God, this reminded me of how many hours I spent on "RS232 gymnastics" back in the 80s and 90s. It's fun to watch someone else do that now, but I really don't miss having to do it for work.
The old British Sci-Fi comedy series Red Dwarf shows a fictional book titled "the joys of the RS232-C interface" or something similar. That interface has a reputation...
You comment makes mine unnecessary. I too had to pore over RS232 tables to get CAD 286s talk to a pen plotter and printers. Just remembering it makes me shudder😒
This one is probably modified by AT&T for tier 5ESS PABX systems. I vaguely remember to see them even with the brown suitcase. They where used to program these PABX systems.
20:00 I do believe that modern health and safety practices would require anyone working in an office full of the old teletype machines to wear hearing protection. You can see why Ti made a big deal out of their silent version.
And it's not even like you could easily isolate the person from the noise. IBM line printers are insanely loud, but they can run inside a padded MDF and plastic enclosure to make it a tolerable experience.
It's amusing that AT&T were still buying and using these into the 1990s. I worked for EO, a company part owned by AT&T, and we were developing the first smartphone (or at least one of several contenders for that title) a couple of years before this teletype was built. So for a short time you could buy both a smartphone and a teletype with AT&T branding. If you can find an EO 440 or 880 (same thing but bigger) I think it'd be an interesting device to make a video of.
I freaking love this series, where Marc and his friends have restored of the old teletypes through to these "modern" digital interfaces. I studied in University in the era of this transition, we accessed mainframes with teleprinters against 300 baud modems where we strapped a phone handset into a cradle with a velcro, programmed with ibm punch cards, outputted to paper and punch tape. Our first crt's were little more than line printers, operated much like this machine, but with a 40 line x 80 character display. The first microprocessor we programmed was a 8080 with a hexadecimal keypad and 8 character lcd, we outputted paper punch tape from punch cards containing our assembly code . Our project was to simulate a traffic light controller, timed to display green yellow and red lights in each direction represented by the segments of the 8 characters, and if you pushed certain buttons on the keypad, it simulated hitting the "walk" button. We learned and threw around terms like baud, ebdic, ascii to describe our technology, but until Marc's rebuild, I never understood the connection between semaphore signals which were sent within eyesight, morse code and today's internet. I learned from Marc, and It's amazing, there is a direct line between the Morse code and the telegraph (1830s's around when electricity was first harnessed to send electrical disturbances along direct current wires laid next to railways) at first to signal train traffic, to the teletype which were open dc circuits across a switching network which grew world wide, to the first digital computers, to today's age and "The Internet."
Until electricity was harnessed and sent along wires about 180 years ago, distance communication required smoke, or humans, horses, birds, or ships moving paper in a coordinated way.
@@zincfive or large wooden semaphore telegraph towers! In a big line, visible from one to the next! Those are my favourite. Not every nation had them, but the term telegraph predates the electrical telegraph 😁
@@kaitlyn__L Neat! So people manned the towers, and flagged each other? Imagine that! I'm reading about it now, called "optical telegraph", the first experiment was put in place in the 1760's to announce the results of a horse race in an hour. Later, a Frenchman named Chappe, implemented a system at the time of the French Revolution, 1790's and criss-crossed france by the 1830's. Very Interesting, Thanks!
If you're really interested in silent printers, you could do worse than look at the Olivetti JP101 spark jet printer. Yes, it printed by burning the paper with a tiny spark. Very silent, but no keyboard. I had one for while to go with my Olivetti M10 laptop. It was pretty fast, very light and never caused a fire. It used normal paper on a roll, no thermal coating. I liked it, but them it was free to me, I worked for Olivetti at the time. All this beggaring about with bit rates and parity certainly takes me back a long way - experience told me that you had a speed problem when I saw those FFs and FEs on the protctol analyser.
Back in the eighties, I worked for a credit bureau organization where our branch offices in the region used these and we provided support remotely. Ours were not AT&T but specifically Texas Instruments running at 1200 baud. Obviously, for us at that time, these were first class machines for our needs. Great little devices!
ASR wasn't a name, but the description of the capabilities: Automatic Send/Receive, the ones without the paper tape were designated KSR - Keyboard Send/Receive. Teletype also offered a Receive Only variant, designated RO.
I think you are on to something. That would explain it. And to your point, it says AT&T on the front, not TI. And the model sticker on the back says MODEL 703 TERMINAL, AT&T. Must be a custom machine indeed.
@@CuriousMarc Yes, you're on to something. That's exactly what I was going to say. AT&T has such a large purchasing power that they got TI to customize the 703. AT&T used specialized equipment to do maintenance in the central offices, and in customer premises. The technicians were given equipment in a suitcase for doing things such as routine maintenance on the backup batteries. These were extremely important because the customer's whole network depended on the datacom equipment being up 24/7/365. So the tech would come to the customer's premises and check the datacom equipment and backup batteries and print out a dated hard copy to prove maintenance was completed. So the suitcase most likely had a handheld tester and cables, and the 703 to print out the hard copy.
@@acmefixer1 given the fact that we currently have customized firmware/software on PCs and terminal I don't think it was a big trouble for TI to do some white label job for AT&T.
