I think he ran that joke in an earlier video, actually. I want to say it was one about the PDP-11. As for the "catch fire" board, it looks like the bit around 11:20 came awfully close . . . 😅
People keep giving me advice on how to slow the tape drive down, and I'm over here trying to figure out how to make the rest of the machine keep up with the tape drive!
@@UsagiElectric Well, maybe some way could be devised to drive the machine with conventional TTL signals through a voltage level translator. Perhaps a stream of data from a serial port would work. This wouldn't be for ultimate use but it would serve for debugging. Maybe the machine really can keep up with the drive but you can't know without a speed test.
@@UsagiElectric Gotta reduce the values of your decoupling capacitors. With only 24 volts the time constant of the RC circuits will be stretching the pulses! Remember Alan Turing got his Pilot Ace to run at around 1MHz!
I think that would require another adjustment of the photo diode bias, as when the paper tape reader was moved from the bench to the machine. Blocking ambient light out of the reader might help improve stability.
Once it's executing the right code, we'll definitely do that! Get it nice and dark in the room, get the camera setup right, and let it blink, blink, blink away!
Really a lot more went right than didn't, Congrats... you are close! A shorted power bus, logic levels wrong, a Halt circuit with a flip flop that flopped... LOL, it's all par for the course. I've built a 4-bit TTL computer home-brew on a ton of breadboards, I had wiring problems like the solid core wires would break after flexing but it would look like they were still connected... so it's normal in the bring-up to have problems.
Yup! More went right than I was actually expecting, and I genuinely think that in the next episode, we can actually get it churning along, executing the right instructions! Building homebrews are tough, and it's always the things least expected that get ya.
@@UsagiElectric Well, you're doing it the right way by bringing up each section and testing it. What I found out, besides the silly wiring errors, was timing problems. For my control logic I programmed ROMS to output the clocks and enables to registers and things on the bus. The inputs to the ROMS were the Instruction Register, T-State Generator, and some outputs from the ALU to work the branching instructions. The outputs would glitch on the T-State transitions and it acted like I had a squirrel brain running my processor. LOL I had to add a latch to buffer the control signals on the falling edge and I needed to rework the microcode a bit... I got it to work, sort of... I had some instructions execute but not all of them. I wish you the best of luck and success on your next attempt.
This stuff is important. Technological dead ends such as The Baine Clock or Doble steam car featured inventive imagination lost to history, or at least in the case of Baine. The method he used to create the three pole magnet in one metal bar seemed beyond the technology of the 1850s...and the way he did it is lost to history. Your determination is remarkable and not as useless as some may think. All the time, we need invention to keep society going. Thank you.
Dude, huge congrats! It's such an awesome thing having people like you in the world who put so much heart and effort into a hobby like this. That feeling you get when the project you've been working on for so long finally shows signs of life is one of the best feelings in the world and it never gets old - just a couple steps away now and we all know it will work beautifully once those last couple bugs are fixed!
What a beautiful machine! Built to perfection! Cool episode too! You managed to over-clock your tube computer and started solving the halting problem in the same episode! Watching your channel is my little Monday morning ritual and I love it!
The UE1 close to working is so cool indeed 🙂 I like how you let us see the first try of tests so we can share in the suspense and excitement of successes and failures. I liked the bunny break. I totally hear ya on the overall current requirement bringing it up and how nail biting that was. We are kindred spirits 😎⚡️ Reading the other comments, I love the "overclock" ones!
Not only that, I'm fairly certain I've got the nibbles backwards too! I'm gonna have to sit down and have a proper think about how to wire this thing up correctly.
Why don't you put a DC dimmer on the motor? By controlling the voltage, you control the speed... To be more 'old-school', you can use a wirewound potentiometer...
The old-school method may also involve a segmented contact connected to the shaft that would create a kind of PWM, with a capacitor in parallel. That would also help to fight local speed alternations to some extent because when speed goes down, the time the contact stays for the next phase gets longer, and that naturally increases the speed. But one should, of course, expect sparks and wear, so that's not an ideal solution.
