Model 15 RO Teletype available (OT) [telecom]

Some time ago, I posted a notice seeking sources for old Teletype machines. I've just had a call from a man who has one, and in fairness, I feel obligated to publish his info.

Mr. Dave Armon dave.armon.at.gmail.com cell 917-863-1008

Model 15 RO, probably with "66 speed" gears. Formerly used by UPI. This is a preasure-feed unit, i.e., it takes regular rolls of paper, not the pin-feed kind.

Model 15D motor 250V

Serial bpi82/97 Motor S/N 53476

The unit is in Manhattan and available for pickup there. There is no table.

(Filter QRM for direct replies)

Bill Horne

Reply to
Bill Horne
Loading thread data ...

I always thought it'd be neat to have such a machine as an I/O unit connected to a PC. There were companies that made interface boxes so as to connect from the PC I/O to the Teletype's I/O, though I don't know if such boxes handled the ASCII-Baudot conversion. (That's tricky because Baudot requires shifts for letters and figures, so it's not a one-to-one character conversion.)

Unfortunately, space limitations prevented me for acquiring a Teletype. Also I figured that the machines would need periodic mechanical maintenance which I'm not skilled at.

But the image of the Teletype clacking out important messages or news has mostly faded away from the popular conscious. Years ago email replaced it and today cellphone texting supplanted that (see separate thread).

J.

***** Moderator's Note *****

I learned to type on a Model 19 Teletype connected to a Ham radio transmitter at the M.I.T. radio club, W1MX, in the Sixties[1]. The machine required both a sense of timing and strong fingers: the keyboard was locked during the time when the previous key-choice was read by the mechanism, so producing consistently high speeds meant synchronizing your typing speed with that of the machine, and it took more force to _start_ a key in motion than it did to _complete_ the trip: each green keycap had a spring under it to absorb the typist's force as the key bottomed out.

I think the converters that Jeff speaks of were current-loop to RS-232 devices, which only changed the "physical layer" signalling from that used by telegraph[2] lines to the bipolar interface specified by RS-232 for computers. Converting from the ASCII code most PC's use to the ITU2[3] or other codes used by Teletypes is almost always done in software, and it's a surprisingly complicated process: the Teletype can't print many of the characters available on computers[4], so the programmer must make choices about which ones to substitute. I know: I coded an interface like that once (although it was for an Anderson-Jacobson machine that used EBCD, not a Baudot Teletype), and although I was able to impress my professors by handing in papers with justified right margins, it took me a week of effort to get CP/M to do the trick.

  1. Although I sometimes wish that I had acquired skill on a Dvorak keyboard instead, they weren't generally available until well into the 1980's. C'est la vie.
  2. Yes, that's right: telegraph. Teletype machines were invented to give Western Union and other telegraph companies a way to replace their high-priced telegraph operators with machines that any dolt could operate: being automated out of a job is, you see, nothing new. Since Teletype machines were used to replace telegraph operators, they use the same signalling methods that the telegraph used: "Mark" (key down) means current flowing, and "Space" (key up) means no current flowing. Ergo, the "current loop" signalling interface that is standard on all Teletype mahines.
  3. According to Wikipedia, Baudot's code was never used by Teletype or similar machines that had a "typewriter" keyboard: a man named Murray adopted Baudot's code to decrease wear-and-tear on the tape punching mechanisms, by assigning the most often used characters to "mostly spacing" codes, since only "marks" were punched. Western Union modified Murray's code to suit itself, and the result became the CCITT International Telegraphy Alphabet No. 2 (ITA2) which is used on TELEX machines to this day. Like many other historical terms in the telecommunications field, "Baudot" has survived as the common name, although some purists insist it should be called "Murray code" or "Baudot/Murray code", even though neither term is any more accurate than just saying "Baudot".
  4. There is, for example, no "@" (at sign) on the Teletype keyboard (although the symbol has been integrated in the International Morse Code), and no "\" (backslash).
Reply to
Lisa or Jeff

Are you referring (above) to the Model 19?

