Repair Log
I started trying to repair boards in November of 2001. I've done a bunch of really stupid stuff, but I've also learned quite a bit. This page is intended to document some of the fixes I've made, and the excruciatingly slow learning process I've undergone.
-
Game: Pac-Man
Status: FIXED - 7/24/99
This repair is documented in more detail here. Basically, I had some intermittent garbage on the screen (lines), and I got a tip from RGVAC where to look. There was a 74LS174 at 1-F which had a couple of lines which would occasionally get stuck high. Replacing this fixed the board.
-
Game: Time Pilot 84
Status: FIXED - 1/8/02
I got this board off eBay as an "Unknown Konami PCB", which I knew from the description and picture was Time Pilot 84. Got it in the mail, plugged it in, and got a garbaged video signal and a solid tone on the speaker. I noticed that 3 of the socketed chips on the main board were inserted backwards compared to the soldered chips. To make it harder, the manufacturer had ground off the identification numbers from the two CPU chips, so I couldn't tell what they were, although I suspected they were Z80s (turns out that there are two 6809s and one Z80).
To make a long story short, I verified that the CPU chips were indeed inserted correctly (they were getting a clock signal on the proper pin). But when I compared my board with a picture of another PCB on eBay, I noticed that the socketed eprom was installed upside down. I turned the chip around, and fired up the board, and it worked perfectly. So, to summarize, Time Pilot 84 has 2 socketed CPU chips that should be inserted "backwards", but none of the other socketed chips should be.
-
Game: Gauntlet 2-Player
Status: FIXED - 5/01/02
This board really has me frustrated - it's a prime example of how "repairing" a board can lead to more problems than it solves, if you aren't careful.
My Gauntlet board had a cold solder connection which would occasionally cause the "grunts" to be drawn grey (missing colors). By flexing the board, I found that the problem was with the EPROM at 1A, and reflowing the solder on the pins fixed the problem.
However, I also noticed that I couldn't coin up the game. Additionally, running the game in test mode yielded a "Sound CPU Error". I tried reseating all of the socketed chips, and ended up with some garbage on the display. By wiggling the RAM chips, I found a couple of bad sockets, which were totally corroded inside. Replacing both of these sockets fixed the garbage on the display, but my coin-up problem remained.
I traced the coin-up signal into the board, and found that it was never getting latched onto the bus. Tracing this even further, I found that the 6502 (sound CPU) wasn't running (constantly resetting), and the 6502 is what's responsible for polling for coin insertions. I pulled the 6502 and found a bunch of the same corrosion in the socket that I found on the RAM chips.
At this point, I got stupid. I had some bus wire, and started poking it into the various holes in the 6502 socket to try to clean out the corrosion. Of course, I wasn't careful and stupidly left the power on. To make a short story even shorter, I accidentally touched the wire to the 5V pole, and the game went from "mostly working" to "mostly not working". The game just sat and reset, over and over and over. Argh!
I checked all of the eproms, and found a corrupted EPROM at 16S (which may have been causing my sound CPU problems) and a totally dead EPROM at 9A. I burned new EPROMs for both, and now the board works a little bit more - the screen clears for a second before the CPU resets, and if I run the game in test mode, the screen never clears, but the board also never resets (so the code must at least be tagging the watchdog location). I also found that one of the IMS1420s was getting really hot to the touch, so I pulled it, and I'm waiting for a replacement from multigame.com.
1/12/02: I'm basically stuck now - I don't really know what more to do. I'm going to try to trace through some of the bus lines just to make sure none of them are stuck (I'm sure they aren't), because there aren't any stuck lines at the CPU. I'm hoping that maybe the missing RAM chip is the problem. Barring that, I'll just start going through the board and looking for stuck lines, I guess.
1/13/02: After playing with the board tonight, I tried swapping some of the RAM chips around, and voila, I got a test mode "RAM Error". It turns out that the RAM chip at 7E was the one that was burnt out, and test mode crashes if the RAM at 6E or 7E is missing. Swapping in working RAM chips into those locations allowed the test code to continue long enough to white out the screen and display a RAM error for the missing chip. I have some hopes that the board will work fully once my replacement RAM arrives.
2/5/02: My replacement RAM arrived, and after plugging it in the game comes up to the attract mode, and the test mode works, except I get a "Sound CPU Not Responding" message. A few of the lines on the 6502 are floating, so it's pretty clearly just a bad socket. I'm not looking forward to desoldering a 40-pin socket though...
3/01/02: I removed the old 40-pin socket, but broke one of the traces on the board in the process (scratched a chunk off with the soldering iron). No big deal - I wired up a replacement trace using some wire-wrapping wire on the bottom side of the board, and soldered it to the appropriate pin of the replacement socket. Fired it up, and it worked great. All the self-tests passed, I can coin up the game, and I could even hear a little bit of sound coming out of my speaker (Gauntlet requires amplified sound, and I don't have an amp on my test bench). Of course, once I play-tested it for a bit, I quickly realized that I hadn't really fixed the original problem with the board (cold solder joint at eproms 1A/2A). I went through and reflowed fresh solder on every pin of both eproms, and it seems to be working fine now. I have one final repair to make (the video connector is a little dodgy, and colors keep going out unless I wiggle it just right - I've already resoldered the pins on the board, so it's probably something wrong with the connector itself). Once that's fixed, I'll start work on hacking it to run both Gauntlet I & II!