6:18 hmm I don't know about ‘it doesn’t work so I’m not going to explain it’. You have clearly managed to build an audience of people who enjoy watching you work on tech mysteries of the ‘why doesn’t this work anymore’ variety, I could see ourselves also enjoying mysyeries of the ‘well here’s how it’s supposed to work but so far it doesn’t’ kind. Plus, you’d get the benefit of rubberducking your ideas? Sometimes, just explaining the thing makes something stand out. (Just a random comment from a random person on the internet - I don’t my take my comment too seriously as I don’t produce videos for a living or at all really, if this made for dreadful watching i wouldn’t know it)
I was pretty sure I was on the right track to figuring out the problem, I just ran out of time to troubleshoot it in this episode. I wanted to wait until I knew what was wrong, so in the next episode I could explain what the design was, what went wrong, and how I fixed it. And today, I did actually get it squared away, and the failure was a routing error and a resistor value error, but the why behind the resistor values being wrong is actually really interesting, and I'm glad I'll get to dig in deep on that particular problem in the next episode.
"It just executed its first instructions, though they were absolutely the wrong instructions" --Usagi, this very video "Bloody instructions, which, being taught, return/To plague the inventor . . ." --Shakespeare, _The Tragedy of Macbeth_ I.vii.9-10
I'm intrigued. Never intended to watch about the UE1, but now I want to see it completed. I came for the Centurion, but am stuck on UE1. BTW I acknowledge you.
I absolutely love your energy and joy even when something goes wrong. One of the best channels on UA-cam right here. I hope you get everything fixed soon.. much love from Germany! :)
I absolutely love ist too. This is so satisfying. I run a small museum in the corridors of my office. People are always amazed at the resources that were used back then. A tube computer like this would be a gem in my collection.
Congrats Dave! I call this a win dispite all the pain you've suffered along the way and especially here at the end. Do NOT let this mess up your Christmas holiday season, sit down with the parents, the critters and a whole lot of rum-loaded eggnog. You done good work.
When he said swedish fish I thought; What is he talking about, until I saw the candy. I have never seen that candy, and I am from sweden!! Great content as always btw!
Hey David! I work in cellular communications and I have access to some really high output rectifiers, I have -48VDC @ 50 or 75 amps and +24V @ 100 amps. They’re 1/4 width, 1U rack mount size. I also have a lot of random cut lengths of fabric sleeved high temp cable (it looks straight out of the 1940s) and multi conductor cable, up to 24x 18GA control cable. Let me know if you’d be interested in any of it, I use the -48V rectifiers to build a massive 96V (-48 to +48), 1000 watt, 4 channel bench power supply. It comes in handy when you need 6.3V at 50 amps lol.
Well, that is debuging the hard way. Propsfor the Tilt 5 T-shirt and I am sure you will figure it out in the end. Sometimes it just needs a good nights sleep. Thanks for sharing.
Who knew that a spectacular short could be so entertaining to you?! You face challenges with impressive charm. Can't wait for the next update. Good luck buddy.
Powering up new circuit without current limit = Madlad! (Or smell-oriented design.) It seems to me that the paper tape is oscillating somewhat in the direction perpendicular to its surface. Maybe some more accurate tape guidance could improve the stability of the bit detection.
Great work! Pushing down the problems for future self to deal with, that is indeed relatable. Good luck on the halting problem, you might be eligible for noble price after you solve it 🎉
I used to heat my house with a room full of servers I bought from my company's surplus department. Hopefully there are plenty of nearby fire extinguishers🚒I wish I still had them, they would be worth so much money today :(
You exemplify the phrase, "Failure is OPTIONAL." You keep on plugging away in spite of set backs and the reason I continue to enjoy your excellent videos.
for the tape reader, you may have better luck using a motor from a fish tank pump. It would be quieter and have much lower rpm's. A smaller one that hangs from the side of an aquarium to be exact. The motor spins a small finned rotor, so may work well with a plastic pulley setup.
