Don't worry, I can't be the only computer nerd/car guy who likes gate level digital design in your audience. This is awesome, and I am only two minutes in.
check out Rory Mangles TIM 8 relay computer. he did it with a minimal number of relays. He implemented a serial ALU. might give you some ideas on how to put together your mechanical computer
Ngl Rinoa, your username got me curious about your channel. Your pfp made me think it might have been related to gaming, as you well know I was incorrect lol (mostly). You come off as a relatively smart person in your videos, so I decided to subscribe. Have a nice day!
I think horsepower will just increase the "transistor count limit". To increase frequency, you would need a gearbox to drive the "clock" at a higher RPM.
No, It is not possible for two reasons (even if his idea is beautiful): 1) The more number of mechanical pieces in use, the more likely to crash. We already have lots of crashes, not want more of them. 2) The latency and timing that sucks!!! Even the electrons and light emissions have unacceptable delays in some complicated circuits.
Not even 3 rules - all computation can be reduced to a single rule: NAND. You can build all gates using NANDs, and you can build FFs using NANDs too. It's feasible to build an entire Turing complete computer using NANDs only.
Yea, a bit like Babbage's Analytical engine - except this will be binary so some of the 'mill' parts will be simpler although the analytical engine would have been able to calculate with 20+ decimal place accuracy. The 'store' part is still going to be massive though. :-)
@@dylanzrim3635 It pains me to explain the joke, but I'm pretty sure the point was "a modern ECU, implemented as a mechanical computer instead of an electrical computer"
9:03 although its worth noting that nasa has actually been considering giving it a look for venus-based rovers, due to the fact that electronic circuits dont hold up well at all under the venusian atmosphere, whereas mechanical devices should hold up far better and remain working for longer. we may actually see mechanical computers make a come back, at least for venus missions
while that might be vaguely true, the amount of processing that has to be done on an modern rover is large enough that rover might have to be much bigger than anything else built before.
@@psd993 true, and the idea of mechanical circuits being used is still only a consideration for now, rather than something that's been decided on. If we can make electrical circuits work, then that's obviously a better choice, I mean there's a reason we moved away from mechanical processing for the most part. It's just that NASA is at least deeming it a worthwhile consideration incase electrical circuits aren't a viable option. Doesn't hurt to have a backup plan after all.
@@revampedharpy09 perhaps mechanical sensors and instrumentation outside of a thermally insulated electronic processing unit will be the compromise they need.
You would still need to convert radio to mechanical energy, and you would need to make it REALLY compact. It would be better if they used the heat from the atmosphere to power a powerful cpu cooler.
It may not be efficient to a electrical CPU, but it gives way to looking at a CPU's in a different light. I think that is valuable. Being able to see things from different angles, always proceeds with progression.
This is awesome!! I think the final design will be excellent for classrooms in order to teach kids how logic gates work, especially if it can be made in such a way that the gates snap together and be 3D printed.
@lucaraza5613 There are tons of explanations, Ben eater, Sebastian lague, etc., you probably just didn’t find any of them until now. In fact, I think Sebastian Lague’s is better, because you can follow along for free (you can follow along with Ben Eater’s but most of the time he just explains the gates and uses chips because real life is limiting. Also, real objects are expensive, so you also need to purchase a kit to follow alongs)
1) Yes, it is possible to build a mechanical logic processor. 2) The frictional losses in a mechanical logic processor would make the entire thing melt into a pile of slag if it ran at anything close to a usable speed. 3) If you're really committed to doing this, you should use pneumatic gates or hydraulic gates instead of purely mechanical gates.
I don't think a "usable" speed is required for this kind of project. It's not really going to accomplish any significant work, just getting it working is an amazing feat
@@sixhundredandfive7123 Heating, cooling, and different lubrication issues are very different. There is no "air" to carry away any heat. Parts exposed to extremes of temperature.
Main problem i see is friction. With enough gates in series you won't be able to move the inputs, since a single input will move the entire machine simultaneously. What i think would be better is to have a mechanical clock input. This input would be what drives the outputs so that all the inputs do is decide whether the "clock" makes the output move or not. For example the AND gate: set A -> reset B -> pull clock -> nothing changes set A -> set B-> pull clock -> output changes I'm bad at explaining my ideas but i hope you get what i'm saying. The only drawback with the mechanical clock input that i see is that you would have so sequence it. For example with your current design as soon as one input is changed the entire machine moves and the result is near instant, but with this mechanical "clock" you would have to strobe each section step by step, sort of like a wave traveling through the gates. But then if you connect each section to the clock but with an inverter in between each section your clock it should be enough to sequence your clock in a "tick-tock" kind of way. And thinking about it you only need two clocks i total, one inverted, the other normal and just connect every second gate to one and the rest to the other clock. Anyways do what you do, will be interesting to see what you come up with.
