electromechanical engineering is an artform that has been almost completely lost. It's very rarely used today because it's become cheaper to let a computer figure everything out
Back that days you "only" needed knowledge and a lot of time to build a computer. You could make it even a 100% electro mechanical based on relays. Today near nobody can build a computer without a computer anymore and time is much more expensive. IMO that knowledge of engeneering should be preserved no matter how much it costs. It may look overcomplicated and overengenered but it solved the same problems which are solved today by microcontrolers but in a mechanical way. I have a lot of respect for that people who only needed their brain, pen and paper to make this work.
I spent many Saturday mornings in the mid-70's playing with a 029 keypunch when i was a pre-teen, my dad would bring me along to work with him, and while he worked i made my best effort to pretend to be a programmer like my dad. 40 years later i'm still working away as a 2nd generation programmer. Miss you Dad!
Nowhere near the complexity. Modern Chips have MILLIONS of transistors, and the schematics are so intricate that people can’t manage them - the design must be computer-assisted.
The really neat feature (I don't remember if the 026 had this as well) was the program drum. You could set it up to do things like automatically skip to column 7 when you fed a new card, duplicate columns 8-60 from the previous card, and then go into numeric shift for columns 61-70. And endless variations on that. An amazing machine.
RIght, and auto numbering in the last colummns. For COBOL it was mandatory numbering the all the cards on the first six columns, but not in FORTRAN. You regretted not having numbered your cards when they were spread all across the floor.
This reminded me of a course I had in high school called "Introduction aux Machines de bureau" where we would dismantle and reassemble typewriters, photocopiers, and a big machine called IBM Selectric III Typewriter. This thing had thousands of mechanical parts, along with quite a few solenoïds and microswitches... It was a big job to take it apart, and a nightmare to put it back together! Especially for a 14-year-old boy that I was. Luckily we worked in teams. The goal was to have it work again. It seldom did, without help from the teacher. No internet in those days, and only a worn service manual with complicated schematics. Anyway, very interesting video on incredible engineering by IBM. Thanks for sharing!
@@b43xoit Never did see such a terminal, but I'd heard of it. BTW that golf ball system for printing was "digital". Well, mechanically digital anyway! Quite impressive engineering.
as part of my education in the 90s we had to learn to draft by hand. We all complained that we had to do it by hand when we otherwise did everything on CAD. The hand drawing class was only for six months and none of us ever had to do it by hand after that. When I then did my workterm (as part of the education), part of what I did while learning their CAD system was to redraw old handmade drawings in the computer, which was otherwise only done when we had to make corrections to these drawings as most of them were so old the original paper drawings were gone and we only had microfiche left. But anyhoo, making those exploded assembly drawings were a lot of fun (on the computer), and an absolute pain to do by hand because it required carefully positioning drawings on a light table to copy the individual parts, which we of course first had to make an isometric drawing of in the scale we needed for the assembly drawing. That is likely also why your instructions and the drawings don't quite match up. Typically the person who writes the documentation and instructions has nothing to do with the drawing, and also works from notes written by whichever engineer figured out how to take the machine apart and put it together again. My first job out of school was to make drawings for the documentation, which included removing any details that would turn into a blob when the drawing was shrunk down to fit in between the text. And all I got to work from was a list of drawings and a figure number so I knew where the drawing had to go in the text. So basically, I understand why your instructions were less than stellar, and apologize to anyone who had to use what I made, because it was probably equally as vague
I know this caused some trouble of the restoration team, but I respect you for making such devices accessible to visitors. As someone who wouldn't have had a chance to use such a device otherwise, but is very interested in the early days of computers, I thank you for your efforts.
The gravity disassembly reminds me of a mistake I made once, taking off a 4 slot toastmaster toasters case upside down. All of the tines that protected the elements fell out. I didn't get a good look at how they went in, and now had to put the ~32 tines back in one at a time guessing how they fit, and get them to line up with the case as I slid it back down over the tight fitting and delicate knobs, a real brow sweating under the leg handshake. When I went to test it, I realized that I'd put them in upside down... and I had to go back in.
I was an IBM CE for 15 years and worked on both the 026 and 029. We never worked on the keyboards -- just sent them back to the factory for repair. Good luck! Same story on the 1052 keyboards. However, I did have to work on the Selectric typewriter mechanism in the 1052 console. It was a beast. We had several operators who actually "pounded the keyboard" and broke the keys.
I started writing Assembly, Fortran and JCL for the IBM system 360 using both the 26 and 29 card punches. You get in a rhythm with the timing of the punch. If you try to type faster than the punch cycle you just end up waiting for each key to unlatch. Much faster than the Model 33 Teletype with it's motor driven mechanical latching cycle running 10 char per second (110 BAUD). At 10 keys per second you really slow down and wait on the each key. There was no concept of the deep multi key type ahead buffer of a modern computer. Thanks for the memories of waiting for the keys to drop!!
I was using Model 33s during the same period I was taking typing in high school. I think being forced to find a rhythm on the 33s helped me in the typing class! But even at 10 characters per second, the timesharing computer those Teletypes were connected to (an XDS Sigma 7) still needed line buffers to keep from missing characters.
IBM loved relays. In fact, I'm typing this comment using an all-metal, reproduction IBM F77 keyboard (IBM PN #6019303), which includes a relay that actuates on each key press. The sound and the vibration from the relay are wonderful and make the whole typing experience divine! Three cheers for the brilliant engineers of IBM !!!
Nice. I used 26 and 29 keypunches in my first two years of university. One got pretty slick using them. Editing cards was a neat trick. You can do insertion, overwrite and deletion with care. I don't think any of us had any clue the punches were so sophisticated mechanically.
This reminds me of the NCR technician I knew, who was called to fix a problem on one of those early electromechanical bookkeeping machines. After disassembly and fixing and cleaning and lubricating and reassembly, the machine did not work. After a second round of disassembly, a secretary noticed his frustration and walked over. After hearing his pained rant, she looked around and asked, "Have you turned on the wall socket?" Here in South Africa we have a switch for every wall outlet.😂😂😂
I'm really surprised your visitors could damage that keyboard. We had three KP 029 punches at my highschool in the 1970s and we (grade 10 students) were not abusive but we were not gentle. I don't recall any of the punches becoming damaged and unable to be put into service. Those things were tanks. As always, love your videos Marc and Crew!
