I thought I would add my two cents to this post.
I am the customer Chris is talking about. I fought with this exact issue on a customers LOTR for a solid week and began to feel pretty helpless even though I am what would be considered a pretty competent tech. I changed the coil, rebuilt the mech, rebuilt the connector at the board, replaced the mosfet and the 74HCT273 on the power driver, swaped out the CPU, burned a new CPU ROM, swaped the PAL, replaced the RAM, swaped the driver board and painstakingly checked continuity on all chip circuits as well as check for correct activity on all chips with a logic probe. Not all necessarily in that exact order, but that is the list non the less. My tech friend (with 50 years experience) and I swaped the CPU board into an Elvis that I had and it did the exact same thing with the coil that was associated with that driver on the Elvis (Q3 operates the left VUK on the LOTR and the left Drop Target bank reset on Elvis) in both machines it wouldn't fire at full power in game mode but worked great in test (the activation pulse is longer in test, which results in a stronger coil pulse). The only caveat to that test though, is that the coil on the Elvis drop targets is significantly larger than that on the LOTR VUK, so it is not overly surprising that it wouldn't activate it correctly. We gave the CPU a full prostate exam and decided to call Chris as a lifeline because if he hasn't seen it, it probably doesn't exist. Unfortunately, however, he had not seen it before and we had already touched on every possible solution Chris could think of. As a result, I finally decided to put a little more powerful coil in, so I replaced the 26-1200 coil specified in the manual with a 25-1000 and it works like a charm. I do not like solving this problem this way, but up sizing the coil should not cause a problem since the same Mosfets are used to drive much larger coils in the game without issue. I tried a few different coil sizes, testing to find the balance between the VUK being able to consistently kick the ball up into
the wireform and not beating it to death and breaking the mounting tab welds, before I landed on the 25-1000. The best idea any of us have come up with at this point is that a combination of slightly worn mechanical parts, combined with a little higher resistance from aging wire in the harness, has resulted in enough of a difference for the coded activation pulse for the VUK coil to no longer be sufficient to fire it in game play. I hope this helps someone out there. Thank you to Al Williams and Chris Hibler for all the effort they put in in this with me.