Retrotechtacular: A Mechanical UART

We’ve heard it said that no one invented the old mechanical Teletype. One fell from the sky near Skokie, Illinois and people just duplicated them. It is true these old machines were similar to a modern terminal. They sent and received serial data using a printer instead of a screen. But inside, they were mechanical Rube Goldbergs, not full of the electronic circuits you’d think of today.

Teletype was the best-known name, but there were other mechanical monster terminals out there. [Carsten] recently took some pictures of his 99 pound Olivetti mechanical terminal. According to him, there’s only one electronic component within: a bistable solenoid that reads the data. Everything else is mechanical and driven with a motor that keeps everything at the right baud rate (110 baud).

Like the Teletype, it is a miracle these things were able to work as well as they did. Lacking a microcontroller, the terminals could respond to an identity request by spinning a little wheel that had teeth removed to indicate which letters to send (TeleType used a similar scheme). Things that are simple using today’s electronics (like preventing two keys pressed at once from being a problem) turned out to be massive design challenges for these old metal monsters.

Turns out that when [Carsten] last fired the terminal up, a capacitor finally gave up its magic smoke. He plans to fix it, though, and as long as it isn’t a mechanical problem, we bet he will.

We’ve talked about Teletypes a few times in the past, including using them for text messaging and even Twitter.

44 thoughts on “Retrotechtacular: A Mechanical UART

  1. Finally! Something I can relate to. I repaired Model 33 Teletypes in college, and developed an appreciation for their intricate mechanical solution to receiving and printing serial data.

  2. My first access to the internet was via teletype. I found a line in the dorm room that we snuck into the computer lab and patched it into one of the unused rs232 ports on the DEC VAX-11 then via several telephone closets across campus jumpered unused phone lines back to the room. 110baud login and access to UUnet with a broken teletype that we repaired.

    Later that year someone liberated a Dasher D200 terminal from somewhere and we were on the crazy train to high tech!

    1. I really like that story! Thank you.

      It reminds me of a guy I knew in college. He somehow scored a VAX in 1992. He kept it in a bathtub in the dorms. Since it was the dorms, he didn’t have to worry about electric bills.

      He was so proud of this thing, in a way that would infect me as I got older. He had changed which program opened one of the common commands, and was juiced: “I have that power!”

      I have more unplugged computer power sitting on my desk than we had on that computer engineering dorm floor in 1992. I still miss those days.

      1. A few years later I was in college studying computer science myself. Judging by a lot of my classmates I was still the odd one out among that segment of the population to use a bathtub for it’s intended purpose.

    2. Similar story here. I had a Teletype in my dorm room. Spent $350 on a 300 baud modem, and phone calls to the campus mainframe were free. As I worked for the computing center, I had an unlimited login. I even programmed the HERE IS drum on the TTY to enter my username and password.

      When I got into grad school, that job ended, but I worked at DEC over the summers, and built up a VT05 terminal from scrap parts. From 110 baud to 300 baud! Those were the days :-)

      1. @Greenaum and TimGray1 – I’m pretty good at W.A.G. – tell me if I’m in the ball park on “how’d he do that…” Someone, probably Timothy Gray, was a volunteer student-assistant to the IT professor in charge of the VAX ROOM (Data Center or DC). This person had the master-key (or bribed the janitor to help out) to open up the MDF near his dorm room and the MDF near the DC, Then after doing the re-jacking of the RJ45 UTP-CAT-5 cables on the MDF racks (which BTW drives the telecom guys nuts!) or they may have been using BNC cabling back then. Then same deal with the DC connection to the VAX. “Hey Tim what are you doing back there?!” “Oh nothing sir… I just lost my dad’s gold plated Cross pen. It rolled back here somewhere…” “Oh OK Tim just make sure you turn off the lights when your done… OK?” “Sure thing teach! (giggles maniacally while plugging in his bootleg plug)…”

        I did the same thing at private industry and got caught every time by telecom guys!!! :-P

        1. Yep but that doesn’t explain how he got a user account. You have to have something to type in at the Login: prompt. On Unix that needs Root to create an account for you, I’m sure Vaxen have a similar concept.

          Then again, physical access… Maybe it had a single-user mode you could boot it into, without needing a password.