The slot car motor won't reliably run slower? You might want to consider an old casette deck motor. They often had adjustable servo-feedback motors that can run at steady speeds
An Edison phonograph motor would look nice. They have a flyball type governor that opens some contacts when the motor gets to speed, closing them when it slows down. They are fairly large, and the ones I have seen run on like 200 volts DC. Usually some on ebay.
The fix is right behind your left shoulder. That is an early prototype of the first ROCKWELL ENCABULATOR. It is a little heavy on the power use but extremely reliable!
Hook a PWM driver to the tape reader motor so you can fine-tune the speed but retain all of the torque. Yes, the "period correct" way would be with more gear reduction, but I don't think we're going to excommunicate you for taking a shortcut for testing purposes!
With that “halt board” not working, there must be a joke about “solving the halting problem” and a reference to Alan Turing somewhere…
Yes - if Dave solves the Halting Problem it'll be a world first!!
Exactly the joke I was thinking about.
What about the “catch fire” board?
I think he ran that joke in an earlier video, actually. I want to say it was one about the PDP-11.
As for the "catch fire" board, it looks like the bit around 11:20 came awfully close . . . 😅
That was my first thought too... Computer science humor at its nerdiest
"You ever overclock your computer?"
"Yup. I got mine up to over 200 Hz, baby!"
"You can’t break the laws of physics" - Scotty
People keep giving me advice on how to slow the tape drive down, and I'm over here trying to figure out how to make the rest of the machine keep up with the tape drive!
@@UsagiElectric Well, maybe some way could be devised to drive the machine with conventional TTL signals through a voltage level translator. Perhaps a stream of data from a serial port would work. This wouldn't be for ultimate use but it would serve for debugging. Maybe the machine really can keep up with the drive but you can't know without a speed test.
@@UsagiElectric Gotta reduce the values of your decoupling capacitors. With only 24 volts the time constant of the RC circuits will be stretching the pulses! Remember Alan Turing got his Pilot Ace to run at around 1MHz!
"If I can solve the Halt problem" ... yeah, that should be super easy, barely an inconvenience.
Is it not proven to be impossible?
He's done it before and will do it again!
Oh really!?
@@hamaljay Tight!
At this rate he'll solve P=NP before year's end.
Would it be possible to add a bell to UE1? It sparked joy every time It rang on G15.
"Aah! I see you have the machine that goes ping. This is my favorite."
UE1 Garand?
@@v2point0 garand ping best ping
Let's get it working first, but it's definitely possible!
@@v2point0 Devise a pea shooter?
Usagi;s Overclocked UE1
You're going to solve the halt problem?!?! Dude, you'll win a Turing award for that!
The slow-mo anger made my morning, loving watching this project develope.
It's always the unexpected things that you figure would be easy that getcha
That was fun to film, haha.
“Watching you solder pcbs” is one of the reasons I come here. Not boring at all. More soldering pls
I want to see this run with the lights out
I think that's the holiday special
@@shadowgrue Now where did I leave the red, green, yellow, and blue vacuum tubes again...
I think that would require another adjustment of the photo diode bias, as when the paper tape reader was moved from the bench to the machine. Blocking ambient light out of the reader might help improve stability.
Once it's executing the right code, we'll definitely do that! Get it nice and dark in the room, get the camera setup right, and let it blink, blink, blink away!
Such a very Usagi thing to say - it executed its first instructions.. the wrong ones! ... always with a grin, Hellorld!
Exactly. After that it's software debug time.😀
If you decide to use percussive maintenance, we could call it bit banging.😂
That woodgrain is LGR worthy
Really a lot more went right than didn't, Congrats... you are close! A shorted power bus, logic levels wrong, a Halt circuit with a flip flop that flopped... LOL, it's all par for the course. I've built a 4-bit TTL computer home-brew on a ton of breadboards, I had wiring problems like the solid core wires would break after flexing but it would look like they were still connected... so it's normal in the bring-up to have problems.
Yup! More went right than I was actually expecting, and I genuinely think that in the next episode, we can actually get it churning along, executing the right instructions!
Building homebrews are tough, and it's always the things least expected that get ya.
