Bringing An ADM-3A Back To Life

[David] at Usagi Electric ended up with an old Lear Siegler ADM-3A terminal in a trade a couple of years ago. But the CRT face was plagued with so-called cataracts, and the condition of the insides was unknown. The video ( below the break ) shows the restoration process, which went quite smoothly. [David] was relieved that the CRT repair in particular was easy, a fact he attributes to the Texas weather —

ADM-3A Under the Hood

The temperature was 110 F / 43 C when he set the CRT outside to bake in the sun for a few hours. Afterwards, removing the “integral implosion protection” plastic screen went better than expected. Everything cleaned up nicely and the screen reinstalled. Introduced in 1976, the main electronics board is chock full of TTL chips with nary a microprocessor in sight. Fortunately the board was substantially intact, and a single missing chip was found hidden underneath the board. [David] gets the terminal up and running in short order, and is confronted with an annoyance familiar to gray-haired programmers who grew up in this era. Most terminals had different sets of commands to control features such as cursor control and clearing parts or all of the screen. Programs often assumed a certain type of terminal. Some terminals could be configured to behave in different ways, and some programs offered the user a choice of terminals. Today your terminal emulator probably still has a few choices of which kind of terminal to emulate, VT-100 being the most common. And eventually some operating systems provided a terminal abstraction, like Unix’s termcap for example.

If you were around in the era where terminals like the ADM-3A were scattered everywhere, what was your favorite terminal and/or terminal feature? And today, do you have any favorite terminal emulator to recommend? Let us know in the comments below.

22 thoughts on “Bringing An ADM-3A Back To Life

  1. When I was using ADM-3A terminals in school, my favorite feature was “This is not a keypunch.”

    The line editor on the Cyber 7600 really did treat the terminal as a “glass TTY”.

    If I recall correctly, the 3A used “magic cookies” to manage changes in the attributes. Go from plain to reverse video, it costs you a character position on the screen, and another to go back.

      1. The Televideo’s 910, 912 and 914 did have a special mode to make parts of the line and/or screen protected “as in a template was needs to be filled in”. Updating only the fields did make it very fast. Biggest problem was that every type had its own commands :(

  2. I think I still have PTSD from being forced to use one of these in college in the late 80s to write Fortran. Imagine coming from colour home computers and PCs to this horror connected to some big iron. Actually I have no idea what it was connected to. Some sort of IBM mainframe maybe? Never so happy to get back on a PC in final year.

  3. The ADM-3 was a marvel of design. It was one of the teminals we used as examples when we designed the Data General Dasher D200. The goal was under $100 in parts cost, including the display and case. Many (but not all) terminals in those days used purchased modular CRT assemblies which included all the analog and HV circuitry, and were typically powered off 12VDC.

    The ADM-3 uses TTL, we designed the D200 (internal code name “Pegasus”) around a Motorola MC6802 processor. The analog circuitry for the CRT was designed in house by Ken Wolff and Mike Ingemi designed the switching power supply. All on a single PCB. Unfortunately, the CRT anode connection ended up sitting just above the microprocessor, when the case was closed. This did not do much for the reliability. In our defense, it WAS our first attempt.

    1. Ah, I remember those DG blue colored keyboards. There was a DG minicomputer at the local college when I was a boy and because my father taught there he was able to convince the system admin to give me a login. The main thing I remember is that I could add shell commands that everyone would see. Of course I added a shell command for every swear word. In my defense, I was an adolescent boy… The other thing I remember is that the admin smoked constantly and his terminal was so gross I never wanted to touch it.

      Regarding the D200 design flaw. I guess no better way to learn… Did you end up doing something like putting an insulator between the anode and the 6802?

      1. Exactly. I think it was a grounded metal plate between the anode cap and the uP.

        There was a documented way to enter S1/S9 encoded 6800 images into display RAM. We had (sadly, I did not save a copy) Display RAM resident Asteroids and PACMAN games. You needed to install an extra pair of RAM chips to run them though, because we tried and failed to fit them into the production RAM space.

        Attributes were handled by a second 2k of 4-bit RAM, one bit each for Reverse video, underline, blink and…whatever else

        I also wrote code for an APL variant. The first video terminal, to my knowledge, that could display the APL overstruck characters – done using extra character generator images for each of the overstruck characters and some firmware lookup of the incoming character and the character currently in the display RAM — a match would replace the incoming character with the composite overstruck character from the CG, no match would simply overwrite the existing character with the new one. I think they may have sold one or two D200/APLs…APL/VS was not a huge seller :-)

  4. Never had one myself but the ADM-3A is an iconic shape, just looking at it takes you back.
    In the 80s I was using a 6809 based computer running the Flex operating system, and used it with a second-hand Televideo 925 terminal I’d acquired.
    Still occasionaly think I’d like to have something like an old terminal hooked up to a Raspberry Pi or something just for a taste of that old-school terminal using!

  5. Fresh out of college in 1977 I scored a job in the refurb center of Western Union Data Service in Mahwah, NJ. About one quarter of the refurb center was dedicated to repair of the ADM-3A and piles of them came in and out of the plant on a daily basis. I worked on other gear for them, Teletype 32, 40, GE Terminettes, cassette data recorders, and much more. Never did work on the ADM-#A but they were sure popular.

