So, I've had this Marshall Valvestate amp since sometime in high school, probably around '98 or so. Some time in late 2001 or early 2002, it developed a strange problem. While playing, the sound would sometimes become suddenly softer and fuzzier, and not the good kind of fuzzy. This would happen maybe every 20 minutes or so, sometimes more. The cure was always a mild kick or tap, but it would come back again. The only thing I could think of to do was replace its lone tube, a 12AX7, but that did not help. At some point someone suggested to me that it could be a worn out capacitor, and that launched a whole investigation into multimeters, which culminated in buying a pretty nice one and then reaching new levels of laziness by refusing to even open up the battery compartment (it required a screwdriver, you see) for many many months.

Recently I decided to renew my efforts, as I'm hoping to start putting on musical performances soon and will need as many amps as I can get. My current technical advisor (my boss) was very skeptical about the burned out capacitor theory, and I hoped he was right, because it turns out one cannot test for a dried out capacitor merely by measuring its capacitance. Instead one must measure something called Effective Series Resistance, which is actually even more complicated than it sounds. One measures it by buying a tool which costs from approximately $100-$200. One can also, according to one web page, roughly replicate what the expensive meters do by pumping a square wave through the cap and examining the output for slopes that are too steep or shallow in the cycle of charging and discharging. This did not sound like something I would be particularly fit to carry out, and the number of capacitors in the amp's circuitry was also a bit daunting.

My technical advisor suggested it was probably just the same thing that is the problem 99% of the time, a connection. This was also slightly intimidating, because there are a lot more connections in there than capacitors, and I had already jiggled all the ones that seemed most in danger of coming loose. A picture for reference:

But I went back to it, and jiggled some more. And sure enough, there it was.

When I pushed this ceramic resistor forward, bad. When I pulled it back, good. Why? Because one of its two contacts (impossible to photograph well enough to see) had no solder at all on it. Nice job, Marshall Quality Control.