So I have a CV on route, but honestly this has been an issue even when it was in my collection.
I've had this game 14 years but this issue started within this year.

The neon quits working about once a month and I pop in a new ULN2803 and all is fine for a while. This transistor array controls a few other things, notably the RM eddy sensor.

Is there anything anyone can think of that would be killing one array in the chip? The board is socketed now and the chips are cheap, but just trying to not have to replace them any longer. Is there maybe a beefier version?

That makes sense. I'll get one and replace it and see if that helps. Thanks.

WPC95 at U3 on the driver board that controls the neon, still going through ULN2803A chips in the order of one a month. Just replaced another today. This time I replaced the entire neon ramp with a different one so it will have a different transformer and neon tube. The chip is from a different manufacturer this time.

If this one goes bad, the only other thing I see in the circuit are some 470 ohm resistors where the Vcc goes to each transistor array. Could one of those be bad? Other than that I've got nothing.

Nothing in the circuit other than 5V through the resistor to the transistor to switch the transistor. Then 12V to the neon transformer and then the transistor grounds it to turn it on.

Odd that the ringmaster eddy sensor is also controlled by another transistor array in that same chip and that never fails, just the pair that controls the neon.