Sorry to hear of your problems. I've done a number of M-1.0t mk II opt 2 upgrades. I found a lot of ways to screw them up, but once working properly, they all produced rated power until the dummy load resistors started to melt. And if the thump problem is bad enough to scare you and your dog, consider it a symptom of some other problem, not your normal annoying (but fixable) M-1.0t thump. Here are a few suggestions.
1. If its running out of gas at high power levels, there is likely a problem in the commutator circuits. Note that the high rail commutators are triggered by a single set of components. So a failure there will cause both channels to loose the benefits of the high rail commutator. Suggest you put a scope on emitter of Q139. (I'll use examples from the positive leg of the left channel, but you might want to look at both legs of both channels.). At idle the emitter should be at the middle rail voltage. As you apply an input signal you should see some gradual switching to the high rail. Use two channels of your scope. Trigger off the output of the amp. Use the second channel to watch the emitter. No need for a load at this point, the commutators will work without it. Its also educational to look at the collector of Q127. It should sit at the low rail voltage, rising as the signal requirements dictate.
2. Check switching diodes. Shorts/open/backward diodes cause all sorts of confusing symptoms. D125 thru D132.
3. You didn't mention if the loss of power was in one or both channels. That's an important clue. Loss of power in both channels is a strong indicator of a problem in the high rail switching since the control circuits are shared between channels.
4. Horrific thumps. Put a scope on the positive and negative side of each rail of the power supply. One at a time is OK. Turn the amp on and back off. Observe the decay time. If one drops like a rock while the other just drift down, it's a strong indicator of a short somewhere. Asymmetric decay of the supply voltages is the primary cause of thumps. Look at the 12v supply while you are at it. They too should drift down together.
Note that the high rail switching transistors (Q139 - Q142) used in the mk II upgrade have been known to be counterfeited. They fail at the slightest provocation. Also, you must use the TIP35C/TIP36C. Not the TIP35/TIP36.
Based on my experience the use of a mag coil instead of a real transformer is not the root of the problem.
Please don't be offended by my suggestions that upgrade errors may be the source of the problems. I've done a lot of mk II upgrades and about the time I felt I could do no wrong was when I committed the most bone headed errors.
Oh, and I think you are in Southern Arizona somewhere. I'm in Scottsdale. If these suggestions don't work out, I would be happy to take a look at the amp. Either ship it to me or stop by when you find yourself in Scottsdale.