An amusing story about a bug in my MB-SID

[Wilba]

I share this for the amusement factor only, I have fixed the bug already  :)

I was wondering why SID #4 was playing some patches different to the rest, "Techno Saw" had some off-pitch tones in there, but "Techno PWM" seemed to play OK. Playing each oscillator by itself seemed to be fine. So begin the usual debugging process....

Problem with SID or Core? Swap SIDs over, problem stays with the Core, not the SID.

Firmware got messed up? Upload the MB-SID app again to all the cores, no change.

Buggy PIC? Swap PICs #3 and #4, problem stays with the PCB. Aha!

Test connectivity to PIC pins that go to the SID. All good.

Hmmm...

Then after starting at the tracks for a while, I see a faint sliver of silver crossing two very close tracks, the tracks connected to pin 27 (the touch sensor output pin) and pin 24 (the SID chip select line). Yep, they're shorted. (Before anyone asks, my PCBs are not solder masked. Lesson learned.)

It amazes me that the SID played anything at all in this state, and somehow managed to play some patches with no noticable difference, and some with only a minor injection of random noise.