Back when I was in Highschool I had a similar idea for a clockwork computer, using the direction of spin as bits. I drew up the plans for the gates and a potential processor, unfortunately, 3D printers didn't exist quite yet back then.
Sometimes the "inefficient" way is objectively better suited to a particular task. For example reading in a settings file "inefficiently" allows for much greater flexibility which ultimately ends up saving time, but it still takes more clock cycles to complete.
I was about to disagree with that statement, mainly as a knee-jerk reaction, because efficiency in design inherently has a beauty on its own. It may not produce rhythmic sounds, or have good-looking patterns, or even look interesting to the naked eye, but a well-designed integrated circuit is beautiful, as well as a plane that flies well, or any other simple and elegant solution to a problem. May be my inner German speaking or something, but there is an undeniable beauty in efficiency, a great satisfaction in seeing the best possible solution to a problem... Why then I do not disagree at all? Because there is one human activity strongly associated with beauty that is, speaking on a purely utilitarian term, an utter waste of resources: That activity is PRODUCING ART! Art is probably the opposite of efficiency. It has zero utilitarian value, and still, it has existed longer than engineering... or probably art and engineering have coexisted forever, because the beauty of engineering comes from its likeness to art: Rhythm, for example, is something humans really like. The basis of music is rhythm, and a pleasant and beautiful thing machines tend to do is producing rhythmic sounds... In rhythm we can find the beauty of a well crafted machine, because an engine with a problem will not produce rhythmic sounds. Electrical equipment on the other hand produces monotonous hums and is not pleasant to listen to. Nobody finds beauty on the sound of electricity. I consider the best testament to human ingenuity the ability to take the beauty of efficiency and the beauty of art, and combine them into one single design or machine. As you can imagine, I really, really like watches, and clocks, and engines, and trains, and railways, and propeller engines...
I would actually disagree with electronic hums and droning being disliked. That is after all the basis of the Hurdy Gurdy instrument, a thousand years ago it was an instrument in nearly every church and was used everywhere because many societies at the time found beauty and perfection in an uninterrupted melody or note, and this produced what is basically a a wheeled violin that can be cranked to create droning near violin noises and the notes played on the strings are played by keys on a box over the fretboard. Electrical hums can harken back to that view on perfection and can even be seen in modern EDM music with droning undertones that act like bass and texture to the notes on top of it. Nevermind the fact that people play music with pure electricity with Tesla coils so even the sharp unpleasant noise of electricity can make music as well.
@@ComotoseOnAnime I absolutely love the sound of a hurdy gurdy - it also has some low-pitch clicking noises as the fret keys are operated. Very rich cello sort of sound, with a continuous drone in the background. Beautiful.
By the way, there is a practical application for this. On the surface of Venus, electronics don't work, everything needs to be mechanical. NASA is/was running a contest for a venus rover that uses no electronics and communicates by positioning pointers to be read from orbit with a camera.
That is interesting to hear about. I think micro-scale selective compliance mechanisms is probably the best way to go about it, and with metals they could withstand VERY high temperatures. www.nasa.gov/feature/jpl/nasas-venus-rover-challenge-winners-announced
its probably going to have to be light enough to be shot in to space, unless it only needs to be 4 bit, he almost has a finished product who the heck would want to go to Venus anyway
more info initial contest announcement: www.nasa.gov/feature/jpl/nasa-wants-your-help-designing-a-venus-rover-concept winners announced: www.nasa.gov/feature/jpl/nasas-venus-rover-challenge-winners-announced Scott Manley video: th-cam.com/video/1DEvcJgBy0c/w-d-xo.html
@@fluffmiceter1846 active refrigeration, vaccum tubes _may_ work, but I wouldn't want to run them through the vibration of launch and landing, let alone the thermal stress of code of deep space followed rapidly by the heat of reentry and the surface of venus (and there is a good reason we don't use them for computers any longer, their reliability is poor) And since we don't manufacture large numbers of tubes any more, setting up a production line for them would not necessarily be any cheaper than developing mechanical computers, and the result may not be any smaller.
