Ignore the people that post negative comments about you not actually working on something. You rambling on is still pretty informative a good portion of the time.
I bet these are the same people that bicker around when some anime has a bunch of "story" episodes. Oh, the bickering in the conference episodes or the current festival arc of Tensura. 😆
"Why can't you/we just *do something*" is always the cry of a fool, no matter the context. All real engineering and science is primarily done in the head. The tools are there to test and verify to refine those mental models, so that you can communicate them in a reliable way in the future, even to yourself - i.e., in the context of fixing something. A real master mechanic usually minimizes his wrenching to what is necessary to get the job done, after spending much mental time and energy optimizing and refining the tasks, often inventing his own theories and practices along the way.
The paperclip tape reader is almost exactly what I have built some 40 years ago for my TI-57. I was not reading bytes, I was establishing a contact between the calculator's keyboard rows and columns wires to simulate key presses. Long story short, it worked pretty well. I could punch programs on bristol cards (much more resistant than paper) and have them loaded in the calculator's memory in 10 seconds, almost every time. I had performed a sort of trepanning on my calculator, which had several DIN plugs that I could plug into my card reader. The reader was made out of wood, and the contacts were copper wires that had been curled (not paperclips) so they touched a single sheet of aluminium foil glued on the bottom plate. It worked because the calculator's chip was doing proper debounce filtering on the keyboard input, so no debounce logic was needed on the reader itself. I'm pretty sure I still have this thing somewhere in my attic.
I used to volunteer at TNMoC at Bletchley Park and I was always amazed by the tube-based design of the PTR on the Colossus rebuild. it's incredible that it actually works and more so manages it at the speed the tape flies past the reader.
I've seen it once in real life when I went to VCF GB in 2010. I seem to recall that someone told me the computer itself is capable of running much faster, but it's limited by the physical strength of the paper which would just tear if it were run too fast. A PET film like Mylar might help that, but that hadn't been invented in the 1940s so it wouldn't be period-correct.
@@Zeem4 I wonder if they could use film stock? (Though being nitrocellulose film stock at the time... that could be exciting at high speeds. Exothermically exciting.)
This "machine" was the reason why I started watching your channel 🙂After realizing what you else are doing, I watched all videos one after the other and became a Patreon 🙂 Keep going!
Thank you so much! Now that we have a design I actually really like on the tube computer, I'm finally fired up on it again. I'll get to designing a new block to machine out and in the next episode in the series we'll get the mill fired up again!
Photo detector is sentive to infrared light so will be overloaded when the tape picked up heat from the lamps I found this when repairing 16 mm sound projectors you should include a fan to keep it cool in use for a long time
I'm currently planning a vintage computer to monitor my property systems. Generator output, run time, bio gas production, solar output, well and cistern data, along with all the error messages I'd want printing out on an old dot matrix printer with the endless paper. I just wanna say how thankful I am to this channel and the patrons who support it for providing such detailed explanation and theories about these vintage machines right down to the basics.
I appreciate meeting some of my fellow High Scorer and 4-bit patrons, and the many emotional-support fans (just as important as us money-bags lol), at some of the events where we appreciate the hobby. And people like Bob at System Source who provide the floor space and electricity to let us get together and enjoy each other's company.
The USAF was using a optical tape reader ,as late as 1979, to input data into a TEST SET. This test set checked the status of a W-62 Nuclear Weapon that sat atop the Minuteman III missile. The optical reader alignment was very problematic and was constantly causing us problems. Very interesting video!
Worked in a manufacturing facility (of a LARGE multi-national manufacturer) where we were loading programs into 16-bit minicomputers on the factory floor with paper tape well into the late 1980's. Our readers were flat-bed units (I *think* made by a company called "Trend") which had no spools, so we'd place a roll of tape onto a screwdriver shank and hold it by hand to let it feed into the reader and then onto the floor. The tape had to be re-rolled by hand. Dirt and static usually combined to clog the holes somewhere resulting in the need for several attempts before getting a good read all the way through. We ultimately replaced the tapes and readers with IBM PCs to emulate the readers RS-232 output for quite some time before the minicomputers were finally replaced with modern computers. Good times.
@@yohannwilkerson6058 beats me why more companies didn't use the fanfold paper tape favoured by DEC, since that neatly stacked itself downstream of the reader, ready to be read again, without any re-rolling needed.
To the channel Members: We hope you feel appreciated, because you certainly are! This is a great channel. Thank you VERY much for keeping it rolling! All good wishes.
Use mylar paper tape - I used that in the 1960's on a PDP-8 - lasted 6 months used multiple times a day. The only down side is the mylar will shorten the life of the paper tape punch device - debug with regular paper tape and switch to mylar after all is good.
This. I used a french 'Logabax' computer with a small tape loop to set the vertical tabbing on the printer (it ran with payslips and other oddly-spaced forms). That was mylar tape as it ran continuously when printing. We also had a manual punch that was a drilled block of perspex and a punch needle. The index holes might have been prepunched. A loop of tape reinforced with a kapton layer might also work if you don't need the tape itself to be period-correct.
There was an 1959 NASA computer that was in the "dungeon junk room" at NCSU in 1970 that had a mylar paper tape loop attached to a Friden Flexowriter paper tape reader - but it could have been from the early 1960's. NCSU had a few old NASA computers - I did learn to program the analog TR-48 that they provided to NCSU. I did not start touching computers until 1963 and they were punched cards and not paper tape.
I appreciate the backstory and ramblings as this gives me an idea of what you're thinking of doing. This sort of context also helps why you decide to choose $procedure and decided to stick with it, as it is a logical decision-making process which do make a lot of sense, and this also helps with cutting out wasting time with experimenting with stuff, only to find it doesn't work, or it takes up too much room etc etc Keep it up! 👍
Looots of germanium semiconductors were photosensitive anyway - that's why their glass housings were painted black :-) But that specific photo-diode you found most probably beats them all
Aw, man, this takes me back. It was content like this that made me subscribe to this channel. I really loved the days when you were more focused on pre-vintage computers, and not vintage computers. Not to say I don't like the vintage computers, but there's many vintage computing channels out there. There was only one pre-vintage computers channel, and it was this one. It made you feel unique. The exposition made you feel unique. The enthusiasm made you feel unique. I'm always shocked when people don't like choices you make, because they make this channel unique. I don't want you to be like every other vintage computer channel, I want you to be unique. I like this unique channel. It's great to see the low-voltage pre-vintage computer stuff again. I look forward to more!
