I still remember when my Dad built a new PSU out of a server PSU for my Amiga. Now that was a proper power supply made to run 24/7 - 365 days of the year.. for at least 10 years.
Amazing video as always. I recommend you should use a small piece of shrinking tube when testing the connector pins. Just put it in the pin and stick the multimeter probe inside.
Daym. I did this for myself maybe 10 years ago, but never thought to document / open source it. The 7805s can fail short and do run hot under normal load, also the caps dry out leading to high ripple. The LM2576 is perfect for this application. Great work opening this up for others to use. I need to work on my internet work.
never yet come across a 7805 short through, go 'noisy' or unstable, yes, i dont think they 'overheat' in these psus, i have two, one for a vic20, one for a plus4, they dont run THAT hot, both didnt work correctly, though, as is, but fault was obvious when bottom was opened, dry joints, once resoldered both worked ok, they can overvoltage if the input and output are still connected, but the ground pin disconnected by a dry joint ...as can be seen in the video, the diodes are the parts that get the hottest, changing them to schottky types should reduce dissipation, although i havent...yet...tried it with mine
Tip for testing pins in a DIN connector: use a cocktail straw (or a regular one if need be) to separate the pins. I forget who I learned that from (some RU-vidr for sure) but it's come in real handy in the past.
Interesting - I'm slowly fixing my C64 that was killed by one of these PSUs (twice) in the 80s and going thru getting it working again, lots of chips to replace unfortunately it seems
The wedge PSU came in many versions, including one with a power switch. There is also a version with two fuses. I can’t recall is there is a version without fuses as well.
The new green capacitors look like some that I have bad experience with, they started bulgeing in my drawer. I hope they are not the same! I repair a lot of the old brown C64 PSU's, mostly just a cable or bad solder joints. When 5V voltage is too high, it's just a bad solder joint on the middle pin of the voltage regulator. Only once have I seen a shorted voltage regulator. Some are easy to break the bottom plate of, others are really hard, like the one you show. Nice touch with the screws, better than glue. I'll remember that next time :)
I would use Buck regulator to convert to 2.5Volt above the target voltage to then feed the analog regulator of the original. I believe a 7805 requires 2.5 volt difference to output a clean 5 volt output. The original designers were wary of low line voltage and would rather burn some extra power rather than have a crash . I have seen 14 unreg DC used to power 5 volt lines at 2 amps approx.
I got my first C64 a few years ago due to a pawn shop find. It still had the original PSU in it, and I knew of the 5 volt death line, but when I tested it with a multimeter I got 5v every time. I bought a new PSU from a seller in Poland the same day I brought this home, and have always purchased new PSU's for my Commodore devices. Now adays I buy from Ray Carlsen since I am in the US.
just testing with a meter isnt good enough, you need to test it on load, and 'soak test' for half a hour or so , if the psu base is opened you'll likely find dry joints, that can cause the overvoltage if the ground of the regulator becomes disconnected, ..resolder , and maybe add a 1N5339 zener across the 5v output to clamp the voltage if an overvoltage does occur
I wonder how many dinosaurs had to give their lives because commodore couldn't be arsed to fit a switch at the spot where a sane person would... Putting one after the transformer would not have made them fold a day earlier. At least it has two indicators now :D
They weren't great quality to begin with either. If you look in mid-80's computer magazines, there are dozens of different ads for aftermarket power supplies.