ShredQuite some time ago, I asked you what I should restore next, and you voted for a #Commodore #Amiga #CD32 that was showing no picture. This project was waiting in my to-do box for some months now, and I did a bit here and a bit there. Let's use this #RetroWeekend to (hopefully) fix this machine.
The seller mentioned that the CD32 was showing no picture. What he "forgot" to mention was that he has tried to replace the capacitors inside the machine, and presumably damaged it on that attempt.
He had already replaced the three through-hole caps and all the 100Β΅F SMD caps. After that, he stopped though, the other SMD caps are still original. He probably checked the machine and found out that the picture was gone.
I guess he then decided that the video encoder chip got broken. But he probably had no equipment for unsoldering an SMD chip, so he tried to cut it out pin by pin. π±β
Fortunately he stopped after one pin, and abandoned his efforts to refurbish the machine.
I got the machine in that state.
I buy those machines because I enjoy repairing them. However, in this case I thought about rescinding the deal due to fraudulent misrepresentation. It's a difference if a machine turns out to be broken after decades of storage, or because of a botched up refurbishment attempt.
Anyway... I kept it, so let's repair the poor machine. πβ
The encoder chip is very obviously broken now. Also, the recapped SMD cap next to it didn't look too good, so I decided to remove it as well. Around and below the cap, I found green varnish that I had to remove with acetone. (I somehow have the suspicion it's just green nail varnish. πβ)
I cleaned the pads and soldered in a new encoder chip and a new cap. The picture didn't return though. Something more must be broken. πβ
This was when I put the board to the side last October. If you're following me since then, you might remember it.
This π€¬β green varnish is just everywhere, even on places where it makes no sense, like on top of solder joints. It makes it impossible to inspect the traces below for damages. What a mess!
I decided to start anew. I will remove all electrolytic caps, even the ones that have already been replaced. I will also remove all the green varnish, so I can start at a more or less clean point.
The varnish cannot be removed with IPA. When I touch it with the soldering iron, it pulls threads like cotton candy. The only way to remove it is with acetone.
Do you see the solder blob at one side of E627C? It's covered in varnish (of course). When I removed it, and touched the solder blob with the tip of the soldering iron, a fishy smell spread.
It means that on the other side, one of the SMD caps has leaked. But instead of cleaning up the area, he just dropped solder on it, and covered it under a layer of varnish. The electrolyte is still there though, and keeps on damaging the board.
This area needs to be cleaned up thorougly, and generously.
That looks better. πβ Now I will also remove the components on the other side, and all electrolytic caps.
And finally, I will remove all that varnish. πβ
I soldered in new components to the area on the bottom side. At the former solder blob, a trace between E627R and E627C was lost. I had to fix it with bodge wire. Just putting some varnish on it wouldn't have solved the issue, of course. πβ The left channel of the headphone connector would have been silent because of that.
On the top side, I have removed all the electrolytic caps, and also cleaned the area around where the solder blob was on the bottom side. (Foto shows the area before the components were removed.)
Unfortunately I have lost a few pads at the 100Β΅F caps. I'm glad that the previous owner had stopped after replacing them, because I could easily remove all the other caps without breaking anything.
Oh, this is going to be fun to fix that with bodge wire. πβ
But maybe I have found the cause for the missing picture! πβ
I rang the traces where the previous owner had removed the solder mask. C281 has no connection at its left pad. It is a coupling capacitor in the S-Video luma signal path, so the S-Video output had no luminance signal.
C236 and C237 are also disconnected from the encoder's signal output, killing the signal on the composite and TV output.
With just two damaged traces, he mananged to make all the CD32 outputs go dark. πβ
More bodge wire will be needed to fix that. I hope it will bring back the missing video signal, but I can already tell that this board won't win a beauty contest anymore.
So... I have now soldered in all the components and capacitors on the top side, except of a few difficult ones.
Instead of the TH caps, I used big SMD caps. The board offers pads for them, and it also solves an issue with reversed polarity markers that Commodore printed on the silkscreen. I used Panasonic FT series caps with a lifetime of 5000h @105Β°C, so leaking shouldn't be a problem.
For today I'm done. Tomorrow I will solder in the remaining caps, and then do a test run. See you then!
Okay, let's solder in the remaining "problem components". One of them is C214. It sits very close to a delay line (that metal cube), so I cannot reach the pad with my soldering iron.
I don't want to unsolder the delay line, because it's impossible to get a replacement if I should break it.
I decided to use hot air soldering here. I've never tried it on an electrolytic capacitor yet, though.
First I used kapton tape to protect the surrounding components from the heat. Then I used a toothpick to put a drop of solder paste on the pads. Only a very tiny amount is needed. I then placed the capacitor into the paste, and heated it with the hot air gun until the solder was melting. The capacitor magically moved into the correct position. πβ
These old machines were assembled with leaded solder, so I also used leaded solder paste for that.