Oh, this brings back some wonderful memories. High school, 1971, sixteen-year-old me sitting at the 029 writing my very first program, in FORTRAN. I loved that class. I knew I had found my calling.
When I trained as IBM CE in 1978, the product I trained on was the 029 keypunch and the System 32 computer. My service territory included 026, 029, 059, 129's in the keypunch family. Your probably aware that one of the differences between the 026 and 029 keypunches is that 026 used reed relays, while the 029 used wire contact relays. A common fix for the 026 was the thump, the reed relays to unstick them, while on the 029 one would replace the silver wires (two per contact) in the relays when needed. I can't imagine that there exist many of these wires available these days. I still have a tube of them which I would be willing to donate to your museum.
Oh my I so regret wreaking one of these only a few years back not knowing its heritage. I am so bored with our current focus on recycling and reparability laws etc and frankly all one needs to do to see a minimum requirement is to look at those manuals. These days you cant get a circuit diagram let alone something like this, but it is also true that they didn't use firmware etc so in some respects a hardware manual is limited. I love what you guys do to keep this stuff alive.
This brings back fond memories of submitting sheets of input to the punch card department and getting my deck back ready for an overnight run. We had a room full of those punch machines, all operated by the prettiest girls working at the company (in my eyes at least!). The racket in there was incredible as were the walls covered in posters of topless men! Not politically correct of course, but we referred to hardware, software and pinkware in those days.
Kinda reminds me of working on electro mechanical pinball machines. Relay banks, solenoids, reset bars and leaf switches. Surprised the mechanisms are not locked up with old grease!
Nice episode! I used such a keyboard / punchmachine in 1979 when I was a student at the university in Enschede, the Netherlands. We had to write programs in ALGOL. 😀
At the HIO in Eindhoven we had such a machine as wel. We used BEATHE, which was an extension of Algol. Burroughs Extended Algol Technische Hogeschool Eindhoven I thought.
OMG! Nostalgia! You do not want to know how many programs I punched on these; when I was a student and at the beginning of my professional life. We still used these things till the end of the eighties, beginning of the nineties.. IBM did not support these things anymore, so we found a retired IBM customer engineer who did the maintenance for us, not only for these punches but for all the other old IBM equipment we still had around. That guy was really amazing. Apart from punching, these machines were really amazing for the other things they could do; e.g. there was a function to check the contents of a bunch of punched cards by someone else retyping what was supposed to be on the original deck. If the punch found a difference, it would tell you. Like another youtuber said in a video about an old jukebox: electrical simplicity in combination with extreme mechanical complexity.
I remember that keyboard from the card punch machines I used at the University of Maryland and at my summer job at the Naval Research Laboratory. It's fun to finally see how they worked!
University of Maryland senior project - Write an IBM assembler in Fortran. Run it on Univac 1108, assemble a given program and have it run. Stuff of nightmares and 16 hour days. At one point dropped the stack. Lesson learned: Number the cards...
My dad (a lifetime IBMer) used to sit me down in front of one of those keypunch machines when he would go in on Saturdays to do some extra work. Despite their complexity, those things were pretty robust beasts. They stood up to a maniacal 8-year-old amusing himself wasting card stock for hours and hours, anyway. Odd to think only a few years later I'd be learning to program early switch-flip personal computers, then later things like the TRS-80...
4:06 my dad has the same or almost identical press you have on the right there. He inherited it from his father, and it has a plaque on it that says TO FRED - FOR ALL OF YOUR PRESSING NEEDS.
I remember using an IBM model 29 keypunch whilst in grade 6 (11 years old) on a field trip to a computer centre at the local Uni. I was hooked! I still have the two punch cards I made for the program they were running as a demo. Those mechanical systems and shift registers are super confusing! Amazing demo Marc! Thanks!
In the old days I spent many a happy hour punching program and data cards on both of these machines. Thanks for revealing the intricacies of this marvelous machine.
I was at the museum back during the summer and saw the key punch machine available for visitors to use, I was confused as I thought you lot must have accidentally left it out and running so I didn't touch it. Shame to see someone else didn't think the same...
Hello there. Just discovered your channel and subbed. Loved watching you guys fix stuff, very interesting. I have not been blessed with the brain to do this kind of stuff although I did use the IBM keypunch during my military service, 1965-1969. I see you guys do alot of Apollo stuff too. The ship I was on (uss yorktown, cvs 10) picked up the Apollo 8 after splash down, which as you know orbited the moon but did not land on it. They put the capsule on the hangar deck and built a special "ramp" and opened the door of the capsule so that we could take a look inside. I will never forget that. I wondered after seeing it how 3 guys could be so cramped for all that time. Hey thanks for your geniuses and the interesting things you do. I might be weird, but I find it relaxing.
Good job, although not as disastrous as I was anticipating, I was waiting for more bits to fall out after you took out the wrong screws, at least you had some documentation.
So nostalgic. My first programs were in FORTRAN on punched cards. Your video also explains the unique rhythm of the keypunch. Every key press generated a double ka-chunk accompanied by the constant whirring of the card motion motors.
spent many hours on 026s (after waiting for them to warm up) and 029s in early to mid 70s, programming an IBM 1130. now I've seen the insides of the 029!
I could fall in love with the mechanical sound when key-typing. Back in the early to mid-80's I had a Brother AX-10 electronic typewriter. I fell in love with its sound, and would get carried away while listening to pop radio in the background (synchronicity). I called it the Groove! I learned to type without any manual, and never received an adequate explanation for the qwerty system. The whole scheme of improvisation isn't bad but when I look at the "Read The Manual" for fixing these inventions.... I have to laugh! Someone had to know it all first and last, right??
Beautiful work, it kind of reminded me of the mechanics of a teletypewriter, although it didn't punch cards, but tapes and was able to read them. The adjustment was also a lot, it had to be precise, another delicacy was the rotary print head. A special measuring weight had to be used on some contacts so that the exact contact pressure could be set. Nice day 🙂 Tom
I'm sure glad IBM Selectric wasn't built that way. It's mechanical keyboard decoder to binary output was much more elegantly in designed. Credit is due to anybody that can work on that data entry beast.
Turning on the 029 and that squeak from the drive belt and motor always brought a feeling of comfort and relaxed the hair on the back of my neck -- she was going to work.
