I’ve been delving into CPT machines recently, and found the CPT4200 from the early 1970s. It is a console that attaches to a modified IBM Selectric typewriter and allows text to be saved and loaded on one of two cassette tape drives in the 4200. The text can then be edited and reprinted as required from the tape. The second tape can be loaded with text as well, for example names and addresses, and the two combined and printed out, giving a mail-merge facility.
This machine is implemented in TTL logic. No microprocessors.
Start your trip down the rabbit hole here:
Everything I’ve collected is here:
Information on these machines is most welcome. Schematics of the 4200 would be a good thing to find.
Just a small update about the CPT4200. There’s some information on the web, including a training manual for the machine which is enough to be able to find out how to use it. After the previous post I was examining the machine I have and noticed that the key switches on the front of the machine had a patent number. Some of the keys on my machine are broken, so I have to make some new keytops, and I thought a drawing would help, so I looked the patent up. It’s not a CPT patent, but does have a drawing of the keys. I didn’t use that in the end, but modelled my own key tops. We’ll see how that goes. What I did do, though, was to search for CPT patents, and I found two. One describes the modifications that are made to an IBM Selectric typewriter to make it work with the 4200, and the other describes the 4200. I think there’s probably enough information to be able to do the modifications to a Selectric, if I wanted to go down that route (I’m really not sure that I do, but it would be a journey),. The 4200 console patent has many pages of schematics, and description of how the machine works. This is a long way to having schematics for the machine. Patents are a good place to look for information on old machines. It’s interesting to note that I can find no patents for CPTs later machines, as the attitude towards patents changed around the 80s and much less hard information appeared in them.
DO you have a complete 4200? I am guessing you don’t have the printer? I don’t have any documentation (so the training manual will help) but the pictures of the CPT 4200 on the BinaryDinosurs site are from me and mine is mostly complete except for the keys that are falling apart. They are like eggshells and so is the CPT 4200 outer case, to be honest.
Good luck with your project. If I can help in any way, let me know. I have not yet tried to power mine up but I have had it for a number of years. I also have a couple of original cassettes.
Hello,
I do not have a complete 4200, no. I have the console but no Selectric. After seeing the patents, I thought about converting one, I think there’s enough information there to modify a Selectric, just about. I’m assuming when you say printer, you mean the Selectric? It’s also used as input to the console, to control the operation.
I have seen your photos on the BinaryDinosaurs site, and I’ve been talking to Adrian about his CPT8500, as I have a set of PCBs from one that I want to build into a working system. So far I have designed a backplane and a debug diagnostic card, both are on their way from China as we speak.
My 4200 keys are also disintegrating, one broke this morning, actually.
I have designed a V1.0 set of replacement keys, and they are being made in China now, and should be with me in a couple of weeks. I fully expect there to be problems and have to make a V2.0, so don’t have any made from my models yet. The idea is to make them in one of the JLC white resins, the legends are embossed in the models and should be able to be filled with paint or air dry clay. In the past I’ve used air dry clay, but looking at the original switches paint might be better.
I powered my unit up this morning, after having a look over it. It appears to be in very good condition. The PCBs look like they were made last year. Everything is there and in place, no cut wires or anything nasty like that. There’s quite a few tantalum capacitors, I may replace them, or cut one leg and put an SMD capacitor on the PCB, so the pretty coloured bits are still there.
My unit came with a cassette, but it was an assembler/editor for an Ohio Superboard I think, so nothing to do with the CPT. Are you able to record your tapes into a WAV file? It should be possible to look at the encoding then and see how it was done. the patent gives some details I think, but I’d like to see the waveforms.
The unit is made to be worked on. The PCBs are hinged , as is the front unit, and they open up for access to the backplane (wirewrapped) and the cassette mechanisms.
I had to re-attach the Selectric cable as the nut inside the machine had undone. there’s actually a gap in the case that allows you to do the nut up from outside, which I did.
While it was powered up I took the opportunity to investigate the 50 pin D type connector that plugs into the Selectric. I have some idea of what the signals might be. I think I have enough to build a Pico-powered PCB that can drive signals and read signals and should then be able to interact with the console. Fortunately there are a lot of GNDs on the connector and not as many signals as you’d think. I’ll put that in the repo.
With the patents it should be possible to work out how it communicates. I have CPT8000 schematics and on some of those there is an interface to the 4200 and that has signal names. I don’t think that interface ever shipped, and it is different to the 4200/Selectric pair as there is one connection to the Selectric and one to the 4200 console, the two aren’t connected together. I think the 4200 console link was intended to provide a data storage facility for the 8000 and the Selectric a printing facility.
Yes, when I said “printer”, I meant the Selectric typewriter. Sorry for the confusion. The keyboard in the typewriter was completely replaced as I am sure the patent mentions and various solenoids put in the right places to make things print properly. That would be a tough one to do but good luck.
The red transparent lens in front is fairly simple. It’s just a rectangular cut piece glued in from the front. It is very hard to see through it unless you hold it up to the light.
Luckily most of my keys are in decent shape. I do have a 3D printer if you want me to test out your design. I would love a set when they are finalized. I am pretty sure more than one are ready to crumble. Here’s a close-up of mine.
