On the CPU Board.

The main drive transistor is Q41 Tip 102, Q33 is the predriver 2n4401. I would replace those two items along with the Diode 1N4007 on that coil.

I love to try learning things reading these posts. If swapping in a different PPB board eliminates the problem, why would a fix on the CPU board be needed? I ask because problem vanishing after swapping PPB boards would have made me think (perhaps erroneously) that the problem must be on the PPB.

You may

Just reread and saw you did this. I missed it initially.

Have you looked over every inch of the bad board? Look for solder bridges, a stray piece of metal shorting something it shouldn’t, a broken fuse holder with one side flipped towards the one next to it shorting out.....

How did you check the diodes and transistors? If you have a coil locking on I’d suspect a bad TIP36. Not sure which one on a LW3 but I’m finishing up two of these boards for other data east games. Each had bad TIP36 devices. Most of the time a diode check on the meter can help weed out bad parts. It’s not perfect though. Any parts that I even suspect as being bad I’ll pull and use a transistor checker to compare against a known good part. You can get some decent ones on eBay for under $30.

I follow the rule that a (BJT) transistor that tests bad is bad but one that tests good is not necessarily good. I have a TIP102 that passes the diode test on a DMM but when used in a board is shorted from collector to emitter.

If you've checked everything else and can't find a measured failure then I would remove the TIP36C and replace it. You can also get incorrect readings from a component that is in-circuit. Ideally you should test the component when isolated. That board is a single sided board so there's no danger of pulling through holes as there are none.

Based on what I read, I'm stumped too. Long shot here, but your headers look new. Any chance there's a solder bridge underneath from some drip through...or does this problem pre-date those? My brain keeps heading back to a solder bridge. Have you checked continuity across those pins?

Left back drops are Q41 as chad mentioned. Q40 is on the right bank drops.

So why had you replaced Q40?

Does R74 to R76 measure close to the same resistance?

You made sure you did not have a solder bridge between any of the transistor legs after soldering?

A little late to the party... but for some reason DE was real loosey-goosey with their pad spacing. I've had several of their power supply boards eat themselves because of high-potential lines being a bit too close to other lines. Once the board scorches a little, it starts to conduct a little... then it burns more and starts conducting a little better. I thought it was just their power supplies... but I've recently had a PPB where J2 had 2 pads were arcing and kept taking out Q5. Had to get a little blade and scrape away the carbonization. There were no "real" connection between the 2 pads, but when I ohm'd it out it was @1.5K. After a scraped the carbon away, it went back to open. Worth checking out if this is still an issue...