Velodyne HGS-12 Shaking the House on Power-up

The electric power in my neighborhood was out for a couple days recently, and when it came back on, my HGS-12 subwoofer started really shaking the house for many 10s of seconds then stopped. Thanks to the schematics posted on this board, I was able to track down the problem and fix it. First thing I did was open up the case, disconnect the speaker and connect a dummy load. I didn't have an 8-ohm load big enough for 1200 Watts. Instead, I used a 120V 1500W halogen work light. This light had a cold filament resistance of about 6 ohms (close enough), and had the added benefit that it lighted up if there was output signal. Before going on, a word of caution if you work on one of these speakers. The speaker output is not isolated from the power mains. Also, the power amplifier uses +/- 160V DC (US model -- 0-320V EU model) and it has multiple internal grounds (one at -160V; one at zero). And when you unplug the power, the high voltage capacitors in the amplifier remain at dangerous levels for several minutes. You really have to be careful when working on this amp! The problem.... On the control board (also used in the HGS-15/18 models), the circuitry around the NE572 compander is unusual. Normally, this chip has a 2.5V bias on the output. In the application note for this chip, the example uses an op-amp differential amplifier to shift the output to 0 volts. In this circuit, however, it's done quite differently. Instead of letting the output have a 2.5V DC bias, they use an op-amp to drive the GND pin of the NE572 to -2.5V in order to bring the output level to zero. This pin wasn't going to -2.5 promptly as it should. Instead, it started at -8V, slowly ramped up to -6V, started oscillating erratically between -8.5 and -2.5, then, after about 30 seconds, finally settled to -2.5 and stayed there. This behavior was repeatable -- just turn off the power for about 10 minutes and turn it back on. One thing a little questionable about this circuit is that the output of the op amp (U3-C) is connected directly to a 4.7uF capacitor to ground. According to the datasheet, the LM324 can withstand shorts to ground, but they warn not to short the output to the power rails. The capacitor is in-between these cases (initially looks like a short to ground, but later, to some voltage) so not clear. But arguably, this configuration may stress the output. Whether or not it's a design flaw, the problem was the LM324. I carefully protected the neighboring components with aluminum foil, used a hot air rework station to remove the old part, cleaned the pads, applied solder paste, stuck the new part and used hot air to solder it. This fix worked -- amp is back to normal now. Hopefully, my experience will help someone else out.