Awesome video Marc, it's always lovely to explore and repair an unknown piece of electro mechanical marvel like this - and I did titter when things didn't go entirely as planned. My golden rule, known by you and your colleagues but maybe not by all visitors, the application of force is rarely the answer.
I just landed on this page, and I can't wait to dive into it. I've been a fan of IBM keyboards since the original IBM PC, which was perfected in the Model M. I'd heard that some of the earlier typewriters from IBM had an insane number of adjustment points, so this will be quite an interesting video I think.
DEC called those documents IPB - illustrated parts breakdown. It's an immense help to fix things. It was 1976 that I took a Fortran class and we had to use those keypunches. The person with the cards then had to stand in line outside of the computer room and wait until he could get the operator to run the cards through the card reader, and run the program on the Burroughs B-2700 and get the printout. Thanks, Marc for bringing back the memories.
Insane was when I had a peep inside my mom's adding machine from sometime in the 70s. electomechanical and packed full of lever action. Wish I could find one.
I came to know the 029 (and the Univac equivalent the 1710) quite well, while aboard the USS Jason AR8 (Auxiliary Repair) in the early 1970s. My main job as a "data processor (DP)" rating was to keypunch/keyverify punch cards off of work orders and supply part lists for repair carried out by the USS Jason work force on ships - usually destroyers - berthed alongside. Cards were input into our discrete transistor/diode/core memory logic AN/UYK 5V computer (a militarized Univac 1218), via programs launched via teletype input with four tape reader/writers, to produce various supply reports by a mongamous chain printer, but only when tied up at the dock! At sea everything was powered down.
Brought back memories of keypunching Fortran IV programs on a IBM model 28 (?) back in high school in the 70's . We ran our programs on the local university IBM 360.. The program I wrote was a lunar lander program where you guessed the fuel rates and saw how deep of a new lunar you would probably make when you crashed... good ol' days !
When I first started programming the IBM 370/165 in Cambridge, I had to use punched cards. Working with the card punch machine wasn’t fun as any mistakes meant throwing the card away and punching another. That was 1979. Fortunately the next term they brought in a whole load of VDUs to give interactive access and program entry using the nice ZED editor.
First machine IBM UK trained me on. Never sent the keyboard back to the plant always repaired on Customer premises. Yes very complex but you got used to it and learned the common faults quick enough. The operators were hard task mistresses they did not like a 'slow' keyboard and a good idea to get their approval of a repair. The key - punch cycle had to be memorised, all the actions from pressing the key to a hole being punched. After 45 years it has faded from my memory. Wish I had an 029 / 129 today.
The IBM Selectric typewriter uses a similar interlock to permit only one key to be depressed. However, instead of washers between plates, think of ball bearings in a long tube. There is only enough spare space for a single keystem to be inserted, forcing all the balls together such that no other stem can enter.
That's just what I was thinking when I saw that clip. I always wondered why IBM made typewriters but I can see how they would already have all the engineering needed to make them, and make the best.
@@mikafoxx2717 I think it was really the other way around. They built electric typewriters first, and then figured they would make great I/O devices for some of their computers, which came along later. But I can definitely see some cross-breeding that resulted in the Selectric typewriter.
Waouh, I'm impressed. Most of the logic's was electromechanical by then. It was a real technical challenge to design and assemble such computers. The price of this set must have been considerable, not to mention the team that had to make it work. Surely you had to have a very important need to start buying (or renting) such a system.
I had to use an 029 for a programming course in my first year at University in 75/76. They broke down fairly often as many of us students would use them to punch out large dot-matrix signs and many of the "characters" required punching out 8 holes at a time ! 5 years later I was working at big blue and running an ancient 1430 as the processor for a logic card tester...
I'm glad they never wanted me to work on those... nearest I came was at Sudbury Towers when I got asked by a colleague to watch his back as he had to dig around the main electric motor, with it on, so I was ready to pull the plug !!
Hey Marc. Great video! One question: When you punched your data card, the last symbol you typed (after REPAIRED) was an exclamation mark (!). A close look at the card in freeze frame shows it had 3 punches on the last column: alpha row 2, numeric row 2 and numeric row 8. This does correspond to the "!" special character. However, your printout from the computer shows "REPAIRED-". What happened to the exclamation mark? Just curious, Marc. 😀
Correct. The 1403 can only print the few special characters from the 026, not the ones from the later 029. Modern witchcraft, those special characters.
Hi Marc. I've just watched this again and noticed at 21:59 where you are comparing the punched card to the pdf document, there are discrepancies in rows "7" and "8" for the "special characters". In the pdf, row 7 has 4 holes and row 8 has two spaces, but in the punched card, row 7 has 3 holes and row 8 has 3 spaces. (I may be wrong about row 8 though). Don't know if anyone else noticed.
@@CuriousMarc I went there last June, it’s huge, and full of the type of equipment you would love to see. And pretty much unknown as I was the only one there for a couple of hours. Definitely worth a visit.
Another travel back in time ... In the early 1980s, in my times at the University of Technology in Vienna, there were still a few of these IBM keypunch machines around. They were located on the corridor in front of the computing center in the Electrotechnical Institute building. As a student, I had to use one occasionally. To sign up for certain tutorials or practical exercises, we had to punch our matriculation number and name on a card and drop it into a dedicated letterbox. The tutors collected these cards and produced the list of participants from the stack. Until the mid 1980s, these keypunch machines were gone.
The system console for the IBM System/360 mainframes, the IBM 1052 Printer-Keyboard, had exactly the same mechanism as the 029 keyboard here but with a different case to accomodate some extra buttons on the sides. It was a mechanically and electrically separate unit but physically attached to a 1053 Keyboardless Printer unit. That's why the 1052 doesn't look quite like a regular Selectric, the distance between the top row of keys and the printer is longer. I have the FE service manuals for the 1052 and 1053 which includes servicing this amazing keyboard on my documents-to-be-scanned pile.
@@CuriousMarc OK sure I will bump the priority :) Been looking for a 1052 myself for years and years but I'll never find one, so have been working on a replica based on a Selectric I've split apart into the keyboard and power frame components, and generated the case shells from photogrammetry on 1052 images.
Oh there's some beautiful engineering in there. I especially love the interlock mechanism with the row of washers, so simple and yet so effective, as all engineering solutions should be! I hope I'll get the chance to see that machine in person someday.