My TTY ASR33 definitely had a "@" as shifted above the "P" as can be seen in the following photo which is identical to the one I had at home since I worked for Tymshare 1967-1972:

The "@" was an important character since it was used in email addresses such as these (for example): rms@mit-ai, thad@sri-kl, etc.

The 33 didn't have a "\" because there wasn't any real need for it since Microsoft didn't exist back then. :-)

Reply to
Thad Floryan

I believe that in the telegraph world almost all transmissions were prepared ahead of time off-line by pre-punching a paper tape, then transmitting the tape at full speed. This of course also allowed for editing and correcting of typos. Every description of machine usage refers to this practice.

When Teletypes began to be used as computer terminals to time sharing systems, often times programs were pre-punched off line to save on expensive connect time. But during debugging the users could work 'live'. Also, as rates went down it wasn't as critical to work off- line.

Some computer time shared systems used full duplex but there was a slight delay in the echo-back of the character. It was disconcerting to be typing and having the character print a moment after you typed it.

Even the newer Teletype 33/35 keyboards were not easy to use. Some of the numeric shift characters did not match standard typewriters and the keyboard was 'heavy'.

Telegraph messages were extremely terse. The header record to a message consisted of a few letters, each character a code representing something. In the 1960s when telegraph systems (like Telex) were first connected to computers to get information, the command langauge was also extremely terse. Businesses and newspapers that were heavy telegraph users made use of standard or company shorthand codes, for example POTUS meant President of the United States. Overseas cables were heavilly coded to save characters.

While today I like the user-firiendliness of modern Internet screens, some websites are far overloaded with bloat, and even with fast connections they take a long time to appear. For some reason the websites of TV networks are especially loaded with bloat (eg extensive "flash" and animated graphics) and are slow to use. If your PC's software isn't up to date, stuff doesn't work correctly.

Yes, I was referring to RS-232 to current-loop conversion. (I remember getting a notice in the phone bill some years ago announcing they would not longer support local loops with current-loop service, which meant any remaining old style Teletypes would have to go modem or be cut off.)

Anyway, I believe a company called "Black Box" made interfaces. I don't think a code conversion was necessary since the newer models

33/35 TTY used ASCII. But your program had to shoot data out slowly.

Looking at my surviving printouts from my TTY 33 days* I find that the print quality wasn't so good even though it was an essentially an electric typewriter. Many high speed computer printouts from other than IBM 1403 printers also look poor with poor letter registration or unevenness. IBM Selectric-style terminals had better quality.

Oslin's WU history implies that their network was wholly teleprinter by 1950. However, other references, such as the WU employee newsletter, suggests here and there Morse operators remained in service on a few trunk lines until the early 1960s. I know WU wanted to paint itself as a modern high-tech company so it wasn't gonna mention it still used Morse operators.

The WU Tech Review**, on discussing ASCII, was concerned about the excess mechanical wear the new code would cause on its machines (two extra bits per characted). At that time (1963) WU was taking a cautious look at ASCII and thinking it's Murray/Baudot would remain the code of choice for a long time.

  • I had a huge box of printouts of old BASIC programs but space limitations (and the reality I'd never use them again) forced me to trash it. I did retain the first program I ever wrote. (Ran clean on the first shot, but it's been downhill from there.)
**Available on the archives of this newsgroup.

J.

Reply to
Lisa or Jeff

I have a friend who is an absolute nut of a Dvorak promoter - he will bend your ear for an hour if given the chance. Seems like I have read recently, however, that the studies showing the superiority of Dvorak were flawed and biased, so there may be no advantage at all.

In any event, it's always been my feeling that with few exceptions the amount of time anyone spends actually pounding keys is insignificant compared with all the other things a person does, so an increase in typing speed doesn't translate into increased productivity.