@@UsagiElectric Well, you're doing it the right way by bringing up each section and testing it. What I found out, besides the silly wiring errors, was timing problems. For my control logic I programmed ROMS to output the clocks and enables to registers and things on the bus. The inputs to the ROMS were the Instruction Register, T-State Generator, and some outputs from the ALU to work the branching instructions. The outputs would glitch on the T-State transitions and it acted like I had a squirrel brain running my processor. LOL I had to add a latch to buffer the control signals on the falling edge and I needed to rework the microcode a bit... I got it to work, sort of... I had some instructions execute but not all of them. I wish you the best of luck and success on your next attempt.
This stuff is important. Technological dead ends such as The Baine Clock or Doble steam car featured inventive imagination lost to history, or at least in the case of Baine. The method he used to create the three pole magnet in one metal bar seemed beyond the technology of the 1850s...and the way he did it is lost to history. Your determination is remarkable and not as useless as some may think. All the time, we need invention to keep society going. Thank you.
“I hope nothing catches on fire.”
Immediately starts arching.
Dude, huge congrats! It's such an awesome thing having people like you in the world who put so much heart and effort into a hobby like this. That feeling you get when the project you've been working on for so long finally shows signs of life is one of the best feelings in the world and it never gets old - just a couple steps away now and we all know it will work beautifully once those last couple bugs are fixed!
I'm amazed at the amount of work you put into these machines, so interesting to follow, incredible to watch
18:54 "the halt circuit is gonna break me" - idk, sounds like the halt circuit can't break anything at the moment *rimshot*
The cutest house warmer in the world... just in time for the holidays. :)
WAIT, there's a HALT problem ?
It looks so cool and I agree, it's executed code. ❤
I don't mind watching you solder, as long as we hear a good story.
Same. Always love a soldering time lapse / montage with a voice-over.
Unforeseen consequences time! As entertaining as starting the Bendix.
"You can't give up just yet, Dave! Stay determined..."
Eli did not like that...
@@GegoXaren 𝓭𝓸𝓬𝓽𝓸𝓻 𝓯𝓻𝓮𝓮𝓮𝓮𝓶𝓪𝓪𝓪𝓪𝓪𝓷
What a beautiful machine! Built to perfection! Cool episode too! You managed to over-clock your tube computer and started solving the halting problem in the same episode! Watching your channel is my little Monday morning ritual and I love it!
Congratulations! Your grit is astonishing.
The UE1 close to working is so cool indeed 🙂 I like how you let us see the first try of tests so we can share in the suspense and excitement of successes and failures. I liked the bunny break. I totally hear ya on the overall current requirement bringing it up and how nail biting that was. We are kindred spirits 😎⚡️ Reading the other comments, I love the "overclock" ones!
Your paper tape is big vs little endian. One of the classic computer quandaries. 😅
And the debate continues!
Who would've guessed even 1-bit machines would have endianness!
that's the way to do it. persistence despite failure 👍
@@andrewdunbar828 Next on Usagi Electric, The Little Endian In The Cupboard.
Not only that, I'm fairly certain I've got the nibbles backwards too! I'm gonna have to sit down and have a proper think about how to wire this thing up correctly.
that is such a cool project. congrats that is a fascinating way to learn about computer when you clearly see all components
Why don't you put a DC dimmer on the motor? By controlling the voltage, you control the speed...
To be more 'old-school', you can use a wirewound potentiometer...
The old-school method may also involve a segmented contact connected to the shaft that would create a kind of PWM, with a capacitor in parallel. That would also help to fight local speed alternations to some extent because when speed goes down, the time the contact stays for the next phase gets longer, and that naturally increases the speed. But one should, of course, expect sparks and wear, so that's not an ideal solution.
was thinking the same. GMTA!
Short answer: loss of torque.
USAGI will salve the halting problem!
love how you take problems/challenges in stride with a smile and a laugh! I want to be more like you
6:18 hmm I don't know about ‘it doesn’t work so I’m not going to explain it’. You have clearly managed to build an audience of people who enjoy watching you work on tech mysteries of the ‘why doesn’t this work anymore’ variety, I could see ourselves also enjoying mysyeries of the ‘well here’s how it’s supposed to work but so far it doesn’t’ kind.