@@WesleyKagan Have you seen @Foone on twitter? Their test, for both salvaged retro computers and hacked up modern embedded devices, is "will it run Doom?".
@@WesleyKagan bruh i build a mc turing machine and steve mould uploaded water computer basically same day. its been a few weeks and i sucessfully built 8 bit by 8 bit long multiplication in mc. i think computers in diffrent mediums might be a meme on the rise
As a recent comp sci grad, I found this fascinating. We had to build a simulated four-bit computer from basic logic gates, and we had to use Solidworks to model mechanical things, but we didn't put them both together! LOL! Also, nice (and very fitting) xkcd reference.
Guy is basically starting over with computing and going through the same minimization issues that original larger computer such as tube based computers ran into. I'm sure there were quite a few mechanical computing concepts back in the day. I mean the fact that certain fancy typewriters had caps lock, key memory, and more - its arguable this was all done before. BUT!!! I think its cool to start over like this as just cause it was done before by people in the past doesnt mean every permutation of possible solutions was experimented with, leaving some more interesting inventions on the table perhaps. We too often find what works through trial and error, move on, and rarely come back to that part of the problem unless the solution we found has a low efficiency or durability. Here I think computers have very much moved on from larger mechanical components however I do still think certain insights can be gathered by revisiting this whole field. Glad to see people are still experimenting and thinking about this stuff.
As an electrical engineer and a person who loves creative DIY projects I really like this video. I did want to point out that building gates for computers and such, they're actually designed as NAND gates and NOR gates as the simplest form. This has to do with the logic of how transistors work whether in series (for NANDs) or in parallel (for NORs). To create either an AND gate or an OR gate requires adding a NOT gate (which is just another transistor). Which is kind of counter intuitive when you're first introduced to gate logic in an intro to digital logic both because you're taught AND and OR gates first and because the symbols seem to indicate that you're building from AND and OR gates. At any rate, I'll have to check back in and see if there any updates on this computer project. It'll be interesting to see what kind of optimizations you can make to your gates and see if it actually works.
interesting thing I can already see is that the gates aren't exactly binary - since you can push them the gate inputs to different degrees. You've created some weird binary-analog abomination and that's pretty cool.
As one of my EE professors used to say, 'All circuits are analog if you try hard enough.' Electronic circuits have the same issue - a typical 5V logic chip will always see an input under 0.8V as a '0' and an input over 2.0V as an '1'. This leaves a range (0.8V - 2.0V) where the behavior is 'undefined'. (It may be off, on, or even weakly on or weakly off) Some amount of effort is put into avoiding this undefined range.
Technically yes, but the energy needed to drive it likely would render it useless. It's much easier to move a few electrons for a few nanometers than a few metal rods for a few millimeters, I think.
@@OnFight1997 Computation is basically just wasting a whole bunch of energy to gain the tiny amount of information entropy. I mean, you wouldn't want your computer to just yeet away while doing stuffs so... Everything would turn into heat eventually. And this kind of mechanical logic gate is huge (a few million nanometers vs a few nanometers in process level), which means you need loads more energy to drive it in order to achieve the same computation power.
@@FlameRat_YehLon when he said "cool" he probably meant "amazing to look at while performing some calculation", I guess he wasn't referring to temperature.
I don't know really how or why I got this channel recommended to me, but I am sure glad I saw this video! I think this is a real neat project and I can't wait for more videos!
I was enjoying the vid up until that point. To me it seemed a shallow pretentious statement, and completely misunderstands what art is and why it exists. Fortunately that really doesn't matter in a vid about making mechanical computers :) It's a truism to say that art is in nature - after all everything is in nature - nothing profound there. The idea that technology somehow reveals 'where' art is is just nonsense - completely meaningless... unless you are shortsighted and need optical technology, or its nightime and need artificial illumination, or you are lost on the way to the museum, and need some cartographical tech... but those are just practical details. As far as the binary logic - that's cool. It would be neat if there was a simple mechanical implementation for a NAND gate, because that is all you need - every other type of logic can be build from NANDs
@@oldmossystone I completely disagree with your blanket, and quite frankly derogative, statement. A statement that is just as shallow as you claim that the original statement is. Because: There would be no flute without the technology to drill holes in bones or forming wood. Not to mention metalworking, but that is much later. Body paint works as sunscreen, and extracting the paint and mixing it is also technology. That statue that always was inside that block of marble? It required many skilled hands to turn the marble rock into a block, and even more technology to create the tools used in all of the processes from extracting a slag of stone to polishing the final statue. Making a strict divisions between art, science and technology is dangerous. Art, including stories, have always inspired scientists and engineers to create new things, while new technology and scientific breakthroughs made new kinds of art possible. The links between technology and science most definitely goes both ways as well.