Being from the TRS-80, Vic20 era it’s amazing to realise how historical these tube computers are. Something like 20yrs from the first computers to when I was born.
Makes me feel really old... I started with computers back in the paper tape age, 1967 to be exact. I still have a few of my programs from back then on tape, but, alas, no reader. Back then, we had optical readers that could reach speeds of 2000 bytes a second, 5 meters of tape a second. It was hooked up to a transistor based computer, germanium transistors like the OC141and SFT228. Thanks for your efforts to bring all this to life !
That brings me back also. I still remember building a simple radio with germanium transistors like the OC13 and later the OC72. I still can remember scraping off the black paint from a OC13, because it made the transistor sensitive for light. I wanted to build a light detector that way, but went for LDR's later on (although they had a terrible response time). I still have a few of them laying around dating from all those years back.
I know it's very weird, but I use OpenSCAD for 3d modelling because I prefer its descriptive approach to the regular drag-and-drop sort of paradigm. Also worth mentioning that it is integrated into FreeCAD, although I just use it standalone.
I second this. OpenSCAD is a great 3d modeler for software developers because of it's C-like syntax. When you need the parts to be in exactly the right position, it's the perfect tool for the job.
OpenSCAD is great. It matches the way I think when designing objects (previously I used to dabble in PovRay). The other advantage is that it works well with version control tools such as git because the text diffs are human readable.
OpenSCAD is great if you're a programmer, terrible if you actually want to machine things with machineable features like minimum radii for example. Just gets very cumbersome very quick. FreeCAD is the closest to the industry standard.
I really like OpenSCAD. It works well for moderately complex functional models. But it does it limits when things get too complicated. It can really slow down, and some shapes are just not doable. I recently had to help my middle schooler find a 3D model program, and we realized that OpenSCAD would have been a very poor fit for the project. A lot of online resources recommend Fusion360, but licensing terms and reduced features in the hobbyist-friendly edition make this much less of a good option than what it might have been years ago. So, instead, we selected OnShape. Super impressed how easy it is to learn even for a middle schooler, how similar it is to how I am used to working with OpenSCAD, and how powerful it can be. Also, the fact that it runs completely in a browser and can be used from a basic Chromebook is huge. The only real downside is that you can't export "source code" in a portable format that you could manage in git or import into other CAD software. But then, that's a shortcoming that OnShape shares with most of the other available tools.
I'm a rather recent subscriber, and I've gone back thru your older videos out of pure interest, and I must say, even with my major ADHD, I can sit through even your longest videos without much urge to skip around, simply because your energy and enthusiasm is absolutely infectious. Not a single moment isn't electrically charged. Don't let the naysayers get you down, that's a them problem!
I’m always blown away at your clear and precise explanations. I find it important to hear your ‘lectures’ before you proceed with construction. Run with it! Love your content.
Really impressive how practical your choices have been. I like how you revisit problems and don't get overly fascinated by your initial attempts. Seems like a successful strategy with time enough to employ it.
For two years every Monday evening is my time to watch new episode you have posted. Thanks to you I have go back to electronics after 10 years and I feel joy when working with that. It is an honor to be patreon and I respect all other patreon legends being with you much longer than me!
Awesome project! (and awesome Patreon patrons, too). Last time I worked on a homebrew paper tape reader, the collimator blocks were just made out of suitably drilled copper-clad SRPB circuit board. The light source was a 12V motor car festoon lamp and the detectors were germanium diodes with the black paint stripped off. The tape drive was built using a long capstan and extra-wide pinch roller similar to a magnetic tape player. The whole was in a block 4" wide by 3" high and 2" deep. It was being used to archive a large library of vintage 7- and 8- level paper tapes.
Gee, that tan based photo-detector tube at 14:14 gave me a throwback. My grandfather is a ham radio operator, and is old enough to still design equipment around vacuum tubes. About 12 years ago, I was poking around his barn and came across one of those. I'd never seen anything quite like it, and I don’t remember what his response to me finding it was. He probably still has it knocking around somewhere.
Been working with computers for 35 years but never really understood how it all started with vacuum tubes or rather how it works in practice. This is starting to clear that up and might be the most interesting thing ever. Oh, and more dog!
19th August 2024. At this point I catch up with your videos, having started at the beginning a few months ago. Thanks, David, for a fascinating journey. Long may it continue. 🙂👍
A mask in front (and possibly behind also) the paper tape would increase the contrast seen by the photodiode. Illumination won't be crucial, so long as there is some illumination - even passive room light would probably do. You could make the masks by drilling some PCB copper clad board at the correct spacing. Alternatively, just drill a block of opaque plastic or aluminium and insert the photodiodes into it. That would guarantee that they would only "see" the area covered by the tape hole. Maybe mill a channel the width of the tape into the side opposite the photodiodes too.
As a viewer who's been watching since day one in this I truly appreciate your candor, and I'm giving you a thumbs up just because you were honest hey yo we're doing another review of the whole thing we're going to have no progress no build and no happy result at the end Skip this vid if you must thanks for viewing as for everybody else keep watching. I love it thank you for respecting your viewers.
@UsagiElectric I use Fusion 360 for my solid modelling. But then I was doing this sort of thing professionally for years before I retired. Fusion has nearly all of the features that were on Solidworks and Inventor when I used them, but for free with a hobbyist licence. But... It's a full on engineers program. A lot to learn, but you will have a lot of fun doing it.
re 3D modeling, as an old school (now retired) programmer, I like OpenSCAD, where you define your objects via a simple language where you can add and subtract shapes (a hole would be a subtracted cylinder). There's a decent OpenSCAD editor that gives you near realtime preview of your results.
Thank you to all the Patreon patrons for letting David just be his awesome self, letting him create and learn, and letting us be inspired and educated along the way.
This project really makes you appreciate what it took to design the old vacuum tube computers. These days people just take it for granted to just use transistors, logic chips, etc... I can't wait to see this work.
Given that I specialize in the modern computers made after 1980, this video helped me understand some fundamentals of your massive beast machines. Much appreciated!
There was a design i have seen that used glass germanium diodes in a reversed biased configuration on an old projector to read optical audio. Essentially when light hits the diode it conducts. Essentially an early photodiode.
Hi. Great project. An idea for you. Use the large tube to read a single hole then offset the boles for each bit so the vac tubes can be next to each other think helical scan VCR tape so each 8 bit sequence will be staggered along about 8 inches of paper tape allowing each tube to simultaneous read 8 bits.