I have this variant as well - the AT&T System Administration Terminal. Interestingly, the manual makes no mention of the baud rate, from what I can see. However, given that the intent was that you'd use it with an AT&T System 25 PBX*, which ran at 1200, I suspect AT&T decided that baud rate was something to not worry about. And, yes, I ran into the same issue, assuming "oh, it's a Silent 703, it'll be 300!" (* Not sure if the System 75 or 85 - which became the Definity line - were sold with this, or with other terminals.)
Very interesting, you're starting to get into computer history I actually lived through :) I would have been quite excited to get one of these in the early 80s.
Portable back then meant "luggable". I used a Silent 700 in the 80's and early 90's to document ladder logic programming on Allen-Bradley PLC-2 programmable controllers. Allen-Bradley had a dedicated "portable" programming terminal for these controllers, the 1770-T3. Anyone that used one called it "the boat anchor". The manual for the T3 said it was "less than 35 lbs."
Marc - at some point, your videos will be shown at tech-school history lessons - to explain "how did we get here". Your adventures teaches so much more than reading about our old tech from the 60s and forward - and seeing someone respecting these devices (machines?), knowing how to take care of them, how they were used and bringing them back to life, is a much better teaching foundation than old dusty books. I hope you make your manuals and equipment all available at the Computer History Museum as one day I would love to browse the actual thing for real.
1:49 That's the model of Silent 700 that my father used to carry home from work every day when he was at GE in the 1970s. My first interactions with a computer were over one of those.
Fun video - thanks for sharing. The later 707/1200 model supports 1200 and 300 via four dipswitch settings. There is also an "Auto Access" cartridge with data terminal options - the cartridge slots into the blue socket. I've used one model with an option for batteries to be fitted beneath one of the flaps underneath the terminal and spare cables to be stored in the second bay.
Marc - See if you can find an original IBM Proprinter. At the time, IBM needed a low-cost printer to sell alongside their PC and found the models being sold by the Asian manufacturers (Epson FX-80, etc.) were cost-competitive only because of the low cost of labor (they used over 150 individual hand-assembled parts). So IBM designed the Proprinter to use no fasteners so it could be assembled by robots - all the parts just snapped together. There was only one step on the assembly line that needed a human to adjust a clearance. The story was that the robots weren't ready in time for production to start, so IBM hired a lot of temp workers and discovered that their simple design made it even easier for people to assemble the product.
I worked with RS-232 for a college district for more than a decade. You can usually get by with only 6 wires in the null modem cable. Each end connect 6, 8, and 20 together. Pins 4 and 5 cross over. Pins 2 and 3 crossover. Pins 1 and 7 straight through. Or you can connect 4 to 5 on each end and use 4 wire cable with software Xon/Xoff flow control.
See my other comment on this arcane subject. The cable you describe (4 cross connected to 5) is one that enables the use of RTS/CTS flow control. The one shown here actually disables RTS/CTS and solely relies on DTR/CTR, which is more of a printer thing.
I see there are multiple comments about buffer and possible support/usage of XON/XOFF to flow control. For some of the older printing terminals, there was also the technique (used by X.25 to Async Packet Assembler Disassembler (PAD) devices I used to work on in the early 80s) of having the computer/device send additional NULL characters and/or delay after CR, LF and FF characters sent to the printing terminal. That gives the printing terminal additional time to move the carriage down or left so the actual printable characters aren't lost. In X.3/X.28/X.29 PADs, that was a parameter you could set to add a specific number of NULLs after specific control characters. This can be simulated when you go DTE to DTE in a PC running Tera Term by adding per line or per character delay. It was really cool to see the test mode at 3:30. I used to run those kinds of tests all the time on terminals, printers, etc. to check for alignment, if the ribbon and print head were OK, etc. and had forgotten about them.
Have one of them. Patiently waiting for you to make a video since you showed it couple of week ago. Typed part of high school paper on it on local copy. Think i managed to interface it with the XT PC or ZX spectrum at the time with limited success and use. Lovely little machine. Who needs a monitor.
Not only do you get rare retro equipment but they also seem to be very rare modified, pre production, prototype or specialist designed models. Good score and thanks for sharing.
I bought one of these at a hamfest for $10 back in the mid 90's, and just had to connect to a BBS with it. Which I did, and poked around in some message forums (very slowly) until the novelty wore off and I used up all the included paper.
Whenever I start a video and I hear "Hello and welcome back!" with that hint of a French accent I know I will learn something about vintage computers today.
That thermal printer mechanism with the "move out of the way" feature is pretty slick. And I had to chuckle with the auto captions. "We have a bald rate problem" - could have been a perfect segue into a Keeps promo 😉
As it happens, I also have the AT&T version of the 703. Thanks to Marc for the video, else I would still be trying to figure out why it did not run at 300 baud.
I got out the manual that came with the terminal and it clearly states that the host needs to run at 1200 baud. That is not in the ti version of the document.
so my guess is. Its an ATT/Lucent tool used to talk to the 5e Phone switch and others. Att just slapped the a ticker on it. And ordered them special. And the teletype used at the end. Is like my dad used in a switch room.
That's a rather odd version of null-modem and I'd be interested to see what happens when you run at speed. It should be CTS that controls flow and others (like DTR, DSR and CD) only change when things go on and offline. We can sense how terrible it is for you to suddenly gain another HP instrument to repair.
