EEVblog #1169 - TI 1972 Computer Interfacing

How did you connect to a time sharing computer in 1972?
Teardown of the Texas Instruments TI Silent 700 model 745 terminal with acoustic coupler modem.

Bitcoin Donations: 38y7DE8HEHNj8fGDtUr4PkCn9nWxiorvvy
Litecoin: ML7oQokTwB38bgzzjLDbRV97HKAHuwRfHA
Ethereum: 0x11AceA38DCA9DbFfB4F35f3F746af65F9dED28ce

Support the EEVblog through Patreon!

Buy anything through that link and Dave gets a commission at no cost to you.

Stuff I recommend:

Donate With Bitcoin & Other Crypto Currencies!

[rbennett]

I was involved with IC design at Honeywell when I had the pleasure of using one of these terminals. The DEC VAX was the mainframe (betcha can’t use just one!) and everybody connected via phone line. It was Berkeley SPICE 2G, if I recall. Netlists were text only. We managed to design analog ICs using these things. The BEL command was useful if you wanted to get the attention of somebody on the other end of the line. You would hit BEL 3 times and hope that there would be a response, so you could chat about a problem or some other problem.
I designed my own 30 baud FSK modem (to connect to my Radio Shack TRS80 - model 4), so my wife and I could play Colossal Cave (a text adventure game) on the company’s VAX after hours when the kids were in bed. Fun times!

[plaws]

^G is still the terminal bell. ^H is still backspace. ^J and ^M are still LF and CR ... Nothing has changed.

[VideoFlyer]

This really brings back memories! I worked for TI in the early to mid 1970's, first at the DCD (Digital Circuits Division) in Stafford TX (outside Houston), where I repaired the IC test systems, and then transferred to DSD (Digital Systems Division) in Sunnyvale, CA, where I worked on the mini-computers (960 & 980), and Silent 700 series terminals. The print-heads actually had the driver transistors built into the die, and if I recall correctly, the final transistor was the one that supplied the heat to the paper. Your terminal may still have some life left in it. Try increasing the contrast pot to see if you can get a little more heat to the paper, which may have lost sensitivity to heat over the years. A new roll of paper, or as was mentioned earlier, unroll some paper from the existing roll may also help. As was mentioned below, cleaning the print-head is an excellent idea as well. Use 97% isopropal alcohol and a cotton swab.

[martin.pokorny]

It's respectable well-designed old technology. And the documentation is really breathtaking.
Thanks for sharing.

[hrnekbezucha]

The manual is absolute gem. I had no idea things were made _this_ clear. Beautiful.

[sbalogh]

I used to have an old PMG 300 baud modem in a grey metal case. I modified it so the modem could originate or answer depending on a toggle switch on the front panel. I used to download files from a couple of online BBS systems and save the audio to cassette tape. When I played the tape back into the modem a copy of the file would appear on the serial interface. This is how I stored programs for a home made computer I build before I managed to get an old 8" floppy drive. I had an old IBM Selectric typewriter which I tried to convert to a printer by building power drivers for each key. I never did get that beast working. Those were the days.

[darikdatta]

You're not a real hacker if you haven't held carrier by whistling into your phone.

[BrightBlueJim]

A number of questions can be answered if you realize that the Silent 700 series was meant to be a portable replacement for an ASR-33 Teletype - you see it has current loop and RS-232 inputs in addition to the modem. Since most people using time shared computers were set up to use Teletypes, the software was optimized accordingly, expecting that the printing terminal kept a record of transactions. Being backward-compatible with the ASR-33 also meant that the standard communications mode would be 110 bits/sec 11 units/character (8 bits, no parity, 1 start, 2 stop), and the high speed would be 300 bits/sec (8 bits, no parity, 1 start, 1 stop). The base model would be upper-case only, and that empty socket may have been where a lower-case character generator ROM would be added, but that's just a guess. I think that the 300 baud, 30 character/second mode was pioneered by GE in their TermiNet 300 series terminals, which used (I kid you not) chain printer mechanisms, and were about as big and heavy as Teletypes. These were not meant to be portable, so the Silent 700 series was also meant to serve that market. GE was also one of the pioneering timeshare companies, so having 30 char/sec printers was a big selling point!
A special Teletype ASR-33 function was the "Here Is" key. On an ASR-33, there was a mechanical drum on which you could program up to 20 characters of ID message, which were transmitted when you hit the "Here Is" key. For most machines, this was not programmed, and it just sent out 20 nul characters (ASCII code 0) (which was convenient for punching leaders on paper tape). Whether TI implemented this on the Silent 700s is outside my knowledge, but I think some terminals just used that key to generate a single nul character, so that may be the case here.

