I can't believe it took UA-cam this long to recommend me this channel. Beautiful work, I love your attention to the aesthetics of the processor in addition to the functionality. Gotta watch all the previous episodes now
Man, this is amazing. The fact that you've build a processor (even if it's just 1 bit) out of freaking vacuum tubes is simply incredible. Congratulations.
Thank you so much! Once we get some memory and program control I think it'll get really exciting to watch it churn out the Fibonacci sequence or something!
This is so incredible. I should have gone to bed like 5 episodes ago but I just had to see the CPU finished! Even the remote control alone is a work of art. I can almost feel the switches clicking! This whole project is a flawless synthesis of art and science.
Thank you so much! I'm super happy with how the remote control turned out! Unfortunately, as the computer grows I'll have to build new remote controls, and I can't find any more of those switches. They're by far my favorite toggle switches, I kind of want about 100 of them for various projects.
Hi I'm a 100% Layman on this topic but I wanted to comment on your video skills and bringing the audience in I found it hard to understand however you made it feel like everyone was welcome and that you were talking directly to your audience. Thank you.
You and me both! I fully expected to plug everything in and for none of it to work, haha. As it sits, I have one bad VFD and need to change the resistor on the soft start and that's it!
Magnificent build! Everything is so clean and structured! I understand your jubilation, because when I finished building and testing my MINIVAC 601 and saw the correct output on the lights it justified all the hours spent in it’s construction. CONGRATULATIONS!
Thank you so much! It really came together better than I was expecting, and I'm just blown away that it all seems to work together really well. I just looked up the Minivac 601, that's a really cool kit! I love the idea of building logic elements out of relays, and one of my all time favorite computers is the FACOM relay computer built by Fujitsu. There's just something so satisfying about hearing the relays click away while computing.
Thanks for checking the videos out! I'm gonna go into this in a lot more detail in the memory videos, but unfortunately, I'm most likely just going to brute force some memory with vacuum tubes. There's a ton of reasons I'm doing it this particular way, but the big one is addressing. Since I'm only working with a 1-bit data bus, that means I need to address a single bit at a time. In comparison to addressing and read/write control for each bit, the physical size of the bit, even if brute force built from tubes, is relatively small. Core memory is fantastic for larger memory sizes, where the amount of control circuitry is balanced by the amount of memory you have, but for the limited physical space I have here, I just can't make it work. Now, having said that, I have two more homebrew computers I really want to build. I want to make an 8-bit machine using core memory from discrete transistors, and I want to build another tube computer that's much more capable using Williams tube memory, so I will definitely be diving into special memory types in the future!
@@UsagiElectric maybe throw in a wire delay line? Or if that's too fast, just a tape loop? You can hack a reel to reel, swap the record and play heads and loop the tape around.
Thank you so much! Only for the time being, the next three parts are going to be equally as big and then she'll be getting pretty large and in charge, though not quite Colossus size. Still, my electric bill will probably disagree, haha.
Amazing! I just wish you had used filament bulbs as limiting resistors on the valve-boards so we could see the states of every logic gate as it worked.
If I counted right, this CPU uses 182 tubes, not including thyratons in soft start and VFD-s. Pretty mindblowing, and great demonstration of how much compactness semiconductors offer
Really close! It actually uses 188 tubes not including the soft start thyratrons or VFDs. With the soft start, it makes it a nice round 190 tubes (by pure coincidence). It's crazy how much more compact this could have been made even with just discrete transistors. I've done a bit of fiddling around with discrete stuff, and using the exact same construction techniques as the tube computer, I can stuff a NOR gate in an 18mm x 15mm space, whereas with the 6AU6, that same NOR gate takes 40mm x 20mm. That's 270 mm^2 vs. 800 mm^2. That means that I could realize about a 67% reduction in size just by going to discrete transistors and not changing anything else. With dual sided pcbs and thinner traces, that could easily be made even smaller!
Thank you so much! It's so hard to convey just how amped up I was inside, I was losing sleep for like 5 days straight while I was finishing up the final touches. I was both excited and terrified, haha.
Thank you so much! Oh man, if I crammed 300V down it, I'd have a whole lot of exploded heaters, haha. Somewhat unconventionally, I'm running the heaters off the primary +24V B+ rail or -12V rail. This greatly simplified construction, but really limits what voltages it can be run at or even which tubes can be used. But, all is not lost, I'm already thinking of ways to set up program control so I can really push the limits of the computers and see just how fast it can run before it gets unstable!
Congratulations, you did a heck of a job building that and it works! As a person who is not an electronics expert, I'm completely lost on how it works, but I still think it's cool! Good job! 😂👍
Nice cable management with the power supply, great solder joints on the boards and it's aesthetically pleasing looking at all those vacuum tubes. Great job overall!
Thank you so much! I think I'm going to take a short break from the tube computer and tackle some other projects that have been stacking up, but I can't help but be excited about the next stages of the project!
Thank you so much! That night shot really turned out excellent didn't it! It was a 20-second exposure, but it was the closest representation to how it actually looks in person.
Wow . . . I can truly appreciate the magnitude of such a project, but um . . . well, maybe a bit too much time on one's hand. But hey, it is cool, replicating motorola 14000 series logic and processors with tubes. When I was very young (4th or 5th grade), I remember being taken on a field trip to a USAF base (Wash), and got to see a digital tube computer. It literally filled a multi-story building, floor to ceiling on all floors. A tech was swapping logic boards, meanwhile the tour guide had the system playing a simple audio tune through a speaker . . . a really big deal for a demo.
Your videos are FAR more entertaining than Avengers End Game. Congrats on a fine job. That power supply case you made is the prettiest I've ever seen. Your troubleshooting skills are impressive. Good for you because I have designed an instruction set and architecture but not had the time or nerve to build it. MIne is a 12-bit bus with 16 instructions in the set. All good wishes!
Thank you so much! You're too kind though, I still have a long ways to go in figuring out a good way to tell the stories of the builds, but I like to think the videos are definitely improving. A 12-bit bus is a really interesting bus width to choose! Then again, here I am with a 1-bit bus, so interesting bus widths are right up my alley, haha.
Wow, i am sitting here on the edge of my chair. Love to see you explode of joy with achieving this great milestone right on time to find a good ending for this year.
Amazing work! Thank you again for sharing this with us... I think the most understated aspect of this is how well this was planned ahead... knowing it would all fit within the mechanical layout (and it actually doing so) and confidently knowing exactly what signals would be required where.
