Recently, I was lucky enough to receive a big haul of retro computer gear from a friend who was emptying out his garage. Even better, the haul was almost entirely old Amiga gear — my favorite computing platform of all time. Upon returning home, I gleefully sorted through the boxes, powering things up one by one. Amazingly, everything worked… except for one lonely Amiga 500+. I was greeted by a dull grey screen. This wouldn’t do, so naturally, I got to work.
It seemed like a shame to be opening the machine, as after almost 30 years of life, this one still had its warranty seal intact. Regardless, nothing ventured, nothing gained – the Torx bits were at hand and the screws were coming out.
Inside, I found much what I’d expected. The electrolytic capacitors were all clean and holding up well, but there was significant corrosion caused by the clock battery. It was thankfully localised to a small cluster of traces near the Gary chip, which provides glue logic and floppy drive functionality. Looking closer, I also noticed that Gary himself was looking rather corroded. Commodore were kind enough to socket the custom chips on all Amiga boards, so I popped it out – only to find that several contacts in the socket had broken or corroded away.
The job was looking like a frustrating, yet achievable one. I cut the battery off, then began patching the traces. One by one, I traced out where they were supposed to go, then scraped back a small amount of soldermask in order to have copper to solder to. Then, a thin patch wire was tacked in place to restore the connection, before verification with the multimeter.
Fixing the Gary situation was going to be more difficult. The entire socket would have to be removed. I cut the socket away pin by pin, so I could deal with each one individually. Patience is the name of the game here, as being too rough in removal can lead to damaged vias or traces. I used a spring-loaded solder sucker to remove as much of the solder as possible from each pin, and then heated the via once more while pushing the remains of the pin out with a toothpick. I found this technique to be relatively successful, far more so than former attempts using metal wire as a pusher. The hardest pins are always those with connection to the ground plane – there are no thermal reliefs, so the solder sucker doesn’t work particularly well there. With perseverance, I was able to remove all the old pins, and solder in a fresh socket for Gary. At this point, I felt accomplished. I’d solved the obvious problems on the board, and thus would clearly be rewarded with a working machine.
Alas, it is rarely so simple. Upon powering up the machine, the screen was wavy, lacking any video sync, except to occasionally flash up a solid colour now and then. Blue, red, green – these are all various error codes the Amiga can display. It seemed unlikely I was suffering all of them, but I was unsure what to do next.
After some further research, I’d heard of another Amiga fan who had issues with the extended RAM present on the Amiga 500+. Using jumper settings, they were able to disable the extra RAM and get the machine to boot. As this seemed easy and didn’t involve serious work on my part, I gave it a try – to no avail. I was rather petulant that my blind stab in the dark didn’t work, and set the machine aside for a week while I brooded over my next move.
A friend then suggested, given that I had a working Amiga 500, to swap over the custom chips in order to identify if any of them in particular are faulty. After some careful work with a flat-bladed screwdriver, and shelling out $15 for a tool to remove the Agnus PLCC from its socket, I eventually figured out that all the custom chips from the 500+ were fine, barring Agnus, which had a version incompatibility between the two machines.
At this point, I was at a loss. I was losing faith in the machine, and was reticent to spend hours blindly probing around for more broken traces when the board looked, for all intents and purposes, fine. It was then that a gift miraculously appeared before me – the Amiga 500 Troubleshooter, courtesy of fellow countryman Ian Perry. It’s a table which shows which areas of the machine should be investigated for given symptoms. I noticed that my wavy screen without sync was likely caused by either a faulty Agnus chip or a failed 74HCT245 in the video output circuit (U41).
I had a hunch that Agnus was fine, so decided to spring for a few 74HCT245s at 90 cents each, plus ten dollars shipping. U41 was not socketed on the motherboard, so it was a case of carefully cutting away the chip pin by pin, before again pushing the remains out with toothpick and soldering iron. As is the custom, I installed a socket before putting in the fresh chip, just in case.
With bated breath, I powered the machine up, connected to the classic Commodore 1084S monitor. The screen was clear and synchronized, but grey… before turning a shade of blue. It continued to do this in a regular cycle. While the machine was still non-functional, my confidence was buoyed. The Amiga was now behaving consistently and showing just one error message – the blue screen representing an issue with one of the custom chips – Agnus, Gary, or Denise.
It was at this point that my luck came home. While sitting there, quietly proud that I was getting closer to a solution, I keyed in “Amiga 500 Plus blue screen” into the search bar. The first result was an eight-minute video that I wasn’t quite in the mood for, so I decided to check the description. It mentioned offhand a broken trace that was hard to spot with the naked eye, linking pin 4 of Gary to pin 1 of U12. A quick check with the multimeter showed no connection on both my working and broken machines, but my interest was piqued, so I decided to look up a schematic.
Finding the service manual on archive.org, it was clear that pin 4 of Gary goes to resistor R113, which then goes to pin 1 of U12, a 74LS244. Checking my machines again, the working Amiga 500 had the connection, while the 500+ did not. Suddenly I found myself rather excited, and hurriedly soldered in a patch wire from Gary to the resistor in question.
The tension was palpable as I plugged in the machine, and flipped the power supply on. There was a grey screen, followed by a new sound – a few clicks from the floppy drive, before the glorious sight of the 2.0 ROM Insert Disk screen sprang forth. I had finally done it!
I was rather ecstatic, for having attempted to fix many retro machines over the years, this was the first one that had been at once both completely successful and without causing any major collateral damage to the machine in the process. Plus, I now had a hotrod Amiga 500+ as a part of my small collection.
The process could certainly have been quicker if I wasn’t so reticent to do the hard yards and check more connections on the PCB, but the biggest help was definitely the troubleshooting matrix. It’s an incredibly useful tool that boils down years of experience into a quick reference guide.
Overall, the repair was a win, and I’m looking forward to firing up some Lemmings before hooking up a couple machines over serial for a multiplayer game of Stunt Car Racer. Most enjoyable!