Watch on YouTube:

How Atari ST Service Engineers Diagnosed Faults In 1989 - Rare Hardware Hands-On!


This rare “engineer use only” diagnostic hardware is very kindly on loan from the North West Computer Museum (
). It features DE-9, DB-19 and DB-25 connectors to hook up to the Atari ST’s joystick ports, ACSI hard drive port and parallel printer port, and works in conjunction with a diagnostic cartridge to run a series of tests on the ST’s hardware.

Although the basic cartridges are fairly common, setting up an Atari Authorised Service Center was an expensive business with the complete diagnostic kit running into the thousands, so it’s rare to see the more advanced specialist kit out in the wild in this day and age. But how does it work? Find out within, where I hook it up to my Atari 260ST and Atari STE!


If you had an Atari ST - like this one - in the late 80s or early 90s and it developed a fault, one option would be to take it to your local Atari Authorised Service Centre, a global network of specialists who’d paid a fair bit of money for a load of specialist kit, training and manuals directly from Atari themselves.

These test cartridges designed for use at those centres are reasonably easy to come by, but I recently got my hands on a much rarer piece of the puzzle that can be used in conjunction with one of these to perform some extended tests, and it’s that that we’re going to be taking a look at today, and trying it out on some actual faulty STs to see what the process is all about.

But first up, I just wanted to give a big shout out to the North West Computer Museum in Leigh, just outside Manchester, who had this item donated to them and had no idea what it was and they’ve very kindly let me show it off on the channel which I’m more than happy to do because, well, I bloody love a good computer museum and I’m happy to do anything I can to support them and raise awareness of them.

That said, they’re not actually open yet but will be very soon and I’ll be covering them on the channel when they do, there’s some more info on the museum and a link to their website down in the description and they’ve literally just announced that they’re having a jumble sale on the 22nd of this month - that’s the 22nd of October 2022 - to raise some funds and I’m planning on being there to hopefully grab some bargains, help them out a bit financially and maybe even put together a video on that day too for those of you who can’t make it.

So I was contacted by Jack, one of the co-founders of the museum and a fan of the channel, and he sent me this rather interesting box. Now I’ve been collecting Atari stuff for 20 years now and I’ve never seen one of these in the flesh, let alone in the original box with all the original paperwork, so let’s get it open and see what it’s all about.

So first things first there’s the original receipt which is in Dutch and it looks like this was originally sold by Atari Benelux - that’s Belgium, Netherlands and Luxembourg to you - on the 26th of July 1989 to a company called Kat & Korsch, which doesn’t seem to have a direct translation and sadly no longer seems to exist. There’s no price on here, but I’ve read elsewhere that a complete setup from Atari could run into the thousands so I suppose it depends on the scale of their operation and whether they just bought this board or a fully kitted out setup.

I checked out the address on this invoice and Evertsenstraat is actually in Rotterdam, it’s a very small residential street and I can’t see any signs of a computer shop there now, just a load of fairly new looking houses, these old industrial units down at the end - which might be where they were based - maybe where this music shop is now - and, oh, there’s also this very cool A-Team van. The things you find on Street View eh.

Anyway, moving swiftly on, there’s a leaflet for something called a Hyper Speed connector, and for those of us who don’t speak Dutch, here’s Google Translate. Oh, except this is actually in German… And that’s a pretty useful translation actually. Seems it was an add-on board for a second CPU or maybe a coprocessor for the Mega ST machines that plugs into that range’s proprietary expansion connector. Sadly I don’t have this board but maybe one to look out for for a future video.

The next piece of paper is the pinout for the RGB monitor port, and it looks like this is a photocopy from one of the service manuals. As you can see it has some scribbles on it - again in Dutch of course - and it looks like someone was trying to put a SCART cable together, it’s easy enough to buy an off the shelf cable for an ST in this day and age but I don’t think Atari offered an official option so that makes sense and maybe if it belonged to a computer shop they were looking at putting these together to sell or maybe for their own internal use.

This is cool, it’s a diagram of all of the RAM chips - looks like it might be for the STFM - and I guess when you run the memory tests it will tell the engineer which chip is at fault so this would help to track them down inside the machine a bit more easily.

And this is written on the back of a price list from a place called Microtech Roos - and it seems to mainly be Amiga stuff. I guess nobody wanted that if they were using it as scrap paper for their Atari ST notes.

