So I had gotten some semi-repaired boardsets off KLOV last week. I've been wanting to get another Donkey Kong boardset for a while since I plan on getting a cabinet soon. Both boardsets did not work, both having different errors. When hooked up, it booted to nothing. First thing I noticed was that a bunch of chips had been socketed. None of the sockets were in good shape (one was an old nasty machine pin socket, another was a damaged dual wipe, and there were two terrible single wipe sockets), so they needed to be replaced. After doing such, I finally got a picture on screen.

Afterwords, I saw that it wasn't booting, so I was able to make use of my recently fixed Fluke 9010a. I used it to run RAM and ROM tests, all the CPU ROMs and CPU RAM tested good. So that meant that something controlling the video board was goofy. I tried writting to the start of the VRAM (Address 7400) and found I was unable to write the data and read it back.

So I set the fluke to constantly write FF to address 7400 in hopes of trying to figure out where the disconnect was using my logic probe. I found that for some reason, all of the data lines were just floating, occasionally having some activity. This means that the buffer at 6A was not passing the data by. It was getting data in, but it wasn't outputing, implying that it wasn't being enabled, meaning it was stuck in triastate which makes sense for it not to be passing data. After looking at the enable line, it looked as if there was some activity. But after looking at it on my oscilloscope, the line wasn't being pulled low enough in order to enable the buffer. The chip that controls this is a 3 to 8 line decoder (74ls139) at 2A. After replacing it and running the ram check on the VRAM (7400 to 77FF), it passed. From there I ran the game and it booted up.

The game worked just fine. Only problem was the colors of the characters were goofy. They would basically change color half way through being drawn. So this had to be a color selection error since it was being drawn properly otherwise. Basically I landed on 5M, a 2 to 1 multiplexer (74ls257). By this point I had just been measuring the inputs with my oscilloscope just to see if they were there but too low to be recognized, and while checking pin 2 on 5M, it had fixed itself. I had to think about it for a minute as to why this worked. I was looking at it and the ground clip had come off, after the clip came off the signal started acting goofy again. At that point it had all made sense. Inside my scope probe, there is a 22pF capacitor. This acts as a filter to get a better reading on the oscilloscope. By measuring it, I was actually slightly filtering the signal. 

This can be seen below, I had added a 68pF capacitor as that is what I had on hand. Notice in the first image, there is a lot of wobble. If the wobble gets too high, it is basically tristate, which means mux may not interpret the signal properly.

Ultimately a small 68pF capacitor fixed the issue.

No capacitor:

With 68pF capacitor

After this was done, I now have yet another working Donkey Kong boardset. This one was a bit more challenging than the other nintendo boards that I've worked on. Especially with that wonky signal, I found that on accident ultimately. I'm just glad in the end it worked out.

If you have any questions, feel free to ask in the comments, or if you are uncomfortable asking it publicly, shoot me a message and I will be more than happy to attempt to answer it.

Update: chips outputing a noisy signal like that of which I fixed with a small capacitor are an indicator of a failing chip. After taking a look at the schematics again, I found that the noisy signal came from the 74LS245 buffer at 1S, after replacing that, the video board worked perfectly without the capacitor.