• Please review our updated Terms and Rules here

Interfacing IBM Selectric to IO Port?

If you had walked into a Heathkit store to purchase a brand new H-11 back in the late 70's, the hard copy device they could have sold you with the system would have been the H36 which was a rebadged Decwriter II (LA36).

I have a Fall 1978 DEC Sales Catalog and the list price of a bare-bones LA36 DECwriter II back then was from $1595 upwards. With optional extras of a 14-key keypad, 20mA current loop interface, paper-out sensor, extra ribbon and 90-day on-site warranty it was $1825. The faster LA120 DECwriter III was even more expensive starting at $3072.
I have no idea what was Heathkit's price for the rebadged unit, but according to the inflation calculator $1595 is over $5800 in today's dollars so it's no wonder people were looking for all sorts of cheaper printing alternatives.
 
I have a Fall 1978 DEC Sales Catalog and the list price of a bare-bones LA36 DECwriter II back then was from $1595 upwards. With optional extras of a 14-key keypad, 20mA current loop interface, paper-out sensor, extra ribbon and 90-day on-site warranty it was $1825. The faster LA120 DECwriter III was even more expensive starting at $3072.
I have no idea what was Heathkit's price for the rebadged unit, but according to the inflation calculator $1595 is over $5800 in today's dollars so it's no wonder people were looking for all sorts of cheaper printing alternatives.

Ya no kidding! It must also be the reason why the Heath 11, and it's companion peripherals are so rare in general. The floppy was $2595 assembled back then. The computer was $1295 in kit form...then hardcopy for another $1600 and terminal for $800. All together you could put a down a sizable down payment on a house, or get a fully loaded H-11. That's not even including memory expansions, and extra cards you would want/need, and the operating system! Thankfully, at least the operating system came with some decent languages, like a very capable BASIC.

I suppose technically you could get away with buying the base unit, using HASSL (ironically named) assembler, and a kit paper tape machine with an homebrew I/O selectric. You then would need one parallel card, one serial card and a printer terminal. That would still be expensive, but you would have some crazy computing power in comparison to most individual people back then, but most didn't do that. Point being, if you got one of these...price was clearly no object, and it shows in the examples of the machine which remain. Almost all of them I have seen are nearly fully populated bus systems, fully expanded memory, with terminals, disk drives, paper tape, and a slew of other goodies. Probably because they were going to buy a DEC and got wind they could get one for thousands less from Heath, so they went nuts and bought all the extras.
 
I knew of a few H-11's sold in northern New Jersey before 1980 and all used third party printers. The fully loaded complete H-11 was a good value at the time, not that much more than a S-100 system with dual 8" floppy drives.
 
I knew of a few H-11's sold in northern New Jersey before 1980 and all used third party printers. The fully loaded complete H-11 was a good value at the time, not that much more than a S-100 system with dual 8" floppy drives.

I guess when you put it that way, ya. Did most people finance these computers? I saw a few companies were doing that in magazines. 78$/month for 12 months or whatever.
 
Selectric IO

Selectric IO

Around 1979 I purchased a worn out IO Selectric used, I think, by the airlines to print tickets. It had solenoids under the keyboard to trip the "interposers" (think that is the term for the cross rods under the keyboard). The tilt and rotate actuation did not match a standard desktop selectric so standard type balls were not compatible. I rearranged the parts under the keys that had the interposer/tilt-rotate coding (mechanical tangs) so that it was standard. Built a driver board to interface the 48V solenoids. Then wrote a CPM program to convert from ascii to tilt-rotate codes. It worked. Might have the code around somewhere but it was pretty simple - just keep track of the shift state and then a lookup table. I think I made it part of the bios (on a BBII) but don't remember all the details. But the unit was very worn and never worked well. I tossed it out in a move many years ago but still have the allen type wrenches (bristol?) and the wheel for manually turning over the selectric mechanism. No longer any use to me so should look at selling them. Was quite an interesting project.
 
Wang Laboratories used their own patented modification to IBM selectrics for printers on their programmable calculators and other computers. You might be able to find those around, I have seen them occasionally. They should not be too difficult to interface to a parallel port. Most allow keyboard to be used (offline), some even as a complete terminal (full I/O). They were called "Output Writers" or "I/O Writers" in Wang documentation.
 
I probably over-simplified the use of a Wang OutputWriter. The parallel port interface would be relatively simple, but the software is not so much. The device uses a 6-bit "rotate-tilt" code (not ASCII) plus the computer needs to manage SHIFT state and, of course, delays for mechanical operations (as I recall, the one handshake signal does not cover the time it takes to perform the mechanicals). I worked out most of that for the Wang simulations I did (no mechanical delays required), but it would still be "interesting" code to write for CP/M. "Fun", or not, depending on your point of view.
 
I probably over-simplified the use of a Wang OutputWriter. The parallel port interface would be relatively simple, but the software is not so much. The device uses a 6-bit "rotate-tilt" code (not ASCII) plus the computer needs to manage SHIFT state and, of course, delays for mechanical operations (as I recall, the one handshake signal does not cover the time it takes to perform the mechanicals). I worked out most of that for the Wang simulations I did (no mechanical delays required), but it would still be "interesting" code to write for CP/M. "Fun", or not, depending on your point of view.

Many of the designs I have seen use an EPROM to convert from ASCII to tile/rotate code. There are 4 rows of 22 characters on an 88 character ball, so that's 2 bits for the tilt and 5 bits for the rotate. The spare bit can be used to trigger the other functions....
 
I was reading this thread and looking at the links and thought it would be cool to have a golf ball typewriter for a terminal running focal on the ohio scientific 560z. After doing a pile of research I decided the easiest way would be to use a ibm electronic 50-85 series typewriter. They are a hybrid that squeezes in between the completely mechanical selectric and wheelwriter. After searching ,a local fellow answered a add i had , he had a electronic 50 he wanted to give away. When I picked it it up he also had a selectric ii from 1974 so I grabbed that too. After going on some typewriter groups I soon discovered that those 50-85 series are very unreliable .hard to fix and parts are unobtainium. But to my surprise that selectric ii had a computer interface already attached to it.
It is made up of 14 solenoids and 17 switches that control all the mechanicals. These typewriter are a mechanical nightmare, and super complicated arrangement of clutches ,gears and pulleys. It so happen we have a guy around here who still repairs these things, so I took it to him to get it working. He said it not worth fixing your better off buying a refurbished machine.well I decided I'm gonna fix it and I did, after hours of reading ,hours of videos ,hours of cleaning and de greasing, some new parts its almost as good as new. Next step was to figure out how to interface to the computer, there is a control board on the unit for solenoids and switches. I designed up a circuit to connect to the computer and wrote some firmware. I am just waiting for the cable and I'm ready to test.
A.jpg
B.jpg
C.jpg
D.jpg
E.jpg
F.jpg
 
Well done! That's no small feat!

Any idea if the computer interface was standard equipment, or from some third-party? any info on what the pin-out, and protocol, is?
 
Back
Top