So that’s what that kind of paper is for! I remember it just being scrap paper for drawing on at school. Old electro mechanical technology is the best.
23:23 I would expect some indication on which column you are typing. Or is that an expensive feature? Or is it visible enough in the card slot itself 🤔
I guess it was visible. Later, I saw punches for 96-column cards and they had a two-digit display for the column number, because it didn't start punching until you hit release, if I remember correctly. At one school I went to, they had an IBM 3 and there was one keypunch. If two people wanted to be punching cards, and if there was nothing else the computer had to be doing, one person could use the computer as a card punch. It had a keyboard for just that purpose, and I remember for quite sure the column number display on that.
Loved visiting the CHM on my last trip to SF. I dutifully punched my personal information into a punch card using the key punch visible at 0:18 and I still have it hung on a note board behind my desk.
Is that lay flat lacing tape I see on the loom? Gotta love that stuff, but my finger joints didn’t. I preferred the black pvc coated variety. Gave a nicer finish IMO.
There is a special place in Hell for people destroying museum equipment. Btw I have used such a 029 on many occasions in past. At the university we had 6 Sperry video terminals connecting to a remote Sperry Univac mainframe. Most students use these terminals, but there were never enough terminals so one often had to wait for another student to finish before you could use one. But there was also two 029 punches sitting unused in a small room next to the big terminal room - and a card reader connected to the remote mainframe. It did not take me long to find a stack of fresh punch cards, fire up the 029, and learn to use it. It turned out that it was actually a very effective method to quickly enter and/or make a change to a program, put the cards in the reader with 3-5 control cards to start a job, push the start button, and after 20-30 seconds the line printer would start spewing out your result (if you were lucky) - or a register dump (if you were not). Using the punch had another advantage: It was free! Using the online terminals meant that CPU time was accruing to your personal account, and each student was only allocated a certain amount of CPU time at the start of each semester. One could apply for more when it ran out, but it was a hassle. Editing a card, as in fixing a line of code, offline was free :) And as I said, very effective once you learned to use the copy function on the punch.
It's the list of punches it will punch on the card, as I explain in the video. That particular control character does not have a printable equivalent, so they just list its punch code.
Still had some of those at my university in 1983. I could get on one of those and complete my programming assignment in batch before other students could get on a terminal and log in to an interactive session.
I'm curious if this shares any mechanics with the selectric typewriter. Considering that used binary, albeit mechanical, it probably could have been a similar mechanism, although it does look quite different.
IBMs documentation standards are legendary, leading to the in house joke: Q) "How many IBM engineers does it take to change a lightbulb?" A) "Five, one to screw it in, and four to write the documentation 'Incandescent Light Source Replacement Procedure'".
My first interaction with a computer took place through an 029. The cool kids in the computer science lab already knew how to make program cards for the KP.
"...the days when mechanical engineers roamed and ruled the Earth." Classic.
It's such an art too. I'm so in love with this tech, to think it was so close ago, but so far away
electromechanical engineering is an artform that has been almost completely lost. It's very rarely used today because it's become cheaper to let a computer figure everything out
Back that days you "only" needed knowledge and a lot of time to build a computer. You could make it even a 100% electro mechanical based on relays. Today near nobody can build a computer without a computer anymore and time is much more expensive. IMO that knowledge of engeneering should be preserved no matter how much it costs. It may look overcomplicated and overengenered but it solved the same problems which are solved today by microcontrolers but in a mechanical way. I have a lot of respect for that people who only needed their brain, pen and paper to make this work.
I spent many Saturday mornings in the mid-70's playing with a 029 keypunch when i was a pre-teen, my dad would bring me along to work with him, and while he worked i made my best effort to pretend to be a programmer like my dad. 40 years later i'm still working away as a 2nd generation programmer.
Miss you Dad!
same pre teen experience❤miss dad too
Heh... I just added a similar comment a day later. Hail fellow childhood keypunch "operator"! 😋
Raised by my grandparents. Forty years later I am still pretending to be like him. I think I will get it right on my last day.
I sure miss him.
"..When mechanical engineers roamed and ruled the Earth..."
until the deadly meteor of THE TRANSISTOR hit and wiped almost their entire race of the face of the earth.
@@juango500 Who builds all the precision positioning machinery?
@@milantrcka121, yes, a few went underground and were saved. 😄
Let's bring them back and repopulate the species!
@@metatechnologist speaking as an old m.e., all i can say is that is a bad idea!!!
All the complexity of modern chips existed in prior generations, it was just expressed in a multitude of creative mechanical ways.
Nowhere near the complexity. Modern Chips have MILLIONS of transistors, and the schematics are so intricate that people can’t manage them - the design must be computer-assisted.
@@JordanOrlando True, true, I wasn't really referring to the density of gates or circuits, but the creative process to solve a problem.
@@JordanOrlando "millions"? Oh sweet summer child ....
@@8BitNaptime Too low, right?
@@mpbgp Yeah, of course you're right. It's a fascinating comparison either way.
The really neat feature (I don't remember if the 026 had this as well) was the program drum. You could set it up to do things like automatically skip to column 7 when you fed a new card, duplicate columns 8-60 from the previous card, and then go into numeric shift for columns 61-70. And endless variations on that. An amazing machine.
Yep, both machines had program drums. As did the 024, which is probably a verifier if I recall correctly.
See the program drum in operation here: ruclips.net/video/YnnGbcM-H8c/видео.html . Definitely my favorite feature too.
Starwheels keep on turnin'
Cardboard keeps on burnin'
Rollin'
Rollin'
Rollin' through a readuh.
RIght, and auto numbering in the last colummns. For COBOL it was mandatory numbering the all the cards on the first six columns, but not in FORTRAN. You regretted not having numbered your cards when they were spread all across the floor.
@@FernandoLichtschein..and it is said that a _real_ Fortran programmer can write Fortran in _any_ language.
This reminded me of a course I had in high school called "Introduction aux Machines de bureau" where we would dismantle and reassemble typewriters, photocopiers, and a big machine called IBM Selectric III Typewriter. This thing had thousands of mechanical parts, along with quite a few solenoïds and microswitches... It was a big job to take it apart, and a nightmare to put it back together! Especially for a 14-year-old boy that I was. Luckily we worked in teams. The goal was to have it work again. It seldom did, without help from the teacher. No internet in those days, and only a worn service manual with complicated schematics.
