I wouldn't worry about the cleaning.

Power supply failures for the DMD are so common that I have a whole box of components to repair this problem.

If you don't want to go through the component level repair, I like this solution from Pinball Life:

https://www.pinballlife.com/williamsbally-dmd-high-power-replacement-board.html

If you DO want to go through the component repair, there used to be a few places you could get a kit of the parts you need, but all I'm finding now is:

ebay.com link: itm

Which has more than two weeks potential shipping delay.

If you are inexperienced in this kind of repair, put in all the parts. Don't try to replace parts only in the one section where you find a problem, this is a common mistake. You replace a part or six that you find are bad, and then you plug it in and blow up the parts you just put in. Then you'll need another kit to do the repair! It's a whole lot easier to just put in all the parts that come with the kit.

My guess is a bad bridge BR1 on that board even though you said it tested fine.

I had a Whitewater that had a snowy display, so I went to Marco and ordered the kit for 17 bucks. I read on their site it takes about 20 minutes to replace the parts. Must be a typo, they missed adding a 1 in front of the 20. So I replace the parts and each component was neatly labeled in a tiny bag. Last component to go in was MJE15031 (Q6) as the bag was labeled. Out of habit I looked at the actual transistor. It was a MJE15030. Luckily, I had bought 2 kits because we have many WPC's as a back up. In that kit, the same wrong transistor in the correct bag. So I alerted Marco of the wrong part in right bag and they said they would check their inventory.
Once the correct transistor came in, I installed, fired it up and the snow effect was gone. One thing I did do was to really get the power resistors away from the board and use heat shrink on the wire ends to prevent shorting.

^^^ THIS ^^^

The heat damage to the board substrate is from the 4.7k Ohm 5W "cement" resistor.

See image 7 in the post https://pinside.com/pinball/forum/topic/thermal-regulation-ignorance-is-bliss-or-show-me-the-heat for the temperature of this sucker. 118.9C (Celsius) = 246F (Fahrenheit).

A resistor that hot is going to dissipate heat in every direction. If the resistor is up against the board it will dissipate its heat directly into the board substrate. Eventually this will weaken the substrate. Worse still given that it is SO hot ... it will dissipate its heat to other nearby components as well.

Install (at the very least) that resistor elevated (raised) off the board. How much you elevate will dictate how much heat is dissipated into the surrounding air rather than the board or other components. Of course, you don't raise it "crazy high" off the board. There's no right or wrong with the amount. You want at least some amount for convective cooling into the surrounding air. Some of these axial form resistors have what could be considered standoffs to raise the body of the resistor off the board but there may still be a point of contact with the board. I install these resistors with no board contact at all. I want all the heat dissipated into the surrounding air as much as possible.

You can eliminate all this heat generation by using a low voltage display - either ColorDMD, PIN2DMD or an LED DMD display. If you opt to go down this route then you can simply remove the two fuses to disable the high voltage regulation and thereby remove all heat generation and dissipation. I am very slowly working through my machines to remove all the plasmas and replace with low voltage monochrome DMDs that I build myself.

Yeah it's going to take way longer than 20 minutes. I've rebuilt probably 20 of those boards over the years and it takes me an hour to desolder, populate and then check continuity once I'm done.

Those kits are to rebuild the high voltage components that tend to get burned up. Not only should the cement resistors be raised up but also the diodes...especially the zener diode. Really the biggest problem is damage to the traces either from the heat or from removing components. Inevitably you're going to need some jumpers on the back side of the board to reconnect things back up.

The two large blue capacitors do not need changed out unless you're having a vertical scrolling display. I don't think I've ever had to change those on a WPC driver board yet.

I raised mine about 1/2 inch. But again I insulted (spellcheck) then I insulated the leads with heat shrink.

Sounds cool, how many 20 minutes' does it take to do that?
Could be a viable product, no?

Agree with this. I raise the 4742/4758/4759 zener diodes off the board. There is another 4742 in the 12V section that you don't necessarily need to raise off the board.

Don't forget the 47k Ohm 1/2W "bleed" resistors. Those can generate some heat well.

Agree with this too. These rarely fail. The other issue with this is that the specification is 150uF @ 160V axial but these are nearly impossible to find. A good substitute is 220uF @ 160V that Ed sells but unfortunately as most people know Ed's store is not open.

https://pinside.com/pinball/forum/topic/dumbass-test-and-reproduction-pcbs/page/12#post-6578684

It takes about 12x 20 minute sessions I would say. It's surface mount as well so it's not a lot of fun unless you like playing around with tiny components that you can easily flick off the table by accident.