          1. The professor probably gave him a user account so Tim could do routine backups, or write code off-hours so he didn’t have to be there to type his credentials every time Tim got the itch to write code or something like that. Or maybe he looked over a user’s login at the DC acting like he was cleaning up the workstation behind him? Or better yet he mounted a CCTV camera over the keyboard or dusted the keyboard for prints using a UV lamp. Straight out of that ridiculous TV show SCORPION??? 8-D

  3. “Like the Teletype, it is a miracle these things were able to work as well as they did” Nonsense, these devices were designed at a time when things were designed by engineers, not sw hacks. Designs were tested before committing to production, there was no “we can do a sw update later” mentality
    I fixed KSR33’s used as console printers on Univac III as well as terminals. Also worked on Tab equipment, again mechanical wonders that were for runners of electronic computers. I often wonder where those designs skills went to, certainly a lost mind set..

    1. We had a KSR33, connected by an acoustic modem to the local college’s Univac 1110. Flat out it could handle about 30 interactive terminals, but even 110 baud was better than hand writing coding sheets, getting them punched (inaccurately) to card, correcting and bug fixing via batch jobs (and a 1 week turn-around).

    2. It’s still very impressive that a UART, decoding from serial to alphabetic, and doing the same the other way round for the keyboard, was all done in mechanics. It’s amazing really! I bet there’s not much around now that’s as complicated mechanically. How would you store the incoming bits for one of those?

      1. Seems more complicated than that. The old mechanical devices were designed that way because it was either the only way to do something or it was cheaper. Engineering rigor was required to make the mechanical assemblies reliable and manufacturable, with little POSSIBILITY of future changes or upgrades.

        Today it is different. Integrated circuits are extremely inexpensive, mechanical devices are not so inexpensive, so there is different price points being aimed for.

        People and companies still use 20 year old thinkpads as dummy terminals, for instance.

        Planned obsolescence is a scourge on our society, but engineering is much more sophisticated now than it was in the past, and not all device are designed for obsolescence – especially in the price range of this old equipment being discussed. Look into the mentality going into a 8-bit home computer of the day. It was certainly more cynical, for a biz point of view, than even planned obsolescence.. IE “WE GOTTA GET SOMETHING ON THE MARKET, WHO CARES IF IT SUCKS SO BAD WE BASICALLY ARE RIPPING OFF OUR CUSTOMERS”

    3. I laugh at the romanticized stories about the good-old Teletype days. Gerritv says that Teletypes were designed back when things were built right, but he also says that he fixed KSR33s. Somebody else above ALSO fixed Teletypes. We had four ASR33s in a math classroom at my high school, and about once a month a tech had to come in and deal with some problem with one of them. If these were such great machines, why are there so many people who remember having to repair them? It really IS a miracle that these things were able to work – they were mechanical nightmares that relied on precise lubrication (aka “dust collection systems”), and NOBODY missed them when the likes of the Datapoint 2200, Hazeltine 1500, and LSI ADM-3a stepped up to provide a faster, quieter, more reliable solution.

      I think that it’s only BECAUSE these machines were such mechanical nightmares that they tend to produce nostalgia in people who had to deal with them. And the same goes for the IBM Selectric typewriter, I have to say.

  4. Not only were they workhorses, they also were a bear to fix when paper got jammed. We had one in high school for sending our punch card RPG2 and COBOL programs to the mainframe downtown. My lab partner (we were in charge of sending the jobs and putting the right cards with the right program) decided one day to play stupid and caused a massive paperjam….

    5 hours later we had it cleaned out and running. Just took every prying tool we could lay our hands on to dislodge all the paper that the innards digested.

  5. My days of learning DOS also involved learning modems and terminal emulator software for calling out to land line BBS systems.
    THESE are among the terminals emulated by that software.
    I have never had the privilege of working on or even playing with an electro-mechanical terminal. They were my older brother’s toy’s not mine unfortunately.
    I do recall; however that more than one of my brother’s friends had an old mechanical terminal they rescued from one place or another sitting in their garage or attic. They were prized by the “olde guarde”.
    I still see them once in a while, usually in an advanced state of decay at a flea market.

    1. No, that other article refers to a daisywheel of some sort, which was a completely different beast.

      I /had/ one of these Olivetti terminals, I believe that a lot had been scrapped by a UK bank or building society and some ended up in a garage in Bletchingley. Where the Teletype ASR or KSR 33 had a bit of flimsy bent metal these things had a turned or milled component: very /very/ well made but abominably heavy, from one of their technicians I understood that the maintenance training course was 12 weeks.

      The mechanical Teletype, Friden, Olivetti and Marchant terminals (listed in approximate order of popularity) really did belong in another era, and one I have mixed feelings about. On the one hand was the sheer ingenuity with which problems were solved- very often multiple times to avoid each others patents. On the other was the fact that they were slow, noisy, and in some cases- I’m looking at you, Burroughs- would piddle oil all over the carpet when upset.

  6. Bad capacitor??

    First rule of powering up old equipment like that.. do it slowly.

