Its 2 milliseconds per vertical grid line, so about 8.4 ms per pulse which is 120 Hz. Which would go along with some kind of full wave rectification of the 60 Hz AC input. As far as the magnitude of the available current
behind the spike, I would say it is very large given the small impact that large capacitor has on it. I'm sure there is some calculation that would give a numerical answer, but I don't have it.
We've already spent $500 on getting the thing repaired. Its unfortunate that I didn't do this before it failed so I could insist the repair facility ensure that ripple was within specifications. At this point I'm not sure its worth taking it out and shipping
it back for what would probably be an expensive evaluation by the shop, and possibly a BS line about how it was all normal. I think the ultimate solution is a more modern unit when it fails again.