I believe it was our Data General(or Honeywell or Lockheed) papertape readers that put a glass/plastic rod between the light source and head to convert the light into a thin row of light. A rubber drum was used to pull the tape through the reader. After heavy use, the drum would accumulate dust and begin to slip. A little alcohol would clean it up.
I have no idea about paper tape reader but this was the first idea that came up: Have 8(or 9) tiny metal plates flying at 0.5mm over the tape. On the opposite side same setup. Basically we build 8 parallel capacitors where the paper tape is the electrolyte. We can use the capacitors to build 8 parallel low pass filter. You will get a 1 whenever a hole travels between the metal plates.
Thanks for the video and update. I can now understand why it took you so long and it would be the same for me. 3 days to come up with the design, 1 to build the prototype breadboard and then 2 months playing with the paper tape and light watching the needle go back and forth 😂😂😂
I never seen those tiny photodiodes before, excellent, now you got A LOT of mechanic work to do, please film everything, watching the construction of a paper tape trader in 2024 , it's gonna be unique! Something i never thought i will see, CAN'T WAIT
I started following your channel after watching your tube designs a while ago. I love all the other content, still, I am so happy to see you came to the 1-bit computer 😊
I like the rambling though! It helps me understand what is actually going on once you start building stuff. Also, thank you Patreon supporters! I would contribute, but unfortunately I am not in a place where I can. But I would if I could, so thanks patrons for covering me.
Glad you spent a good amount of time petting the dog, stuff like this is important when you're talking about vacuum tubes and primitive computers. (Not sarcasm, dogs are awesome.)
I literally found your channel because I wanted to know if it would be possible to create an original NES that only used vacuum tubes, how big that would be, and how much energy it would take to run. 😂 that was YEARS ago! You rock. You projects rock. You don’t ramble and building or talking I appreciate all your content.
Oooh damn, that was fun. I cannot wait for the next part. Everytime I watch one of his videos, I cannot help but think a) the most humble mad genius ever and b) thank the universe that he devotes his talents to good rather than becoming a Bond villain or something. And I'm sure someone's already mentioned, tinkercad would be trivial to design a quick block to hold the diodes.
I love that you show the enthusiasm of just playing with things like we all secretly do.. "lookatit go! lookatit go!" 😁 Because we all do things like that.
Absolutely love this project thanks. As a 'single bit' of information for you - I too have built a optical tape reader using tiny surface mount photodiodes and the ubiquitous Arduino. I discovered that the holes in a prototyping perf board have the exact same spacing as the paper tape holes. I used the perf board as the mask... Now on to the really hard 'bit' - How to source the blank tape and how to apply the perforations; I haven't solved that yet...
Just a word of encouragement. Almost all of the early business calculators back in the mid to late 70's ( before even digital I.C. became common and certainly before the first microprocessors ) were bit serial machines. I used to work on them. Of course they utilized shift registers ( or some other exotic serial memories ) as storage media. Of course as you pointed out even this might be a herculean stretch for your single bit tube type computer. Just way too many gates and combinatorial logic, even if you used diode logic ahead of the tubes logic. I just want to say that I applaud your dogged pursuit of this project. You Go USAGI ! ! !
It would be a treat if you could mold and cast the Bendix photo diode mount and reuse it. It wouldn't get more period correct than that! Also, I'm here for the rambles, don't stop! In fact, if you could give a little extra ramble on how you are calculating those resistor values, that would be great! Thanks Patreons!
According to the schematic, your vacuum tube voltmeter also doubles as a radio ;-) Love to see the old test equipment. It gives a lot of flair to your videos!
Holy cow! This project came along nicely, just caught up watching the entire series. I am almost tempted to start something similar :), I really like to push tubes to do digital logic at 10+MHz, and I like the double triodes so maybe that's a path to explore. Regarding the paper tape, conceptually for me the magnetic tape solution is easier, because the tape player is already done. There is a bit of analog circuit needed to filter and amplify the tape signal, to filter the '0' bit frequency vs the '1' bit frequency, then an 8 bit shift register is needed. It definitely takes more tubes than for the tape reader through... All in all, awesome project! Well done!
20:11 David! Thanks!!! Now I can revisit my papertape to rs232 device project. Wanted to build that back at VCC/Dallas Makerspace but couldn't solve this issue. Hope to see you at VCF East to share notes and give you one of the first production runs.
That is a fascinating project! I have worked on some mechanical tape readers - they used tiny plastic five-point "star wheels" set in the ends of sprung arms, packed over the width of the tape, with contacts above them. When a hole passed under a wheel, a point dropped in, the arm moved down and allowed the contact to close. When there was no hole, the wheel skidded along the tape on two adjacent points, so the arm stayed lifted and the contact opened. These were on Devlieg CNC milling machines, which were operated by modified "Minic" minicomputers. Not a good system though; the star wheels wore out over time & they could not handle normal tape splices; any change in tape thickness stopped them from reading data. Optical is definitely the way to go. For a 3D modelling program, have a look at the free "Designspark Mechanical" program from RS components. That is excellent for doing mechanical part models. Just be sure to set the STL export quality to high; it defaults to low, if I remember correctly. Mylar tape would be far better than paper for long term use, as others have said. I have plenty of that and a couple of punches, if you need any tapes.
Regular diodes are light sensitive as well so you could construct an array of diodes next to each other with a reflector behind it and light barriers between to have a makeshift tape reader.
But that's why regular diodes are sold in opaque cases these days. You can use regular LED's as photodiodes, but their sensitivity is probably worse than a purpose built photodiode
Super cool. Fwiw, I worked with a Varian Data Systems mini-computer in 1976. It booted from punched tape. Low level diagnostics and maintenance programs were also on tape. The reader was in a Model 33 teletype system console. Iirc, Model 33s use mechanical probes as sensors. High-use tapes were all punched in mylar. These lasted for years of daily use.
All diodes and transistors have the photo effect, that's why they come in black plastic usually. Silicon's bandgap makes it most sensitive to infrared. That's why optoisolators always have infrared LEDs in them.
You need a roll of the tape we used in the 70s for control loops on our line printers - that was basically a thin paper tape laminated to sandwich a strip of mylar, so it could go round and round the mechanical taper reader in the back of an ICL1933 for weeks before it had to be re-punched. As an aside, the H1, H2, etc. in the FORTRAN write statements referred to the holes in the tape, with channel 1 being the top of form by convention and the number of sprocket holes to the next channel 1 hole determining the length of the form feed - for 11" paper there would be a single hole 1 on the loop but smaller forms like cheques would have several to bring the loop up to a manageable size. The rest of the holes could be used on pre-printed forms to allow fast tabbing vertically, though I cannot remember ever having a loop that used more than channel 1, 2 and 3 on our system. Ah, happy days.