Whether the Morse code telegraph operators were high-priced is open to question. Certainly they would be today, but back then...The book by Edwin Gabler, "The American telegrapher : a social history 1860-1900" states there was always an oversupply of Morse telegraphers, tending to drive down wages. It was denounced as unethical for a telegrapher to teach telegraphy to anyone but an immediate family member; yet many telegraphers did it as a way to boost their income and acquire a helper cheap. Some of the operators were a lot better than average and did command higher pay in consequence.

Yep, ITA #2 is Murray's code, not Baudot's. However Baudot certainly deserves a lot of honor as the inventor of the basic idea of a fixed- length binary code for telegraphy. In efficiency it was a step backward from Morse code; but in making an efficient machine printer possible it was a great leap forward.

While we're at it, Morse code is actually Vail's code; and a lot of people think Vail was robbed of the credit he deserves. However I think we have to cut Morse some slack. Morse did realize that electromagnetism could be the basis of a practical long-distance single-wire telegraph. He did realize there was a market for the service. And the modern concepts we have of technology-driven corporations having a stable of engineers and patent assignment contracts were unknown in Morse's day.

Reply to
Jim Haynes

Most high traffic machines were ASR, where the keyboard punched a tape that was fed through a distributor. They did not require the ham-handed attack on each key that manual keyboards did. The touch was much like that of computer keyboards. It also provided a steady and uniform transmission that could not be achieved by manual typing, perhaps making the real output almost double what you could achieve on a manual keyboard. The teletypewriter required an operator only at the sending end. Morse operators were required at the receiving end, too. There were many more people with typing skills than ability to use code.

Wes Leatherock snipped-for-privacy@aol.com snipped-for-privacy@yahoo.com

Reply to
Wesrock

When a Morse telegraph line was converted to teleprinter, did the physical line need to be upgraded to properly handle the bits from the teleprinters? Even at the slow speed of 60 words-per-minute that's a lot of bit* per second to move over a coarse open wire with ground return. Were the relay repeaters sensitive and fast enough to replay teleprinter bits?

(Teleprinters require bits with a good waveform--as the bit waveform distorts the teleprinter can't make it out and errors occur.)

  • Let's see, that's 1 word per second, or 5 characters per second, or
25 bits per second, right? (not counting start stop bits)
Reply to
Lisa or Jeff

He's talking about _any_ TTY that usedd '5-level" code, as distinct from '8-level' (ASCII), like your model 33 used.

There were many more models, and numbers of units, made by Teletype Corp over the years, that were 5-bit devices, than 8-bit ones.

Reply to
Robert Bonomi

It appears that, while the part of the company handling customer messages was all teleprinter, there was still some use of Morse by the maintenance people. I've heard the same about AT&T - at the telegraph test boards they had keys and sounders and the older employees knew how to use them. When you're doing testing and maintenance the simplicity and portability of the key and sounder are worth something, if you still have people who know how to use them.

The railroads continued to use Morse right into the 1960s. In contrast to Western Union, telegraphy was a tool in running a railroad rather than the reason for existence of the company. And again the simplicity of the hardware was worth something. The railroads did use teleprinters extensively on circuits where the traffic warranted.

There is also the relationship between Western Union and the railroads. In my home town, after the W.U. office closed for the evening, you could send and receive telegrams by going to the railroad station. These would be handled by Morse, with a connection at some place to a W.U. office that was open all hours. In some smaller places the railroad station was the W.U. office.

Reply to
Jim Haynes

No, its actually 6 characters or 30 bits/word. You have to count the space between words. At the typical 66 speed, thats 66 bits/sec.

Reply to
Rich Greenberg

While in high school, I ran several miles of single conductor wire through trees to a friend's house. We ran a couple [of] model 15 Teletypes over that line to ground. I quickly discovered that T=L/R on that circuit. I used a low voltage power supply and adjusted it for 60mA loop current. We got garbage when we typed. We then used the power supplies that came with the Teletypes, which were something like

150VDC with a large series resistor limiting the current to 60mA. They worked fine. The time constant had kept the selector magnet current from increasing to a high enough current during a mark bit time for the bit to be detected properly. With the higher voltage and the series R, it worked great.