Next problem: This corner had been through a lot. At least one capacitor had leaked and damaged the PCB. The previous owner scraped off the solder mask, replaced three 100Β΅F caps, and sealed the corner with the green varnish. I removed the caps and the varnish again. As a result I lost two pads.
But I'm lucky. For these caps, a through-hole alternative can be used. This makes it much easier to fix the damaged pads and traces with bodge wire.
As I said: This CD32 board will not win a beauty contest again. This fix looks ugly, but I think it's the least intrusive way to repair it. At this level of mess, one has to be creative.
Okay... These were the last components. The board is fully assembled again, and I bodge-wired all broken traces that I have found.
Let's find out if this machine is working again. I'm nervous... π β
β
β Check correct polarity of all caps
β
β Inspect the board for solder bridges and other unwanted residues
β
β Connect S-Video and power connector
β
β Turn on power
Aaaand... No, still no picture. πβ
Voltages are all fine. The chips are getting as warm as I would expect them. It should work, but still... No picture, neither on S-Video nor on component.
Time to get the scope and check the signals. But I need a break now. This machine does not spark too much joy at the moment. πβ
I checked with the scope. On the address and data bus, there is action for a fraction of a second, and then nothing for two seconds or so. This pattern is repeating. It seems that the machine is trying to boot, but then crashes and tries to boot again. I have different theories:
- The PSU isn't original, but a cheap no-name brick that has once powered an external harddisk.
- One of the chips might be damaged, maybe due to ESD (possible that the previous owner did not wear a wrist strap).
- The naked CD32 board fails to boot because peripherals are missing.
I will keep you updated. But I'm afraid this #RetroWeekend won't have a happy ending yet.
Let's go on with the patient. The question is: Do I get not picture because the machine does not start at all, or because something is still broken in the video area?
The DiagROM will hopefully bring an answer. I ordered a 27C800 EPROM for the CD32 DiagROM.
To connect to the CD32 Aux port, I cut an old PS/2 extension cord in half and used one of these MAX3232 based TTL to DB9 converters that are sold on online marketplaces for a few bucks. It is connected to my PC with a serial to USB converter.
YES! I get output on the serial console! πβ So the system is running, and should produce a picture.
I now scope the signals at the encoder chip.
At the RGB inputs there is a signal, but the /CSYNC input is flat.
The /CSYNC signal is still present at TP9, and I could follow it up to this via. But it ends on the other side of the PCB. This via is clearly damaged. I must have missed it when I inspected the board.
There is a blotch of this magic green varnish at the bottom side. The previous owner must have found a problem there, but just covered it instead of repairing it.
48kRAM@shred Have you checked the ROM and it's bus connections to the CPU? Maybe it's just unable to kickstart?
ΔΈurth@shred too bad :( but good luck in the future
@shred Ah, nuts. But this is gonna feel AMAZING when it finally comes to life. πβ
@MyKomment@shred
Nice one !
This fellow might benefit
https://flipping.rocks/@colinpurrington/110677978436143680
Nice one !
This fellow might benefit
https://flipping.rocks/@colinpurrington/110677978436143680
Colin Purrington (@[email protected])
Attached: 1 image Anyone have experience repairing Canon flashes? I fried the bulb last year and was hoping I could just swap in a new one. But after taking it apart the soldering looks a little beyond my skill level. Also read that one should discharge the capacitor, a word that I know but I'm not sure how one does that. It's a 600EX-RT. Ten years old. Alternatively, can anyone suggest a replacement, ideally one that accepts feed from an external battery pack? Because I have a birthday next week and I've been good. #photography #camera #repair #electronics #ExternalFlash #flash #canon
mnemonicoverload@shred
Hopefully any capacitor leaks near those delay lines weren't too bad. I've heard they can get ruined if electrolytic wicks up into them.
Hopefully any capacitor leaks near those delay lines weren't too bad. I've heard they can get ruined if electrolytic wicks up into them.
Olem@shred Excellent work! Seems like a tricky one. Fingers crossed for the smoke test.
@shred OMG! Why?!? Come on people - notoriously fiddly and fragile gear like AGA amigas is no place to "get your feet wet" with PCB rework. Good luck!
derSammler@shred Make sure he corrected the alignment of the through-hole caps, as you probably know that Commodore made a "mistake" on the pcb there. π
Anyway, I wish you luck with fixing it. π
I completely revised my #CD32 in 2015 already.
@derSammler Thank you! The TH caps have been replaced by the previous owner, and he actually aligned them correctly. (I first thought that this id10t couldn't even do that correctly, but then I learned that Commodore did it wrong. πβ)