Friendly trolling aside - the ST was hugely popular in the more Eastern parts of Europe and Scandinavia in the 80s and early 90s, particularly as a business machine for word processing and desktop publishing and was very much seen as a great alternative to the Mac and IBM PC rather than perhaps the gaming machine that it was marketed as in the English speaking world. For example my Mega ST is a German machine because it just wasn’t a big seller in the UK, and I also have a Dutch 260ST which we’ll have a look at shortly, and that very early model wasn’t even available in the rest of the world.

The final bit of paperwork is a printed copy of the relevant part of the service manual - and this has been available online for a long time now so it’s not a new discovery, but still cool to have an old printed copy like this, and it just talks about the test procedures and how to use the hardware and diagnose faults.

Finally, the big reveal we’ve all been waiting for - the test harness itself. It looks like this is based around off-the-shelf logic chips - I don’t think there’s anything too exotic on here - and we can see the Atari part number which is CO26108.

There’s a 4MHz timing crystal, and as far as ports go there are 2 DE-9 joystick ports on one end and DB-25 and DB-19 ports on the back which will connect to the ST’s parallel printer and ACSI hard drive ports respectively.

There’s no power connector, and it seems from the traces on the bottom that this draws its power from the joystick ports, which of course provide 5V - you can see the thicker traces connected to pin 7 on those ports and they go through the board via vias to the top as well.

There are also holes in the corners so this can be mounted in some kind of enclosure - but I don’t think there was ever an official option available as I’ve only ever seen the bare boards, but perhaps someone in the comments knows these things better than me and can shed some light on the subject.

Now, the board doesn’t work on its own and requires a test cartridge which of course goes into the ST’s cartridge port. There wasn’t one included, so I decided to buy one on ebay. The label really got me thinking about whether this was original or a modern clone because - well - it’s obviously printed on an inkjet printer which I very much doubt Atari would’ve done in 1989 - so to satisfy my curiosity of course I had to take it apart - as you do.

The case is definitely injection moulded and seems to be the same type of plastic as the ST itself, and it seems the top case has the Atari part number on it which is CO70182 so that’s a very promising sign. The bottom case is much the same and has the part number CO70183 moulded into it so that all looks proper as well.

Finally, the board itself, and it definitely has that 80s Atari PCB air about it. It features a couple of Atmel 256 kilobit one time programmable EPROMs which contain the menu system and diagnostic routines, and on the back we can see that the ground plane around the outside of the board has started to bubble up in that way that these old PCBs do.

So I’m pretty sure this is an original with a rather dodgy looking sticker on it. Not that it really matters as long as it works of course, but it’s nice to know. It also came with a copy of the same manual pages that were included with the test rig, and a lovely printed diagram which I assume is from the same manual.

So with the cartridge reassembled, let’s get this all hooked up and see what those diagnostic tests were all about. I’m going to try this on my STE initially as I know it all works fine and it’s already set up and ready to go - this is the machine that I’ve had since I was 8 years old by the way - and I’ll just try the cartridge on its own first and make sure that’s all working.

That all flashed by very quickly, and it’s running in 60Hz mode which my camera isn’t set up for, thankfully I can switch that in the menu as it’s determined by the OS in the ST rather than the hardware, and the Atari monitors support both rates.

But first let’s slow things down. Although this is menu driven, the cartridge runs a few tests on boot, so let’s have a look at those.

The first test immediately after hitting the power button mentions bus error handling, so that’s a very low level hardware test that flashes by very quickly.

Next it tests the MFP, GLUE chip and Video. The MFP is an MK68901 chip that handles all of the I/O, and the GLUE is a custom Atari ASIC that bundles together a load of different functionality and manages bus communications and whatnot, and it’s important to note that this is different in the STE to the earlier models as they integrated it and a load of other functionality into a new chip called the GST, and this difference will become apparent in a second.

This is a version 4.3 cartridge from 1990 that supports the ST and Mega ST - no mention of the STE, and we’re getting a load of errors and a warning about an unrecognised TOS version, which in this machine is TOS 1.62 - a revision 2 STE TOS from that same year.

Pressing V switches us into 50Hz mode and gets rid of that horrible flicker - which is all well and good - but I don’t think it’s worth proceeding with this machine as it’s not officially supported and that may screw up the outcome of some of the tests.

So I think a more appropriate choice is my 260ST - and was certainly the kind of early machine that this setup was designed for. Incidentally, if you want to know more about the story behind the 260ST I made a video all about it which I’ll link in the usual places.

