Searching for IBM S/23 Datamaster users

[RetroAND]

Preparing for a 3-week period where I won't be able to experiment with it, my 5322 has been enclosed again. Still, design work on prototypes continues and I expect before the end of the month to make another boards order.

For in case anybody is wondering, I will open my designs in time; however I would like to feel the satisfaction to have them working first. My goals with my patches, replacements and extensions is not to make money with them, but to widen the lifespan of the Datamasters. Actually, I'm helping actively another owner to bring back his unit from the dead.

 

[RetroAND]

@NeXT I am curious about your unit. I see through some pictures that you have the later dark-brown one, but I would like to know how is your RAM structured: do you have a single 64k board or two 32k boards?

If you need anything feel free to ask, I will share the few things I know.

By the way, are the two case supports alright? Mine came with one broken and had to print a replacement. If you needed one, I could share my crude design in stl.

 

[nebular]

Hello. I have a Datamaster that I pulled out of e-waste a year and a half ago. I posted about it on reddit here and here. It seems to be in very good condition. Is there anything I can do to help out?

 

[NeXT]

@NeXT I am curious about your unit. I see through some pictures that you have the later dark-brown one, but I would like to know how is your RAM structured: do you have a single 64k board or two 32k boards?

If you need anything feel free to ask, I will share the few things I know.

By the way, are the two case supports alright? Mine came with one broken and had to print a replacement. If you needed one, I could share my crude design in stl.

I'll double check but I could of sworn I took board photos of my machine when it was apart for cleaning.

Support wise everything seems to be great. The only issue is the harness-to-board connection for one of the floppy drive broke in half when I was removing it from its socket. >.>

 

[NeXT]

Looks like the best photo I got since high-resolution board shots already exist online is this.

 

[RetroAND]

I have seen three Datamaster boards so far: a first one, green; this one, light brown (has patches underneath?); and a dark brown without patches (mine).
Are all your ROMs from 1980? If so, could I ask to dump them when the occasion favours it? Some of mine were revised in 1982 while most were from 1980. This means the original set from 1980 is still incomplete.

I see you have 96KB RAM. That's very nice.

We could check the parts that you need to generate the appropriate files to replace them if needed.

 

[RetroAND]

Hello. I have a Datamaster that I pulled out of e-waste a year and a half ago. I posted about it on reddit here and here. It seems to be in very good condition. Is there anything I can do to help out?

Hello and welcome,

I read you in a couple of occasions and was wondering if I would see you around sometime.

First, congratulations for finding and rescuing such an important part of history. This achievement is already a contribution.
There are many ways in which to help:

• If you found documentation or software, or if you had to repair it explaining what you found and how it was resolved.
• If you find software listings for it, scan them.
• Assist users with broken units. There are only a few Datamaster owners around and in my opinion we are also responsible of the future of this system.
• Design parts for broken/incomplete units.
• Assist its reverse-engineering for a better comprehension of it.
• Help dispelling myths and misconceptions.

But whatever you decide to do (or not to do), the most important thing is both to use it and to talk about it. Old systems are brought to life most of the time by this conjunction alone. I saw it on another occasion, where I had a computer where only a group of three had (mine was an even rarer variant) and just by talking about it other units began to appear; at the end a small community was created around the system and its last remaining creator. I wish something similar would happen here.

Thank you for coming

 

[RetroAND]

Regarding my patches and replacements, soon will be a prototyping batch. For this reason, I decided to share publicly the schematics for my modern 64KB memory board. I would like to ask is someone would want to revise it before the order is placed, please. I have the feeling this board is part of the future of the system, as its components are cheaper and easy to find and stockpile. Everything is THT in order to ease maintenance. As I only want this design to contribute to the system, I restrict its usage to non-comercial and share-alike, as well to give credit to the autor/s. I have used lots of resources from people who shared them in order to reach the point I am at and this is a way to thank them for their work.

I hope you like it.