I see your Conar vacuum tube multimeter and raise you my passive AVO meter which also has a humungous, mirror-backed dial. 😜 Appreciated the "cook's tour" of your vacuum tube computer's architecture and origins, as well as the summary of your investigations into designs for your own paper-tape reader and how each might or might not meet your design constraints. P.S.: Glad you didn't decide to use mercury delay lines as WW][ era DRAM, rofl!
I've always wanted to build a paper tape or card reader. I thought about using a bunch of photo resistors and a light connected to something like a Raspberry PI. I never did it, but seeing this video, I may try designing one now. At worst, I get out the breadboard and fry something. At best, I finally finish a project I've thought about. Either way, I will hopefully learn something. Thanks for the video!
Back in the early 70s when I was working with paper tape we used a mylar version of "paper" tape to make loops. It stands up to repeated use much better than the standard oiled paper. The mylar's only fault is that not all punches will cleanly punch the tape. I know some ASR33s (which I think you have) will work depending on how dull the punches are. I did a quick Internet search any couldn't find a source, but you have much better contacts than I do.
Coincidentally, I just stumbled across UK national treasure Tim Hunkin's old television series on TH-cam, search for "The Secret Life of the Word Processor - Remastered" and marvel at the mighty Pegasus computer with its paper tape reader.
My pappy paid for my college back in the '70s operating a Gerber cutting table for GM. He'd lay up a couple dozen layers of fabric on the table, check everything, then load a punched tape into the machine that sent the knife along a 2 1/2 dimensional path. (x, y, and angle). He would occasionally bring these tapes home. They weren't the paper tapes I used in college, but they were a robust metalized plastic. Y'all want to go touring with those metalized plastic punched tapes instead of paper.
Thanks to myself and other 4-bit patrons for this interesting deep dive. I think Autodesk Fusion is the best for what you're trying to accomplish, but I'm sure you could also get OnShape sponsorship for your channel to get a license (both have free non-professional licenses). Fusion has a built in CAM feature which would help with CNC machining, it could output GRBL if that's what you use already to mill your PCB's. I also wonder why you didn't consider air and reed switch sensing for paper tape like the player pianos, that way nothing is touching the tape besides air flow, I am guessing it would be too slow.
This is a project I'm very keen on as I'd love to have my own paper tape reader. I was thinking optical as well (since I don't have such period-correct requirements), and while photodiodes are hard to get at 2.5mm, LED's are not and they can function in the same capacity, even if they have a diminished output. So that leaves how to move the paper and given that I am not mechanically minded, that is a pretty big stopping point for me. It'll be cool to see how you tackle it.
17:30 if you omit the shift part, but just use a register and build the Nibkow disc out of a PCB with rotating contacts where you use PCB traces to connect the photodetectors output to the register input exactly when the hole is in front of the lamp for that specific bit? Mechanical demuxing!
Just a thought about the tube based optical tape reader using a focusing lens-I think it might be simpler than it appears! If you imagine the system as a camera focused on the paper tape, its image should be wide enough to allow each tape dot to focus on one photodetector in the camera image plane. It is basically doing the same thing as a microscope camera, taking an image that's larger than the thing photographed. There may already be micro lenses for something like a medium format camera (using 6cm film) that would do this job.
Freecad/Ondsel works. There's a bit of a learning curve, but it's definitely a good tool for the job. If you're already familiar with Blender, cadsketcher for Blender is another possibility. If you're more comfortable with code than with 3d modeling, OpenSCAD is also a possibility. My personal feeling is that OpenSCAD would be the hard way to do it. In any case, I recommend 3d printing your model before sending it off to be machined. For that matter, you could probably build the final read head that way. Plastic was a thing in the 1950s.
Perhaps create a little board with LEDs that can 'fake' a tape when against the sensor - just for fun :D A webcam watching the tape go past and some software to process the image would also be fun! a very interesting programming problem!
Ignore the people that post negative comments about you not actually working on something. You rambling on is still pretty informative a good portion of the time.
I bet these are the same people that bicker around when some anime has a bunch of "story" episodes.
Oh, the bickering in the conference episodes or the current festival arc of Tensura. 😆
? The rambling is the working. Its development after all.
"Why can't you/we just *do something*" is always the cry of a fool, no matter the context. All real engineering and science is primarily done in the head. The tools are there to test and verify to refine those mental models, so that you can communicate them in a reliable way in the future, even to yourself - i.e., in the context of fixing something. A real master mechanic usually minimizes his wrenching to what is necessary to get the job done, after spending much mental time and energy optimizing and refining the tasks, often inventing his own theories and practices along the way.
Negative comments only for politicals and social issues not 4 technology
@evaDrepuS who is posting negative comments
“That is going to work absolutely flawlessly… I think.” Such confidence.
I've gone through enough design revisions that I'm not sure of anything at this point, haha.
Me before most pushes to prod: “this is guaranteed not to tank prod, mostly, I hope”.
@@UsagiElectric that's my default mindset. Expect the worst, cherish if it works. :D
The paperclip tape reader is almost exactly what I have built some 40 years ago for my TI-57. I was not reading bytes, I was establishing a contact between the calculator's keyboard rows and columns wires to simulate key presses. Long story short, it worked pretty well. I could punch programs on bristol cards (much more resistant than paper) and have them loaded in the calculator's memory in 10 seconds, almost every time. I had performed a sort of trepanning on my calculator, which had several DIN plugs that I could plug into my card reader. The reader was made out of wood, and the contacts were copper wires that had been curled (not paperclips) so they touched a single sheet of aluminium foil glued on the bottom plate. It worked because the calculator's chip was doing proper debounce filtering on the keyboard input, so no debounce logic was needed on the reader itself. I'm pretty sure I still have this thing somewhere in my attic.
If you can find it you should totally try to upload a video if it running (if it still works)
I used to volunteer at TNMoC at Bletchley Park and I was always amazed by the tube-based design of the PTR on the Colossus rebuild. it's incredible that it actually works and more so manages it at the speed the tape flies past the reader.