[Man, this is so confusing, I had to rewrite this 3 times!] CTS is an input on a DTE, and is an output on the DCE. Since there is no DCE in a null-modem DTE to DTE connection, the CTS signal has to be generated somehow, in this null-modem by looping RTS back. Same goes for CD, which is really a modem-only signal. This is a pretty standard null-modem arrangement (and is shown in the Silent manual). It does not rely on RTS/CTS control flow. It uses DSR/DTR as the control flow - the common way to deal with serial printers, which the Silent is. It will actually prevent RTS/CTS from hanging up the link and get communications going, which is quite convenient. You'd have to cross connect RTS to CTS of both CTEs to enable that flow control. By the way, the RTS/CTS flow control is a modern bastardization of the original flow control mechanism which was intended for half-duplex DTE-DCE modem links. It was not meant for full-duplex DTE-DTE, although it is now bent into doing that. Which makes things horribly confusing.
@@CuriousMarc Does it honour in-band XON/XOFF control flow characters? (I assume it doesn't implement HP's Enq/Ack protocol, which I thankfully haven't had to worry about since I stopped having to work on a HP2392.)
@Frank Wales No, I think only DSR/DTR, and maybe RTS/CTS but it would need to be wired with the different null modem that crosses 4 and 5. Haven’t tried yet.
Oh man, the discussion about how to wire up the control signals to allow two DTR devices to talk to each other directly took me WAAAAYY back to my days of trying to get an IBM 7171 "ASCII Device Attachment Control Unit" to talk to hard-wired VT220s. (cheaper than genuine IBM 3178s!) The 7171 was the first thing I'd encountered that DEMANDED the RS-232 control signal protocols be obeyed exactly, and I worked out how they'd have to be looped back and crossed over to simulate a set of nonexistent modems between the two devices. I was overjoyed to see that same exact diagram in Marc's notebook here! We made up a bunch of hacked-up cables and called them "null modems" because they faked an entire data signalling system. Later we were a little embarrassed when things like the little dongle in the video became available, and were named the same thing. These RS-232 videos really take me back in time! Thanks so much for them.
No reason to be embarrassed. That just shows how your ingenuity and problem solving skills and even the naming of the device aligned with the solutions of the greatest minds at the time. It is something to be proud of even if you weren't the first to think about it.
It's also funny how people played fast and loose with the standards. I remember having to set flow control type (CTS/RTS, DSR/DTR, or software XON/XOFF) because every system seemed to have a different idea of how to control the flow of data. Then came the era of the HP48 calculators, where flow control pretty much disappeared.
You're not the only one who dealt with the 7171; I did just like you did. We had to make up 128 adapters to connect the 7171 to the Gandalf port selector. And we were supplied with 128 full 25 pair cables that we had to pull under the raised floor. The IBM 4361 mainframe was used to teach Fortran and Ada, and it was absurdly underpowered. It would take 8 minutes to compile an Ada program with one single user and slow down another 8 minutes for each additional user. This wasn't acceptable for a college class, so it was replaced after only a few years. I still have some 7171 manuals stored somewhere.
We had one of these to communicate with the Ontario Hydro Research Division PALC Auxilary Logic computer. It used Motorola 6809 processor if memory serves. A rack mounted alternative to a to a bunch of relay logic...the 700 just worked when you plugged it in and hit the right keystrokes...
Maybe it used the X/On and X/Off command sequences to control the sending computer when the buffer gets near full? That’s usually how async printers in this era worked. Put your scope on the Tx line to see if it’s sending them?
Ah the hours I spent connecting VT terminals to VAX terminal servers back in 1986... I would have liked one of those 4957s back then. Oh, and DSR is "Data Set Ready" if I remember correctly.
Ahh cool! Along with the 707/1200, I have come across one other 703 that was fitted with a replacement ROM allowing it to be switched between 300 & 1200 baud. While it could send ok at 1200 baud, the internal buffer filled up pretty quickly when receiving data causing it to start printing gibberish. The 707/1200 just prints super fast at the expense of it not being quite as silent :)
I remember that start-up sequence, and the idiosyncratic head movement! Your 703 looks like a newer one than I had access to... I also remember going through all of the combinations and permutations of RS-232 wiring :-( Thanks, Marc.
I've got a TI DS990 Model 1, which is essentially a board upgraded TI Model 770, which is basically a fancy CRT based Silent 700 series data terminal. I wish the DS990 Model 1 was this easy to get up and running! I've been trying off and on for years!
serial rs-232 comms can be fun and madening at the same time! I remember spending hours of time researching and setting dip switches and rotary encoders for fixed clock UARTS... Fun times! only certain ports could do certain speeds and protocol combinations, because of the way they were feed the divided clock timing generator.
Boy, this brought me back a few years... especially that null modem adapter; I've still got a half-dozen of them floating around here. Never know when you might need them! 😋
Agreed! The 19 is loud, but has this strange rhythmic and soothing quality. The ASR 33 on the other hand sounds like it’s hyper, it drunk too much caffeine! I always half expect seeing springs and screws being shaken out of the mechanism...
@@CuriousMarc The Model 19 was my first machine as a teenage ham! It used to shake the house! There was something warm about the big rectifier when the tubes would flash over purple and I knew there was 60 mA loop current available and off I went onto the airwaves with RTTY!
12:47 you wrote 1994 in the captions, but it says 1984 on the chip ;) But still surprised how many full-size components are in this device, mixed in with the ICs. So used to microscopic surface components in modern stuff, you don't see the full-size versions much anymore on boards with ICs
Yes you are correct, it's 1984! Which makes a lot more sense. But then it's an early machine, not a late one, and as the other commenters found out, it's an AT&T special that has 1200 baud but would not have the double-speed head. Weird animal it is!