We added some current detect relays in series at each end with capacitors across them. When the current was off for a long time, the relay would drop out and turn off the printer motor. We had a power supply at each end. We could turn on the supply at one end or the other, start the motors at both ends, then leave a message. We'd then turn off the supply. The printers would "run open" for a second, then shut down.

These were running 60WPM and were real tolerant of our line. Of course, by telegraph standards, this was a pretty short line.

I read once that when experimenting with undersea telegraph, they originally put a big roll of wire in a vat of saltwater and tried to send a message through it. It appeared that undersea telegraph was not going to work because it was so slow with even this relatively short wire. It was later determined that the big coil of wire was a pretty good inductor.

Also, finally, on Morse... In 1969, I went to the FCC office in San Francisco to take the exam for my radiotelephone license. The PA system in the FCC office used Morse. They'd use tone over speakers to send the name of the person and the line they needed to pick up on the phone. I don't know when that was taken out. Perhaps when the FCC office moved out of San Francisco.

I use Morse ringtones on my cellphone to ID who's calling me. For my next phone, I'd like Morse text entry for text messaging and email.

Harold

Reply to
Harold Hallikainen

For an 8-bit data code with 1 start and 1 stop bit -- i.e. 10 bit per character, 'words per minute' was the sam as the bit-rate per second. 10 bits/char * 6 characters (5 printing, plus the space) exactly compensates for the difference between 'per second' and "per minute".

IIRC, ttys used 1 start, and 1.5 stop bits. giving a total 'frame' of

7.5 bits. If that's correct, 60 WPM was the equivalent of 45 baud, and 66 WPM was 49.5 baud. That also 'looks right' -- I remember shaking my head over the `half bit' in the speed specification. :)
Reply to
Robert Bonomi

45.45 bits per second for 60 wpm teleprinter.

Certainly you wouldn't use regular Morse hardware, such as the old Morse relays, in teleprinter service. But as for the line itself, Western Union for a long time operated time division multiplex over duplexed, ground-return wire circuits. This was typically four channel multiplex. I don't know the bit rate off the top of my head but seems like in the range of 100-200 bits/sec. This lasted into the 1950s, when they converted to frequency division multiplex carrier systems using electronics. One of the problems of the conversion was that they had to achieve essentially voice-grade lines, metallic pairs, and they had all these single-wire ground-return lines so it was hard to find a pair of wires on a pole line where both wires of the pair had the same wire gauge and material and position on the crossarms for the length of the run.

I have some information dating from about 1930 on W.U. operating tables, showing teleprinter operation but with a key and sounder to use if the teleprinter failed. So the relays and repeaters of that time frame were good enough for teleprinter use.

Reply to
Jim Haynes

Depends greatly on the application. And the 'urgency' of the message.

Wire-service "5 bells" stories _were_ often typed 'live'. They were too urgent for the latency of off-line composing.

On lightly-loaded remote circuit -- e.g. train orders on a low-traffic section -- were also sometimes keyed live. There was 'bandwidth' to spare, and it saved the time of the tape run, after the original keying.

When one wants to maximize throughput, and can tolerate the latency, off- line prep is a necessity, Practically nobody can maintain even the 60/66 words/min of the medium-speed machines, *with* the uniformity of keystroke spacing required. And, as line speeds got higher, it became more nearly impossible for direct keying to keep the line busy.

An absolute non-issue with any 5-level TeleType(r) machine I ever encountered. They were physically not capable of full duplex operation.

Standard manual typewriter keyboards didn't match standard electric ones either. Not to mention the fact that typewriter companies, like IBM, Remington, etc. would sell you custom keytops with whatever symbols you wanted on them. IIRC, IBM had well over _100_ 'off-the-shelf' substitute key-tops for he 'Model B' -- above and beyond the 'standard' sets for any of the large number of languages they sold, and would make 'anything" on special order.