5/01/02:Redid the video connector, and the problems disappeared - I finally had a working game again! Also burnt a new set of double-sized eproms to allow me to run Gauntlet and Gauntlet II on the board, with the high address line wired to a switch so I could swap between them easily, following the directions at this page. Works like a charm!
-
Game: Venture
Status: FIXED! - 2/5/02
I got this board in from eBay, and started making a JAMMA adapter for it. It generates positive sync, but there's a trace you can cut to generate negative sync instead. Likewise, you can cut a trace to invert the video signal if that's necessary (I have yet to determine this, as the board isn't running fully yet).
I wired up the board, and it seems to be working because I'm getting some video (an occasional "zoom" effect from the attract mode screen, when Winky enters a room in the dungeon). I noticed that the socket on the audio board that the ribbon cable plugs into is faulty (one or more of the pins aren't making good contact, so I need to replace that). Additionally, one of the 10-pin interboard-connectors (P3, I think) is broken, and is missing a bunch of the little metal leaves, so I need to replace that as well. Bob Roberts has these, but they are only 8-pin (not 10), but I'll make it work somehow. I've got a new socket and the 8-pin connector on order now.
1/31/02: I pulled the metal tabs out of one of the 8-pin connectors Bob Roberts sent me, and inserted them into the 10-pin connector (they were kind of like Molex connector tabs) and resoldered it onto the board. Now I get full video, although inverted because I had cut the trace to invert the video signal (so, the lesson is, don't cut the video invert trace, as the video signal already has the proper polarity). The board initially was giving me a "Check Audio Board: 80 80" error, denoting an error sending data from the audio board back to the logic board, but this problem went away when I started trying to debug it. I only need to wire up the controls now.
2/5/02: I finished wiring up the controls and the rest of the ground/power lines, and it all seems to work now.
-
Game: Mega Force
Status: FIXED! - 1/17/02
This was another board that I did more damage to than I intended. Occasionally parts of the background would drop out, to be replaced by solid blue, which I'd figured out was due to a loose eprom or bad solder connection. While re-seating the EPROMs, the solid blue background problem went away, to be replaced by a problem where the front screen logo was garbled (instead of saying "Mega Force", it said "MegMegorce". I figured this was some kind of a problem with a stuck address line, and started playing with the lines (dragging them to ground and 5V) to see if I could isolate the problem. Of course, I touched 5V to a /CE line that was tied to ground, and smoked a trace. No real damage seemed to be done, as the game came up and was playable, but all the fonts on the screen were garbled (the characters were wrong).
Playing around with the logic probe, I found that one of the address lines was stuck high for a set of the EPROMs, and tracing this further along I found a 74LS273 with a stuck pin. I've ordered a replacement, and I'm hoping that everything will work just fine once I've swapped it in. I suspect I'll still have the "MegMegorce" problem on the title screen, though.
1/17/02: Replaced the 74LS273 at 3J, and the board works fully again! Even the "MegMegorce" problem went away.
-
Game: Neo Geo 1-slot (MV-1F)
Status: FIXED! - 8/6/02
I saw this board listed on eBay as having graphics problems - I love auctions like this, because they scare off the majority of the bidders. The seller claimed there was a problem with the video ground, and showed a screen shot where you could see the game playing, but the graphics were not centered on the screen (they were wrapping around the left/right/top/bottom edges of the monitor). He claimed he had damaged it when trying to hook it up to a Playchoice monitor.
When the board arrived, sure enough, the video ground lead was missing (assumedly it had burned up). I had some leftover conductive paint from overclocking my AMD CPU, so I fixed that quickly enough and fired up the board. The video display was totally garbled. My logic probe showed I was getting a video signal on the RGB lines, and when I pressed the 1-player button I could see the image change on the monitor as it cycled through the various test screens, so the board itself seemed to be working, but I wasn't getting any video sync. After a couple of days of unsuccessfully searching for schematics on the net, I just decided to try to trace the sync signal manually to see if I could figure out if there was a faulty component. I tracked it back to a 100 ohm resistor, which had a pulsing signal on one side of it, but no signal on the other. I measured the resistor in-circuit, and it read 700K ohms. Replacing this resistor fixed the problem, and the board worked perfectly.
-
Game: Mario Bros
Status: FIXED! - 1/23/04
I bought this board on eBay "as-is/untested". Brought it home, wired up a JAMMA adapter for it, and it worked great on my test harness. But when I installed it into a cabinet, I started seeing some graphics glitches (the fireballs were drawing in two places at once, and the pipe at the lower right corner of the screen was drawing near the top of the screen).
After messing around with the board for a while, I noticed that it worked fine when it was first powered on, then slowly the problem got worse, which pointed to some sort of an overheating issue. A couple of cans of freeze spray from Radio Shack let me isolate the problem down to one of the RAM chips on the B board - a 2148 RAM chip at "B-2". As a bonus, this chip was already socketed, so after I purchased a replacment (Thanks Bob Roberts!) the game was up and working again.