Anyway, very interesting video on incredible engineering by IBM. Thanks for sharing!
Did you ever see a 1050 terminal? It used a Selectric golf ball, but could send and receive over wires. There was a slow card reader for it, too.
@@b43xoit
Never did see such a terminal, but I'd heard of it. BTW that golf ball system for printing was "digital". Well, mechanically digital anyway! Quite impressive engineering.
@@AlainHubert Yeah, I read that the only electric thing on a Selectric is the motor.
as part of my education in the 90s we had to learn to draft by hand. We all complained that we had to do it by hand when we otherwise did everything on CAD. The hand drawing class was only for six months and none of us ever had to do it by hand after that. When I then did my workterm (as part of the education), part of what I did while learning their CAD system was to redraw old handmade drawings in the computer, which was otherwise only done when we had to make corrections to these drawings as most of them were so old the original paper drawings were gone and we only had microfiche left. But anyhoo, making those exploded assembly drawings were a lot of fun (on the computer), and an absolute pain to do by hand because it required carefully positioning drawings on a light table to copy the individual parts, which we of course first had to make an isometric drawing of in the scale we needed for the assembly drawing. That is likely also why your instructions and the drawings don't quite match up. Typically the person who writes the documentation and instructions has nothing to do with the drawing, and also works from notes written by whichever engineer figured out how to take the machine apart and put it together again. My first job out of school was to make drawings for the documentation, which included removing any details that would turn into a blob when the drawing was shrunk down to fit in between the text. And all I got to work from was a list of drawings and a figure number so I knew where the drawing had to go in the text. So basically, I understand why your instructions were less than stellar, and apologize to anyone who had to use what I made, because it was probably equally as vague
Thank You, CuriousMarc, for featuring this 029 keyboard in this video. I was hoping for such a fantastic close-up example of one of these...
My college computer room from 40 years ago had a lot of similar gear. I'm glad some of that stuff survived and is preserved and loved.
Marc to gamer, "Call _that_ a mechanical keyboard? _This_ is a mechanical keyboard!"
I know this caused some trouble of the restoration team, but I respect you for making such devices accessible to visitors.
As someone who wouldn't have had a chance to use such a device otherwise, but is very interested in the early days of computers, I thank you for your efforts.
The gravity disassembly reminds me of a mistake I made once, taking off a 4 slot toastmaster toasters case upside down. All of the tines that protected the elements fell out. I didn't get a good look at how they went in, and now had to put the ~32 tines back in one at a time guessing how they fit, and get them to line up with the case as I slid it back down over the tight fitting and delicate knobs, a real brow sweating under the leg handshake.
When I went to test it, I realized that I'd put them in upside down... and I had to go back in.
I was an IBM CE for 15 years and worked on both the 026 and 029. We never worked on the keyboards -- just sent them back to the factory for repair. Good luck! Same story on the 1052 keyboards. However, I did have to work on the Selectric typewriter mechanism in the 1052 console. It was a beast. We had several operators who actually "pounded the keyboard" and broke the keys.
I’ve used this! They had that keypunch at my school in the 1970s - for programming their 1130. I remember it fondly.
Same here! (And same era!)
1991 we hat still spirit duplicator at school and watched documentations on 16mm film (West Germany, elementary school). ^^
I started writing Assembly, Fortran and JCL for the IBM system 360 using both the 26 and 29 card punches. You get in a rhythm with the timing of the punch. If you try to type faster than the punch cycle you just end up waiting for each key to unlatch. Much faster than the Model 33 Teletype with it's motor driven mechanical latching cycle running 10 char per second (110 BAUD). At 10 keys per second you really slow down and wait on the each key. There was no concept of the deep multi key type ahead buffer of a modern computer.
Thanks for the memories of waiting for the keys to drop!!
I was using Model 33s during the same period I was taking typing in high school. I think being forced to find a rhythm on the 33s helped me in the typing class! But even at 10 characters per second, the timesharing computer those Teletypes were connected to (an XDS Sigma 7) still needed line buffers to keep from missing characters.
IBM loved relays. In fact, I'm typing this comment using an all-metal, reproduction IBM F77 keyboard (IBM PN #6019303), which includes a relay that actuates on each key press. The sound and the vibration from the relay are wonderful and make the whole typing experience divine! Three cheers for the brilliant engineers of IBM !!!
Nice. I used 26 and 29 keypunches in my first two years of university. One got pretty slick using them. Editing cards was a neat trick. You can do insertion, overwrite and deletion with care. I don't think any of us had any clue the punches were so sophisticated mechanically.
This reminds me of the NCR technician I knew, who was called to fix a problem on one of those early electromechanical bookkeeping machines. After disassembly and fixing and cleaning and lubricating and reassembly, the machine did not work. After a second round of disassembly, a secretary noticed his frustration and walked over. After hearing his pained rant, she looked around and asked, "Have you turned on the wall socket?" Here in South Africa we have a switch for every wall outlet.😂😂😂
Chapeau! I take my hat of to you and your team as you all seem to be born to venture into the most abstruse rabbit holes.
I'm really surprised your visitors could damage that keyboard. We had three KP 029 punches at my highschool in the 1970s and we (grade 10 students) were not abusive but we were not gentle. I don't recall any of the punches becoming damaged and unable to be put into service. Those things were tanks.
As always, love your videos Marc and Crew!
First rule of museums: there is nothing that visitors can’t break…
Oh, this brings back some wonderful memories. High school, 1971, sixteen-year-old me sitting at the 029 writing my very first program, in FORTRAN. I loved that class. I knew I had found my calling.
When I trained as IBM CE in 1978, the product I trained on was the 029 keypunch and the System 32 computer. My service territory included 026, 029, 059, 129's in the keypunch family.
Your probably aware that one of the differences between the 026 and 029 keypunches is that 026 used reed relays, while the 029 used wire contact relays. A common fix for the 026 was the thump, the reed relays to unstick them, while on the 029 one would replace the silver wires (two per contact) in the relays when needed.
I can't imagine that there exist many of these wires available these days. I still have a tube of them which I would be willing to donate to your museum.
