Helping out a friend with a Stern Flight 2000 pinball, and we've got a problem that has got us at a total loss. The game was playing fine until one evening a coil decided to lock on and smoke while the game was in attract mode (Right top kicker). The game was shut down as soon as we realized the coil was locked on. The next time the machine was turned on, the transistor for that coil (Q17) failed violently and spectacularly as soon as the switch was turned on.

We pulled the driver board and I replaced Q17, C12, CR17, and R42 with brand new parts. (See highlighted section of attached schematic.)
We then replaced the melted coil and verified that the diode was wired correctly.
When the game was turned on, the coil seemed to partially energized, and R42, the 330ohm attached to the base of Q17, burned up in short order. (Red arrow in attached schematic)

Since then, we have replaced the coil again with a known good coil, pulled the driver board, replaced R42, tried again with the same result.
I then completely rebuilt the entire Q17 circuit again, replacing Q17, C12, CR17, and R42 a second time, again with brand new parts.

Same issue. As soon as the game is powered up, the top right kicker partially energizes and R42 burns out after about 10-20 seconds. As long as J5 (playfield connector

R42 is a 330 ohm, 1/4 watt resistor. If R42 was connected directly between +5V and Ground without CR17 or R41, it would still only be drawing 0.076W. That's well withing the 1/4 Watt rating.

At this point I see only two possible sources of the problem:
1. Capacitor C12 is shorted / leaky and applying +43V to R42 or
2. There is AC voltage present at J5 Pin 7 that is crossing C12 and flowing through R42 and burning it out.

The issue with both of those is that I have replaced C12 twice with brand new caps of correct value and voltage rating, and I have verified with an oscilloscope that there is no AC voltage present on J5 Pin 7. (Which, if there were AC there, it should be on the whole solenoid bus and would be burning out all of the 330 ohm resistors on all the driver transistors.)

So unless I got a bad batch of caps, I have no idea what is going on. Any ideas?

Problem (resized).PNG

Yep. We've checked out U4. It's a CA3081 transistor array. It checks out fine. If it were shorted, or even partially shorted, though, it would effectively bypass R42. If it were to be in a permanent non-conductive state, Q17 would never fire. The other thing there is that U4 is still in the 5V section of the circuit, so even if there were something bizarrely wrong with it, we're still back in the territory of 5V directly across R42, which is not enough power to burn it out.

Powering on the board without the playfield connected, the resistor does not heat up. But I have also probed out all the voltages on the playfield and made sure there are no shorts there, or incorrect voltages.

The original diodes are 1kv, .002uf. The caps I'm using are from a Pinball Resource Bally Driver board rebuild kit. 0.0022uf 2kV rating. The resistors have been from both the PBR rebuild kit and from my own stock that I ordered from Mouser.

TIP102 from PBR rebuild kit the first time, second time, TIP102 from Mouser.

With J5 disconnected, the voltages are what I would expect. Don't remember exact readings at the moment, and nowhere near the machine, but I checked them and made sure they were well below the turn-on threshold of Q17. And yes, O/R wire is on non-banded side of diode, double yellow on banded side. First thing I checked. Also replaced diode, and then coil to be certain that there weren't any internal shorts. Last night, I rebuilt that entire section of the driver board, including the U4 CA3081 as I managed to kill the transistor element that drives Q17 by forgetting to put a resistor on my test lead and jumpering
J4P11 directly to +5V. on my workbench.

After rebuilding the circuit last night, I powered up the driver board, set one of my bench supplies to +43V, Connected that supply's ground to J3P23/24 and connected the +43V to a coil then to J5P7 and powered everything up. R42 did not burn out this time, so maybe all is well now. Going to be later this weekend before I can get the board back in the machine and test again.