@@57thorns The implication in your original statement is that art is hidden and needs technology to reveal it. If this were true, all art would require technology to reveal it. This would exclude e.g. singing from art. QED. While various forms of technology can be exploited in the creation and dissemination of art, there is no essential need for technology in the pursuit of art. There is no need for a 'division' in order to make a distinction between technology, art and science either. Division would be bad, distinction is necessary in order to understand what each of them is, and to help us understand their unique value in our culture(s). Returning to your example of the Block of marble. The statue was not always inside the block. The statue is a result - a side effect - of an expression of the artists state of mind during that creative process. The important point here is that if there was no stone, the expression could still have occurred in some other way - the Block of marble is not the essence of the art. The statue itself isn't even the essence of the art - otherwise any statue would be art (clearly not the case, many are just ornaments with no greater value). The statue or song or painting is a conduit - a medium through which to communicate. A book is a useful analogy here. A book is a physical thing with pages and words. But the book is not the story, and the words are not the story either the story is an abstract construction of ideas that the reader decodes by reading the book. A statue is like a book, it is a medium through which an artist has encoded some abstract meaning. If the statue is art, it is because it contains some deeper meaning encoded by the artist well enough that it can be decoded by the 'consumer'. In the same way a story doesn't have to be in a book, it could be spoken word, or a movie, the essential meaning encoded in a statue could potentially have been encoded in a painting, or some other art 'form'. Not all available encoding forms require technology.
@@oldmossystone And yet, you can't accurately retell a story over time without a book (or other media that conveys words, including, but not limited to wooden staves, loose sheets of paper, movies, computers, clay or stone tablets).
Awesome video! Once in a while, I daydream about alternate implementations of logic gates. One fun one would be water-driven gates next to a waterfall.
I love that you are trying this. I had one thought though. The problem with doing this mechanically is that unless each piece requires little or no force, as it adds up you will have to use springs, grease or put the whole thing in a vat of oil to keep the force levels down. Kind of in a similar way where they made the first computers, they had to deal with current losses as more chips are added. This was done by having each chip be self powered and specifying what the maximum current draw could be for outputs and minimum for inputs. So I think that might be a challenge that you would need to overcome. Mechanically I think grease/ oil / springs might be a good way to help, but im a computer engineer not a strong mechanist.
Well this is getting interesting, I work on turbine engine fuel controls for a living, essentially a hydromechanical computer, although much more analog than binary. Its impressive what can be achieved without the use of electricity, these take into account the throttle position, barometric pressure, temperature, and current engine rpm along with a few other parameters to meter the correct amount of fuel at any given time. And all designed in the era of slide rules no less! A digital system would likely be better, but they still do the job. I look forward to seeing where you will take this.
My friend used to work on pneumatic logic for tube trains doors and aircraft wings or something like that. Again, no electricity, but I do thing it was binary decision logic stuff
@@WesleyKagan Check out the vacuum logic setup of an 80's era Honda and imagine having to troubleshoot it: www.autoweek.com/car-life/but-wait-theres-more/a1860746/run-away-screaming-1985-honda-cvcc-vacuum-hose-routing-diagram/
Great video! I was surprised you didn't mention hydraulic computers. Lots of older automatic transmissions are, very basically speaking, just sets of logic gates that are controlled using hydraulic fluid and vacuum.
Have alwasy thaught about this. Compliant designs seem a realy good way of doing this as are far more simple. Would also allow for extreem minaturisation as could be manufactured using steriolithography similar to a silicon chip.
Fun fact is that one of my professors basically worked on a design kinda like that on the silicon level. There was a poster outside the door when I went to go have her give me one more submission time cause I submitted the wrong assignment. It used a relay design and the price bends with a compliant design on it if I remember correctly.
Ya beat me to it :) I was going to suggest the same thing for the same reason. Honestly I'm surprised micromechanical computers aren't a niche market already. Seems like they'd be useful in nuclear disasters, to go into high radiation areas where electronics don't work.
