MegaST board with strange CPU mod

[dml]

...so that was a repair I didn't realise I needed to make.

I started off trying to repair a dead board and discovered my working/reference board was also dead. It was working when I last packed it away!

Symptom was a black screen / no signal on SM124 and an empty white raster on a 1084 RGB CRT. No activity from hitting reset button.

Poked around quite a bit, checked voltage rails, re-seated the socketed ICs but no joy.

Could see video sync and RGB signals on the monitor socket (because: white screen, so..) but no interesting activity.

Scope on the ROM socket address/data pins - nothing. Just deadness after power on.

Checked the RESET signal to 68k... seems ok. Checked the BERR signal... same.

Checked HALT - it's low. So the HALT circuit is broken or more likely the CPU said HALT because of extra complicated stuff going wrong elsewhere. :/

Followed that around but nothing really interesting.

Randomly decided to swap the GLUE and retest. Now half the time HALT is low, the other half its high. o_O

Now put the old GLUE back in, now HALT is high. Great, random glitchy physical fault on the board.


Now I put it under the microscope and scan around for a bit. Nothing interesting for 99.9% of the board.... except one solder pad is green on the bottom side of the board. Hmm. It's a CPU pin.

Very crusty, green pin. All other pins look fine. No obvious reason.

Screenshot 2025-06-11 at 15.32.06.png

[dml]

So I try to clean it up and it won't clean. Solder won't melt, flux isn't helping, lots of passes at scraping, scrubbing with IPA, flux, scraping & IPA... eventually I get this.

Screenshot 2025-06-11 at 15.32.22.png

...which still looks terrible but visible now that the darkened bit of track had just disappeared below the solder mask.

[dml]

More scrubbing & cleaning, this time with an unsafe combination of random chemicals - WD40, servisol, IPA, acetone - everything. It did not want to de-crustify. But I got it clean and managed to bodge a track back on.

Screenshot 2025-06-11 at 15.32.34.png

Machine decided to boot to the desktop after that. Now what was I doing? -_p oh yeah - a different MegaST.... maybe this evening I'll have a look.

[Maximilian]

:bravo:

[dml]

I had a quick look at the other MegaST - the one I was originally trying to fix - and it's just very dead. Nothing going on anywhere, except the reset line looks ok. Not a lot else. No activity of any kind on monitor port or bus.

The CPU and MFP are also getting a bit warm, more than the other machine. Will spend a bit more time on it at the weekend and might just take the CPU & MFP out and socket them since the CPU had lots of hand-soldering done all over it and I can't really see if there are any problems under that structure, the way it is currently.

[dml]

So this evening I had a go at removing that CPU+socket-squared contraption and fit a normal socket.

Maybe a bit overconfident with that - got the thing out but took 3 pads and 2 tracks with it. That was with many desolder-gun passes with new solder+flux each pass and then hot air at the end. Heavy copper tracks sinking the heat in a few places - and so many pins.

rail-carriage.png

Fortunately it wasn't an insane level of damage and I was able to repair that. It was a hassle though. This is an early (blurry!) stage of me putting the traces back... original trace pinned back down with green soldermask with one end taped into/through the via. Tacked down onto the track on the back side of the board.

oops.png

[dml]

Same repair continues on the top side of the board....

repair-front.png

...and on the back...

repair-back.png

[dml]

After finishing that I fitted a new socket and fitted a different CPU. The old one is cleaned up a bit now but haven't tested it.

Just as I was setting up to re-test the board though with the restored CPU, I noticed this next to the GLUE.... o_O

blegh.png

...4 interrupted tracks in one place. No idea what must have happened there... didn't notice it when I looked at the board the first time. The same area on the working board looks like this...

other-mega.png

Around that point I started getting that 'rabbithole factory' feeling - that I'd keep finding more problems from previous attempts or reversed modifications - and it was never going to work.

[dml]

Fixed that mess next...

trackrepair.png

....and then noticed this....

diode-hax.png

o_O ....and decided to ignore that completely. Pretend I didn't see it.

[dml]

...which turned out to be the winning decision!

undead.png

I had to disable the blitter with the jumper pads as the socket is empty. Will look at that another time. Probably there are other issues waiting here but it's good enough for now. :p