    For simple AC powered stuff without power regulators (like antique radios) just plug it into a variable transformer. Turn the power up slowly. For things with regulated power supplies built in you may have to disconnect the power supply. Power the supply, by itself slowly. Meanwhile connect a variable supply to the device and power it up slowly too. For more complex stuff you may have to just desolder the capacitors and power them up slowly with your bench supply.

    What happens is there are aluminum plates inside the capacitor. Those plates are supposed to have a layer of aluminum oxide which acts as an insulator. Over years of not being powered the oxygen gets released turning the oxide back into aluminum. This lets more current through, your capacitor and possibly other electronics get cooked.

    Powering up the capacitor slowly caused that layer of oxide to reform.

    Of course.. depending on what level of authenticity you are going for another popular method is to just replace all the capacitors with new ones BEFORE you apply any power at all. That gets you the best working device in the end because those old capacitors probably aren’t very reliable or on-spec now anyway. You would want to do this before you power the device up to make sure you aren’t breaking any other components.

    1. The cap that blew is not an electrolytic capacitor. Electrolytics lose their oxide layer. But it can take a couple of hours to reform that layer, and you can only do that by limiting current (not voltage). Variable transformers are useless by themselves, since they have no current limiting. Often it seems to work, but that’s mostly because old electrolytic caps aren’t necessarily very bad.
      The cap that blew is a paper-type. Those get leaky and can’t be fixed. Nothing that can be done to prevent that.

  7. The first computer I ever programmed used a Friden Flexowriter for its primary I/O device. High speed mass storage was a 100 character per second (or so) paper tape reader/punch the size of a refrigerator.

  8. My parents had a big yellow beast of a teletype just inside our foyer. A monsterous thing they used almost daily. Received or transmitted messages could also be recorded on ribbons of paper that could be later replayed. I remember very little of it since I was far more interested in emptying out the little tray that held the chad from the tape. But it was far bigger than even the TV we had at the time and we treated it like a piece of furniture. My parents only required repairs once, calling a technician out for the repair since the machine was insanely heavy to move.

    I asked my mother about it years later and my mother recalled that it was a military surplus they obtained years before. They got rid of it when suitcase models became attractive in price (some time around mid 1980’s). As near as she can recall AT&T took it.

    I know nothing about it and have only seen its like once. About five or so years later at the Bureau of Land Management sitting in a corner in one of the offices. It was covered in a gray dust cover. But the sheer size, color, and shape was unmistakable as well as the chad from the recently dumped tray on the floor.

      1. @Greenaum – From the clues SavannahLion gives I will take a WAG at it:

        BLM DoI employees with the BIA division out mid-western USA somewhere. Maybe in the Dakotas (i.e. Wounded Knee, etc)? Probably an ASR teletype dedicated to link back to DC for bureau updates and such. Probably a AT&T (Ma Bell) copper pair dropped down from local regional office in Belle Fourche? That would mean his parents where maybe BIA Commissioners, Superintendents, or Police? Indian Agents were sunset by POTUS-26. During the 1980’s that place was literally under sige by FBI and AID. Google: Leonard Pelletier who is STILL in jail over that! Mr. Obama MAY pardon him (or not), Not even close SavannahLion ???


  9. @Al Williams – Believe it or not the FIRST Teletype machine was invented by a very unusual guy named Royal Earl House in 1830’s. This thing was amazing! He had learned of the telegraph that used Morse Code. This guy some how devised a machine of amazing intricacy that could send 56 alphanumeric characters from a 28 key marked piano keyboard with a shift key to a distant receiving terminal over only TWO WIRES! Normally we’d think he did it with 28 or more wires but he used the standard telegraph wires of the day. Some attribute BINARY coding concept to the famous Alan Turing, but Mr. House was using a similar 0 and 1 scheme in mid 19th century. He even actually invented a audio telephonic device BEFORE Alexander Graham Bell but lost the lawsuit over that. He and his nephew invented some pretty interesting things that seem very out of place for their time even better than Tesla’s gadgets of the same period. Here is the patent: US4464

    Here is a picture of it at the Smithsonian:

    His bio can be found here: Type in “Royal E. House, the Electrician” on GOOGLE BOOKS. Look for January 4th 1889 article in the Telegraphic Journal and Electrical Review, Volume 24. Note preceding Mr. House’s invention was the simple printing telegraph which only printed dots and dashes to a strip of paper. His invention printed complete words (English characters) on a piece of paper just like a teletype does. He even invented repeater stations to boost the signal over longer distances. Others came along and expanded on his idea and made it better but not more fascinating as his original model.

