The first repairs I attempted were in the speaker system, which most obviously needed repairs. Two amplifiers didn't work at all. One of them I didn't even have to plug in to know that it had failed -- it had obviously caught fire. There's a photo at right that should show some of the scorch marks. I decided not to try repairing the amplifier as it was: because I was at my parents house I couldn't get at my inventory of parts and I didn't have time to order replacement parts piecemeal. Besides, I could see that this design was prone to failure and I didn't know if its performance was all that great in the first place. So, instead I replaced the entire amplifier circuit. The result is shown at right. I used two LM3886 chip amps wired in parallel as discussed in this application note. (I've wanted to try using these chip amps for a long time.) I'm not happy with the length of the leads to the chips, but it was unavoidable given the arrangement of the heat sinks. The result sounds just fine though, which suprised me a bit, since I didn't have anything more sophisticated than a multimeter for troubleshooting the new design. I also later added an adjustable preamp to increase the overall gain.
I'm not entirely happy with the resulting amplifier, but it was the best I could do in the time I had and with the resources I had available. I don't know if the amplifier will be stable all the way up to full power; I must admit I didn't try testing it. The speaker connected to this amplifier gives off a loud "THUD" when power is applied, which is caused by the amplifier charging up the output capacitor. My preference would be to build a new amplifier entirely with a bipolar power supply, but I didn't get the chance.
The other failed amplifier was a newer model, a Rodgers S101-D (instead of a S100-C). This one didn't have any visible damage, and in fact when turned on it produced some faint sound. It turns out only the output capacitor was bad. (All of these amplifiers use a single-ended power supply, so the amplifier output must go through a capacitor to block the DC offset (Vcc/2) between the output and ground.) Replacing the capacitor was easy enough. Interestingly, the only replacement I could find that had the same diameter (so that it would fit in the same spot in the amp) and capacitance also had a higher voltage rating and was less than half the height of the original. So I guess there's been some advances in capacitor technology in the past thirty years.
The speakers themselves also required some attention. I found some loose connections to the speakers in the M13 cabinets. When I checked the big woofer (the one with two speakers in it) I found that it's two speaker cones had been wired in parallel, giving it an impedance of only two ohms. That's lower than the amplifiers are designed to handle, so that could have caused one of the amplifier blowouts. I rewired the speakers in series. I also found crossover networks inside both woofers that weren't actually wired to anything. They weren't necessary, since each speaker has its own amplifier. (Still, it's odd.)
I used the crossover components to connect the odd horn to one of the amplifiers. I used the amplifier connected to the output channel that includes the Krummhorn stop, since that's what the organist guessed the horn was meant for. The horn doesn't sound very exciting though; I think to get any real volume it will need its own amplifier and output channel for just the Krummhorn.