I've seen it once in real life when I went to VCF GB in 2010. I seem to recall that someone told me the computer itself is capable of running much faster, but it's limited by the physical strength of the paper which would just tear if it were run too fast. A PET film like Mylar might help that, but that hadn't been invented in the 1940s so it wouldn't be period-correct.
@@Zeem4 I wonder if they could use film stock? (Though being nitrocellulose film stock at the time... that could be exciting at high speeds. Exothermically exciting.)
@@timwilliscroft9615: Read-once memory.
This "machine" was the reason why I started watching your channel 🙂After realizing what you else are doing, I watched all videos one after the other and became a Patreon 🙂 Keep going!
Thank you so much!
Now that we have a design I actually really like on the tube computer, I'm finally fired up on it again. I'll get to designing a new block to machine out and in the next episode in the series we'll get the mill fired up again!
Same here. Came for the unique tube computer, stayed for the "big old iron"
I first saw the Ti99 4 videos... didn't realize that's almost his newest machine
Photo detector is sentive to infrared light so will be overloaded when the tape picked up heat from the lamps I found this when repairing 16 mm sound projectors you should include a fan to keep it cool in use for a long time
Didn't understand 1 bit of this video but i liked every bit of it.
My life playing Caves of Qud
I'm currently planning a vintage computer to monitor my property systems. Generator output, run time, bio gas production, solar output, well and cistern data, along with all the error messages I'd want printing out on an old dot matrix printer with the endless paper. I just wanna say how thankful I am to this channel and the patrons who support it for providing such detailed explanation and theories about these vintage machines right down to the basics.
Man, that end result is really promising. Looking forward to the next phase of UE-1
Thank you so much!
Time to start designing up a block and get the old mill fired up again!
I appreciate meeting some of my fellow High Scorer and 4-bit patrons, and the many emotional-support fans (just as important as us money-bags lol), at some of the events where we appreciate the hobby. And people like Bob at System Source who provide the floor space and electricity to let us get together and enjoy each other's company.
It's always a pleasure to meet y'all in person!
The USAF was using a optical tape reader ,as late as 1979, to input data into a TEST SET. This test set checked the status of a W-62 Nuclear Weapon that sat atop the Minuteman III missile. The optical reader alignment was very problematic and was constantly causing us problems. Very interesting video!
Worked in a manufacturing facility (of a LARGE multi-national manufacturer) where we were loading programs into 16-bit minicomputers on the factory floor with paper tape well into the late 1980's. Our readers were flat-bed units (I *think* made by a company called "Trend") which had no spools, so we'd place a roll of tape onto a screwdriver shank and hold it by hand to let it feed into the reader and then onto the floor. The tape had to be re-rolled by hand. Dirt and static usually combined to clog the holes somewhere resulting in the need for several attempts before getting a good read all the way through. We ultimately replaced the tapes and readers with IBM PCs to emulate the readers RS-232 output for quite some time before the minicomputers were finally replaced with modern computers.
Good times.
@@yohannwilkerson6058 beats me why more companies didn't use the fanfold paper tape favoured by DEC, since that neatly stacked itself downstream of the reader, ready to be read again, without any re-rolling needed.
"Take a book out of everybody else's page"? LOL, that's a good one :-) Great to see you back on the UE14500!
I didn't even notice that one until after the final render, lol.
@@UsagiElectric Love the fact that you've included it as a section title too 🙂
To the channel Members: We hope you feel appreciated, because you certainly are! This is a great channel. Thank you VERY much for keeping it rolling! All good wishes.
Use mylar paper tape - I used that in the 1960's on a PDP-8 - lasted 6 months used multiple times a day. The only down side is the mylar will shorten the life of the paper tape punch device - debug with regular paper tape and switch to mylar after all is good.
This. I used a french 'Logabax' computer with a small tape loop to set the vertical tabbing on the printer (it ran with payslips and other oddly-spaced forms). That was mylar tape as it ran continuously when printing. We also had a manual punch that was a drilled block of perspex and a punch needle. The index holes might have been prepunched. A loop of tape reinforced with a kapton layer might also work if you don't need the tape itself to be period-correct.
I used mylar punch tape back in the 90's. We were trying to support some old equipment. The problem was that we couldn't find new stock.
I remember that the tapes you used over and over, e.g. bootloader, used to sometimes be mylar so they would last longer.
Is it period-correct for the late 1950s?
There was an 1959 NASA computer that was in the "dungeon junk room" at NCSU in 1970 that had a mylar paper tape loop attached to a Friden Flexowriter paper tape reader - but it could have been from the early 1960's. NCSU had a few old NASA computers - I did learn to program the analog TR-48 that they provided to NCSU. I did not start touching computers until 1963 and they were punched cards and not paper tape.
I appreciate the backstory and ramblings as this gives me an idea of what you're thinking of doing.
This sort of context also helps why you decide to choose $procedure and decided to stick with it, as it is a logical decision-making process which do make a lot of sense, and this also helps with cutting out wasting time with experimenting with stuff, only to find it doesn't work, or it takes up too much room etc etc
Keep it up! 👍
Looots of germanium semiconductors were photosensitive anyway - that's why their glass housings were painted black :-)
But that specific photo-diode you found most probably beats them all
Hooray a new Tube computer episode. 🤠🎉
It's back baby!
@@UsagiElectric 🎉🎆🎇🤠
Aw, man, this takes me back.
It was content like this that made me subscribe to this channel. I really loved the days when you were more focused on pre-vintage computers, and not vintage computers.
Not to say I don't like the vintage computers, but there's many vintage computing channels out there. There was only one pre-vintage computers channel, and it was this one. It made you feel unique. The exposition made you feel unique. The enthusiasm made you feel unique.
I'm always shocked when people don't like choices you make, because they make this channel unique. I don't want you to be like every other vintage computer channel, I want you to be unique. I like this unique channel.
It's great to see the low-voltage pre-vintage computer stuff again. I look forward to more!
I am here for the hand waving just as much as the hand-building! It's neat to see your tube computer getting some love again.
I thoroughly enjoyed all of the "hand waving"! Looking forward to the next episode. 🙂
Being from the TRS-80, Vic20 era it’s amazing to realise how historical these tube computers are. Something like 20yrs from the first computers to when I was born.
Makes me feel really old... I started with computers back in the paper tape age, 1967 to be exact. I still have a few of my programs from back then on tape, but, alas, no reader. Back then, we had optical readers that could reach speeds of 2000 bytes a second, 5 meters of tape a second. It was hooked up to a transistor based computer, germanium transistors like the OC141and SFT228. Thanks for your efforts to bring all this to life !