    His next idea was this device powered by steam rather than by wet cell batteries.

    1. If anyone can figure this thing out then good luck to them. Evidently it worked and worked well. It could type up to 40 wpm. A Mr. Phelps mass produced it out ofTroy NY and many concerns used it successfully. There was even a telecomm link using it between Philadelphia PA and NYC in late 19th century. Imagine if POTUS Lincoln secretly used this thing to coordinate military ops during Civil War. This could have been used to communicate with Union spies in the field. All they needed was a telegraph pair of wires dedicated to their station. Using a 18th century POTUS Jefferson Cipher Wheel (steganography) it could have been encrypted too, I just learned about this thing. Imagine what else we were never taught in school!

      GOOD LUCK figuring this thing out. It’s mechanically complex:

      1. Mechanically complex, but conceptually simple: it works using a pulse-position code. Drums on the transmitting and receiving ends are synchronized together (didn’t read closely enough to catch how that was done on this device, but on Teletypes this was done by a clutch activated by the start bit). Each key on the transmit end was “sampled” by a pin on the rotating drum, so the length of time between the synchronizing pulse and the “data” pulse was determined by which key was pressed. On the receiving end, the print wheel rotated in sync with the transmit drum, and the “data” pulse activated the hammer that pressed the print wheel onto the paper. MUCH simpler than the later Teletype mechanisms, but also much slower.

    2. Here is a good bio on Royal E. House (his obituary in 1895):

      It contains the ONLY known photo of him as he was a very private fellow. This paragraph was inspiring:

      Mr. House possessed keen powers of observation, great originality of mind, and extraordinary tenacity of purpose. He was a man of vigorous physique and attractive personality. He was in full possession of his faculties to an advanced age, and retained in his memory the minutest details of his diversified and eventful life. His first patent bore the early number of 1,200. His last was No. 533,600.

      Royal would have LOVED HaD and been one of it’s foremost posters!!! :-D

  10. The mechanical beauty inside teletypes are a marvel relatively unknown to the younger generation, like IBM Selectric typewriters. I’d like to see more of a resurgence in preserving both before more of the remaining examples are carted off to the scrap bin. I’ve a non-functional Selectric, and it’s above my current knowledge level (perhaps intellect as well) to work out its intricacies.

    1. packrat – IBM Selectric Daisy Wheels STILL fluster me to no end! I have yet to figure out the magic behind how they work. Admittedly I haven’t tried that hard though. The way it spins up and impacts the ribbon and paper just at the correct character baffles me. I used to think the old IBM mainframe Nixie tube displays where baffling but this takes the cake!

      I remember back in the 60’s playing tic-tact-toe on a IBM mf at a science museum in Philly. Next to the exhibit was the MA BELL future exhibit, They had a videophone connection going from Philadelphia (where I was) to San Francisco. The caption said “Everyone in 1975 will have one!” NOPE never happened… However I just watched a girl have lunch at KFC while having a videophone conference with her boyfriend! That was November 2015.

      1. I realize I’m probably picking a nit here, but in case you weren’t aware, only the last versions of IBM Selectrics used a daisy wheel. The “traditional” Selectric design used an interchangeable “golf ball” style typeball with four rows of type on it. Similar to the Teletype model 33, which also had four rows of type, but where the Teletype moved its typehead up and down and rotated it on a vertical axis, the Selectrics rotated in two axes.

        And by the way, the multi-level type cylinder or ball was not invented by either IBM or Teletype, but by George Blickensderfer in 1892. It even had interchangeable type cylinders to allow it to be easily converted from one language to another.

        And yeah, 2015 – 1975 = 40 years of error. This probably means it won’t be until 2040 that we get the personal jet packs and flying cars we were supposed to have by 2000.

  11. I have read here numerous stories about ASR33s and KSR33s, and similar. Anyone here familiar with the old Model 14 or Model 15, that used five-level Baudot code instead of ASCII?

    1. When I was a member of W5YD (Miss. State Univ. ham club) we had a 15 and a 19 and I used to keep them running. The 15 was build during the war and most of the parts in it were made from zinc to save steel for the war effort. Those parts were not as reliable as “the real thing”. While typing this, I kept instinctively trying to find the FIGS and LTRS shifts ;-)

  12. In my TRS80 Model I days, I disassembled the rom and dumped the output to a a teletype that I trash picked. it was so noisey I had to wad for the week my mom and dad when on vacation. Ran day and night for nearly 4 days. The result was a four inch stack of printout that was a hacker’s heaven! The only time it had a problem was when it ran out of fanfold paper and went the vibrations jarred the serial connection lose.

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.