I have not yet tried to read the cassettes. I am actually in the process of trying to read some other data cassettes for something else so I will gladly read them and post them somewhere for you to review. They look like this if you haven’t seen a picture of them.
I will post a link to the cassettes when I’ve read them. I’ll finish up on the other project I’m working on an perhaps fire things up and see what happens once I’ve done a thorough inspection. I did one a while back but it never hurts to do it again. Glad to have finally found someone with one
Ah, yes you have a Selectric. I wondered if you had one of the ‘Rotary’ CPT printers.
Thanks for that photo, I’ll get some filter material and put it in.
The keys I’m printing are for a resin printer, if you have one of those then the files are in my GitHub repository, or should be. If you don’t have one then I’ll send you some once I know the model works.
I think I have seen a photo of a cassette somewhere on the web. It would be nice to examine the files sometimes, thanks for offering to do that.
I have noticed that some capacitor vent plugs seem to have degraded, so.i may swap.some.out.
You’re going to force me to start up another project soon. I have three or four on the go now. Looks great (other than the caps). I guess the “Legal” or “Normal” under the Paragraph banner might be for paper orientation? (legal or landscape)? I haven’t looked at the manual yet.
On another note, take a look here is you want to take a look at my cassettes. The data is in strange locations unless the spools move really fast. For a moment, I thought the tapes were one sided but there is data on both sides. vintagecomputer.ca
Let me know if you take a look at them. I’d love to know what you find out.
For the record, my outer case looks like an eggshell with several cracks all over it. It’s still holding together and still seems quite sturdy given all of the cracks.
“Legal” might be for paper size: 216 x 356 mm (8.5 × 14 in). In contrast “letter” is 216 x 279 mm (8.5 × 11 in), so it is the same width but a bit shorter. It is more popular, so “normal” might be a reasonable label for it.
Ah, another project would be fun for you!
Yes, my unit looks good, apart from no typewriter, but then it takes up quite a bit of space.
I think the switch is paper size, yes, or maybe characters per line and lines per page? I also need to read the manual more closely…
I had a quick look at the cassette files, thank you for them. I cant really see any data on them. There’s amplitude changes, but the frequency seem smor eor less fixed aroud 120Hz. There are a couple of peaks at 600Hz and 1200Hz approximately, but I can’t see that in the waveforms. I’ll check the docs and see if I can find how they encoded it. I’m wondering if they made their own tape decks and the tracks are in non-standard positions?
The machine is very modular and easy to work on. The only cramped area is where the fans are, between the logic PCBs and the front panel, but that highes up as well, so even that isn’t too bad.
The unit still works as it did before the removals and capacitor replacements…
That relay is missing on mine. No little PCB on there. I finally made some room and fired mine up. I am able to rewind both tapes (top and bottom) and if I play one of them, the typewriter tries to type.
I’ve spent most of my time trying to free up various mechanical parts in the Selectric. I can’t get the carriage to move much. The keys on the keyboard activate something but the carriage only moves 1/2 inch at most and then is stuck. I looked at some videos and I don’t have mechanical keys to try a TAB or RETURN. Anyway, for whatever I can get going, it’s semi-working. The Selectric is the part that needs work for me.
On a couple of occasions, I did start the system, read a tape and it tried to type something. It failed because the carriage didn’t move but at least it tried.
The tapes have data on them. Look at them in Audacity, if you are not using that and you can see where the data is. I used it to determine how best to record the data.
It looks like we have two working consoles between us. I think yours is working a bit better than mine.
I’m going to create a Pico powered gadget to plug in to simulate a typewriter
If I get it working, I can send you one. You can check the console is working then, hopefully.
I did use audacity to look at the tapes. I couldn’t really see anything I could decode. Normally it’s a frequency shift keying signal, but this doesn’t seem to be. I’ll investigate.
I think cot replaced the Selectric insides for solenoids, so you may have an easier job freeing up the typewriter than if it was standard. I’ve seen the insides of one. Complex.
Thank you. I look forward to progress on your project.
I am pretty sure I burnt out the Selectric motor. I freed up quite a bit but without knowing enough about Selectrics, I’m not sure what should be moving and what should not when powered up. I will put it away for now until you have your “gadget” ready for some testing.
Hi,
That’s bad news about the Selectric. Are you sure the motor is damaged? Even if it is, I suspect there are a lot of spares for that typewriter available.
I have parts on order for the typewriter simulator, I’ll update as things happen.
I have looked at the tape files again and have found data. The patent describes the tape structure as one track for ‘codes’ and one for text data. I think thats what you have when you record the two sides of the tape. The data is reversed, I think, and encoded in some form of FSK, maybe without phase compensation. I’m not entirely sure exactly what encoding it is. The patent says the codes are recorded as N pulses for the marks, the formula for N for a mark of X (the seven segment number) is 8*X+4. An example of 02 is given as 20 pulses. I think these pulses are at 4.8kHz.
I’m going to have a go at decoding the text, once I work out the encoding. Well, if I do…
The replacement keys are back. There’s some problems but they are just about usable as is, apart from the wide switches that require some sanding as they are about 1mm too wide. The locating nubs on the inside are 90 degrees out, too, so they are all a bit high and don’t positively locate. But usable and definitely worth a v2. The light shines through differently but ok.