One of the reasons these early machines were so nice and compact compared to my later STE were the external power supply and floppy drive, so let’s get that PSU hooked up.

Now, on first boot it runs through the same tests as the STE and we can see that the errors are gone. One interesting thing about this machine is that it’s running the very early bootstrap software that predates TOS - a story I covered in detail in that other video - and the cartridge actually detects this as a Spanish OS version 0.0.

The video timing in these very early machines is a little bit ropey to say the least and I couldn’t get my camera to lock on to the slightly out of spec refresh rate in either 50 or 60Hz modes so sorry about the slight flicker, but we’ll persevere.

I’ll just quickly run through a few of the tests - to be honest most of them aren’t very exciting to look at, and for some reason the audio test doesn’t output any audio on either this machine or my STE, which is a bit odd.

There’s a function to modify memory which looks like this, I’ll leave that alone, and we can set the baud rate of the serial port.

There’s also an option to test high resolution output, and I happen to have a high resolution monochrome monitor, so let’s give that a go - and as you can see, that cycles through a few different test patterns, so I guess we can call that one a success.

Last time I tried to use this particular machine was on the RMC show and tell and there seemed to be an issue with the floppy controller.

It would be interesting to see whether the diagnostics pick up on that, as it didn’t work with any of my known working floppy drives or indeed a Gotek, and yet again it seems that there’s nothing happening with these drives. It’s a shame that the cartridge can’t seem to differentiate between a bad drive and a bad controller but hey, there we go.

But of course I promised to show off how this thing works - so let’s get it hooked up!

…and it’s not working!

Which is odd, because I was testing it a few days ago and it all worked fine. I thought maybe the 260ST power supply was too weak to handle it so I switched back to the STE, but that also refused to power up.

It was then that I started to suspect the brand new DE-9 joystick cables I’d bought on Amazon just for this video. Swapping back to the rather cobbled setup I had before - which involved a DB-25 to DE-9 converter on one cable and a CGA Y-cable on the other, and it all worked again.

So there we go, don’t bother buying fancy cables just because you think they’ll look good in a YouTube video.

One thing I can’t test is the serial port, because that needs some kind of special dongle that I don’t have, and I’ve also hooked up a MIDI cable, that just goes straight from the in to the out port so I can run that test too.

So here goes!

And that’s sorted it. The DMA port test passes with flying colours, so we know the hard drive connector works. But the printer and joystick test is a bit of a mixed bag - I tried it on a few systems and it was very inconsistent about whether it passed or failed even when jiggling the cables and running the same test repeatedly, so I’m not sure if it’s those cobbled together joystick cables or maybe something else on the test rig getting a bit flaky with age.

But, we’ve seen how it all works now, so that’s good enough for me.

Oh, and if you’re wondering how the completely dead 520ST got on… Yeah. I think that one has bigger issues - although it’s doing something so it looks like it might just be a video output problem. Definitely one to revisit in future.

I’m actually really tempted to try to recreate this test rig and make a modern clone and get it manufactured by my channel sponsor and long term PCB manufacturing partner JLCPCB. As with my previous open source hardware projects it will just be a case of drawing it up in KiCAD and sending the files off for fabrication using their easy to use website, and I may even design a nice enclosure for it and use their 3D printing service as I’m not lucky enough to have a 3D printer at home.

If you have a similar project in mind, 1-4 layer PCBs start at just $2 with all sorts of customisation options and if you’ve never used them before, there’s a link down in the description for new users to get $54 off their first order, so go and check them out, and hopefully - as I always have - you’ll find that they offer a very quick turnaround, great customer service where you can talk to real people, and a very high quality product.

See what I did there?

So I hope you found this old and rare piece of hardware interesting and I’ve certainly enjoyed getting the opportunity to check it out so big thanks to the North West Computer Museum - details about them are down in the description. Thanks as always to my patrons and channel members for your support, their names are up on screen as I speak, and I’ll hopefully see you again next time.

Support the channel!
Become a Member:

Relevant Links:
North West Computer Museum:
Pixel Addict:
History Of Atari Benelux:
Atari ST Diagnostic Cartridge From:
Service Center Book / Map Image Credit:
Atari Service Manual Image Credit:

Further Viewing:
260ST Story:
Atari Show & Tell With RMC:

If you liked this video please consider subscribing to ctrl.alt.rees on YouTube!