Im in the process of making a serial data cable for my Yaesu ham radio. Im really glad they dont use all the ready to send stuff. Only have to wire TX RX and ground, and its still a PITA.
Nice Paladin cable tester, I have the Mac one that does SCSI, ADB, and a few others. Mostly use it to test ethernet patch cables these days. Got it for $1 when a former employer went out of business ;)
Having a serial port witht 1200 baud does not necessarily mean that the print mechanism can cope with that ~120 chars per second. This was either be managed by software flow control (Xon/Xoff) or hardware flow control (RTS/CTS if I remember correctly), depending on the device. I used a TI Omni800 a while ago which was capable to print ~130 chars/s and had a interface running up to 9600 baud.
It probably accespts at 1200 bps, and if things go too fast, hardware handshaking takes over. It also needs to have DSR active (this is on all TI silent 700's)
Those vestigial battery bays present some opportunities for non-invasive retrofitting. You could put in a single-board computer with BASIC and tell people it's a lost prototype microcomputer.
That could have been helpful in 1983 for me - when was it released - as the computer I was using did not have a local printer and I depended on the mainframe chain printer to send my results back, which were crap as I missed out decimal points in the input data for calculations
Just got one of these myself, the serial port has been ripped out from it so I need to replace that, but weirdly mine had a footnote with it which says: This Texas Instruments Model 703 Data Terminal has been quipped with a special PROM, which allows a data transmission of both 300 and 1200 bits/second (300 & 1200 BAUD). It also gives the ability to use the DC1/DC3 (or X-ON/X-OFF) Ready/Busy Protocol. NOTE: You can not use this PROM in conjunction with the Auto Access Cartridge, or any other cartridge. To select the various options: Simultaneously hold down CONTROL and SHIFT and press: 1 - Speed = 1200 BAUD 2 - Speed = 300 BAUD D - DC1/DC3 ON N - DC1/DC3 OFF But... mine is missing this PROM chip, so looks like it was pulled to go into another one perhaps when this was replaced.
nice and silent machine, just bit slow and papers tend to wipe out itself in hot, moist or sun, so it was better to make copy when possible :) , there was also bigger unit with bubble memory and 300 modem....😁
It's most probably a custom model for AT&T, although I wouldn't be suprised this particular model later led to the 703/1200 model later on (assuming the timelines match)
Try leaving a printout from one of these in a car during the summer! But the printer were cheaper and smaller, and the printouts very temporary. Old prints would fade considerably over time even in a controlled climate.
Regretfully, I lack any self control and so about 18 hours after first seeing this video... I snapped one up on ebay. Not exactly sure what it is, but seemingly the exact later-model 700/1200 baud modem and RS-232 as well! Hopefully I'll find an adapter that works when it arrives...
Im always curious when I see an old ad with an phone number on it what happens if you call that number today - is it still active… is someone else using it meanwhile - is it dead 💀
God, this reminded me of how many hours I spent on "RS232 gymnastics" back in the 80s and 90s. It's fun to watch someone else do that now, but I really don't miss having to do it for work.
The old British Sci-Fi comedy series Red Dwarf shows a fictional book titled "the joys of the RS232-C interface" or something similar. That interface has a reputation...
You comment makes mine unnecessary. I too had to pore over RS232 tables to get CAD 286s talk to a pen plotter and printers. Just remembering it makes me shudder😒
Breakout boxes worked wonders to figure out how your “flavor” if RS-232 worked.
@@joelongjr.5114 I still have my breakout box, somewhere
Had to help a buddy "serial" into a commercial washing machine once. Some skills are surprisingly universal.
This one is probably modified by AT&T for tier 5ESS PABX systems. I vaguely remember to see them even with the brown suitcase. They where used to program these PABX systems.
20:00 I do believe that modern health and safety practices would require anyone working in an office full of the old teletype machines to wear hearing protection. You can see why Ti made a big deal out of their silent version.
And it's not even like you could easily isolate the person from the noise. IBM line printers are insanely loud, but they can run inside a padded MDF and plastic enclosure to make it a tolerable experience.
It's amusing that AT&T were still buying and using these into the 1990s. I worked for EO, a company part owned by AT&T, and we were developing the first smartphone (or at least one of several contenders for that title) a couple of years before this teletype was built. So for a short time you could buy both a smartphone and a teletype with AT&T branding.
If you can find an EO 440 or 880 (same thing but bigger) I think it'd be an interesting device to make a video of.
I freaking love this series, where Marc and his friends have restored of the old teletypes through to these "modern" digital interfaces. I studied in University in the era of this transition, we accessed mainframes with teleprinters against 300 baud modems where we strapped a phone handset into a cradle with a velcro, programmed with ibm punch cards, outputted to paper and punch tape. Our first crt's were little more than line printers, operated much like this machine, but with a 40 line x 80 character display. The first microprocessor we programmed was a 8080 with a hexadecimal keypad and 8 character lcd, we outputted paper punch tape from punch cards containing our assembly code . Our project was to simulate a traffic light controller, timed to display green yellow and red lights in each direction represented by the segments of the 8 characters, and if you pushed certain buttons on the keypad, it simulated hitting the "walk" button.
