So I rebuilt the drop target assembly a third time with the NMP sensors. First was with L base drops, second with T base drops. Neither of those times was the game playable. The third attempt was T base and some brand spanking new springs. I also cleaned up the metal assembly itself. A previous tech had decided to use hot glue to hold old cracked target guides together along with all the springs to the bracket (including over the screw itself on the back of the original drop targets).
Third time around. The results were better but the game still wasn't playable. This time only the queen would fail to reset properly and just do the continual thunk thunk thunk and a blown coil fuse before I could turn the machine off.
I kept the assembly on the machine and started playing around with all the moving parts trying to figure out why 4 were fine and 1 refused to work and what was different with that queen.
From what I could tell, The 'L' base targets with the larger ledge also means that the curved part of the target guides push it back slightly further before the springs pull it back foward at the very top. I believe from an engineering standpoint, Williams may have also utilized the 'T' base as a means to prevent the top of the targets from tilting so far back as well.
That's when I realized that the coil sleeve for the queen was a short one whereas all the others utilized 2 1/16" sleeves. With the shorter sleeve, the queen had too much play in it and that had it tending to stay tilted backwards from the ledge contacting the target guide.
Last night I did a 4th attempt, this time with brand new 2 1/16" coil sleeves all around. The machines drop targets are now working the best they ever have in my time owning it. No more drops going down that don't register. And in probably 10 or so 2 player games I think we had one single case of a target not latching correctly on the first attempt. I will say this, it really changes how you approach those drop targets. You need to be very deliberate in what target you're aiming at now because even glancing one you didn't want to hit yet is a good chance it'll go down.
My recommendation for anyone looking to put the NMP sensors on their AP. I would recommend brand new springs, brand new 2 1/16" coil sleeves and make sure your target guides have a good edge where they latch at the top to the targets ledge.
oldschoolbob I wonder if you did a rev.2 of the NMP-5 if a modification could be made to give the installer two options.
1) Remove the horeshoes altogether, essentially the same as the current board.
2) Also on the backside of the board, keep a copper track, exactly where there was a copper track on the originals for the horseshoes. I believe one of the horseshoe arms is directly in the way of one of the sensors through hole solder points so it would need to be cut off and no copper track on that side. But if the other arm was left on the horseshoe, and there was a copper track on the board for it to run along to prevent wear on the glass/resin of the board itself.... This would give some additional forward pushing tension, making the targets slightly harder to drop, but not so hard as from factory. (might not be possible though with copper being so close to the sensors, I don't know. Copper isn't ferrous but still might trick it?) If there was even one arm of a horseshoe still in there it would make the target less reliant on the coil sleeves for being pushed forward upon reset.