Regards

 

[ggs]

Just started using this website again. ;-)

Just to introduce myself to this group. My first job out of college was for IBM in Boca Raton, FL. The first project I worked on was writing some (a lot actually) of the ROM code for S/23. It was released much later than what IBM wanted because of changing requirements. It was originally supposed to be compatible with the 5110 BASIC but was decided instead to be compatible to S/34 (the delay caused the release of the 5120 which is not much more than a 5110 repackaged in S/23 "skins"). That change of requirements resulted in a "slip" of a year in the release of S/23.

Development on the S/23 was arduous because we had no real infrastructure for building for the 8085 with an EBCDIC character set. :-( For much of the project a complete build took 2 weeks. Someone eventually wrote a cross assembler and linker that ran on the mainframe (S/370) and builds went down to 24 hours but that was fairly late in the project. But before then we were doing "patches" to the ROM code (early systems didn't have much in the way of ROM code...just enough to load a full image into ROM emulator cards...e.g. special RAM cards). Even though all of the code was in RAM, we still had to do patches. Usually we have 30-40 patches that needed to be applied to a build in order to continue testing/development. It was "fun" when a new build came out because we had to see if our previous patches had been incorporated properly into the new build (and of course there were always patches that were developed after the build was submitted and those had to be re-worked for the new build).

A lot of S/23 folks went on to develop the IBM PC. Dr Dave Bradley (inventor of Control-ALT-Delete) had his office across the hall from mine. ;-)

 

[RetroAND]

Just started using this website again. ;-)

Just to introduce myself to this group. My first job out of college was for IBM in Boca Raton, FL. The first project I worked on was writing some (a lot actually) of the ROM code for S/23. It was released much later than what IBM wanted because of changing requirements. It was originally supposed to be compatible with the 5110 BASIC but was decided instead to be compatible to S/34 (the delay caused the release of the 5120 which is not much more than a 5110 repackaged in S/23 "skins"). That change of requirements resulted in a "slip" of a year in the release of S/23.

Development on the S/23 was arduous because we had no real infrastructure for building for the 8085 with an EBCDIC character set. :-( For much of the project a complete build took 2 weeks. Someone eventually wrote a cross assembler and linker that ran on the mainframe (S/370) and builds went down to 24 hours but that was fairly late in the project. But before then we were doing "patches" to the ROM code (early systems didn't have much in the way of ROM code...just enough to load a full image into ROM emulator cards...e.g. special RAM cards). Even though all of the code was in RAM, we still had to do patches. Usually we have 30-40 patches that needed to be applied to a build in order to continue testing/development. It was "fun" when a new build came out because we had to see if our previous patches had been incorporated properly into the new build (and of course there were always patches that were developed after the build was submitted and those had to be re-worked for the new build).

A lot of S/23 folks went on to develop the IBM PC. Dr Dave Bradley (inventor of Control-ALT-Delete) had his office across the hall from mine. ;-)

I never expected a creator of the Datamaster would visit this thread...

I think the fun part is that we both patched the ROMs: you in software and me in hardware. The important part is to keep them running.
Jokes apart, these details are very interesting. Your tale made me wonder if I could ask you a few questions...

First of all, do you know what is the difference between the original set of ROMs from 1980 and the patched ones in 1982? This information could help a lot in repairs.
Second, wouldn't you had a list of I/O ports by casuality? Finding them by hand is a pain, and knowing them could help a lot to their maintenance.
Third, was the miniprobe just a bunch of LEDs like the one I built? This one is just for curiosity.
Fourth, would it be possible to have details about the checksum algorithm and format of the roms used in the Datamaster (especially the header)? This could be useful indeed as there is a 16KB page empty, with a little of luck we could insert new code there.
Fifth, how were the basic statements separated in ROM? I found nearly all of them in the first (or second) ROM but without delimiters.
Sixth, is there any designation of the chips in the board to order them? I have documented part of the hardware and if possible would like to use the original ordering for the components.

I have more questions but I think this post is already crowded.

Sorry if I am asking too much, maybe you wouldn't be able to answer due to NDA. The case is this thing is one of the most obscure microcomputer I have ever seen and at the end secrets attract curiosity.