    1. I worked as a programmer there on the Money Order Central Automation system there during those years. I wonder if I ever saw you. I didn’t even know there was a refurb center there are well. I wonder if they were in the same building.

  6. My favorite vt100 and vt220 feature was based on a security flaw in the way the school had the vax 11/780 set up: /dev was 777. Why? I dunno. They weren’t thinking, I guess. Anyway so I could pipe my processes to /dev/ttyxxx, anyone else’s screen, so I could concatenate a bunch of ^G’s and then cat a nice tasteful ascii nude into the file behind them and then cat the whole file to someone else’s screen and the terminal ward would be interrupted by a high pitched ringing bell sound, and everyone would turn to look at the person making the noise, and see the art scrolling up across the person’s screen, while the person is saying “I didn’t do this! I don’t know what’s happening!”

    The admins pretty quickly locked /dev down, as they should have in the first place.

  7. I used an ADM-3A for a while, along with others. The software company I worked for in the 80s had several Prime 750 minicomputers and we all had Prime-brand CRTs. One feature of the operating system, PRIMOS, allowed you to send a message to another user’s terminal. We got pretty adept at hacking each other’s terminals until the next OS release, when the system started stripping non-printing characters from the message text. A sad day!

  8. ADM-3A’s were available as kits in 1977 for 1/2 the assembled price. I was a student tech for a college chemistry department and built three of the kits. It took 25 feet of solder for each if I recall. I was terribly afraid of causing static damage to the expensive MOS uart chip, so took one kit into the bathroom and took a shower to make the air steamy before plugging the uart into its socket.

    The university I teach at now still has a couple ADM-3A’s preserved on a shelf. They were such a huge improvement on DECwriters.

  9. Clearing out an equipment shelter about 10 years ago, there were three very similar ADM-5 terminals. Only had room to rescue one sadly, and it remains part of my ‘stash’. All three did work for the most part, and were still in use before the site got upgraded

  10. I remember ADM-3As at University of Maryland in the early 1980s. Connected to IBM mainframes, initially line printer mode (better than the Decwriters at least) and then they connected them through a Series/1 for 3270 emulation. My first video terminal ~1982 was a Wyse 100. Absolutely beautiful machine – detached keyboard in a metal case. Later got a multiuser Altos 586 (I have a twin brother so we got a 2nd terminal and a multiuser system) and the Wyse 100 had Word Star EPROM available to make the keyboard work well with Word Star, though I memorized all the Control Codes anyway.

  11. I have a few of these; when I bought them the CRT problems that they had were screen burns, not cataracts. Too bad you can’t cut THOSE off. I was lucky to get new CRT’s (in glorious bluish white, NOT “modern” amber, green, or paper-white) to replace them with. Don’t know if the company I obtained those from (and a few VT-320 tubes, also to save “burn victims”) is still around.

    I’m probably going to sell the ADM-3A’s soon; can’t take them with me. I had intended to design a “screen saver” in hardware for them (someone should do this, it would be fairly easy). I used them rarely due to screen burn paranoia. I’ll bet they have <10 hours on the CRT's.

    Fun facts; when they were new, memory was so expensive the "poverty" model displayed only 12 lines and upper case only! More lines? You needed 7 more 2102's. Lower case? You needed a special lower case version of the 2513 character generator. One of mine did not have lowercase, so I basically copied the 2513 upper case ROM with a homebrew adapter from the one that did, blew that into a 2716 UVEPROM, and wrapped an adapter board for the footprint of the 2513-LC. Those -LC chips were not easy to find/expensive 15 years ago and probably available only in unobtanium now.

    On a 1970's HP terminal, I DID have a cataract problem. That CRT was very oddball, "wide" and very, very expensive to replace even 15 years ago. I built a box out of plywood with which to mount the CRT in case of implosion and I cut the face plate off using a nichrome wire with screws/wooden handles and an XT power supply to heat the wire up with (12 volts). I did not have to get it "red hot", at a certain temp it just cut like butter. That seems safer than the gouging out procedure shown in the video.

    I HATE working on TV's/CRT's but you do what you gotta do.

    Anyway, after cleaning, I glued it back on at the very edge (out of view) with a thin layer of aquarium-grade RTV. I'm not sure that I like the tape used in this video. At the very least, I would use automotive grade double sided tape (like that used for mounting badges and emblems).

  12. I hated these in the late 1970s or early 80s. I don’t exactly recall why, but for some reason they were much inferior to VT-100s or other similar dumb terminals. Maybe it was keyboard quality.

  13. When I was in college in the early 80’s, we had one of these in our ACM-student chapter “club room”. A fun prank was to rotate the connector for the horizontal coil on the CRT yoke 180 degrees, which gave the appearance of seeing the text from “inside” the CRT.

  14. I remember one of these (or an ADM31) in Weil Hall at the University of Florida that someone had hacked to include a two digit LED column display. I assume that was useful for OG Fortran and other position dependent coding.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.