Oh my I so regret wreaking one of these only a few years back not knowing its heritage. I am so bored with our current focus on recycling and reparability laws etc and frankly all one needs to do to see a minimum requirement is to look at those manuals. These days you cant get a circuit diagram let alone something like this, but it is also true that they didn't use firmware etc so in some respects a hardware manual is limited. I love what you guys do to keep this stuff alive.
I am not that old, but I still manage to remember registering for college classes with IBM cards. (1980's)
This brings back fond memories of submitting sheets of input to the punch card department and getting my deck back ready for an overnight run. We had a room full of those punch machines, all operated by the prettiest girls working at the company (in my eyes at least!). The racket in there was incredible as were the walls covered in posters of topless men! Not politically correct of course, but we referred to hardware, software and pinkware in those days.
Well, what was wetware referring to?
Brains!
Remember my step mom was a keyboard operator on one of these machines at the university. Thanks.
Wow.... those manuals really are works of art.
I literally failed mechanical drawing. And then went on to fail many quarters of the calculus. Fortunately, "software engineering" requires neither.
SWE requires the lambda calculus, not the infinitesimal calculus (usually).
Kinda reminds me of working on electro mechanical pinball machines. Relay banks, solenoids, reset bars and leaf switches. Surprised the mechanisms are not locked up with old grease!
2:42
"With just a pencil and a dream"
-Joey Drew
How do mathematicians work out constipation?
With a pencil.
Nice episode!
I used such a keyboard / punchmachine in 1979 when I was a student at the university in Enschede, the Netherlands. We had to write programs in ALGOL. 😀
At the HIO in Eindhoven we had such a machine as wel. We used BEATHE, which was an extension of Algol. Burroughs Extended Algol Technische Hogeschool Eindhoven I thought.
Just love when you repair those beautiful machines.
You have a wonderful sense of humor, Marc! @8:47, just one example. Fine work! Enjoyed the video. Thanks
OMG! Nostalgia! You do not want to know how many programs I punched on these; when I was a student and at the beginning of my professional life. We still used these things till the end of the eighties, beginning of the nineties.. IBM did not support these things anymore, so we found a retired IBM customer engineer who did the maintenance for us, not only for these punches but for all the other old IBM equipment we still had around. That guy was really amazing.
Apart from punching, these machines were really amazing for the other things they could do; e.g. there was a function to check the contents of a bunch of punched cards by someone else retyping what was supposed to be on the original deck. If the punch found a difference, it would tell you.
Like another youtuber said in a video about an old jukebox: electrical simplicity in combination with extreme mechanical complexity.
I remember that keyboard from the card punch machines I used at the University of Maryland and at my summer job at the Naval Research Laboratory. It's fun to finally see how they worked!
University of Maryland senior project - Write an IBM assembler in Fortran. Run it on Univac 1108, assemble a given program and have it run. Stuff of nightmares and 16 hour days. At one point dropped the stack. Lesson learned: Number the cards...
We had a 029 and a 026.
@@milantrcka121 you didn't get the idea to run a diagonal pencil line over the stack ?
@@TheStefanskoglund1 Hmmm. No I did not (1974)
My dad (a lifetime IBMer) used to sit me down in front of one of those keypunch machines when he would go in on Saturdays to do some extra work. Despite their complexity, those things were pretty robust beasts. They stood up to a maniacal 8-year-old amusing himself wasting card stock for hours and hours, anyway. Odd to think only a few years later I'd be learning to program early switch-flip personal computers, then later things like the TRS-80...
4:06 my dad has the same or almost identical press you have on the right there. He inherited it from his father, and it has a plaque on it that says TO FRED - FOR ALL OF YOUR PRESSING NEEDS.
These machines are always so impressive to me. Could you imagine developing that in what had to be the late 50's?
That reveal was really funny. The repair in itself must have taken quite a bit of time. I think you were doing a great job.
I remember using an IBM model 29 keypunch whilst in grade 6 (11 years old) on a field trip to a computer centre at the local Uni. I was hooked! I still have the two punch cards I made for the program they were running as a demo. Those mechanical systems and shift registers are super confusing! Amazing demo Marc! Thanks!
In the old days I spent many a happy hour punching program and data cards on both of these machines. Thanks for revealing the intricacies of this marvelous machine.
I was at the museum back during the summer and saw the key punch machine available for visitors to use, I was confused as I thought you lot must have accidentally left it out and running so I didn't touch it. Shame to see someone else didn't think the same...
Hello there. Just discovered your channel and subbed. Loved watching you guys fix stuff, very interesting. I have not been blessed with the brain to do this kind of stuff although I did use the IBM keypunch during my military service, 1965-1969. I see you guys do alot of Apollo stuff too. The ship I was on (uss yorktown, cvs 10) picked up the Apollo 8 after splash down, which as you know orbited the moon but did not land on it. They put the capsule on the hangar deck and built a special "ramp" and opened the door of the capsule so that we could take a look inside. I will never forget that. I wondered after seeing it how 3 guys could be so cramped for all that time. Hey thanks for your geniuses and the interesting things you do. I might be weird, but I find it relaxing.
Nice! The Apollo 8 command module you picked up is now on display in Chicago, in the Museum of Science and Industry.
Good job, although not as disastrous as I was anticipating, I was waiting for more bits to fall out after you took out the wrong screws, at least you had some documentation.
So nostalgic. My first programs were in FORTRAN on punched cards. Your video also explains the unique rhythm of the keypunch. Every key press generated a double ka-chunk accompanied by the constant whirring of the card motion motors.
spent many hours on 026s (after waiting for them to warm up) and 029s in early to mid 70s, programming an IBM 1130. now I've seen the insides of the 029!
I could fall in love with the mechanical sound when key-typing. Back in the early to mid-80's I had a Brother AX-10 electronic typewriter. I fell in love with its sound, and would get carried away while listening to pop radio in the background (synchronicity). I called it the Groove! I learned to type without any manual, and never received an adequate explanation for the qwerty system. The whole scheme of improvisation isn't bad but when I look at the "Read The Manual" for fixing these inventions.... I have to laugh! Someone had to know it all first and last, right??
Marc, you and each & every one of your friends consistently amaze me.
Beautiful work, it kind of reminded me of the mechanics of a teletypewriter, although it didn't punch cards, but tapes and was able to read them. The adjustment was also a lot, it had to be precise, another delicacy was the rotary print head.
