I'm sure you're just being cheeky, but in case anyone is curious about the full explanation...
I've never seen an issue where a solid state game gets stuck in a bonus countdown, but my first guess was the RAM. I figured that there might be a RAM location where the BonusX is stored and then decremented during the countdown. Maybe that place in RAM was always reading 1 or whatever. Never getting to zero, it would get trapped in a loop.
But I replaced the RAM (6810 chips) and the battery backed-up RAM (5101) but the issue continued. Maybe it was a lack of imagination, but I didn't figure that my ROM would have gone bad so quickly. I burned and verified those chips (Flipper ROM Green1 & Green2 and Game ROM) just last week. It should have occurred to me that one of those had gone bad. I guess that the reason I didn't suspect them was because when a ROM goes bad the results are often very obvious. The game won't run at all because the CPU jumps to nowhere and hangs.
But slochar was right--the problem was with the Game ROM. As he pointed out, the code to run the bonus countdown lives in the Green Flipper ROMs, but the parameters for the countdown come from the Game ROM. Maybe the speed of the countdown, which lamps are used for 1k, 2k, etc., or what sound is played with each 1k rewarded? I'm not sure what other parameters there are.
Just to be sure, I removed each of the 3 ROMs separately and verified them with the programmer. The Green 1 & 2 checked out--they hadn't changed since I burned them. The Game ROM was 1 bit different from what I burned. It's a 2k chip (2716), so there are 16384 bits of information on there, stored as 2048 bytes (8 bits per byte). The byte at location 0x0353 (that's in hexadecimal and equals 851 in decimal) is supposed to be 0x7B (which is 123), but when I read it with my programmer, it had become 0x7A (which is 122).
So the 2716 Game ROM that I just burned had lost a single bit and that was causing the count down issue. Replacing the ROM fixed the problem and I tossed the bad chip. If you expose the 2716s to UV light through that little window, the data on them will all go to 1s. They're only set to 0s if you apply programming voltage (I think it's 28V for these?) in a particular way that burns the little fuses inside. So there was no circumstance other than a failure that could explain this change.
In the end, it was a simple explanation and very straightforward (runtime bug caused by bad firmware), but I just wasn't getting there in my logic. When my ROMs have gone bad in the past, the failure was totally bonkers, you know? In this case, it looked like everything was working perfectly except it just wouldn't finish counting the bonus. So odd.
This Gorgar has proven to be haunted by issues. Every hour that I run it, something else unrelated fails, but I'm working through things.