Thank you so much! It wasn't too incredibly well planned, I just kind of started designing PCBs at the top and worked my way down, totally making it up as I went, haha. I'm really happy with how it came together though! However, the overall size is really arbitrary, which is meaning that I have to put a lot more planning into the next parts of the build because I want all four parts to fit on the same size of backboard (1,000 x 700). I think I've just barely got memory fitting in my early designs though, so I'm hopeful the whole thing will come together smoothly!
I followed your videos about this project, and i must say this is really impressive. Your work - this 1 bit processor - is not only functioning, but looks amazing. Nice and clean job, really a masterpiece! Big thumbs up!
Thank you so much! A big project like this is as much art as it is educational for me. I really wanted to learn the very basic fundamentals of computing, but if I was going to build something, I wanted it to look cool too, haha. I'm so happy with the way it came together!
This is so hot💡it's cool, and I'd think it likely that once you have hit the frequency max of this configuration, you will be able to bump it up even more by cutting the delay capacitor values and attaching more filter capacitors on the power rails close to the circuitry. If you scope those power rails before doing that you'll probably see lotsa noise. You ought to get it at least well into the sonic range, fast enough that most things controlled by it wouldn't care. But getting it to work right at all is a huuuuuge milestone. Congratulations.
Thank you! I think the biggest limitation on speed isn't going to be the processor itself, but rather the program control. I'm going for a rather unconventional build style of reading the program directly from the storage media (1/4" audio tape). That means I'm limited by how fast the tape can be fed through and how many bits of information I can stuff in a single inch. My first goal is to fit one full instruction per one inch of tape, which at reel-to-reel speeds should be around 7 inches per second, or 7Hz operating speed. If I can get that working reliably, I'll work towards fitting a full instruction in 1/2" of tape, and then maybe push for an instruction per 1/4" of tape, which would be 28 instructions per second. Now, I also plan to add an external audio jack so programs can be read in directly from another computer or even my phone (they'll just be audio files), which means I can play with how fast I can push it before my read amplifiers on program control start having problems. Either way, if I can get it above 10Hz, that'll be a total win in my book!
Thank you! It's really wild to see it all working and working well on top of that! I think I'm going to take the month of January to catch up on some other projects, so it'll be a few episodes before we're back onto tube computer stuff, but I've got some designs for memory already and I'm getting ready to start testing and building!
Thank you! That night shot really turned out excellent! It was a 20-second exposure, but it was the closest representation to how it actually looked in person!
New viewer here.... I watch a lot of "build a CPU from TTL" videos.... and I loved an old channel that never finished a relay based CPU... I'm going to love this.
Thanks for checking the videos out! I love relay based computers as well and really want to build one some day in the future. Then again, I also want to build a discrete transistor CPU, and a high voltage tube CPU, and a pure hardware cellular automata, and man the list is getting long, haha.
What you have done is admirable. This is the first video of yours I have known, and looking into all you have done... from Diode.. triode.. very impressive work.... amazing. Thanks for sharing.
Thank you! I'm not quite the first human this century, though. There have been a few other vacuum tube computer projects. The Ena Computer is a pretty recent one that's honestly way cooler than mine! www.ena.computer/ (Though, if I'm being honest, while the Ena Computer is way more powerful and much better representation of how to actually build a computer, I think mine looks cooler, 😁 )
@@UsagiElectric didn’t know about ena… but yea yours is prettiest vacuum tube computer of the century 😂 , your one is cool and minimalist .. I m so tempted to build one on my own but I don’t have to be guts to make such a big commitment , I’m building small breadboard solid state stuff at the moment
I spent a lot of time unsoldering and replacing mini tubes on removable boards in cryptographic equipment I didn't want to see another tube, but now in my 70's I'm a little nostalgic for them.
Absolutely phenomenal job. Your engineering prowess is humbling, your enthusiasm is infectious. I'm very much looking forward to getting back into some analog amp design and builds based upon your efforts here. Thanks, and please continue to do your good work. We very much appreciate it!
This stuff is way over my head, but it for sure makes me appriciate the folk who pioneered computing technology because I certainly couldn't design and build this, not in a billion lifetimes.
Saw a video you might enjoy -- "1957 Digital Technology Using Tubes!" -- it's on the channel Mr. Carlson's Lab. Use of tubes in the HP-521C digital frequency counter ...
p.s. Since comments disappear sometimes, so I'm not sure if you saw what I wrote a few weeks ago - I'm letting you know that next week I'll attempt to reach a former business associate from the 90's whose dad knows Toru Iwatani ...
Yup, I saw Mr. Carlson's newest video! I actually have a very similar model of HP frequency counter! Though, don't tell Mr. Carlson, but he got some stuff wrong about how it actually counts and displays, particularly for the nixie version.
@@ka4dqe606 That's awesome! And yeah, UA-cam handles comments relatively well, but comment replies very often get lost in the notifications. So, the reply may be live and visible, but there's not a great way for me to find it.
Not forgotten ... Still trying to contact Jim Jarocki about his dad Stan (chief architect of Midway's PAC-Man success) - regarding Toru Iwatani. I forget sometimes that everyone is getting older ... decades pass fast ... since my focus is problem solving and building things and family. I'll keep at it, as I'm certain Iwatani would appreciate your electronics discussion in Japanese -- Christopher Watkins
I appreciate how you take us through the entire journey - including errors and tweaks. Every video I watch besides yours, makes it seem so easy. Seemless - painless, no issues. Thats never the case in real life with computers haha! The battle of man vs computer and debugging, troubleshooting to me is 90% of the fun is. Its what makes it interesting. Thanks for these awesome videos. I would love to get some more information on how I can do this myself - maybe you could sell a kit or provide links and instructions. I would love to do this at a school for a yearlong hands on project or etc. Thanks again for the great content and good luck in your endeavors!
Thank you so much! It is really surreal to think that it was just a logic diagram we came up with inspired by an IC. To see it come to life and actually operate is just kind of overwhelming in the good way!
Thank you so much! I'm so happy with how the processor turned out. I've got some designs for memory already and am ready to start cutting and testing boards! Though, I think I'm going to take a few week break from the tube computer to catch up on some other projects that have been piling up, so it'll be a few episodes before back into the tube computer, but I'm quite excited about the upcoming parts!
Just watched the entire build to this point. Just awesome! So inspiring! Going to work my way through the rest of your back catalog while i am snowed in. Keep up the great work!
Well congratulations! And a Happy New Year too! I'm one of those people that missed the first 21 episodes and just watched this one out of curiosity, but I have to say that I'm impressed. I spend a few years working on some actual vacuum tube computers back when I was young, so I'm really impressed at seeing something vaguely along the same lines recreated.
