Amiga 4000 restoration

[Zylark]

My trusty old Amiga 4000/040 have been sitting idle for far too long. 13-14 years at least. Before it got decommissioned in favour of a Windows box, it was not all that reliable either. After a while of being powered on, it started to go slow and then freeze outright. So at the very least it got a heating issue.

This can be remedied by cooling down the CPU some more I assume, and luckily I got plenty of CPU cooler fans lying about. So that should be a simple fix.

A bit more serious is that the old battery have leaked...



(picture in full size here)

It's not spread too far though, and the corrosion isn't terribly bad as these things go. What is affected is the 10k resistor network SIP8 package, the HCT174A flip-flop chip and four of the eight 4.7k SMD resistors. Two of the pins on the rightmost HCT166M shift register chip is a bit dull as well.

Parts are ordered, and whilst at it I went ahead and ordered new caps for the motherboard and the CPU board. And a couple of other bits, like a coin-cell battery replacement and a 4 gig CF card 'Hard Disk' replacement.

It'll be a while until I got all my ducks in a row and can get it all fixed. But that is fine. It'll give me time to go over the rest of the machine with a keen eye out for potential problems.

First on the agenda is having a good look at (and test of) the PSU. I suspect that may need a recap job as well. The floppy-drive will need an inspection. Last but not least, the front plastic cover is now more brown than off-white, so it'll need some treatment.

Plenty to do in other words. Will update as the project snails along over the following weeks.

[Zylark]

Why thank you.

I've been doing a bit of research lately, and heard it mentioned in some YouTube video or other that CBM (or rather their PSU supplier) did not always pay attention to little details like what direction the fan got installed. So yes, I'll be sure to check that out. I'll probably replace the fan alltogether, and add a couple of potentiometers to control the PSU fan and the to be installed CPU fan. The 5 1/4" drive bay face-plate is a nice place to mount those, and perhaps also a few LEDs for the bling factor. Hmmm, maybe an Arduino powered little LCD to report fan speed and temperature?

Anyhow. Fix first, bling later

[Zylark]

Anyone by chance got the pin-outs for the PSU? The 4-Pin molex connectors shouldn't be a problem - though the color-coding of the wires isn't even close to standard. The MB connector however...

Expecting the first batch of parts to trickle in during the coming week, so I thought testing the PSU should be first order of business. Just in case that'll  need some work as well.

I could go with a new PSU, but I want to keep it all original. It's not like if I'm going to stuff the machine full of peripherals that draw a lot of power, so the original will do just fine. All I'm going for is to get the machine up to original condition. Perhaps invest in a Scandoubler at some time, but then I'll also need to replace the Buster to get it up to v11. Though I'm not new to SMD soldering, replacing a chip of that size would be new territory for me.

[Zylark]

It's a Zorro slot card right? As I've understood it, in a 4000 the non v11 Busters have some issues with expansion cards.

Anyway, I and my assistant will research further.



Paws here have promised to supervise my restoration

[Zylark]

Thanks Mechy.

Slight update. A few parts have indeed trickled in. I've received the 4.7k resistors I need, the 10k resistor net and the HCT174 flip flop. Still missing the caps and the HCT166 shift register. Tomorrow is last day of mail delivery until Wednesday here in Norway due to a few days holiday early next week. So fingers crossed...

Anyway, had a look at the PSU. And it is absolutely filthy. Back when this was in use, I was a bit of a smoker, and it is noticeable. On the bright side, the fan looks like it is in the right way and the caps looks good. Well, none of them are bulging at least.

I'll have to dismantle the entire thing, and clean it up as best I can. some work with brushes to get the worst of the crud off, and then soak the PCB in isopropanol to get rid of the rest from the nooks and crannies.

Whilst in there, I had a cursory look at the caps on the motherboard and CPU board. It all looks quite good. No obvious leaks or corrosion (apart from, you know, around the battery). There are some places that look a bit dull, but that may just as well be dirt and grime from my nicotine filled youth. Nonetheless, they'll all get swapped out for new ones.

First thing first, clean and test the PSU

[Zylark]

Finally, got all I need to get started. Shift registers arrived today, together with a fresh batch of flux.

First order of business, getting my 'assistant' well fed and ready for a good catnap. Some Tuna in a can did the trick. No interruptions what so ever for the rest of the evening.

Got on with it by dismantling the Amiga piece by piece until I could finally get the motherboard out of there. Quite an operation in itself and not a design I'd call brilliant.

That done, the battery got unsoldered and removed. Next up the 10k resistor network. That took care of the through-hole parts. Before getting on with the SMD parts, I took a little 'before' picture:


(full size here: clicky click)

Looking kinda' crusty huh?

After removal my heart sank a bit, as the pads where not to be seen whatsoever. Just a lot of crud and stuff.



(full size here: Clicky McClickface)

Added flux, applied some heat, went to town with soldering iron and new solder. Wiped off using solder-wick, cleaned with isoprop and a toothbrush. Rinse - repeat a few times.

Slowly getting better, so after a proper isoprop cleaning and reassured that none of the pads would lift, I went to work with a dentists pick to loosen up the worst of the crud. Once that was done, some careful wiping (with plenty of isoprop on the board) with steel-wool attached to the end of a q-tip stick.



(full size here: clicky'nopornIpromise'me)

Looking much better. Though not entirely done. Some probing have shown that I suspect a trace or two may need some attention as well. Anyhow, I got all weekend to work on it. Now it is late, and I'd probably screw something up if I work more on it now as I am getting a bit tired...

[Zylark]

Aight, a bit more cleaning, probing, chasing down traces and pondering over schematics and the final diagnosis is pretty much done.

Pads are fine, but there is one trace that need be wired (pin 2 from flip-flop to pin 4 on resistor network). Several connections from pad to trace down by the 4.7k resistors need be re-connected by means of a bit of solder and finally I need cover up all the naked Vias' that I had to scrape a bit off to test continuity and make sure was all good.



(full size here)

As you can see in the image, from bottom left, pad 2 got the issue with the trace. Broken in so many places there is no point trying to fix it.

I've gone ahead and mounted the new 10k resistor network. Now I'm just waiting for my iron to get a bit cold for me to change tip to a very sharp and pointy one so I can get on with the more fiddly stuff

[Zylark]

Aaaaaand done - kinda'



(full size here - clicky click)

Yes, I know. I'm not going to win the Solder-Master 2016 award of excellence. It all checks out though through continuity tests, so the resistors being a bit crooked will have to do

Still got to run a small wire from pin 2 on the flip-flop and over to pin 4 on the resistor network. But once that is done, this part of the restoration is over.

Worry ye not however. Still plenty to do.

- Clean up and test the PSU. Been postponing this as it is dirty as all hell.
- Recap the mobo and CPU board.
- Open up and inspect the Floppy drive.
- Install new battery and CF-Card 'Hard-disk'
- De-Yellow the front plastic cover. Hydroperoxide + UV light anyone? Means I need wait until that yellow orb in the sky decide to shine on my fair but woefully overcast city...

edit:

Got the wire installed:



(full size clickymahbob)

Also did a double check of everything and had to do a little touch-up on the third 4.7k resistor from the right. For stable continuity I had to scrape off a little bit of the solder-mask going from the pad to the trace, and expose a bit of the trace and then cover it with solder for a solid contact. Must have been a slight crack that caused a more intermittent fault.