The 33/35, and later ASCII models, had options for an RS-232 interface, or current loop. The 33 could keep up with a 110 baud line, as long as there were two 'null' characters after each [CR}. IT couldn't get the printhead back to the left margin in 1/10 sec, 3/10 was adequate.

"Black Box Corporation', out of suburban Pittsburgh, was (and is) a _big_ player in providing 'glue' devices to allow incompatible devices to inter-operate. Whether it was the physical layer, the data encoding (character-set), or a higher-level protocol issue, they had solutions. Their *real* forte was being able to turn out custom boxes for *ANY* situation -- no matter whether you needed one device, or a thousand -- in short order (a 1- to 10-lot was typically under 2 weeks), and for (all things considered) a very "reasonable" price.

***** Moderator's Note *****

Black Box has a very mixed reputation with me: I still have the RJ-48 jack that almost ended my engineering career, and Black Box sold it to a contractor I hired. The contractor followed my wiring spec, and connected the jack to a Cisco router, as part of a "media converter" between "Type 1" and "Type 3" Token-Ring connectors, in a network I built for a new control center that I was in charge of cutting over. It didn't work, and I was informed of a failed test, with less than 24 hours to go before the cutover, with more than 1,000 technicians and managers committed to work on a Sunday: I had to assemble a "Tiger Team" and get them to the site on less than 24 hours notice.

We got the cutover to work, but the jack Black Box sold turned out to have the colors wired backwards: pin 1 was Brown, pin 8 was wh/blue. The contractor had wired it by color, which was what any technician would do.

Once burned, twice careful: I never bought another part from Black Box, and never allowed my contractors to do it either.

BTW, all Teletype machines I used for ham radio, i.e., Models 15 and

19, were capable of full duplex operation. They were _wired_ for half-duplex, but that wasn't a limit of the design.

Bill Horne Moderator

Reply to
Robert Bonomi

Physically or electrically? The model 15 printers I had had separate keyboard and printer mechanics and electrical wiring. Though the stock Teletype wiring was very complex (allowing for polar relays and lots of other options), the machine came down to two wires to power the motor, two wires to drive the printer selector magnets, and two wires from contact closures on the keyboard. They COULD be wired for full duplex, relying on a remote echo, but were often wired in series for a local echo and half duplex operation.

On punched tape, I used to maintain Alden fax machines that printed weather maps in an FAA flight service station. The people would do a weather observation, punch a tape, and put it in the reader. Later, the reader would be polled from central location, and the tape would be sucked through the reader. That's the only place I've seen polled tape reading, though it probably was used elsewhere.

Harold

Reply to
Harold Hallikainen

While maintaining the speed and uniform spacing on 60/66 circuits by manual keying was indeed virtuallly impossible, most press operators could readily do that speed punching tape, and with no more latency than that required for the small loop of tape under the sensor bar (maybe 1 or 2 seconds, most 5-bell ("Bulletin") material was sent by breaking the tape and starting cold on a new strip of tape. A press operator could do this more quickly than direct keyboarding. Only the most extremely urgent mateiral was sent by direct keyboarding ("Flash"), 10 bells or more. This was an extremely rare occurrence. "Japan surrenders.," e.g. The wire stays idle after a flash, with a bulletin following with a normal text lead, usually within a few seconds, the bulletin often being dictated to the operator. Most press operators could get the tape well below the sensor bar and feeding continuously within a minute or less. Many operators could put the tape on the floor, so far ahead of the distributor that the tape hung down to the floor. Wes Leatherock snipped-for-privacy@aol.com snipped-for-privacy@yahoo.com

***** Moderator's Note *****

The "sensor bar" that Wes refers to was a mechanical arm which hung down beside the tape reader on a Moder 19 or similar machine: if the operator who was typing the original text onto the punched tape wasn't fast enough to keep up with the reader, then the tape would become taut and the sensor arm would rise, interrupting the reader's feed mechanism.

I don't know how an operator could "get ahead" of the reader: the Model 19's keyboard was drivin by the same shaft as the printing mechanism, so I don't understand how it was possible to exceed the reader's speed.