We learned and threw around terms like baud, ebdic, ascii to describe our technology, but until Marc's rebuild, I never understood the connection between semaphore signals which were sent within eyesight, morse code and today's internet. I learned from Marc, and It's amazing, there is a direct line between the Morse code and the telegraph (1830s's around when electricity was first harnessed to send electrical disturbances along direct current wires laid next to railways) at first to signal train traffic, to the teletype which were open dc circuits across a switching network which grew world wide, to the first digital computers, to today's age and "The Internet."
Until electricity was harnessed and sent along wires about 180 years ago, distance communication required smoke, or humans, horses, birds, or ships moving paper in a coordinated way.
@@zincfive or large wooden semaphore telegraph towers! In a big line, visible from one to the next! Those are my favourite.
Not every nation had them, but the term telegraph predates the electrical telegraph 😁
@@kaitlyn__L Neat! So people manned the towers, and flagged each other? Imagine that! I'm reading about it now, called "optical telegraph", the first experiment was put in place in the 1760's to announce the results of a horse race in an hour. Later, a Frenchman named Chappe, implemented a system at the time of the French Revolution, 1790's and criss-crossed france by the 1830's. Very Interesting, Thanks!
If you're really interested in silent printers, you could do worse than look at the Olivetti JP101 spark jet printer. Yes, it printed by burning the paper with a tiny spark. Very silent, but no keyboard. I had one for while to go with my Olivetti M10 laptop. It was pretty fast, very light and never caused a fire. It used normal paper on a roll, no thermal coating. I liked it, but them it was free to me, I worked for Olivetti at the time. All this beggaring about with bit rates and parity certainly takes me back a long way - experience told me that you had a speed problem when I saw those FFs and FEs on the protctol analyser.
Back in the eighties, I worked for a credit bureau organization where our branch offices in the region used these and we provided support remotely. Ours were not AT&T but specifically Texas Instruments running at 1200 baud. Obviously, for us at that time, these were first class machines for our needs. Great little devices!
You have WAY to many toys that each would make us jealous!
ASR wasn't a name, but the description of the capabilities: Automatic Send/Receive, the ones without the paper tape were designated KSR - Keyboard Send/Receive. Teletype also offered a Receive Only variant, designated RO.
When you showed the wiring in your notebook, I thought "Hm, looks like a Null Mod..." - and there you are!
I wonder if these were custom built to AT&T specs - hence, the non-switchable 1200 baud speed.
I think you are on to something. That would explain it. And to your point, it says AT&T on the front, not TI. And the model sticker on the back says MODEL 703 TERMINAL, AT&T. Must be a custom machine indeed.
@@CuriousMarc
Yes, you're on to something. That's exactly what I was going to say. AT&T has such a large purchasing power that they got TI to customize the 703.
AT&T used specialized equipment to do maintenance in the central offices, and in customer premises. The technicians were given equipment in a suitcase for doing things such as routine maintenance on the backup batteries. These were extremely important because the customer's whole network depended on the datacom equipment being up 24/7/365. So the tech would come to the customer's premises and check the datacom equipment and backup batteries and print out a dated hard copy to prove maintenance was completed.
So the suitcase most likely had a handheld tester and cables, and the 703 to print out the hard copy.
@@acmefixer1 given the fact that we currently have customized firmware/software on PCs and terminal I don't think it was a big trouble for TI to do some white label job for AT&T.
Does it use any specialised paper?
@@marcwolf60 thermal paper, if it is a thermal printer
I used to fix one of these for a local Ford dealer, the thermal print head would wear out. It was used in the repair dept so that may explain why!
Fix Or Repair Daily ... ;)
I have this variant as well - the AT&T System Administration Terminal. Interestingly, the manual makes no mention of the baud rate, from what I can see. However, given that the intent was that you'd use it with an AT&T System 25 PBX*, which ran at 1200, I suspect AT&T decided that baud rate was something to not worry about.
And, yes, I ran into the same issue, assuming "oh, it's a Silent 703, it'll be 300!"
(* Not sure if the System 75 or 85 - which became the Definity line - were sold with this, or with other terminals.)
We were always trying to push from 9600 baud to 1.92K so this 300/1200 situation is pure nostalgia for me.... I miss RS232... those were the days.
Ford Dealerships used to use these to check for recalls.
Very interesting, you're starting to get into computer history I actually lived through :) I would have been quite excited to get one of these in the early 80s.
Portable back then meant "luggable". I used a Silent 700 in the 80's and early 90's to document ladder logic programming on Allen-Bradley PLC-2 programmable controllers. Allen-Bradley had a dedicated "portable" programming terminal for these controllers, the 1770-T3. Anyone that used one called it "the boat anchor". The manual for the T3 said it was "less than 35 lbs."
Marc - at some point, your videos will be shown at tech-school history lessons - to explain "how did we get here". Your adventures teaches so much more than reading about our old tech from the 60s and forward - and seeing someone respecting these devices (machines?), knowing how to take care of them, how they were used and bringing them back to life, is a much better teaching foundation than old dusty books. I hope you make your manuals and equipment all available at the Computer History Museum as one day I would love to browse the actual thing for real.
1:49 That's the model of Silent 700 that my father used to carry home from work every day when he was at GE in the 1970s. My first interactions with a computer were over one of those.
Yes I have already done that!
Keep up the great work, Marc. Your channel is a national treasure.