That brings me back also. I still remember building a simple radio with germanium transistors like the OC13 and later the OC72. I still can remember scraping off the black paint from a OC13, because it made the transistor sensitive for light. I wanted to build a light detector that way, but went for LDR's later on (although they had a terrible response time).
I still have a few of them laying around dating from all those years back.
Lovely walk through of engineering constraints and trade offs. 🎉
I know it's very weird, but I use OpenSCAD for 3d modelling because I prefer its descriptive approach to the regular drag-and-drop sort of paradigm. Also worth mentioning that it is integrated into FreeCAD, although I just use it standalone.
I second this. OpenSCAD is a great 3d modeler for software developers because of it's C-like syntax. When you need the parts to be in exactly the right position, it's the perfect tool for the job.
OpenSCAD is great. It matches the way I think when designing objects (previously I used to dabble in PovRay).
The other advantage is that it works well with version control tools such as git because the text diffs are human readable.
OpenSCAD is great if you're a programmer, terrible if you actually want to machine things with machineable features like minimum radii for example. Just gets very cumbersome very quick. FreeCAD is the closest to the industry standard.
I really like OpenSCAD. It works well for moderately complex functional models. But it does it limits when things get too complicated. It can really slow down, and some shapes are just not doable.
I recently had to help my middle schooler find a 3D model program, and we realized that OpenSCAD would have been a very poor fit for the project. A lot of online resources recommend Fusion360, but licensing terms and reduced features in the hobbyist-friendly edition make this much less of a good option than what it might have been years ago. So, instead, we selected OnShape. Super impressed how easy it is to learn even for a middle schooler, how similar it is to how I am used to working with OpenSCAD, and how powerful it can be.
Also, the fact that it runs completely in a browser and can be used from a basic Chromebook is huge. The only real downside is that you can't export "source code" in a portable format that you could manage in git or import into other CAD software. But then, that's a shortcoming that OnShape shares with most of the other available tools.
I'm a rather recent subscriber, and I've gone back thru your older videos out of pure interest, and I must say, even with my major ADHD, I can sit through even your longest videos without much urge to skip around, simply because your energy and enthusiasm is absolutely infectious. Not a single moment isn't electrically charged. Don't let the naysayers get you down, that's a them problem!
I’m always blown away at your clear and precise explanations. I find it important to hear your ‘lectures’ before you proceed with construction. Run with it! Love your content.
Really impressive how practical your choices have been. I like how you revisit problems and don't get overly fascinated by your initial attempts. Seems like a successful strategy with time enough to employ it.
For two years every Monday evening is my time to watch new episode you have posted. Thanks to you I have go back to electronics after 10 years and I feel joy when working with that. It is an honor to be patreon and I respect all other patreon legends being with you much longer than me!
Awesome project! (and awesome Patreon patrons, too).
Last time I worked on a homebrew paper tape reader, the collimator blocks were just made out of suitably drilled copper-clad SRPB circuit board. The light source was a 12V motor car festoon lamp and the detectors were germanium diodes with the black paint stripped off.
The tape drive was built using a long capstan and extra-wide pinch roller similar to a magnetic tape player. The whole was in a block 4" wide by 3" high and 2" deep.
It was being used to archive a large library of vintage 7- and 8- level paper tapes.
Yeah, I wonder, isn't the photo-diodes he got simply uncoated diodes? I think so!
Gee, that tan based photo-detector tube at 14:14 gave me a throwback. My grandfather is a ham radio operator, and is old enough to still design equipment around vacuum tubes. About 12 years ago, I was poking around his barn and came across one of those. I'd never seen anything quite like it, and I don’t remember what his response to me finding it was. He probably still has it knocking around somewhere.
Thanks to all the patreons! I always find old tech fascinating.
Been working with computers for 35 years but never really understood how it all started with vacuum tubes or rather how it works in practice. This is starting to clear that up and might be the most interesting thing ever.
Oh, and more dog!
19th August 2024. At this point I catch up with your videos, having started at the beginning a few months ago. Thanks, David, for a fascinating journey. Long may it continue. 🙂👍
A mask in front (and possibly behind also) the paper tape would increase the contrast seen by the photodiode. Illumination won't be crucial, so long as there is some illumination - even passive room light would probably do. You could make the masks by drilling some PCB copper clad board at the correct spacing. Alternatively, just drill a block of opaque plastic or aluminium and insert the photodiodes into it. That would guarantee that they would only "see" the area covered by the tape hole. Maybe mill a channel the width of the tape into the side opposite the photodiodes too.
As a viewer who's been watching since day one in this I truly appreciate your candor, and I'm giving you a thumbs up just because you were honest hey yo we're doing another review of the whole thing we're going to have no progress no build and no happy result at the end Skip this vid if you must thanks for viewing as for everybody else keep watching. I love it thank you for respecting your viewers.
@UsagiElectric I use Fusion 360 for my solid modelling. But then I was doing this sort of thing professionally for years before I retired. Fusion has nearly all of the features that were on Solidworks and Inventor when I used them, but for free with a hobbyist licence.
But... It's a full on engineers program. A lot to learn, but you will have a lot of fun doing it.
re 3D modeling, as an old school (now retired) programmer, I like OpenSCAD, where you define your objects via a simple language where you can add and subtract shapes (a hole would be a subtracted cylinder). There's a decent OpenSCAD editor that gives you near realtime preview of your results.
Watched the whole video due to the chuntering on. This project is a fascinating piece of work!
Thank you to all the Patreon patrons for letting David just be his awesome self, letting him create and learn, and letting us be inspired and educated along the way.
thanks to all those patreons!
i can't affort to donate, so i'm sure glad you can do we can keep injecting this great info!
thanks!
This project really makes you appreciate what it took to design the old vacuum tube computers. These days people just take it for granted to just use transistors, logic chips, etc...
I can't wait to see this work.
Given that I specialize in the modern computers made after 1980, this video helped me understand some fundamentals of your massive beast machines. Much appreciated!
There was a design i have seen that used glass germanium diodes in a reversed biased configuration on an old projector to read optical audio. Essentially when light hits the diode it conducts. Essentially an early photodiode.
Hi. Great project. An idea for you. Use the large tube to read a single hole then offset the boles for each bit so the vac tubes can be next to each other think helical scan VCR tape so each 8 bit sequence will be staggered along about 8 inches of paper tape allowing each tube to simultaneous read 8 bits.