Bill Horne Moderator

Reply to
Wes Leatherock

A participant in nyc.transit had a comment on the above:

"Well, that was MY office from 1967 to 1995!

The FCC office and that of its predecessor agencies had been in those same quarters since 1912, and the paging system was OLD - it used pre-WW-II vacuum tubes that were no longer available in the early

1970s when the oscillator failed! Each staff member was identified with a single letter, usually related to their first or last name (mine was P). We didn't have to do much training for the receptionists - just a list of who has what letter - and there was a large board on the wall opposite the front desk with the code for each person and an indicator of whether that person was in the office, had not yet arrived or had left for the day, or was temporarily out of the office. The equipment was replaced with a more modern oscillator in the mid 70s.

In 1980, the office was moved from the third floor of the San Francisco Customhouse to the fourth floor, incident to an upgrade of the building, and the "code line" was replaced by a conventional PA system. The building was designated as a National Historical Customhouse (built in 1906 right after the earthquake). I was the agency's representative to the Relocation and Upgrade Committee and I could have written a Master's thesis on what it takes to make upgrades and repairs to such a structure.

In 1986 the office telephone system was upgraded to one that included a paging function and the outboard PA system was retired. We took this system with us when we moved to an industrial park in Hayward, CA (a suburb) in 1990, and when that office was moved to another location in 2000 (after I retired as the District Director) the new telephone system had an integrated paging function as well.

Memories, memories...... and it's OK to repost this story to the group where the original posting came from."

Reply to
Lisa or Jeff

Paper tape storage could be used to save line time or operator time. It saves line time because as you say the operator can't keep up with the transmission for any long period of time. It saves operator time when the circuit is heavily loaded and a central controller polls the out stations for traffic to be sent. This allows the operator to prepare the message and put it into the reader and then go away and do other things.

Then the other use of paper tape was for message relaying without having to re-keyboard the messages.

They were perfectly capable of full-duplex operation - there is no fixed connection between the keyboard and the typing unit. But probably just about never operated where the typing unit was not connected to get local copy during keyboarding.

The 32/33 keyboards were pretty awful, but then those were cheap machines. The 35 keyboard was not bad.

There was an issue in the design of keyboards for ASCII. To make a mechanical keyboard there had to be a simple relationship between the bits of the characters and the characters on the key tops, so that the shift and control keys affected all the other keys in the same manner. A standard was issued which allowed two different keyboard arrangements. One was called "typewriter paired" and had the keys arranged as was typical for typewriters at the time. The other was called "bit paired" and met the requirement for mechanical keyboards. Use of the "typewriter paired" arrangement required some electronic translation between bit patterns the keys could easily generate and those of the ASCII code.

Of course in the modern PC keyboard the keys generate arbitrary binary numers for each key and software translates those into ASCII or any other code desired.

Reply to
Jim Haynes

. . .

A question about the bells, if I may.

On the Teletype 33, the bell sounded at the end of a line (just like a typewriter) or in response to CNTL-G (ASCII)*. However, the bell wasn't very loud and barely audible over the chugging of the machine.

In school, sometimes for a joke we'd write a program to sound the bell a few times as if to sound an alarm. But the machine 'chugged' for every character, printing or not, and so the bell wasn't loud.

I wonder in a room full of Teletypes clacking away (newsrooms usually had multiple machines) if their bell was equally barely audible. As mentioned, the bell would send at the end of every line anyway so it wasn't a unique sound.

  • I believe to this day if you write a little BASIC program to PRINT cntl-G the computer will beep.
Reply to
Lisa or Jeff

Perhaps not the best example. While I can't recall if it was VE or VJ day, a witness in a newspaper office told me that:

a) All the wires; AP & UPI's, Sports, Finance, state, local.. all went silent and stayed that way for several minutes....

b) They all restarted together [They had been patched together...], and simoutanously printed out the same message....

Reply to
David Lesher

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.