Welcome and thanks for checking the video out! That's awesome that you got to see and work on some of the real deal! One of my bucket list goals is to see an IBM 604 in person someday as it's been a massive inspiration. Thank you for the kind words on the project, I'm really liking how it's shaping up!
Congrats! the processor turned out a piece of art. Thank you for sharing this experience. Can´t wait to see how you will tackle the other sides of this puzzle.
Thank you so much! I'm super happy with how the processor turned out! Memory is up next and while it may not be quite as interesting to watch being built, there's a lot of really great stories about memory and how I came to the decisions I did!
Geez, that's more tubes than an LGP-30 from ~1957, a serial 31-bit, 16-instruction machine with a 4K-word drum memory. The LGP-30 blazed along at 0.00024MIPS!
Thank you! I grew up on some of Square's best games (in my opinion at least), like FF6, FF7 and Chrono Trigger, so I like to give a nod to them whenever I can!
Thank you so much for the sub! Memory and program control should be quite fun, and then the computer will get really exciting because it'll be able to execute full programs!
This looks like a project I might have considered in about 1972, and dammit, I had enough 6AU6 tubes to do it too, from years of tearing apart old TVs found in the garbage.
This is really quite awesome. And in the video, you never even mentioned the main cool thing about this - that you're operating way outside the spec of these tubes on these low voltages, and you got them to work so reliably that you can demonstrably build a functioning processor out of them. That is quite a cool feat!
Thank you so much! And I totally forgot to mention that the entire thing is running at super low voltages. I've been working with them at just 24V for so long now, a part of my brain has just normalized it. I often forget that these little 6AU6s are usually meant to have four times as much B+ voltage!
Thank you so much! It came together way better than I could have ever hoped for! I'm really excited about the next parts as well, program control in particular!
@@UsagiElectric where you are now compared to where you started prior to this being built out: do you like this specific processor's design? Do you see things you'd rather do differently (add/remove) to make it "better"? Would you ever contemplate designing a processor from scratch?
@@colinstu That's an excellent question! And truthfully, I've learned so much about processor design in the process of this, if I were to start again, I would do a totally different design. I've actually be toying around with some 4-bit and 8-bit processor designs from scratch that I quite like, and the 4-bit one in particular I think could be made in tubes and result in something not that much physically larger than what I will ultimately have here. Here's a simplified block diagram for the ultra simplified 4-bit design: i.postimg.cc/BQbJK7W1/UE4-S-Simp.png When I finish the 1-bit computer, I may actually try to tackle the 4-bit design with tubes. The 8-bit design I'm quite happy with too, and while I don't think I have the resources to build it out of tubes, I do think it would make an excellent discrete transistor processor. Here's the 8-bit design, and while it's not anywhere near as capable as most "modern" 8-bit processors, it shares a lot of its instruction set with the AMD AM2901 bit-slice processor. So, it's fairly capable as far as ALU functionality, and when combined with the 256 byte RAM, it should be capable of some pretty interesting stuff. i.postimg.cc/zBvxjhqp/UE8K.png
Very cool Hobby...Not much practical purpose to reinventing the wheel, but really awesome that you have such a zeal , and passion for it , and it is very watchable and entertaining. Only several Million more like it , and you can build a computer the size of a 5 story building with the computing power of a 1979 TRS 80.Dont get me wrong , I am not hating or being synical..I actually think it's cool and enjoy watching it .
My Goodness its a thing of beauty !!!!! Can you do a video on your board manufacture process ? thats one serious milling machine !!!!! how did you convert it to CNC ? what program are you using to design those beautiful boards ? Thank you for sharing your fantastic work !!!!!! look forward to each of your videos !!!!!!
Thank you so much! I definitely plan to do a video in the future to fully outline the entire process of making a PCB from idea to finished product! To answer your questions though, it's a Bridgeport EZTrak, which I believe was sold by Bridgeport as a CNC machine. We got it used from a college that was shutting down for... not cheap, but not too expensive either. I design all my PCBs in DesignSparkPCB, but really, any of the modern EDA software should work. I have set a custom design rule (grid spacing of 2.54mm and 1mm trace widths) just to make it a little more forgiving for unevenness or lack of precision on the oversized mill. Once I have the board designed, I export it to gerber, then load it up in FlatCAM to convert it to G Code that can be read by the mill. Then, it's just getting the machine setup and cutting!
@@UsagiElectric Thank You for taking the time to respond !!!! I just learning CAD and I have a small cnc router im learning to use. unlike your bridgport its just got the stones to do PCB and not much more !!!! Man the things you could make with that Bridgeprt are amazing and in steel !!!! but it sure does a nice job on those PCB's Happy New Year to you and yours !!!!!
Amazing, the remote is a nice touch. Maybe a "overlay" PCB that masks the VFD dots and adds a label would be a good addition, however it would obstruct the view of the beautiful tubes. It seems like a great thing for christmas, it lights up nicely and heats your room. I'm really looking forward to where your journey brings you in the next year. Until then, Michael
Thank you so much! The harness and remote (as much as it pains me to say it) are really just temporary pieces to demonstrate the processor itself. Once I get the other portions of the processor built, I'll have to rebuild the harnesses that connect them, as well as way to switch between external control and program control for executing programs. Ultimately, the I/O board should have some pretty cool VFDs on it though showing us human readable output!
I see all this soldering without using flux and it triggers the tech in me lol. Using flux would make the solder "flow" SO much easier and be less "sticky" and make more beautiful joints. Awesome build though! Far above my understanding!
btw, I am looking at how everything connected, do you keep in mind that those header pins are only rated for 2-3A max (per pin continues load) as well as max working voltage of 250V max. Running relays in parallel is actually not very good idea. If one fails, you certainly will blow up the other one. Some can also have balancing issues. Even more so when one switches on (or off) a fraction of a second earlier etc. For the heaters I would just use a low drop constant current source, to make things easy. But anyway, looks amazing, great job! 👍
Thank you so much! Thanks for the heads up on the relays, I never thought about that being an issue. Looks like a soft start V3 may be in my future, haha. Fortunately, I'm actually just under what a single relay can handle as far as current (the G2R1-E I'm using is rated to 16A and I'm spiking to 15A), so I totally could have gotten away with just using a single relay, but I think in the future I'll upgrade to some 30A capable DPDT relays. As for the header pins, they're really the only part of the build I'm not a huge fan of. Fortunately, I'm running everything at really low voltages for tubes. The entire thing is run on just +24V and -12V, so there's no high voltage anywhere. To give the header pins as much of a fighting chance as possible, I have two large power buses that run top to bottom on both the left and right sides. These busses are double wide, using two pins each, so if the pin can handle 3A of current, that's 6A per side and 12A total, which is pretty much exactly what the computer is running at steady state, so I'm really pushing the limits here. The male pins are fine as they're just a nice strong piece of metal and can handle quite a lot of current. It's the female connectors I'm unsatisfied with. Fortunately, I built the design of the buses with the female headers on the removable connection links, so if I come up with a better solution for connecting the male headers to each other, I can replace the connection links with ease. Now, that's just for the vertical power buses on the sides. The boards themselves have power buses that run left to right, ultimately getting their power from the vertical buses. The good news on those is that they share the load from both left and right, and the heaviest load on a single horizontal bus is 20 tubes, which is 1.5A at 24V. That shouldn't overload the individual board to board connections. So, that's a really long winded way of saying I'm pretty confident in the horizontal board to board connections, it's just the large vertical power buses on the left and right I'm a little concerned about. So far it seems alright, but I'll keep letting my subconscious chew on better ways to handle that vertical connection!