Fun video - thanks for sharing. The later 707/1200 model supports 1200 and 300 via four dipswitch settings. There is also an "Auto Access" cartridge with data terminal options - the cartridge slots into the blue socket. I've used one model with an option for batteries to be fitted beneath one of the flaps underneath the terminal and spare cables to be stored in the second bay.
Marc - See if you can find an original IBM Proprinter. At the time, IBM needed a low-cost printer to sell alongside their PC and found the models being sold by the Asian manufacturers (Epson FX-80, etc.) were cost-competitive only because of the low cost of labor (they used over 150 individual hand-assembled parts). So IBM designed the Proprinter to use no fasteners so it could be assembled by robots - all the parts just snapped together. There was only one step on the assembly line that needed a human to adjust a clearance. The story was that the robots weren't ready in time for production to start, so IBM hired a lot of temp workers and discovered that their simple design made it even easier for people to assemble the product.
0:58 love that machine, the sound, the cylindrical keys and the perforated tape!
I worked with RS-232 for a college district for more than a decade. You can usually get by with only 6 wires in the null modem cable. Each end connect 6, 8, and 20 together. Pins 4 and 5 cross over. Pins 2 and 3 crossover. Pins 1 and 7 straight through. Or you can connect 4 to 5 on each end and use 4 wire cable with software Xon/Xoff flow control.
See my other comment on this arcane subject. The cable you describe (4 cross connected to 5) is one that enables the use of RTS/CTS flow control. The one shown here actually disables RTS/CTS and solely relies on DTR/CTR, which is more of a printer thing.
It's 2:30AM Eastern...CuriousMarc posts...I watch!
I got almost no understanding about what you were doing but seeing you solve a problem like this is like a miracle to me!
Repairing the HP display module sounds cool
I see there are multiple comments about buffer and possible support/usage of XON/XOFF to flow control.
For some of the older printing terminals, there was also the technique (used by X.25 to Async Packet Assembler Disassembler (PAD) devices I used to work on in the early 80s) of having the computer/device send additional NULL characters and/or delay after CR, LF and FF characters sent to the printing terminal. That gives the printing terminal additional time to move the carriage down or left so the actual printable characters aren't lost. In X.3/X.28/X.29 PADs, that was a parameter you could set to add a specific number of NULLs after specific control characters.
This can be simulated when you go DTE to DTE in a PC running Tera Term by adding per line or per character delay.
It was really cool to see the test mode at 3:30. I used to run those kinds of tests all the time on terminals, printers, etc. to check for alignment, if the ribbon and print head were OK, etc. and had forgotten about them.
I have one (standard TI 703) and I love it!
The box went emergent "Knock, Knock. Wake up Neo..."
I bought one of these at a thrift store back in 1995...and I still have it in the Forgotten Machines lab! Awesome video, as always!
Have one of them. Patiently waiting for you to make a video since you showed it couple of week ago. Typed part of high school paper on it on local copy. Think i managed to interface it with the XT PC or ZX spectrum at the time with limited success and use. Lovely little machine. Who needs a monitor.
I was really hoping you'd dump a page of ASCII art or something from the PC to terminal to show us what 1200bps looks like on it!
Not only do you get rare retro equipment but they also seem to be very rare modified, pre production, prototype or specialist designed models. Good score and thanks for sharing.
I bought one of these at a hamfest for $10 back in the mid 90's, and just had to connect to a BBS with it. Which I did, and poked around in some message forums (very slowly) until the novelty wore off and I used up all the included paper.
I remember seeing a TI Silentype teletype machine in the mid-80’s in a local surplus shop. Brought the memory back after seeing the video 😀
Whenever I start a video and I hear "Hello and welcome back!" with that hint of a French accent I know I will learn something about vintage computers today.
That thermal printer mechanism with the "move out of the way" feature is pretty slick. And I had to chuckle with the auto captions. "We have a bald rate problem" - could have been a perfect segue into a Keeps promo 😉
I love how at the end of one episode, the content of the next shows up!!
As it happens, I also have the AT&T version of the 703. Thanks to Marc for the video, else I would still be trying to figure out why it did not run at 300 baud.
I got out the manual that came with the terminal and it clearly states that the host needs to run at 1200 baud. That is not in the ti version of the document.
Awesome. Would you mind scanning that doc and make it available? I have not seen any info on the AT&T version of the 703.
that first ad is funny, 'here's a secretary for scale!'
so my guess is. Its an ATT/Lucent tool used to talk to the 5e Phone switch and others. Att just slapped the a ticker on it. And ordered them special. And the teletype used at the end. Is like my dad used in a switch room.
That's a rather odd version of null-modem and I'd be interested to see what happens when you run at speed. It should be CTS that controls flow and others (like DTR, DSR and CD) only change when things go on and offline. We can sense how terrible it is for you to suddenly gain another HP instrument to repair.
[Man, this is so confusing, I had to rewrite this 3 times!] CTS is an input on a DTE, and is an output on the DCE. Since there is no DCE in a null-modem DTE to DTE connection, the CTS signal has to be generated somehow, in this null-modem by looping RTS back. Same goes for CD, which is really a modem-only signal. This is a pretty standard null-modem arrangement (and is shown in the Silent manual). It does not rely on RTS/CTS control flow. It uses DSR/DTR as the control flow - the common way to deal with serial printers, which the Silent is. It will actually prevent RTS/CTS from hanging up the link and get communications going, which is quite convenient. You'd have to cross connect RTS to CTS of both CTEs to enable that flow control. By the way, the RTS/CTS flow control is a modern bastardization of the original flow control mechanism which was intended for half-duplex DTE-DCE modem links. It was not meant for full-duplex DTE-DTE, although it is now bent into doing that. Which makes things horribly confusing.