This is the same mechanical systems that were used on mechanical train signalling systems. These were popular in the 1940's onwards. They are still used on some of the older london underground lines. They are called interlocking machine rooms
I subscribed because of this video. I did something similar when I was young using relays but it never occurred to me to try to make it use no electricity at all. This is a very cool project and I'm really excited to see it get done.
11:34 "Beauty in engineering doesn't necessarily *have* to be efficiency" couldn't agree more. That's why vintage technology still capture the hearts of many. Things like hourglasses, sundials, wind and water mills, those things are still aesthetically enjoyable despite being inefficient technology today.
Looks like a fun adventure! I imagine the resistances of the system are going to stack up pretty quickly though. Electronic circuits can be built with gain, to increase the strength of a weak signal. Can you achieve the same with your mechanical system? Where does the energy come from to do so?
@@geemy9675 a hydraulic computer would be the easiest kind to build that I can think of, you could use air, water, mercury whatever lol most electrical circuits could be copied into a hydraulic circuit anyway. Would be slow but also would be cool, my optimized guess for core speed is like 100 hertz, possibly 1 Mhz lol but I think it'd be really hard to get that fluid to move any quicker and still retain accuracy in the system
Matt Parker built a calculator out of domino pieces. But it can only do one calculation. I guess the pieces hold potential energy which is good. Other people built calculators with "marble gates". You could let marbles roll down and then lift them up on several places to add energy again. Are marbles too slow? You could power each mechanical gate and only let the connected gates direct the energy instead of providing it on their own. That's how real electronic computers work, don't they? Maybe push some rod upwards and another rod can connect it to another gate or disconnect it, like a clutch.
This is a "CMOS NAND Gate": en.wikipedia.org/wiki/CMOS#/media/File:CMOS_NAND.svg It connects a high voltage / push to the output, if at least one input pulls and it connects a low voltage / pull to the output if both inputs push. So, in the end it "not-pushes" exactly if input a AND input b push, like a NAND gate should. (IDK if that helps anybody, sorry. I was pretty intrigued by that difference between the flow of the 1s and 0s and the flow of the voltage in computers when I first learned about it.)
"Charles Babbage. Paging Mr. Charles Babbage! Please come to the white courtesy phone. Somebody on TH-cam who obviously didn't take computer science has never heard of the Analytical Engine."
I can't wait to see were this project goes. I really wanted someone to build a functioning mechanical computer that was programable. Thank you for giving me a view on what that is like, and hope to see more.
Can you imagine Babbage, as the world's first computer geek, coming up with these things all on his own? It's no wonder he never finished the analytical machine, kept thinking up better ways to do the calculations.
Whether it's the sound of a mechanical computer tapping away, or the pony express delivering letters, both will be loud. Pony express for delivering bits might be faster though.
@@WesleyKagan Speaking of the pony express, cars and planes carrying hard drives and/or sd cards is still faster by total download time than anything else. And it may always be, those things keep storing more and more things in less space, even while Moore's Law slows down in other respects. (And hey, while I've got you, I highly recommend building out the full common logic gate set, not just using the minimal complete set (e.g. the AND, OR, and NOT gates you mention, or you could've just used NORs and nothing else, like Minecraft). That's all well and good for Minecraft (I made one, it was fun), but having XOR, NAND, etc, will let you have fewer components, lower friction, etc., all while keeping to the spirit of the project. May also want to make custom Muxers, Flip Flops, etc? E.g. google the wikipedia article "List of 7400-series integrated circuits". The low-complexity stuff on that list is probably good inspiration for what you might want to consider making specialized modules for.)
Looking forward to seeing where you go with this. If you want some inspiration, the Zuse Z1 is a general purpose mechanical computer, and lots of Zuse's notebooks are available online, although sometimes it's more fun to figure it out from first principles yourself.
11:18 In my opinion, art is anything that somebody keeps for reasons beyond practicality. That's exactly what your computer is, and we're better off for it.
I think a good way to make a NOT Gate could be to put a simple gear in the middle, that grabs both pistons on opposite sides, so when the INPUT piston is moved inwards, as a 1, so is the OUTPUT piston, as a 0, and vice-versa.