A special measuring weight had to be used on some contacts so that the exact contact pressure could be set.
Nice day 🙂 Tom
I'm sure glad IBM Selectric wasn't built that way. It's mechanical keyboard decoder to binary output was much more elegantly in designed. Credit is due to anybody that can work on that data entry beast.
Nice job and extremely clear explanation as usual! By the way if you think your computer is noisy, just hear this one 🙂
Thanks for this nice video!!! i have worked with IBM 029 many times punching cards for programs writen in assembler 360 computer language
Turning on the 029 and that squeak from the drive belt and motor always brought a feeling of comfort and relaxed the hair on the back of my neck -- she was going to work.
Awesome video Marc, it's always lovely to explore and repair an unknown piece of electro mechanical marvel like this - and I did titter when things didn't go entirely as planned. My golden rule, known by you and your colleagues but maybe not by all visitors, the application of force is rarely the answer.
I just landed on this page, and I can't wait to dive into it. I've been a fan of IBM keyboards since the original IBM PC, which was perfected in the Model M.
I'd heard that some of the earlier typewriters from IBM had an insane number of adjustment points, so this will be quite an interesting video I think.
DEC called those documents IPB - illustrated parts breakdown. It's an immense help to fix things.
It was 1976 that I took a Fortran class and we had to use those keypunches. The person with the cards then had to stand in line outside of the computer room and wait until he could get the operator to run the cards through the card reader, and run the program on the Burroughs B-2700 and get the printout. Thanks, Marc for bringing back the memories.
Insane was when I had a peep inside my mom's adding machine from sometime in the 70s. electomechanical and packed full of lever action. Wish I could find one.
I also got a CZUR book scanner for archiving manuals like that!
I came to know the 029 (and the Univac equivalent the 1710) quite well, while aboard the USS Jason AR8 (Auxiliary Repair) in the early 1970s. My main job as a "data processor (DP)" rating was to keypunch/keyverify punch cards off of work orders and supply part lists for repair carried out by the USS Jason work force on ships - usually destroyers - berthed alongside. Cards were input into our discrete transistor/diode/core memory logic AN/UYK 5V computer (a militarized Univac 1218), via programs launched via teletype input with four tape reader/writers, to produce various supply reports by a mongamous chain printer, but only when tied up at the dock! At sea everything was powered down.
Brought back memories of keypunching Fortran IV programs on a IBM model 28 (?) back in high school in the 70's . We ran our programs on the local university IBM 360.. The program I wrote was a lunar lander program where you guessed the fuel rates and saw how deep of a new lunar you would probably make when you crashed... good ol' days !
Awesome, so good. Thanks for the post!
When I first started programming the IBM 370/165 in Cambridge, I had to use punched cards. Working with the card punch machine wasn’t fun as any mistakes meant throwing the card away and punching another. That was 1979. Fortunately the next term they brought in a whole load of VDUs to give interactive access and program entry using the nice ZED editor.
First machine IBM UK trained me on. Never sent the keyboard back to the plant always repaired on Customer premises. Yes very complex but you got used to it and learned the common faults quick enough. The operators were hard task mistresses they did not like a 'slow' keyboard and a good idea to get their approval of a repair. The key - punch cycle had to be memorised, all the actions from pressing the key to a hole being punched. After 45 years it has faded from my memory. Wish I had an 029 / 129 today.
The entire video my face was switching between 😐 and 😟. That thing is at once absolutely beautiful and an eyesore.
The IBM Selectric typewriter uses a similar interlock to permit only one key to be depressed. However, instead of washers between plates, think of ball bearings in a long tube. There is only enough spare space for a single keystem to be inserted, forcing all the balls together such that no other stem can enter.
That's just what I was thinking when I saw that clip. I always wondered why IBM made typewriters but I can see how they would already have all the engineering needed to make them, and make the best.
Western Electric multi line phones used a similar interlock to prevent connecting two lines together.
@@mikafoxx2717 I think it was really the other way around. They built electric typewriters first, and then figured they would make great I/O devices for some of their computers, which came along later. But I can definitely see some cross-breeding that resulted in the Selectric typewriter.
Hello Carl! We miss you and your IBM well of knowledge!
After realising it's Carl I read this comment with his voice in my head.
Waouh, I'm impressed. Most of the logic's was electromechanical by then. It was a real technical challenge to design and assemble such computers. The price of this set must have been considerable, not to mention the team that had to make it work. Surely you had to have a very important need to start buying (or renting) such a system.
That’s perseverance! Thanks, Marc 👍
I had to use an 029 for a programming course in my first year at University in 75/76. They broke down fairly often as many of us students would use them to punch out large dot-matrix signs and many of the "characters" required punching out 8 holes at a time ! 5 years later I was working at big blue and running an ancient 1430 as the processor for a logic card tester...
I'm glad they never wanted me to work on those... nearest I came was at Sudbury Towers when I got asked by a colleague to watch his back as he had to dig around the main electric motor, with it on, so I was ready to pull the plug !!
I loved the feel of the 029 keyboard. I want a keyboard that feels and sounds like that to use with my computers today.
Hey Marc. Great video! One question: When you punched your data card, the last symbol you typed (after REPAIRED) was an exclamation mark (!). A close look at the card in freeze frame shows it had 3 punches on the last column: alpha row 2, numeric row 2 and numeric row 8. This does correspond to the "!" special character. However, your printout from the computer shows "REPAIRED-". What happened to the exclamation mark? Just curious, Marc. 😀
The IBM 1403 Printer used doesn't have a ! in its 48 character set and - is presumably the default for unsupported characters
Correct. The 1403 can only print the few special characters from the 026, not the ones from the later 029. Modern witchcraft, those special characters.
Oh goodness, flashbacks of FORTRAN programming using that terminal at UC Davis in the early 70's!
With all the Apollo content, I totally forgot why I found this channel in the first place 😅
Hi Marc.
I've just watched this again and noticed at 21:59 where you are comparing the punched card to the pdf document, there are discrepancies in rows "7" and "8" for the "special characters".
In the pdf, row 7 has 4 holes and row 8 has two spaces, but in the punched card, row 7 has 3 holes and row 8 has 3 spaces. (I may be wrong about row 8 though).
Don't know if anyone else noticed.
Is there a name for that dot-pattern metal plate material, as seen at 20:17? Is there a reason for the dot pattern?