@@UsagiElectric Thanks for the clarification. I forgot you were running those tubes on low voltages. A standard practice is to just double up on the header pins. (so use two pins instead of one) Which also doubles up the current capabilities. By standard the male and female part should be rated the same.
@@p_mouse8676 Yeah, I often forget that these tubes usually have four times as much B+ voltage as I'm running here! I know the male and female parts should be rated the same, but the female part construction is not nearly as robust as I would have liked, but then again, the connectors were super cheap, so it shouldn't be all that surprising. If I do come across some more robust female connectors, I'll totally remake the power bus connectors, which should help with voltage drops from top to bottom a bit!
Fantastic! Once you're completely done with the entire computer, you should calculate its MIPS per watt. For ENIAC it was 0.00289 MIPS @ 140,000 watts. 😁
Thank you! Oh man, MIPS per watt is a terrifying prospect! If I get my goal of 10Hz, that would be 10-instructions per second, which is 0.00001 MIPS. Each tube pulls 1.9W in heater power, and currently I'm at 190 tubes, but I think by the time I finish, I'll be closer to 500 to 600 tubes, which should put power draw right around 1,000 watts, which isn't all that bad comparatively!
Wow! You did it. For the next step, maybe you can cheat a bit and use an EEPROM as a program store and clock out the input sequences you did by hand? That would allow a “real” program to run on the tube computer without too much more effort. Your workmanship and design is remarkable. All the best for the New Year!
Thank you so much! I'm trying to avoid using any modern components anywhere on the machine, even if only for temporary use. I even feel bad when I use an LED to check something right quick, haha. But, using an EEPROM carries with it an additional problem of needing some voltage level shifting. Even though I'm running the tubes at incredibly low voltage, 24V is still a bit too high for most modern-ish stuff, so I would need to go through and build a custom level shifter to get the EEPROM output levels up to the right values without damaging the EEPROM itself. Also, without any memory on the board, the instructions have to move at a glacial place so I can "read" the output and "store" it in my brain-case. Once we get memory though, we can start running programs a little quicker and then just focus on reading the final output. So, it might still be a little while before we run a full program on it, but we're slowly getting there!
I can't believe it took UA-cam this long to recommend me this channel. Beautiful work, I love your attention to the aesthetics of the processor in addition to the functionality. Gotta watch all the previous episodes now
Thank you so much!
I know. it was almost like a fyi if you really need it.. I have been blown away! I even had to start from the beginning and can't stop watching!
Same! Amazing channel
Likewise, this is fine work
Man, this is amazing. The fact that you've build a processor (even if it's just 1 bit) out of freaking vacuum tubes is simply incredible. Congratulations.
Thank you so much!
Once we get some memory and program control I think it'll get really exciting to watch it churn out the Fibonacci sequence or something!
@@UsagiElectric Mine some BTC with it 😂
This is so incredible. I should have gone to bed like 5 episodes ago but I just had to see the CPU finished! Even the remote control alone is a work of art. I can almost feel the switches clicking! This whole project is a flawless synthesis of art and science.
Thank you so much!
I'm super happy with how the remote control turned out! Unfortunately, as the computer grows I'll have to build new remote controls, and I can't find any more of those switches. They're by far my favorite toggle switches, I kind of want about 100 of them for various projects.
Hi I'm a 100% Layman on this topic but I wanted to comment on your video skills and bringing the audience in I found it hard to understand however you made it feel like everyone was welcome and that you were talking directly to your audience. Thank you.
Holy crap... this was just loaded into my YT suggestions. Holy Crap! Now I have to start at the beginning. So many questions.... That's a work of art.
Thank you so much!
And it's good to hear that The Algorithm is trying to get the thumbnail out there!
ohohohohoh this turned out a lot better than I expected.
You and me both! I fully expected to plug everything in and for none of it to work, haha.
As it sits, I have one bad VFD and need to change the resistor on the soft start and that's it!
Magnificent build! Everything is so clean and structured! I understand your jubilation, because when I finished building and testing my MINIVAC 601 and saw the correct output on the lights it justified all the hours spent in it’s construction. CONGRATULATIONS!
Thank you so much!
It really came together better than I was expecting, and I'm just blown away that it all seems to work together really well.
I just looked up the Minivac 601, that's a really cool kit! I love the idea of building logic elements out of relays, and one of my all time favorite computers is the FACOM relay computer built by Fujitsu. There's just something so satisfying about hearing the relays click away while computing.
Power consumption?
I hope you are going to use something like "core memory" along with tubes. That would be so nice!
knitting cores and tube computing! perfect for winter hobbying.
@@thedogwooddandy no, please, buy a second hand memory!
Thanks for checking the videos out!
I'm gonna go into this in a lot more detail in the memory videos, but unfortunately, I'm most likely just going to brute force some memory with vacuum tubes. There's a ton of reasons I'm doing it this particular way, but the big one is addressing. Since I'm only working with a 1-bit data bus, that means I need to address a single bit at a time. In comparison to addressing and read/write control for each bit, the physical size of the bit, even if brute force built from tubes, is relatively small.
Core memory is fantastic for larger memory sizes, where the amount of control circuitry is balanced by the amount of memory you have, but for the limited physical space I have here, I just can't make it work.
Now, having said that, I have two more homebrew computers I really want to build. I want to make an 8-bit machine using core memory from discrete transistors, and I want to build another tube computer that's much more capable using Williams tube memory, so I will definitely be diving into special memory types in the future!
@@UsagiElectric maybe throw in a wire delay line? Or if that's too fast, just a tape loop? You can hack a reel to reel, swap the record and play heads and loop the tape around.