@@CuriousMarc Does it honour in-band XON/XOFF control flow characters? (I assume it doesn't implement HP's Enq/Ack protocol, which I thankfully haven't had to worry about since I stopped having to work on a HP2392.)
@Frank Wales No, I think only DSR/DTR, and maybe RTS/CTS but it would need to be wired with the different null modem that crosses 4 and 5. Haven’t tried yet.
its been some time since I used a NULL modem cable
Oh man, the discussion about how to wire up the control signals to allow two DTR devices to talk to each other directly took me WAAAAYY back to my days of trying to get an IBM 7171 "ASCII Device Attachment Control Unit" to talk to hard-wired VT220s. (cheaper than genuine IBM 3178s!)
The 7171 was the first thing I'd encountered that DEMANDED the RS-232 control signal protocols be obeyed exactly, and I worked out how they'd have to be looped back and crossed over to simulate a set of nonexistent modems between the two devices. I was overjoyed to see that same exact diagram in Marc's notebook here!
We made up a bunch of hacked-up cables and called them "null modems" because they faked an entire data signalling system. Later we were a little embarrassed when things like the little dongle in the video became available, and were named the same thing.
These RS-232 videos really take me back in time! Thanks so much for them.
No reason to be embarrassed. That just shows how your ingenuity and problem solving skills and even the naming of the device aligned with the solutions of the greatest minds at the time. It is something to be proud of even if you weren't the first to think about it.
It's also funny how people played fast and loose with the standards. I remember having to set flow control type (CTS/RTS, DSR/DTR, or software XON/XOFF) because every system seemed to have a different idea of how to control the flow of data. Then came the era of the HP48 calculators, where flow control pretty much disappeared.
You're not the only one who dealt with the 7171; I did just like you did. We had to make up 128 adapters to connect the 7171 to the Gandalf port selector. And we were supplied with 128 full 25 pair cables that we had to pull under the raised floor. The IBM 4361 mainframe was used to teach Fortran and Ada, and it was absurdly underpowered. It would take 8 minutes to compile an Ada program with one single user and slow down another 8 minutes for each additional user. This wasn't acceptable for a college class, so it was replaced after only a few years. I still have some 7171 manuals stored somewhere.
We had one of these to communicate with the Ontario Hydro Research Division PALC Auxilary Logic computer. It used Motorola 6809 processor if memory serves. A rack mounted alternative to a to a bunch of relay logic...the 700 just worked when you plugged it in and hit the right keystrokes...
Maybe it used the X/On and X/Off command sequences to control the sending computer when the buffer gets near full? That’s usually how async printers in this era worked. Put your scope on the Tx line to see if it’s sending them?
Ah the hours I spent connecting VT terminals to VAX terminal servers back in 1986... I would have liked one of those 4957s back then.
Oh, and DSR is "Data Set Ready" if I remember correctly.
These were extremely popular for phone men who worked on EPABXs to carry.
Ahh cool! Along with the 707/1200, I have come across one other 703 that was fitted with a replacement ROM allowing it to be switched between 300 & 1200 baud. While it could send ok at 1200 baud, the internal buffer filled up pretty quickly when receiving data causing it to start printing gibberish. The 707/1200 just prints super fast at the expense of it not being quite as silent :)
Thanks for the info. Do you have a way to dump that special ROM?
@@CuriousMarc Unfortunately, I gave it away last year during a clear out :(
15:20 RS-232 Gymnastics - new Olympic sport.
I remember that start-up sequence, and the idiosyncratic head movement! Your 703 looks like a newer one than I had access to... I also remember going through all of the combinations and permutations of RS-232 wiring :-(
Thanks, Marc.
I've got a TI DS990 Model 1, which is essentially a board upgraded TI Model 770, which is basically a fancy CRT based Silent 700 series data terminal.
I wish the DS990 Model 1 was this easy to get up and running! I've been trying off and on for years!
serial rs-232 comms can be fun and madening at the same time! I remember spending hours of time researching and setting dip switches and rotary encoders for fixed clock UARTS... Fun times! only certain ports could do certain speeds and protocol combinations, because of the way they were feed the divided clock timing generator.
Years ago i had the version of the 700 that had an acoustic couple modem. Kinda cool
Boy, this brought me back a few years... especially that null modem adapter; I've still got a half-dozen of them floating around here. Never know when you might need them! 😋
As cool as this is, i still prefer the chattering of the model 19 TTY.
Agreed! The 19 is loud, but has this strange rhythmic and soothing quality. The ASR 33 on the other hand sounds like it’s hyper, it drunk too much caffeine! I always half expect seeing springs and screws being shaken out of the mechanism...
@@CuriousMarc that is the most accurate description of the 33 i ever read! 😆
@@CuriousMarc The Model 19 was my first machine as a teenage ham! It used to shake the house! There was something warm about the big rectifier when the tubes would flash over purple and I knew there was 60 mA loop current available and off I went onto the airwaves with RTTY!
I have the same! beautifull machine!
I wonder if it was an in house developmental example machine or something that walked home with someone eventually.