I am wondering if you ever went to the computer museum in Namur.. looked
like a dream place for a curious Marc.. :)
What? I did not even know there was a computer museum in Namur! Count me in during my next visit to Belgium!
@@CuriousMarc I went there last June, it’s huge, and full of the type of equipment you would love to see. And pretty much unknown as I was the only one there for a couple of hours. Definitely worth a visit.
The key interlock is very similar to the Selectric type writer, which is a tube full of ball bearings, with slots for each key.
A repair, with a free reverse-engineering thrown in at no extra charge!
Another travel back in time ...
In the early 1980s, in my times at the University of Technology in Vienna, there were still a few of these IBM keypunch machines around. They were located on the corridor in front of the computing center in the Electrotechnical Institute building.
As a student, I had to use one occasionally. To sign up for certain tutorials or practical exercises, we had to punch our matriculation number and name on a card and drop it into a dedicated letterbox. The tutors collected these cards and produced the list of participants from the stack.
Until the mid 1980s, these keypunch machines were gone.
Nice. I have an old selectric that is quite satisfying to use.
The sweet spot of electromechanical engineering imho.
Selectrics are awesome!
The system console for the IBM System/360 mainframes, the IBM 1052 Printer-Keyboard, had exactly the same mechanism as the 029 keyboard here but with a different case to accomodate some extra buttons on the sides. It was a mechanically and electrically separate unit but physically attached to a 1053 Keyboardless Printer unit. That's why the 1052 doesn't look quite like a regular Selectric, the distance between the top row of keys and the printer is longer.
I have the FE service manuals for the 1052 and 1053 which includes servicing this amazing keyboard on my documents-to-be-scanned pile.
oh please scan it! I will have a 1052 to repair pretty soon.
@@CuriousMarc OK sure I will bump the priority :) Been looking for a 1052 myself for years and years but I'll never find one, so have been working on a replica based on a Selectric I've split apart into the keyboard and power frame components, and generated the case shells from photogrammetry on 1052 images.
@@1944GPW contact me through the link in the video description. We’ll find you a 1052…
@@CuriousMarc That... would be supremely awesome. I've just started levering out the rusty staples now.
👍 👍 👍 👍 👍 👍 👍 🍀 Wow! What a video! Vintage devices are really so fascinating! I'm also doing sometimes sth about old electronics! 🤪🎥
Oh there's some beautiful engineering in there. I especially love the interlock mechanism with the row of washers, so simple and yet so effective, as all engineering solutions should be! I hope I'll get the chance to see that machine in person someday.
This looks like my day, only you actually made progress. I only got burns.
So that’s what that kind of paper is for! I remember it just being scrap paper for drawing on at school. Old electro mechanical technology is the best.
23:23 I would expect some indication on which column you are typing. Or is that an expensive feature? Or is it visible enough in the card slot itself 🤔
I guess it was visible. Later, I saw punches for 96-column cards and they had a two-digit display for the column number, because it didn't start punching until you hit release, if I remember correctly. At one school I went to, they had an IBM 3 and there was one keypunch. If two people wanted to be punching cards, and if there was nothing else the computer had to be doing, one person could use the computer as a card punch. It had a keyboard for just that purpose, and I remember for quite sure the column number display on that.
The position counter is located at the bottom of the drum, visible through the window at the center of the machine.
I still used these for my Cobol and Fortran classes at the U of A early eighties. Tucson.
Loved visiting the CHM on my last trip to SF. I dutifully punched my personal information into a punch card using the key punch visible at 0:18 and I still have it hung on a note board behind my desk.
Is that lay flat lacing tape I see on the loom? Gotta love that stuff, but my finger joints didn’t. I preferred the black pvc coated variety. Gave a nicer finish IMO.
Wow, the history lessons are great.
There is a special place in Hell for people destroying museum equipment.
Btw I have used such a 029 on many occasions in past. At the university we had 6 Sperry video terminals connecting to a remote Sperry Univac mainframe. Most students use these terminals, but there were never enough terminals so one often had to wait for another student to finish before you could use one.
But there was also two 029 punches sitting unused in a small room next to the big terminal room - and a card reader connected to the remote mainframe. It did not take me long to find a stack of fresh punch cards, fire up the 029, and learn to use it. It turned out that it was actually a very effective method to quickly enter and/or make a change to a program, put the cards in the reader with 3-5 control cards to start a job, push the start button, and after 20-30 seconds the line printer would start spewing out your result (if you were lucky) - or a register dump (if you were not).
Using the punch had another advantage: It was free! Using the online terminals meant that CPU time was accruing to your personal account, and each student was only allocated a certain amount of CPU time at the start of each semester. One could apply for more when it ran out, but it was a hassle. Editing a card, as in fixing a line of code, offline was free :) And as I said, very effective once you learned to use the copy function on the punch.
In that place in Hell, you have to code in trinary on a machine that drops half the trits.
oooh, that washer interlock mechanism is really slick!. Can we please return to an electromechanical world?
Can someone tell me please what is the 0-8-2 over the key T? At 18:52 you can see many more of these 3 number combinations over the keys. Strange.
It's the list of punches it will punch on the card, as I explain in the video. That particular control character does not have a printable equivalent, so they just list its punch code.
@@CuriousMarc Thanks a lot, Marc!! I should have found it out, though :) Congrats for taming this beast!
Awesome ❤️
Still had some of those at my university in 1983. I could get on one of those and complete my programming assignment in batch before other students could get on a terminal and log in to an interactive session.
I'm curious if this shares any mechanics with the selectric typewriter. Considering that used binary, albeit mechanical, it probably could have been a similar mechanism, although it does look quite different.
18:08 I see the @ symbol in the left corner. What was this used for before email?
"27 widgets @ $1.25 each" and the like
IBMs documentation standards are legendary, leading to the in house joke:
Q) "How many IBM engineers does it take to change a lightbulb?"
A) "Five, one to screw it in, and four to write the documentation 'Incandescent Light Source Replacement Procedure'".
My first interaction with a computer took place through an 029. The cool kids in the computer science lab already knew how to make program cards for the KP.
Second only to Monotype and Linotype! A joy for ever. No NKRO on this one, haha!
Modern witchcraft to some, advanced tech to others... right?
Your daughters are adorable Marc! and mechanical engineers STILL rule the Earth ;)