@@senilyDeluxe Using reel-to-reel is actually how I'm planning to implement reading programs in on program control, so lots to come on that!
She's smaller than a colossus.
Very cool build dude. mad respect.
Thank you so much!
Only for the time being, the next three parts are going to be equally as big and then she'll be getting pretty large and in charge, though not quite Colossus size. Still, my electric bill will probably disagree, haha.
Amazing!
I just wish you had used filament bulbs as limiting resistors on the valve-boards so we could see the states of every logic gate as it worked.
CONGRATULATIONS! I've been here from the start and I'm so happy seeing the processor finally finished! :D
Thank you so much for sticking with me from the beginning!
I also am super happy to see the processor finally finished!
If I counted right, this CPU uses 182 tubes, not including thyratons in soft start and VFD-s.
Pretty mindblowing, and great demonstration of how much compactness semiconductors offer
Really close! It actually uses 188 tubes not including the soft start thyratrons or VFDs. With the soft start, it makes it a nice round 190 tubes (by pure coincidence).
It's crazy how much more compact this could have been made even with just discrete transistors. I've done a bit of fiddling around with discrete stuff, and using the exact same construction techniques as the tube computer, I can stuff a NOR gate in an 18mm x 15mm space, whereas with the 6AU6, that same NOR gate takes 40mm x 20mm. That's 270 mm^2 vs. 800 mm^2. That means that I could realize about a 67% reduction in size just by going to discrete transistors and not changing anything else. With dual sided pcbs and thinner traces, that could easily be made even smaller!
You are so excited that I know you go to reach all objectives what you are thinking!
Thank you so much!
It's so hard to convey just how amped up I was inside, I was losing sleep for like 5 days straight while I was finishing up the final touches. I was both excited and terrified, haha.
This is a nice professional job. It looks like a lot of work.
Now run it off 300 volts and overclock it.
Thank you so much!
Oh man, if I crammed 300V down it, I'd have a whole lot of exploded heaters, haha. Somewhat unconventionally, I'm running the heaters off the primary +24V B+ rail or -12V rail. This greatly simplified construction, but really limits what voltages it can be run at or even which tubes can be used.
But, all is not lost, I'm already thinking of ways to set up program control so I can really push the limits of the computers and see just how fast it can run before it gets unstable!
Very impressed with your work sir, the algorithm has been kind to me today.
Thank you!
This is so cool to finally see it finished
Nice work can't wait for the rest
Thank you so much!
I'm really happy with how it turned out, and I'm looking forward to getting some more functionality built into it!
That is amazing! Seeing this tube processor gives you a glimpse on how incredible complex microprocessors really are.
wow that thing is beautiful in the dark
Right! That night shot was a 20-second exposure, but it was the best representation of how it really looks at night with all the lights off.
Congratulations, you did a heck of a job building that and it works! As a person who is not an electronics expert, I'm completely lost on how it works, but I still think it's cool! Good job! 😂👍
Happy finished project 🎉🎊🎉🎊
Thank you so much! It's so surreal to finally see it all working!
Nice cable management with the power supply, great solder joints on the boards and it's aesthetically pleasing looking at all those vacuum tubes. Great job overall!
Congrats on getting this far. Now go out and celebrate, you deserve it! This has been so much fun to watch, looking forward to the next stage.
Thank you so much!
I think I'm going to take a short break from the tube computer and tackle some other projects that have been stacking up, but I can't help but be excited about the next stages of the project!
Congratulations, this is huge! That picture with the lights out... chef's kiss! Looks like a city
Thank you so much!
That night shot really turned out excellent didn't it! It was a 20-second exposure, but it was the closest representation to how it actually looks in person.
Wow . . . I can truly appreciate the magnitude of such a project, but um . . . well, maybe a bit too much time on one's hand. But hey, it is cool, replicating motorola 14000 series logic and processors with tubes. When I was very young (4th or 5th grade), I remember being taken on a field trip to a USAF base (Wash), and got to see a digital tube computer. It literally filled a multi-story building, floor to ceiling on all floors. A tech was swapping logic boards, meanwhile the tour guide had the system playing a simple audio tune through a speaker . . . a really big deal for a demo.
Your videos are FAR more entertaining than Avengers End Game. Congrats on a fine job. That power supply case you made is the prettiest I've ever seen. Your troubleshooting skills are impressive. Good for you because I have designed an instruction set and architecture but not had the time or nerve to build it. MIne is a 12-bit bus with 16 instructions in the set. All good wishes!
Thank you so much!
You're too kind though, I still have a long ways to go in figuring out a good way to tell the stories of the builds, but I like to think the videos are definitely improving.
A 12-bit bus is a really interesting bus width to choose! Then again, here I am with a 1-bit bus, so interesting bus widths are right up my alley, haha.
Wow, i am sitting here on the edge of my chair. Love to see you explode of joy with achieving this great milestone right on time to find a good ending for this year.
Thank you so much!
It was so surreal to see the entire thing working as expected! The whole processor came together better than I could have hoped!
Amazing work! Thank you again for sharing this with us...
I think the most understated aspect of this is how well this was planned ahead... knowing it would all fit within the mechanical layout (and it actually doing so) and confidently knowing exactly what signals would be required where.
Thank you so much!
It wasn't too incredibly well planned, I just kind of started designing PCBs at the top and worked my way down, totally making it up as I went, haha. I'm really happy with how it came together though! However, the overall size is really arbitrary, which is meaning that I have to put a lot more planning into the next parts of the build because I want all four parts to fit on the same size of backboard (1,000 x 700). I think I've just barely got memory fitting in my early designs though, so I'm hopeful the whole thing will come together smoothly!
My Analog (main)❤ is saluting you with great respect. Fantastic work!
I followed your videos about this project, and i must say this is really impressive. Your work - this 1 bit processor - is not only functioning, but looks amazing. Nice and clean job, really a masterpiece! Big thumbs up!
Thank you so much!
A big project like this is as much art as it is educational for me. I really wanted to learn the very basic fundamentals of computing, but if I was going to build something, I wanted it to look cool too, haha. I'm so happy with the way it came together!
This is so hot💡it's cool, and I'd think it likely that once you have hit the frequency max of this configuration, you will be able to bump it up even more by cutting the delay capacitor values and attaching more filter capacitors on the power rails close to the circuitry. If you scope those power rails before doing that you'll probably see lotsa noise. You ought to get it at least well into the sonic range, fast enough that most things controlled by it wouldn't care.
But getting it to work right at all is a huuuuuge milestone. Congratulations.