Regards

 

[ggs]

This is all deep dark memories from the past. I have 0 documentation from that time. IBM required us to do periodic purges of stuff. :-(

In case you’re not aware, the BASIC code is parsed and converted into an internal form to speed execution. The parser was table driven, so that’s probably what you’re seeing. Expressions were converted into a postfix (in some cases prefix) notation. One of the things that I did was to unwind expressions when listing the program to put them back into something resembling the original (including reparenthesing). I also did all of the BCD math functions as well as all of the print formatting (including the PIC statement).

I don’t know about the different ROM versions but I do know the very early systems shipped with EPROMs rather than mask ROMs as there was a 12 week lead on the mask ROMs.

The address space was divided into 4 16kb regions (as I recall, it could have been more…but it’s a detail I’ve forgotten). Addresses below 32k were ROM, addresses above 32k were RAM. the first “page” of ROM and RAM were fixed. The others were paged. This is what allowed the larger than 64k of ROM and RAM. Nothing in memory was directly addressed, there was always a conversion from the “memory” pointer to a pointer usable by an 8085 instruction. Since in most cases the code didn’t issue call instructions to invoke a routine, the ROM could be patched by simply updating the table.

RAM was also garbage collected so stuff moved around and the code couldn’t keep native 8085 addresses around and had to re-resolved that addresses when it needed them again.

The only special card we used in development was what we called a “camera” card. It had a single push button that when you pressed it caused an NMI and entered a low level debugger ( at one point I redid it to be more capable). But that all depended upon the debugger ROM and the emulated ROM as it allowed directly updating not only the RAM but the ROM contents.

 

[nebular]

Hello and welcome,

I read you in a couple of occasions and was wondering if I would see you around sometime.

First, congratulations for finding and rescuing such an important part of history. This achievement is already a contribution.
There are many ways in which to help:

• If you found documentation or software, or if you had to repair it explaining what you found and how it was resolved.
• If you find software listings for it, scan them.
• Assist users with broken units. There are only a few Datamaster owners around and in my opinion we are also responsible of the future of this system.
• Design parts for broken/incomplete units.
• Assist its reverse-engineering for a better comprehension of it.
• Help dispelling myths and misconceptions.

But whatever you decide to do (or not to do), the most important thing is both to use it and to talk about it. Old systems are brought to life most of the time by this conjunction alone. I saw it on another occasion, where I had a computer where only a group of three had (mine was an even rarer variant) and just by talking about it other units began to appear; at the end a small community was created around the system and its last remaining creator. I wish something similar would happen here.

Thank you for coming

I don't have any documentation other than what's already out there online. When I pulled it out of e-waste it only notable extra things it had was the original box (which was water damaged and not really worth saving), the two custom program disks the company that originally owned it used (Which still work), and the overlays for above the keyboard. I'm not sure if those are already out there, but I can scan them.

I didn't have to do any repairs on it. Other than some yellowed sticky gunk on the fan, probably from years of a minor amount smoking around it, the inside was in very good condition. Just dust, no corrosion or damage. Powering it up went smoothly, except for my son complaining about his ears from the flyback transformer.

I'm willing to dive in a help with any reverse engineering, though my hardware experience is only a bit above novice at best. I've got a soldering station and a multimeter myself and I have access to an oscilloscope, but I'm not very experienced in using it. I'm willing to disassemble, photograph, and test in any way requested, but for the testing I'll probably need to be walked through the process.

Honestly, I just really love the look and feel of this beast (especially that wonderful pre-Model F keyboard) and I'd love to have something somewhat useful to do with it. I had been searching as much as I could for the Word Processor for it, but after reading this thread I now know that it needed a card that I don't have.

One area of reverse engineering I think we should look into is the serial port it has. It would be great to be able to adapt it to standard RS-232 which would allow some kind of connectivity to more modern systems for data transfer. Personally the only 8-inch floppy drives I have access to are the ones in the Datamaster itself (though I have a whole lot of 8-inch floppies from my work that was going to e-waste them). So any software for it I would need to write myself or have someone image a disk using a shugart drive. Maybe a kryoflux or a greasewezel could be adapted to use the proprietary IBM drives as well.