Man that HP at the start is awesome!
12:47 you wrote 1994 in the captions, but it says 1984 on the chip ;) But still surprised how many full-size components are in this device, mixed in with the ICs. So used to microscopic surface components in modern stuff, you don't see the full-size versions much anymore on boards with ICs
Yes you are correct, it's 1984! Which makes a lot more sense. But then it's an early machine, not a late one, and as the other commenters found out, it's an AT&T special that has 1200 baud but would not have the double-speed head. Weird animal it is!
Takes me back to monitoring the old Sat-Nav navigation system on an offshore seismic survey ship back in '82 but then it was just a Silent 700.
Greetings from Germany. Love the videos on this channel.
Those 53904A Displays are very cool.
Im in the process of making a serial data cable for my Yaesu ham radio. Im really glad they dont use all the ready to send stuff. Only have to wire TX RX and ground, and its still a PITA.
Nice Paladin cable tester, I have the Mac one that does SCSI, ADB, and a few others. Mostly use it to test ethernet patch cables these days. Got it for $1 when a former employer went out of business ;)
Used to work with the same printer (In the 80's). Was 1200 bps default.
Fantastic...I don't have a clue...before my time...but it's just fantastic
Could you do a video on those hp display boxes by any chance?
voting for this comment =)
Having a serial port witht 1200 baud does not necessarily mean that the print mechanism can cope with that ~120 chars per second. This was either be managed by software flow control (Xon/Xoff) or hardware flow control (RTS/CTS if I remember correctly), depending on the device. I used a TI Omni800 a while ago which was capable to print ~130 chars/s and had a interface running up to 9600 baud.
It probably accespts at 1200 bps, and if things go too fast, hardware handshaking takes over. It also needs to have DSR active (this is on all TI silent 700's)
Those vestigial battery bays present some opportunities for non-invasive retrofitting. You could put in a single-board computer with BASIC and tell people it's a lost prototype microcomputer.
That's some weird equipment, a huge box that displays a number.
UART: universal asynchronous receiver transmitter. Correct?
Is it using the TI-99 keyboard?
My mom's real estate office had Silent 700s for accessing MLS in the 80s, before the internet was a thing. :)
That could have been helpful in 1983 for me - when was it released - as the computer I was using did not have a local printer and I depended on the mainframe chain printer to send my results back, which were crap as I missed out decimal points in the input data for calculations
crazy and awesome stuff!!!
Amazing video as always :)
hey, I have already done that since the early age of the agc
excellent
Ooooh.... imagine a paper tape on the teletype with a whole bunch of instructions for the numeric displays.... that would be very nice to watch.
Just got one of these myself, the serial port has been ripped out from it so I need to replace that, but weirdly mine had a footnote with it which says:
This Texas Instruments Model 703 Data Terminal has been quipped with a special PROM, which allows a data transmission of both 300 and 1200 bits/second (300 & 1200 BAUD). It also gives the ability to use the DC1/DC3 (or X-ON/X-OFF) Ready/Busy Protocol. NOTE: You can not use this PROM in conjunction with the Auto Access Cartridge, or any other cartridge.
To select the various options:
Simultaneously hold down CONTROL and SHIFT and press:
1 - Speed = 1200 BAUD
2 - Speed = 300 BAUD
D - DC1/DC3 ON
N - DC1/DC3 OFF
But... mine is missing this PROM chip, so looks like it was pulled to go into another one perhaps when this was replaced.
Oh, this was so promising until the last bit. If someone had the ROM it would be so easy to dump it and create clones!
400% overclock, those were the days.
Use one of those to dialup CompuServe back in the day.
nice and silent machine, just bit slow and papers tend to wipe out itself in hot, moist or sun, so it was better to make copy when possible :) , there was also bigger unit with bubble memory and 300 modem....😁
...with bubble memory...
That s something I expect to see on this channel someday.
Come on. You didn’t send 80085 to the HP Displays? 😄
👙
My problem was I ironed all my crinkley printouts and reciepts once, they all went black
If you follow the channel you'll know that Marc will use the term "if you follow the channel" within the first 30 seconds of each video. :-)
It's most probably a custom model for AT&T, although I wouldn't be suprised this particular model later led to the 703/1200 model later on (assuming the timelines match)
Try leaving a printout from one of these in a car during the summer! But the printer were cheaper and smaller, and the printouts very temporary. Old prints would fade considerably over time even in a controlled climate.
That's a nice unit.
Regretfully, I lack any self control and so about 18 hours after first seeing this video... I snapped one up on ebay. Not exactly sure what it is, but seemingly the exact later-model 700/1200 baud modem and RS-232 as well! Hopefully I'll find an adapter that works when it arrives...
Congrats! Well done.
regarding that last bit of fun you did; Imagine having a GPIB program on paper tape...
I liked how this episode ended with a prelude to a future repair. The End?
didn’t the serial analyser have a learn feature to figure out the speed&parity?
I wondered about that too, I thought baud-rate detection was common for test equipment of that time.
1:26 not a bad exchange rate 🙂
This is ideal for logging secutity logs. The ultimate immutable log.
I thought I could just communicate with an ATMega328. I only need hardware handshake and TTL to serial converter and + / - 12 volts.
Im always curious when I see an old ad with an phone number on it what happens if you call that number today - is it still active… is someone else using it meanwhile - is it dead 💀