Thank you!
I think the biggest limitation on speed isn't going to be the processor itself, but rather the program control. I'm going for a rather unconventional build style of reading the program directly from the storage media (1/4" audio tape). That means I'm limited by how fast the tape can be fed through and how many bits of information I can stuff in a single inch. My first goal is to fit one full instruction per one inch of tape, which at reel-to-reel speeds should be around 7 inches per second, or 7Hz operating speed. If I can get that working reliably, I'll work towards fitting a full instruction in 1/2" of tape, and then maybe push for an instruction per 1/4" of tape, which would be 28 instructions per second.
Now, I also plan to add an external audio jack so programs can be read in directly from another computer or even my phone (they'll just be audio files), which means I can play with how fast I can push it before my read amplifiers on program control start having problems. Either way, if I can get it above 10Hz, that'll be a total win in my book!
What a ride! Super stoked that it's working and I can't wait to see the fun stuff in store for next year!!!
Thank you!
It's really wild to see it all working and working well on top of that!
I think I'm going to take the month of January to catch up on some other projects, so it'll be a few episodes before we're back onto tube computer stuff, but I've got some designs for memory already and I'm getting ready to start testing and building!
Well, I'm glad the algorithm showed me your channel, this stuff is awesome!
Thank you! And I'm glad the algorithm decided to throw a little favor my way!
The money shot for me was with the lights out ! well done.....cheers.
Thank you!
That night shot really turned out excellent! It was a 20-second exposure, but it was the closest representation to how it actually looked in person!
At this point, congratulations are in order. Well done! 🙂👍
As you said, an insane endeavour. In the dark it is by far the most beautiful Christmas lights I have seen!!!
I agree 100%, it's the best Christmas tree lighting I could have ever asked for!
New viewer here.... I watch a lot of "build a CPU from TTL" videos.... and I loved an old channel that never finished a relay based CPU... I'm going to love this.
Thanks for checking the videos out!
I love relay based computers as well and really want to build one some day in the future. Then again, I also want to build a discrete transistor CPU, and a high voltage tube CPU, and a pure hardware cellular automata, and man the list is getting long, haha.
What you have done is admirable. This is the first video of yours I have known, and looking into all you have done... from Diode.. triode.. very impressive work.... amazing. Thanks for sharing.
Thank you so much!
And thank you for checking out the videos!
Congratulations 🥳! You are the first human in this century to build a vacuum tube computer
Thank you!
I'm not quite the first human this century, though. There have been a few other vacuum tube computer projects. The Ena Computer is a pretty recent one that's honestly way cooler than mine!
www.ena.computer/
(Though, if I'm being honest, while the Ena Computer is way more powerful and much better representation of how to actually build a computer, I think mine looks cooler, 😁 )
@@UsagiElectric didn’t know about ena… but yea yours is prettiest vacuum tube computer of the century 😂 , your one is cool and minimalist .. I m so tempted to build one on my own but I don’t have to be guts to make such a big commitment , I’m building small breadboard solid state stuff at the moment
It is done. Congratulations 🙂
Thank you so much!
I spent a lot of time unsoldering and replacing mini tubes on removable boards in cryptographic equipment I didn't want to see another tube, but now in my 70's I'm a little nostalgic for them.
Absolutely phenomenal job. Your engineering prowess is humbling, your enthusiasm is infectious. I'm very much looking forward to getting back into some analog amp design and builds based upon your efforts here. Thanks, and please continue to do your good work. We very much appreciate it!
Congratulations on completing this phase of your project!!!
Thank you so much!
It came together way better than I ever imagined it would, and honestly, I'm still totally blown away that it actually works, haha.
This is rad as hell. Looking forward to seeing what you do with the rest of the project
They sound so much better
Tubes definitely sound amazing in audio projects, but if I can hear the tubes on this project, I think something may have gone horribly wrong, haha!
This stuff is way over my head, but it for sure makes me appriciate the folk who pioneered computing technology because I certainly couldn't design and build this, not in a billion lifetimes.
Saw a video you might enjoy -- "1957 Digital Technology Using Tubes!" -- it's on the channel Mr. Carlson's Lab. Use of tubes in the HP-521C digital frequency counter ...
p.s. Since comments disappear sometimes, so I'm not sure if you saw what I wrote a few weeks ago - I'm letting you know that next week I'll attempt to reach a former business associate from the 90's whose dad knows Toru Iwatani ...
Yup, I saw Mr. Carlson's newest video! I actually have a very similar model of HP frequency counter! Though, don't tell Mr. Carlson, but he got some stuff wrong about how it actually counts and displays, particularly for the nixie version.
@@ka4dqe606 That's awesome!
And yeah, UA-cam handles comments relatively well, but comment replies very often get lost in the notifications. So, the reply may be live and visible, but there's not a great way for me to find it.
Not forgotten ... Still trying to contact Jim Jarocki about his dad Stan (chief architect of Midway's PAC-Man success) - regarding Toru Iwatani. I forget sometimes that everyone is getting older ... decades pass fast ... since my focus is problem solving and building things and family. I'll keep at it, as I'm certain Iwatani would appreciate your electronics discussion in Japanese -- Christopher Watkins
I appreciate how you take us through the entire journey - including errors and tweaks. Every video I watch besides yours, makes it seem so easy. Seemless - painless, no issues. Thats never the case in real life with computers haha! The battle of man vs computer and debugging, troubleshooting to me is 90% of the fun is. Its what makes it interesting. Thanks for these awesome videos. I would love to get some more information on how I can do this myself - maybe you could sell a kit or provide links and instructions. I would love to do this at a school for a yearlong hands on project or etc. Thanks again for the great content and good luck in your endeavors!
Congratulations! Seeing something one day was just in the mind working after too many efforts building it is amazing.
Thank you so much!
It is really surreal to think that it was just a logic diagram we came up with inspired by an IC. To see it come to life and actually operate is just kind of overwhelming in the good way!
I was thinking about the "it freaking works!" Quote from Adrians Digital Basement, and so happy to see the tag at 24:54 haha.
Amazing journey man. Be proud of your work is phenomenal. I look forward to what the future holds.
Thank you so much!
I'm so happy with how the processor turned out. I've got some designs for memory already and am ready to start cutting and testing boards! Though, I think I'm going to take a few week break from the tube computer to catch up on some other projects that have been piling up, so it'll be a few episodes before back into the tube computer, but I'm quite excited about the upcoming parts!
Just watched the entire build to this point. Just awesome! So inspiring! Going to work my way through the rest of your back catalog while i am snowed in. Keep up the great work!