Plus, you’d get the benefit of rubberducking your ideas? Sometimes, just explaining the thing makes something stand out.
(Just a random comment from a random person on the internet - I don’t my take my comment too seriously as I don’t produce videos for a living or at all really, if this made for dreadful watching i wouldn’t know it)
I mean, we have just watched a whole episode of him troubleshooting his computer lmaoo
I was pretty sure I was on the right track to figuring out the problem, I just ran out of time to troubleshoot it in this episode. I wanted to wait until I knew what was wrong, so in the next episode I could explain what the design was, what went wrong, and how I fixed it. And today, I did actually get it squared away, and the failure was a routing error and a resistor value error, but the why behind the resistor values being wrong is actually really interesting, and I'm glad I'll get to dig in deep on that particular problem in the next episode.
@@UsagiElectric Great way to build up the anticipation for the next episode! (I see what you did there!) :)
"It just executed its first instructions, though they were absolutely the wrong instructions" --Usagi, this very video
"Bloody instructions, which, being taught, return/To plague the inventor . . ." --Shakespeare, _The Tragedy of Macbeth_ I.vii.9-10
Are you trying to overclock the UE1? :)
I'm intrigued. Never intended to watch about the UE1, but now I want to see it completed. I came for the Centurion, but am stuck on UE1.
BTW I acknowledge you.
I absolutely love your energy and joy even when something goes wrong. One of the best channels on UA-cam right here. I hope you get everything fixed soon.. much love from Germany! :)
I absolutely love ist too. This is so satisfying. I run a small museum in the corridors of my office. People are always amazed at the resources that were used back then. A tube computer like this would be a gem in my collection.
0:38 almost ready to work is me every workday
Same here
Congrats Dave! I call this a win dispite all the pain you've suffered along the way and especially here at the end. Do NOT let this mess up your Christmas holiday season, sit down with the parents, the critters and a whole lot of rum-loaded eggnog. You done good work.
Came to the comments to see jokes about the halt problem and was not disappointed
Let's just hope the halt problem isn't like the Halting problem!
Amazing work. You have serious levels of persistence.
Building, soldering, design I for one am here for it. I learn new techniques that way and it’s never boring. And yes… it is an amazing work of art.
When he said swedish fish I thought; What is he talking about, until I saw the candy. I have never seen that candy, and I am from sweden!! Great content as always btw!
I know so very little about computers. But, I love your videos! I hope this goes in to a museum someday where people can enjoy it!
Iceland Calling, I always look forward to your Episodes and especially ones on deep dives into how it works. thanks!
Hey David! I work in cellular communications and I have access to some really high output rectifiers, I have -48VDC @ 50 or 75 amps and +24V @ 100 amps. They’re 1/4 width, 1U rack mount size. I also have a lot of random cut lengths of fabric sleeved high temp cable (it looks straight out of the 1940s) and multi conductor cable, up to 24x 18GA control cable.
Let me know if you’d be interested in any of it, I use the -48V rectifiers to build a massive 96V (-48 to +48), 1000 watt, 4 channel bench power supply. It comes in handy when you need 6.3V at 50 amps lol.
Well, that is debuging the hard way. Propsfor the Tilt 5 T-shirt and I am sure you will figure it out in the end. Sometimes it just needs a good nights sleep.
Thanks for sharing.
Still counts indeed! Wonderful progress
Who knew that a spectacular short could be so entertaining to you?! You face challenges with impressive charm. Can't wait for the next update. Good luck buddy.
It may be that he laughs when something goes wrong because the only alternative would be to cry every time something goes wrong. 😅😉
Powering up new circuit without current limit = Madlad! (Or smell-oriented design.)
It seems to me that the paper tape is oscillating somewhat in the direction perpendicular to its surface. Maybe some more accurate tape guidance could improve the stability of the bit detection.
I'm not this deep into electronics, but I really enjoy watching these videos
Great work! Pushing down the problems for future self to deal with, that is indeed relatable. Good luck on the halting problem, you might be eligible for noble price after you solve it 🎉
Really enjoying the series. Also, loved the use of Hackerman.