I believe it was our Data General(or Honeywell or Lockheed) papertape readers that put a glass/plastic rod between the light source and head to convert the light into a thin row of light. A rubber drum was used to pull the tape through the reader. After heavy use, the drum would accumulate dust and begin to slip. A little alcohol would clean it up.
I have no idea about paper tape reader but this was the first idea that came up:
Have 8(or 9) tiny metal plates flying at 0.5mm over the tape. On the opposite side same setup. Basically we build 8 parallel capacitors where the paper tape is the electrolyte. We can use the capacitors to build 8 parallel low pass filter. You will get a 1 whenever a hole travels between the metal plates.
Thanks for the video and update. I can now understand why it took you so long and it would be the same for me. 3 days to come up with the design, 1 to build the prototype breadboard and then 2 months playing with the paper tape and light watching the needle go back and forth 😂😂😂
I never thought about using vacuum tubes with low voltage until watching your videos. Thank for sharing!!!…😄
I love using OpenSCAD for 3D modeling. You define everything like to would a regular program: define variables, use functions, etc.
I never seen those tiny photodiodes before, excellent, now you got A LOT of mechanic work to do, please film everything, watching the construction of a paper tape trader in 2024 , it's gonna be unique! Something i never thought i will see, CAN'T WAIT
One of the most educating AND entertaining theory-based videos on TH-cam. Kudos and thank you!
I started following your channel after watching your tube designs a while ago. I love all the other content, still, I am so happy to see you came to the 1-bit computer 😊
I like the rambling though! It helps me understand what is actually going on once you start building stuff.
Also, thank you Patreon supporters! I would contribute, but unfortunately I am not in a place where I can. But I would if I could, so thanks patrons for covering me.
I really appreciate your attention and commitment to period accuracy. I’m looking forward to seeing it run some programs in the near future!
Glad you spent a good amount of time petting the dog, stuff like this is important when you're talking about vacuum tubes and primitive computers. (Not sarcasm, dogs are awesome.)
Those photodiodes are adorable! They remind me of the “grain of wheat” light bulbs. 😃
I literally found your channel because I wanted to know if it would be possible to create an original NES that only used vacuum tubes, how big that would be, and how much energy it would take to run. 😂 that was YEARS ago! You rock. You projects rock. You don’t ramble and building or talking I appreciate all your content.
Oooh damn, that was fun. I cannot wait for the next part.
Everytime I watch one of his videos, I cannot help but think
a) the most humble mad genius ever and
b) thank the universe that he devotes his talents to good rather than becoming a Bond villain or something.
And I'm sure someone's already mentioned, tinkercad would be trivial to design a quick block to hold the diodes.
I love that you show the enthusiasm of just playing with things like we all secretly do.. "lookatit go! lookatit go!" 😁 Because we all do things like that.
Maybe these tiny photodiodes will come in handy in fixing the bendix, if turns out to have a bad diode in its paper tape reader.
Absolutely love this project thanks.
As a 'single bit' of information for you - I too have built a optical tape reader using tiny surface mount photodiodes and the ubiquitous Arduino. I discovered that the holes in a prototyping perf board have the exact same spacing as the paper tape holes. I used the perf board as the mask...
Now on to the really hard 'bit' - How to source the blank tape and how to apply the perforations; I haven't solved that yet...
Just a word of encouragement. Almost all of the early business calculators back in the mid to late 70's ( before even digital I.C. became common and certainly before the first microprocessors ) were bit serial machines. I used to work on them. Of course they utilized shift registers ( or some other exotic serial memories ) as storage media. Of course as you pointed out even this might be a herculean stretch for your single bit tube type computer. Just way too many gates and combinatorial logic, even if you used diode logic ahead of the tubes logic. I just want to say that I applaud your dogged pursuit of this project. You Go USAGI ! ! !
It would be a treat if you could mold and cast the Bendix photo diode mount and reuse it. It wouldn't get more period correct than that! Also, I'm here for the rambles, don't stop! In fact, if you could give a little extra ramble on how you are calculating those resistor values, that would be great! Thanks Patreons!
The hard part is getting a reliable motor drive system for the tape. Can't wait to see your solution.
1 bit ought to be enough for anybody.
Haha! Nice
I was wondering why so many of your old tube videos were getting recommended and now it all makes sense.
According to the schematic, your vacuum tube voltmeter also doubles as a radio ;-) Love to see the old test equipment. It gives a lot of flair to your videos!
Yes, an All American 5 vacuum tube volt meter 🙂
I may have potentially been a little sleepy and copy/pasta'd the wrong schematic!
I'm one who loves this stuff. I might only understand the tip of the iceberg but it still fascinates me.
Holy cow! This project came along nicely, just caught up watching the entire series. I am almost tempted to start something similar :), I really like to push tubes to do digital logic at 10+MHz, and I like the double triodes so maybe that's a path to explore.
Regarding the paper tape, conceptually for me the magnetic tape solution is easier, because the tape player is already done. There is a bit of analog circuit needed to filter and amplify the tape signal, to filter the '0' bit frequency vs the '1' bit frequency, then an 8 bit shift register is needed. It definitely takes more tubes than for the tape reader through...
All in all, awesome project! Well done!
20:11 David! Thanks!!! Now I can revisit my papertape to rs232 device project. Wanted to build that back at VCC/Dallas Makerspace but couldn't solve this issue. Hope to see you at VCF East to share notes and give you one of the first production runs.
Whatever 3D modeling software you choose, parameterize your designs from the start! It makes later changes to your design so much easier.
That is a fascinating project!
I have worked on some mechanical tape readers - they used tiny plastic five-point "star wheels" set in the ends of sprung arms, packed over the width of the tape, with contacts above them.
When a hole passed under a wheel, a point dropped in, the arm moved down and allowed the contact to close.
When there was no hole, the wheel skidded along the tape on two adjacent points, so the arm stayed lifted and the contact opened.
These were on Devlieg CNC milling machines, which were operated by modified "Minic" minicomputers.
Not a good system though; the star wheels wore out over time & they could not handle normal tape splices; any change in tape thickness stopped them from reading data.
Optical is definitely the way to go.
For a 3D modelling program, have a look at the free "Designspark Mechanical" program from RS components. That is excellent for doing mechanical part models. Just be sure to set the STL export quality to high; it defaults to low, if I remember correctly.
Mylar tape would be far better than paper for long term use, as others have said. I have plenty of that and a couple of punches, if you need any tapes.