Thank you so much, and thanks for watching the whole build!
@@UsagiElectric Truly my pleasure
Chrono trigger music at 8:00 great surprise, thanks 👀👀
OMG chrono trigger music, I thought I was dreaming!
Clearly demonstrates your knowledge, very impressive!
Totally awesome 😎 a real cpu built by hand. Insane!
Well congratulations! And a Happy New Year too! I'm one of those people that missed the first 21 episodes and just watched this one out of curiosity, but I have to say that I'm impressed. I spend a few years working on some actual vacuum tube computers back when I was young, so I'm really impressed at seeing something vaguely along the same lines recreated.
Welcome and thanks for checking the video out!
That's awesome that you got to see and work on some of the real deal! One of my bucket list goals is to see an IBM 604 in person someday as it's been a massive inspiration.
Thank you for the kind words on the project, I'm really liking how it's shaping up!
This is an amazing accomplishment! Congratulations!
Thank you so much!
Congrats! the processor turned out a piece of art. Thank you for sharing this experience. Can´t wait to see how you will tackle the other sides of this puzzle.
Thank you so much! I'm super happy with how the processor turned out!
Memory is up next and while it may not be quite as interesting to watch being built, there's a lot of really great stories about memory and how I came to the decisions I did!
What an awesome journey, thank you for all the work you put into making these videos! I'm so excited to see it working!
Thank you so much!
And thank you so much for checking out the videos!
holy fuck dude, this is great beyond all recognition
Thank you so much!
Thank you for making these videos. It was a blast to see your progress. Good luck in the next year!
Thank you so much!
And thank you for checking the videos out, hope to see you next year!
This is an outrageously awesome accomplishment! I bow deeply to you, sensei.
Thank you so much!
It came together much better than I was expecting, considering I was totally making it up as I went along, haha!
Geez, that's more tubes than an LGP-30 from ~1957, a serial 31-bit, 16-instruction machine with a 4K-word drum memory. The LGP-30 blazed along at 0.00024MIPS!
I really like the Square music you put in some of your videos I've seen.
Thank you!
I grew up on some of Square's best games (in my opinion at least), like FF6, FF7 and Chrono Trigger, so I like to give a nod to them whenever I can!
Very cool. Looking forward to the memory. Subscribed.
Thank you so much for the sub!
Memory and program control should be quite fun, and then the computer will get really exciting because it'll be able to execute full programs!
Omedetou gozaimasu! This is a labor of love, dude. Now I've gotta go back and start this series from the beginning. So very well done!
This looks like a project I might have considered in about 1972, and dammit, I had enough 6AU6 tubes to do it too, from years of tearing apart old TVs found in the garbage.
Lovely to watch !
Thank you so much!
One word: BADASS
Thank you, I'm super happy with how it's all coming together!
Unbelievable !!!
How is this channel not known better? I just blows my mind, a DIY tube processor...
Thank you!
Honestly, the channel is growing steadily, which is awesome and kind of wild to me!
Congrats dude !!!!!!!!!!!!!!!!!!
Thank you so much!
This is really quite awesome. And in the video, you never even mentioned the main cool thing about this - that you're operating way outside the spec of these tubes on these low voltages, and you got them to work so reliably that you can demonstrably build a functioning processor out of them. That is quite a cool feat!
Thank you so much!
And I totally forgot to mention that the entire thing is running at super low voltages. I've been working with them at just 24V for so long now, a part of my brain has just normalized it. I often forget that these little 6AU6s are usually meant to have four times as much B+ voltage!
VFD blinkenlights might be coolest interface ever
It's more like art than a processor - and I meant this in a positive way, of course :)
Thank you!
And it absolutely is more of an art project than a processor!
Congratulations! Can not wait to see what's next :)
Thank you so much!
It came together way better than I could have ever hoped for! I'm really excited about the next parts as well, program control in particular!
Amazing !!!
Congratulations.
dude you are super cool. keep up the good work !
Thank you so much!
A beautiful machine and a smart guy, a century late
been a wild ride but we're here! gratz
Thank you so much coming along with me!
There's still plenty more to come!
@@UsagiElectric where you are now compared to where you started prior to this being built out: do you like this specific processor's design? Do you see things you'd rather do differently (add/remove) to make it "better"? Would you ever contemplate designing a processor from scratch?
@@colinstu That's an excellent question! And truthfully, I've learned so much about processor design in the process of this, if I were to start again, I would do a totally different design. I've actually be toying around with some 4-bit and 8-bit processor designs from scratch that I quite like, and the 4-bit one in particular I think could be made in tubes and result in something not that much physically larger than what I will ultimately have here. Here's a simplified block diagram for the ultra simplified 4-bit design: i.postimg.cc/BQbJK7W1/UE4-S-Simp.png
When I finish the 1-bit computer, I may actually try to tackle the 4-bit design with tubes.
The 8-bit design I'm quite happy with too, and while I don't think I have the resources to build it out of tubes, I do think it would make an excellent discrete transistor processor. Here's the 8-bit design, and while it's not anywhere near as capable as most "modern" 8-bit processors, it shares a lot of its instruction set with the AMD AM2901 bit-slice processor. So, it's fairly capable as far as ALU functionality, and when combined with the 256 byte RAM, it should be capable of some pretty interesting stuff.
i.postimg.cc/zBvxjhqp/UE8K.png
@@UsagiElectric excellent writeup! Great to see the gears already turning on future designs, look forward to seeing the next developments.
Loved the video, dude. You look like James Franco's brother who's really into computers. And for that, I thank you.
Very cool Hobby...Not much practical purpose to reinventing the wheel, but really awesome that you have such a zeal , and passion for it , and it is very watchable and entertaining. Only several Million more like it , and you can build a computer the size of a 5 story building with the computing power of a 1979 TRS 80.Dont get me wrong , I am not hating or being synical..I actually think it's cool and enjoy watching it .
Congrats man ! I love to se a project like this giving its first results :D
Thank you so much!
It's really wild to see a project that you've been working on for nearly a year breath into life!
Yaaaay! Just like Tommy Flowers did. 80 years ago .... ☺️
the Thumbnail looks like a city in the night.
really cool
My Goodness its a thing of beauty !!!!! Can you do a video on your board manufacture process ? thats one serious milling machine !!!!! how did you convert it to CNC ? what program are you using to design those beautiful boards ? Thank you for sharing your fantastic work !!!!!! look forward to each of your videos !!!!!!
Thank you so much!
I definitely plan to do a video in the future to fully outline the entire process of making a PCB from idea to finished product!