Dude. This is how I found your channel because I googled “how big would a / could you make an NES with vacuum tubes…” 😂
Tears and Triumphs. Usagi videos have them all!
Your workmanship is stunning
Amazing, you were bringing it up for the first time and *already* overclocking it! 😆
It's a hell of a project! Congrats on getting it this far!
A thermal camera view of the UE1 would be interesting. For the power it should draw some of the supply wiring seems very thin.
First power on executed one code perfectly: "Halt and Catch Fire" 😂😂
It’s amazing to see how far this project has come, great stuff 👍
17:35 very much National Lampoon's Christmas Vacation energy 😂😂
I know you'll get it working! I believe in you!
The UE-1 WILL run!!!
Such an impressive project. Huge respect to you, sir.
The good news . With all those tubes staying warm won't be a problem . Still very cool .
I used to heat my house with a room full of servers I bought from my company's surplus department. Hopefully there are plenty of nearby fire extinguishers🚒I wish I still had them, they would be worth so much money today :(
You exemplify the phrase, "Failure is OPTIONAL." You keep on plugging away in spite of set backs and the reason I continue to enjoy your excellent videos.
for the tape reader, you may have better luck using a motor from a fish tank pump. It would be quieter and have much lower rpm's. A smaller one that hangs from the side of an aquarium to be exact. The motor spins a small finned rotor, so may work well with a plastic pulley setup.
The slot car motor won't reliably run slower? You might want to consider an old casette deck motor. They often had adjustable servo-feedback motors that can run at steady speeds
An Edison phonograph motor would look nice. They have a flyball type governor that opens some contacts when the motor gets to speed, closing them when it slows down. They are fairly large, and the ones I have seen run on like 200 volts DC. Usually some on ebay.
Congratulations, Usagi! It runs!
Very interesting update, the tape speed is faaaaaast great video👍
"The next step is to solve the halt problem"
good luck with that :D
10:15 I nominate “A Beautiful Culmination of Three Years of Questionable Decisions” for the plaque that goes on the paper-tape catcher…
Congrats David! Awesome stuff, love the UE1
The fix is right behind your left shoulder. That is an early prototype of the first ROCKWELL ENCABULATOR. It is a little heavy on the power use but extremely reliable!
It's very nice to see Soviet parts in your computer, in 2024. These buttons are very high quality and reliable.
Great progress!
Lifespan of relay: one million cycles. That computer's relay: i'm done in ten seconds.
I do love this channel and I follow you with interest well done!
You think you're building a computer, but this is art, man! Great stuff!
Awesome stuff!
Great Work! I'm so excited to See this in operation.😁
Totalmente ALUCINADO !!!!!👏👏👏👏👏👏
It failed, but at least it failed gracefully.
Happy Christmas to the bunny household.
New Episode of Usagi's UE1 for my Birthday ❤ #Winning
You are a cool cucumber in the face of catastrophic failure! That astable smoking relay might have caused me to need a pants change.
Most amusing and entertaining episode ever.
It's so close. I'll eagerly wait for the next episode.
Thing is coming along nicely.
Neat, and side comment “filamnent” overlay caught me funny
The first and only overclocked single bit processor, one byte RAM, low voltage vacuum tube computer. Epic.
It's always really interesting and fun to see your videos. Would be really interested in how many animals run around at your home ;)
The fuse track worked as expected
So epic! Can't wait for the next video!
You're a mad lad and I like that
Now this is getting exciting!
It's a piece of Art !!
Some UA-camrs go back to 8 bit for a challenge, he goes back to 1 bit.
Hook a PWM driver to the tape reader motor so you can fine-tune the speed but retain all of the torque. Yes, the "period correct" way would be with more gear reduction, but I don't think we're going to excommunicate you for taking a shortcut for testing purposes!
When you do your cable lacing Make sure you get Wax String. I use to be a Telcom Detail Engineer/Installer and it works best and looks nice.
“Magically fills vacuum tubes with propane gas”
Absolutely fantastic! ❤❤