Regular diodes are light sensitive as well so you could construct an array of diodes next to each other with a reflector behind it and light barriers between to have a makeshift tape reader.
But that's why regular diodes are sold in opaque cases these days. You can use regular LED's as photodiodes, but their sensitivity is probably worse than a purpose built photodiode
Super cool. Fwiw, I worked with a Varian Data Systems mini-computer in 1976. It booted from punched tape. Low level diagnostics and maintenance programs were also on tape. The reader was in a Model 33 teletype system console. Iirc, Model 33s use mechanical probes as sensors. High-use tapes were all punched in mylar. These lasted for years of daily use.
Somebody give this man an honorary PhD in Computer Science!!! 👏🏻👏🏻👏🏻
its very interesting to see your process. Hope we get to see all the machining etc
All diodes and transistors have the photo effect, that's why they come in black plastic usually. Silicon's bandgap makes it most sensitive to infrared. That's why optoisolators always have infrared LEDs in them.
You need a roll of the tape we used in the 70s for control loops on our line printers - that was basically a thin paper tape laminated to sandwich a strip of mylar, so it could go round and round the mechanical taper reader in the back of an ICL1933 for weeks before it had to be re-punched. As an aside, the H1, H2, etc. in the FORTRAN write statements referred to the holes in the tape, with channel 1 being the top of form by convention and the number of sprocket holes to the next channel 1 hole determining the length of the form feed - for 11" paper there would be a single hole 1 on the loop but smaller forms like cheques would have several to bring the loop up to a manageable size. The rest of the holes could be used on pre-printed forms to allow fast tabbing vertically, though I cannot remember ever having a loop that used more than channel 1, 2 and 3 on our system. Ah, happy days.
I see your Conar vacuum tube multimeter and raise you my passive AVO meter which also has a humungous, mirror-backed dial. 😜
Appreciated the "cook's tour" of your vacuum tube computer's architecture and origins, as well as the summary of your investigations into designs for your own paper-tape reader and how each might or might not meet your design constraints.
P.S.: Glad you didn't decide to use mercury delay lines as WW][ era DRAM, rofl!
Thanks for sharing more of your VTC project!
I've always wanted to build a paper tape or card reader. I thought about using a bunch of photo resistors and a light connected to something like a Raspberry PI. I never did it, but seeing this video, I may try designing one now. At worst, I get out the breadboard and fry something. At best, I finally finish a project I've thought about. Either way, I will hopefully learn something. Thanks for the video!
Back in the early 70s when I was working with paper tape we used a mylar version of "paper" tape to make loops. It stands up to repeated use much better than the standard oiled paper. The mylar's only fault is that not all punches will cleanly punch the tape. I know some ASR33s (which I think you have) will work depending on how dull the punches are. I did a quick Internet search any couldn't find a source, but you have much better contacts than I do.
I have no problem with the patron scrolling up the screen. I wish to thank all of them.
If I could afford it, I would be one of them.
Coincidentally, I just stumbled across UK national treasure Tim Hunkin's old television series on TH-cam, search for "The Secret Life of the Word Processor - Remastered" and marvel at the mighty Pegasus computer with its paper tape reader.
My pappy paid for my college back in the '70s operating a Gerber cutting table for GM. He'd lay up a couple dozen layers of fabric on the table, check everything, then load a punched tape into the machine that sent the knife along a 2 1/2 dimensional path. (x, y, and angle). He would occasionally bring these tapes home. They weren't the paper tapes I used in college, but they were a robust metalized plastic.
Y'all want to go touring with those metalized plastic punched tapes instead of paper.
So cool! Excited to see this complete, as I've been here since the beginning!
I was hoping for a new vacuum tube computer video, and finally it comes, and that was very interesting, loved It 👍
I knew of your 1 bit vacuum tube PC, but it's nice to see an update!
That's one cooooooooool journey! Thanks for sharing your progress.
Thanks to myself and other 4-bit patrons for this interesting deep dive. I think Autodesk Fusion is the best for what you're trying to accomplish, but I'm sure you could also get OnShape sponsorship for your channel to get a license (both have free non-professional licenses). Fusion has a built in CAM feature which would help with CNC machining, it could output GRBL if that's what you use already to mill your PCB's. I also wonder why you didn't consider air and reed switch sensing for paper tape like the player pianos, that way nothing is touching the tape besides air flow, I am guessing it would be too slow.
Fantastic! Thank you for all these videos. I love all the projects but the 1bit valve computer has my most interest.
PATREONS: MASSIVE THANKYOU!!!
Seriously y'all, thank you so much!
This is a project I'm very keen on as I'd love to have my own paper tape reader.
I was thinking optical as well (since I don't have such period-correct requirements), and while photodiodes are hard to get at 2.5mm, LED's are not and they can function in the same capacity, even if they have a diminished output.
So that leaves how to move the paper and given that I am not mechanically minded, that is a pretty big stopping point for me. It'll be cool to see how you tackle it.
17:30 if you omit the shift part, but just use a register and build the Nibkow disc out of a PCB with rotating contacts where you use PCB traces to connect the photodetectors output to the register input exactly when the hole is in front of the lamp for that specific bit? Mechanical demuxing!
Just a thought about the tube based optical tape reader using a focusing lens-I think it might be simpler than it appears! If you imagine the system as a camera focused on the paper tape, its image should be wide enough to allow each tape dot to focus on one photodetector in the camera image plane. It is basically doing the same thing as a microscope camera, taking an image that's larger than the thing photographed. There may already be micro lenses for something like a medium format camera (using 6cm film) that would do this job.
When I was a child, I've played with these tiny photo diodes. No others were available in former Czechoslovakia back in the day 🙂
Another step along the way. It take a lot of planning to make project happen. Thanks for the video
Freecad/Ondsel works. There's a bit of a learning curve, but it's definitely a good tool for the job. If you're already familiar with Blender, cadsketcher for Blender is another possibility. If you're more comfortable with code than with 3d modeling, OpenSCAD is also a possibility. My personal feeling is that OpenSCAD would be the hard way to do it. In any case, I recommend 3d printing your model before sending it off to be machined. For that matter, you could probably build the final read head that way. Plastic was a thing in the 1950s.
Thank you to the patrons who make these videos possible! I appreciate and enjoy them….😄
Perhaps create a little board with LEDs that can 'fake' a tape when against the sensor - just for fun :D
A webcam watching the tape go past and some software to process the image would also be fun! a very interesting programming problem!
Normally I skip the vacuum tube videos, but this one illuminated my curiosity.