To answer your questions though, it's a Bridgeport EZTrak, which I believe was sold by Bridgeport as a CNC machine. We got it used from a college that was shutting down for... not cheap, but not too expensive either. I design all my PCBs in DesignSparkPCB, but really, any of the modern EDA software should work. I have set a custom design rule (grid spacing of 2.54mm and 1mm trace widths) just to make it a little more forgiving for unevenness or lack of precision on the oversized mill. Once I have the board designed, I export it to gerber, then load it up in FlatCAM to convert it to G Code that can be read by the mill. Then, it's just getting the machine setup and cutting!
@@UsagiElectric Thank You for taking the time to respond !!!! I just learning CAD and I have a small cnc router im learning to use. unlike your bridgport its just got the stones to do PCB and not much more !!!! Man the things you could make with that Bridgeprt are amazing and in steel !!!! but it sure does a nice job on those PCB's Happy New Year to you and yours !!!!!
Amazing, the remote is a nice touch. Maybe a "overlay" PCB that masks the VFD dots and adds a label would be a good addition, however it would obstruct the view of the beautiful tubes.
It seems like a great thing for christmas, it lights up nicely and heats your room.
I'm really looking forward to where your journey brings you in the next year.
Until then,
Michael
Thank you so much!
The harness and remote (as much as it pains me to say it) are really just temporary pieces to demonstrate the processor itself. Once I get the other portions of the processor built, I'll have to rebuild the harnesses that connect them, as well as way to switch between external control and program control for executing programs.
Ultimately, the I/O board should have some pretty cool VFDs on it though showing us human readable output!
Absolutely wonderful. Fantastic work ! :-)
I see all this soldering without using flux and it triggers the tech in me lol. Using flux would make the solder "flow" SO much easier and be less "sticky" and make more beautiful joints. Awesome build though! Far above my understanding!
Awesome!! That’s amazing!
Man this is so impressive ! ❤️
Thank you!
btw, I am looking at how everything connected, do you keep in mind that those header pins are only rated for 2-3A max (per pin continues load) as well as max working voltage of 250V max.
Running relays in parallel is actually not very good idea. If one fails, you certainly will blow up the other one.
Some can also have balancing issues. Even more so when one switches on (or off) a fraction of a second earlier etc.
For the heaters I would just use a low drop constant current source, to make things easy.
But anyway, looks amazing, great job! 👍
Thank you so much!
Thanks for the heads up on the relays, I never thought about that being an issue. Looks like a soft start V3 may be in my future, haha. Fortunately, I'm actually just under what a single relay can handle as far as current (the G2R1-E I'm using is rated to 16A and I'm spiking to 15A), so I totally could have gotten away with just using a single relay, but I think in the future I'll upgrade to some 30A capable DPDT relays.
As for the header pins, they're really the only part of the build I'm not a huge fan of. Fortunately, I'm running everything at really low voltages for tubes. The entire thing is run on just +24V and -12V, so there's no high voltage anywhere. To give the header pins as much of a fighting chance as possible, I have two large power buses that run top to bottom on both the left and right sides. These busses are double wide, using two pins each, so if the pin can handle 3A of current, that's 6A per side and 12A total, which is pretty much exactly what the computer is running at steady state, so I'm really pushing the limits here. The male pins are fine as they're just a nice strong piece of metal and can handle quite a lot of current. It's the female connectors I'm unsatisfied with. Fortunately, I built the design of the buses with the female headers on the removable connection links, so if I come up with a better solution for connecting the male headers to each other, I can replace the connection links with ease.
Now, that's just for the vertical power buses on the sides. The boards themselves have power buses that run left to right, ultimately getting their power from the vertical buses. The good news on those is that they share the load from both left and right, and the heaviest load on a single horizontal bus is 20 tubes, which is 1.5A at 24V. That shouldn't overload the individual board to board connections.
So, that's a really long winded way of saying I'm pretty confident in the horizontal board to board connections, it's just the large vertical power buses on the left and right I'm a little concerned about. So far it seems alright, but I'll keep letting my subconscious chew on better ways to handle that vertical connection!
@@UsagiElectric Thanks for the clarification.
I forgot you were running those tubes on low voltages.
A standard practice is to just double up on the header pins. (so use two pins instead of one)
Which also doubles up the current capabilities. By standard the male and female part should be rated the same.
@@p_mouse8676 Yeah, I often forget that these tubes usually have four times as much B+ voltage as I'm running here!
I know the male and female parts should be rated the same, but the female part construction is not nearly as robust as I would have liked, but then again, the connectors were super cheap, so it shouldn't be all that surprising. If I do come across some more robust female connectors, I'll totally remake the power bus connectors, which should help with voltage drops from top to bottom a bit!
HAPPY NEW YEAR !
Thank you! Happy New Year to you and yours as well!
Excellent work! Really appreciated the walkthrough of this amazing project.
Thank you so much!
And thank you for tagging along on the build!
those new relays look SO cool
Really Cool, loved it!
Thank you very much!
Fantastic! Once you're completely done with the entire computer, you should calculate its MIPS per watt. For ENIAC it was 0.00289 MIPS @ 140,000 watts. 😁
You might take a dark room shot of it powered on in its upended position. That would look very cool.
Thank you!
Oh man, MIPS per watt is a terrifying prospect! If I get my goal of 10Hz, that would be 10-instructions per second, which is 0.00001 MIPS. Each tube pulls 1.9W in heater power, and currently I'm at 190 tubes, but I think by the time I finish, I'll be closer to 500 to 600 tubes, which should put power draw right around 1,000 watts, which isn't all that bad comparatively!
Incredible work.
Wow! You did it. For the next step, maybe you can cheat a bit and use an EEPROM as a program store and clock out the input sequences you did by hand? That would allow a “real” program to run on the tube computer without too much more effort. Your workmanship and design is remarkable. All the best for the New Year!
Thank you so much!
I'm trying to avoid using any modern components anywhere on the machine, even if only for temporary use. I even feel bad when I use an LED to check something right quick, haha.
But, using an EEPROM carries with it an additional problem of needing some voltage level shifting. Even though I'm running the tubes at incredibly low voltage, 24V is still a bit too high for most modern-ish stuff, so I would need to go through and build a custom level shifter to get the EEPROM output levels up to the right values without damaging the EEPROM itself.
Also, without any memory on the board, the instructions have to move at a glacial place so I can "read" the output and "store" it in my brain-case. Once we get memory though, we can start running programs a little quicker and then just focus on reading the final output.
So, it might still be a little while before we run a full program on it, but we're slowly getting there!
amazing achievement !
Thank you so much!