Brett's STE - Repair thread

[rubber_jonnie]

Makes me wonder if the whole board got flexed at some point and just broke a few iffy factory solder connections on these sockets.

They are terrible for that. The solder literally disintegrates as soon as you remove the metal shielding from the thing. That was one of the main problems I had with development as I would lift them on and off my shelf and that was enough to break them :roll:

Yeah, they're pretty awful for it.

[rubber_jonnie]

For those who are interested, this is the final rewire job:

Bodge wires :excited: :excited: :excited:

By the truckload!

[Badwolf]

Should all be done in the next week or so though.

Fantastic effort @rubber_jonnie!

:2k2:

BW

[Icky]

:bravo: @rubber_jonnie - good effort and was a good read.

[rubber_jonnie]

Should all be done in the next week or so though.

Fantastic effort @rubber_jonnie!

:2k2:

BW

Thank you :)

[rubber_jonnie]

:bravo: @rubber_jonnie - good effort and was a good read.

Cheers :)

[rubber_jonnie]

Ok, so here is essentially he finished product:

IMG_20220131_223051__01.jpg

I promised to do a little summary of the work done to this machine to bring it back to life, so here we go:

1. Removed the existing ROM sockets and reviewed the damage. On the topside, there were 18 vias completely gone, and 8 on the underside.
2. Cleaned up any traces that had torn and were hanging in the breeze to ensure there would be no shorts due to them on reassembly.
3. Repaired the broken pins on the Dual TOS switcher (Who doesn't like a quick win :) ) and tested the original ROMs,TOS Switcher, CPU, Blitter and Shifter in my own STE.
4. Using @sporniket's awesome Kicad schematic, and the original Atari STE schematics, traced all the damaged lines to every possible endpoint, and rebuilt all the damaged traces using bodge wire.
5. Mapped all the address & data lines from the ROMs to 68901, the sound chip, the CPU, to the Blitter, the GST GLUE/MMU combo and finally to the Shifter. All lines tested good.
6. Reviewed the results with an oscilloscope and logic analyser, comparing results to a working machine.
7. Reviewed the track repairs roughly 5 times completely to ensure I hadn't missed anything. As it happened I had missed a connection to the DMA, so this was wired in. At this point I was sure all the physical damage to the vias and traces was made good.
8. Used the LA and 'scope again, but was still not getting results, so ran a series of reset tests whilst probing with the oscilloscope.
9. Had a spot of luck, whilst probing the UDS line of the CPU. I obviously bridged between the CPU and socket finger and the machine booted, which led to realising there was a problem with the CPU socket.
10. Reflowed the CPU socket and tweaked the socket contacts. We have lift off, the machine is booting at last.
11. During testing, it was found that enabling the Blitter caused the desktop to freeze. The socket, like the CPU, was the problem and a socket reflow/tweak was once again the order of the day, bringing back full functionality.
12. Testing was completed using the Atari Diag Cart and port testing adapter. This allowed all ports to be fully tested bar the enhanced joystick ports, since I have no test plugs for them. I'm not sure they even exist.
    Last of all, since it was on my bench, I replaced the bus resistor packs for 2.2K, and added a 10K pullup to the WD1772 as per the mandatory fixes, and replaced W102, W103 & W104 with jumpers as a bit of future proofing.

Overall I've had the machine 2 1/2 weeks, and have probably spent around 40 hours working on it. If that seems a lot, it really isn't for a repair like this, especially when there are unknown other faults like the CPU and Blitter socket to contend with.

So what, if anything, is left to do?

Well, Brett has requested I do a mod whilst I have the board here to add a switch to allow the A & B drives to be swapped with a switch as he has an Internal Gotek and wants an external FDD drive for flexibility.

Personally, I've not done the mod, so it's a good learning experience for me to do it, and it certainly makes sense to do it whilst I have it here.

Once that is done, I will run a final battery of tests before putting some lacquer over the open ends of the bodge wires to seal up against accidental shorts, plus a few spots on any tracks I bared whilst buzzing out the lines to the ROMs.

There is a moral to the tale of this STE.

If you are inexperienced you can kill a machine requiring many hours of repairs.

If you are inexperienced in soldering, stop and think, and if you are not up to the job just take the time to pause. Go away and practice on something less precious until you have the skills you need.

For machines like the ST, damage to the PCB is very easily done by simply manhandling the board, and incorrectly desoldering components or applying too much heat for too long. After all, they are 30+ years old now.

It's a long game, and patience is the winner.

Last but by no means least, there are the tools I used for this job:

1. Duratool de-soldering gun.
2. A solder station comprising of a soldering iron and hot air gun.
3. Various screwdrivers, snips and pliers.
4. A chip pry tool.
5. A PLCC chip puller.
6. ESD safe nylon brushes.
7. A fibre glass pencil.
8. Dual channel Oscilloscope.
9. 16 Channel Logic Analyser.
10. Autoranging multimeter.
11. Large illuminated magnifier.
12. A 1000x microscope.
13. Two bench power supplies (One for each of 5v and 12v).
14. A laptop at my workbench to allow direct review of schematics.
15. Probably some more stuff I can't remember, and disposables like desolder wick, cotton buds, IPA, flux etc.

Why have I added this list?

It's just a reminder that repairs like this require dedicated time effort and specialist tools. It's taken a very long time to assimilate my tools, probably over 10 years or more, so if you are considering modding a retro machine, just be sure that you are equipped to do the job you intend to do, properly.

I've made my own mistakes over the years, including ripping up traces myself, the main thing is that you learn from the process.

[exxos]

Been a adventure that's for sure ! Yeah diagnosing takes days :( normally just cheaper to get a new board :(

Floppy pullups next ? ;) also worth checking P100 is 1.2k..

DAC fix ... :lol:

:hide:

[rubber_jonnie]

Been a adventure that's for sure ! Yeah diagnosing takes days :( normally just cheaper to get a new board :(

Floppy pullups next ? ;) also worth checking P100 is 1.2k..

DAC fix ... :lol:

:hide:

I already checked P100 and it is 1.2K :)

I could go on, but Brett would never get his STE back!!

[BrettRogersUK]

Been a adventure that's for sure ! Yeah diagnosing takes days :( normally just cheaper to get a new board :(

Floppy pullups next ? ;) also worth checking P100 is 1.2k..

DAC fix ... :lol:

:hide:

I already checked P100 and it is 1.2K :)

I could go on, but Brett would never get his STE back!!

You've done an exceptional job mate. I can't really put into words how excited I am to hear that the machine is actually working again.

I will be forever grateful to you for doing this. I can't wait to get it back and back in its case where it will stay from now on.

I see you've added the jumpers as well in place of the resistors. That was another one of things on my list to do. Like I say though, now when I get it back it can go in its case and be left alone. No way am I touching this or any other board again (well not without making sure I've got the proper tools to hand first and foremost). I think I'll leave this kind of job to the professionals in the future.

It kind of makes you wonder just looking at the size of that board, with the size of components rapidly shrinking and being surface mount components as well, just think how big the STE motherboard could be now days compared to that.

Kindest Regards and best wishes,

Brett