[fluxoff]

Hi Dave my Dad worked at Lockheed in yhe early 70's; on family day (open house) they brought out the TI silent 700 terminal for games like hangman and (later) LIFE. The thermal paper was very cheap & non-archival(turned brown with heat & sun exposure) but cheaper than huge 132 column line printer paper which they used for satellite test run procedure documentation. Programmers & engineers would check out the TI terminals to take home & remotely dial in & check on diagnostics & tests. Very popular! Led to a boom in Lockheed remote work on Tymshare (brand) systems for FORTRAN work. Also led to my entry into programming!

[rleeAZ]

Being able to use one of these, instead of going to the college computer center and waiting in line to use a keypunch, saved literally hours. It was slow compared to modern standards, but using the time sharing system instead of submitting batch jobs via cards was an order of magnitude faster.

[ricknelson]

At age 13 in 1975, my friends father who was an Engineer for Corning glass works, had one to connect to the Laboratory computer. I believe it wast whopping 110 baud. He would dial into the computer (rotary dial phone) and let my friend and I play “Star Trek” on it. It was awesome for the day. I’ll never forget how much trouble we got into when we used the last of his fathers paper playing games. Thanks for reminding me how old I have gotten :-)

[m.k.]

I suspect that the old paper is causing the print problems..sometimes old thermal paper does not show print.
Try unwapping a few wraps of paper, then try printing-the inner wraps may work better.
Else, try a roll of thermal fax paper-should work fine.
Also, check the printhead for debris.

[thedevilinthecircuit]

The orientation of the blades on the cooling fan show it's not blowing air into the case; it's drawing air through the entire case and blowing it past the motor housing to the outside. Cool gear!

[Mostlyharmless]

This is just begging to be brought back to working order and hooked into a UNIX system. It should still work, with just a few lines in gettytab. We should be playing zork in no time!

[mikesradiorepair]

Oh yes they do! My aunt was a executive secretary for 35 years and had long fingernails. Damn woman could type twice as fast as I could without long fingernails. Probably wouldn't work with a modern shallow key with little stroke length. The old keys were really deep and had a long stroke length. When she typed her fingernails were just over the edge of the key. LOL, still don't know how she did it.

[radry]

Funny how we are shifting back to "time sharing" computers aka cloud computing on demand.

[steve]

My uncle worked for Microswitch in Freeport, IL. The product was top of the line and much imitated after the patents expired.

[altair]

FWIW, the trick used in V.90 to exceed 33.6kbps wasn't actually particularly sophisticated, but it was clever. Essentially the way it worked was by relying on the accuracy of the DACs installed in end users' line cards in digital telephone exchanges; rather than sending an FSK (or actually, for later standards, QAM) signal, what basically happens is that particular bit patterns (symbols) are assigned voltages, with each symbol being sent as a voltage pulse (Pulse Amplitude Modulation). Your modem negotiates with the line card on the number of different levels it's able to distinguish, which is obviously dependent on line conditions and that's why V.90/92/X2/K56flex modems would connect at a variable rate. The digital data is transmitted, in digital form, to the line card over the network, which requires special equipment rather than a regular modem. In the other direction (upstream), 56k modems actually use 33.6kbps analogue, just like V.34.

[Falcrist]

*_"They don't do theory of operations anymore"_*
You're right, but it's also worth noting that this thing cost about half as much as a brand new car back then. We're talking about $8k-$10k in 2019 dollars
I bet if you bought a similar machine that cost an equivalent amount of money today, the service manual would be really decent.

[henrikjohnsson]

I used one of these in 1986 to debug a Motorola 68000 board I had wire wrapped for a school project (and I have no doubt it was kept in use a lot longer, the owner of the company was rather tight and wasn't about to replace working equipment)
This company built production gear for the automotive industry and the terminal was mainly used for maintenance in the field, and as such it made perfect sense. By then there started to appear "portable" computers that could do terminal emulation but this was cheaper, smaller and lighter than most of those and a lot more rugged, hence perfect for chucking in the back of a car and doing a crazy dash to a car factory where production had stalled.