Well let me know if there's anything specific you'd like me to do with my Datamaster.

 

[RetroAND]

I don't have any documentation other than what's already out there online. When I pulled it out of e-waste it only notable extra things it had was the original box (which was water damaged and not really worth saving), the two custom program disks the company that originally owned it used (Which still work), and the overlays for above the keyboard. I'm not sure if those are already out there, but I can scan them.

I didn't have to do any repairs on it. Other than some yellowed sticky gunk on the fan, probably from years of a minor amount smoking around it, the inside was in very good condition. Just dust, no corrosion or damage. Powering it up went smoothly, except for my son complaining about his ears from the flyback transformer.

I'm willing to dive in a help with any reverse engineering, though my hardware experience is only a bit above novice at best. I've got a soldering station and a multimeter myself and I have access to an oscilloscope, but I'm not very experienced in using it. I'm willing to disassemble, photograph, and test in any way requested, but for the testing I'll probably need to be walked through the process.

Honestly, I just really love the look and feel of this beast (especially that wonderful pre-Model F keyboard) and I'd love to have something somewhat useful to do with it. I had been searching as much as I could for the Word Processor for it, but after reading this thread I now know that it needed a card that I don't have.

One area of reverse engineering I think we should look into is the serial port it has. It would be great to be able to adapt it to standard RS-232 which would allow some kind of connectivity to more modern systems for data transfer. Personally the only 8-inch floppy drives I have access to are the ones in the Datamaster itself (though I have a whole lot of 8-inch floppies from my work that was going to e-waste them). So any software for it I would need to write myself or have someone image a disk using a shugart drive. Maybe a kryoflux or a greasewezel could be adapted to use the proprietary IBM drives as well.

Well let me know if there's anything specific you'd like me to do with my Datamaster.

Eager to start, aren't you? I can't blame you, I was in the same mood when the thing arrived.

So, you are collecting some computer history with your son? That's very relatable to my case, as my father and I have been for some years. If he complained from the CRT, it seems he is too yound to have known that type of screens... something similar happened on my side - when my father showed me his first computers (a Sinclair ZX-81 and a ZX Spectrum), I called him a liar because I thought those were just keyboards... it's been a long time ago though.

If you have the overlays, it should be worth scanning them for preservation.

I would like to use my System/23 for something else than occupying a bench. I would like to control a scale model train with one of my oldest computers (this one or maybe some PET). It's good that you want to achieve the same goal of giving it a purpose.

My skills in reverse engineering aren't the best either, but I have achieved the few things I have by persevering. I let you know, it's a risky process in which you could damage your board; therefore think it very well before acting. I learned many things during my process with the memories, but my board suffered damage I had to repair later. Still, having to pay that price was worth to me as the results I got enabled to repair other systems and provide some security to my own unit. If you do it, always balance the risks and consider it. The failures I induced by accident also gave me an insight on the behaviour of the system in certain cases. Have in account that If something bad happened while you were reverse-engineering, I would help you as much as I could to bring it back. After all the collected data has to serve this purpose of reparation.

About the floppy drives, we (my father and I) tried to enable unit 3. However, that's not an easy task. We considered the adaptation and usage of a gotek, but unfortunately the IBM format drives the magnets of the rotor directly from the controller board instead of having an enable and a direction signals. It won't be easy to have something working for some time.

Talking of the serial port, if you wanted I could give you some files that would help you guide through the board. There is still an unidentified component in the subsystem, I know it's an optocoupler, it's from HP and the timeframe it was in the market; the possibilities with that data and the number of pins left a small list from a Databook but only one will make sense the day its sorroundings are mapped. If you want to explore that area, I can help you.

Still, there are other things you could start with. Could you describe your model, please? If yours has 32KB or 96KB you must have a 32KB board. I reverse engineered the one of 64KB but as I don't have the 32KB one, I cannot study it. If you had other expansion cards other than the floppy controller, that would be worth noting.

Finally, I shall thank you for your decision to study it. Every contribution, for small as it can be, is important.

P.D. I work with Kicad. If you were to start documenting, which cad would you use?