Sometimes electronics stop working when you need them most.  This was the case this weekend, when Adam ran into me at sprout.  Adam plays flute in a hip hop band for DJ Afro DZ ak, and had a gig that night, but his flutes mic/preamp combo had stopped working!  Fortunately, I had a few hours to kill and a room full of test equipment.

Fx-1 mic that like the one attached to the amp

I started by checking to see if the microphone was producing some kind of output.  I plugged in the power and looked at the output on the oscilloscope.  Strangely, it seemed to work.  I decided that this meant that something on the preamp board was borked.  The easiest way to test if the board was working was to give it a known input with a function generator and look at the output it produced.  Again, after some fiddling it seemed to work, leading me to the conclusion that the whole system was working and that adam had a problem downstream of the preamp.  I picked it up to put it back in the case, but I wanted to show Adam that it worked.  So, in mid-air I tried to connect the probes.  Then it stopped working.  Shaking it seemed to make it work intermittently.  It was then when I noticed the long leads snaking down to one of those 9V battery clips.

One of these buggers

Immediately, I was suspicious of the battery clip.  After testing it with a multimeter for continuity, I determined that the ground wire had been worn so much over the TEN YEARS that Adam had been using it, that the stranded wire inside had frayed and broken.  Occasionally it would work, because the wires would make a connection.  Adam further confirmed my suspicions that this was the problem when he mentioned that the amp would connect and disconnect sometimes, making loud popping noises on the speakers.

Once I knew what was wrong, it was easy to solder another clip into place.  Adam rode off into the the sunset (or over to Union Square), and played his gig.  Apparently, the preamp and microphone worked fine once they had power.

This is an interesting case of debugging because the “divide and conquer” approach that I like to use didn’t work.  The initial division was between the board and the microphone, but it was a poor division because the microphone actually depended on power supplied by the board.  I guess I missed an even simpler debugging step: make sure the power is on.