's preamp, and then installing a capsule.

The disaster:


It looks nice, but the two larger electrolytics on the left are too tall to fit inside the shell. So, a new PCB spin will be required to fit inside this body. Fortunately there are other bodies this one fits just fine. (Like WGTcenter's U87ish clone)

 

[Pariah Zero]

It's been interesting. Initial tests using a 'spare' electret capsule from a BM-800 sounded great. The hex booster worked great.

So I hooked up a capsule and... nothing. Since I was short a 220 Ω resistor for the second circuit, I just used a single-sided capsule, and a 3-pin connector.

More testing. Seemed fine. Every time I tested it with a voltmeter, everything checked out. Closed it up, plugged it in... nothing.

Huh. Well, I decided to try it out topless. It worked great, but hummed a lot (go figure...). So, I started looking closer at the mic body, looking for a short. I couldn't find one.

I did find that if I screwed the bottom cap on too tightly, the mic went dead. Ok, then.

The missing resistor arrived, so I decided to switch out the connector, since it seemed a prime suspect for the short. I was going to switch to a 5-pin connector so change the mic pattern in the DAW, vs a switch.

I also disassembled part of the mic body and did some metal work to ensure there was more room around the connector, no place for shorts, etc.

Then I reassembled, and checked thoroughly for shorts in the connector/body before wiring up the PCB.

ALL GOOD!

 

[Pariah Zero]

Except now, nothing worked. Even worse, the voltmeter was telling me things that should have been impossible. I laid out the PCB, so I knew the tracks, right?

If I plugged in only the cardioid cable (no back capsule action), then how in the nine hells was it registering 48V on the output pins!?! There was no electrical path. None.

I started probing each point, and the results just kept getting more baffling. Until, that is, I noticed something... one trace on the ground plane I kept for electret mics didn't seem to connect to pin 1. After looking much closer, I finally found the problem: I had missed a label in the schematic when I renamed one net, giving me this:



Now that's a beautiful ring of solder mask sitting around the ground plane contact for Pin 1.

So, I was hit with a double fault: First, the ground pin didn't go anywhere, which let the rest of the board float. Second, the short circuit I originally observed seems to have provided a working ground path.

... and of course prodding with a voltmeter did manage to fry one of the op amp chips. Fortunately, the $26 hot air pencil I picked up from Amazon made swapping the SMD chip a breeze.

It's finally working, and sounds pretty good. I just haven't had a chance to do any frequency/noise tests yet.

 

[MicMaven]

Except now, nothing worked. ... I started probing each point, and the results just kept getting more baffling. Until, that is, I noticed ... It's finally working, and sounds pretty good. I just haven't had a chance to do any frequency/noise tests yet.

Novice Observation No. 208974 -

Thank you for your detailed troubleshooting story. This tells the Novice mic builder two helpful things - first, that he is not alone in having this type of issue, and second, how to go about finding and resolving the same. James - K8JHR -

 

[Pariah Zero]

Novice Observation No. 208974 -

Thank you for your detailed troubleshooting story. This tells the Novice mic builder two helpful things - first, that he is not alone in having this type of issue, and second, how to go about finding and resolving the same. James - K8JHR -

Thank you. I certainly was taught that issues are normal and that resolving them takes determination and that the 'root' cause is often further down than expected. (Wire is often my #1 culprit, and once you accept wire is a likely problem, well, everything is up for grabs).

After that, it's definitely an XKCD 979 situation - sharing is always for the best.