Subscribed because this promises to be fascinating. Three immediate thoughts: friction, tolerance stacking, hysteresis. They're all factors that are going to be working against each other, and you. Whilst you said "I don't need to de-bounce the clock signal from a 555" my gut feel is that you are going to need some kind of mechanical buffer to sit between every few modules to ensure you keep a clean 'mechanical signal path'. Which is pretty much a mechanical analogue to debouncing an electrical signal. Now if I had any idea about what such a buffer would look like, I'd say...but I've no clue :)
First time running into your channel. I know nothing about cars nor do I wish to, but computing intrests me and this video was really interesting, can't wait for follow-up!
These Jacquard looms are indeed programmable, however, they do not actually do calculations. It represents a mechanism to record a set of instructions and have those instructions performed reliably. But that must then be sent to a General Purpose Processor; a device that can perform calculations via an Arithmetic Logic Unit, and then, as directed, be able to save the results to referenceable locations in some form of referenceable storage. It's a similar situation with player-pianos, which inspired paper-tape memory: they recorded instructions, but the piano does not do any calculation, nor saving, nor retrieving of such calculations in order to execute further instructions.
@@trespire Oh hey, that's right, 1500s does indeed predate jacquard by quite a bit...In that case, that's rather interesting. I know that jacquard didn't really invent programmable looms; he just put together some ideas of others before, but I thought the ideas he borrowed from where fairly recent to him. I'm under the impression that the London Science museum has an impressive collection of looms that demonstrate this progression; I've always wanted to check it out. You have inspired me to do research! :D
@@verdatum It was a small loom, 400 to 500 mm across, was damaged and not in working order. It had very complex mechanisms all out of wood, forged iron and strings. I went back to look at it 3 times, still couldn't figure it all out. It was the most delapidated display in the museum, and yet impressed me the most.
Making a modern mechanical computer was actually an idea I was just thinking about recently. So you can imagine as soon as I saw this video I had to watch, I am happy to say I also definitely wasn't disappointed. Great content and I hope to see more!
In real life, Mechanical Computers are better off as analog rather than digital because analog machines doesn't take as much space as digital machines.
This is totally awesome! I know about Konrad Zuse's mechanical computer Z1, but I never seen befor explained this basic components. Thank you very much, it really helped me to understand it!
THANK YOU for this. In a sci-fi setting I roleplay in, I have a where a race of aliens get into space before they've mastered electronics. However, they are absolute masters of clockwork and so their ships computers are completely mechanical. And this showed me how this could work.
I've built a calculator in Minecraft and just seeing this title made me worried about the size of computer this could result in. Definitely fascinating though, and I wouldn't mind having some mechanical logic gates as decorations.
man this is quite the amazing video. I havent gone through you channel but i would really be interested in hearing more about this project, and your ideas about computers in general. great content.
Dude if you don’t have electricity check if you forgot to pay a bill first, rather than just build a mechanical computer straight away. But on a serious note: brilliant!
Buddy, I just subbed specifically for this project and the fact that you shouted out Ben Eater, right before citing Douglas Adams. This is an awesome project.
Hey man ! Just a tip that might be useful; I call this the "If it floats, it's where the mics goes", which basically means whatever you're wearing, you're mic goes on the most steady part of the cloth that's the most going to move. Like, your shirt's neck not the t shirt in that case lol. Anyway great video dude, subscribed !
I know, I know. Not a car video. Next week the Jaguar is getting some upgrades!
Don't worry, I can't be the only computer nerd/car guy who likes gate level digital design in your audience. This is awesome, and I am only two minutes in.
Turing was an assclown and now the government is using 5g chips to turn us all gay.
check out Rory Mangles TIM 8 relay computer. he did it with a minimal number of relays. He implemented a serial ALU. might give you some ideas on how to put together your mechanical computer
@@Cracked1ce serial architecture is a great idea. It's much lighter on gate counts. It's also slower, but this wasn't going to be a speed demon.
@@jefffrasca4054 Me too! I love the nerd-vids.
When modern society breaks down and electricity is no more, you'll be the only one playing DOOM on your mechanical computer
but does it run crysis?
@@ExplizitDuester Estimated to run at least 60FPW
@@Skinflaps_Meatslapper Does "FPW" mean "Frame per Week"?
@@gweltazlemartret6760 If you prefer FPS that would come out to about 0.00009920634FPS
If modern society breaks down, I assume the rest of us will be playing DOOM for real.
Ok this was quite neat. Just fyi this appeared on my YT home screen and i've never seen you before, so you have good reach.
Same
also same
Same - also, subscribed immediately both for cars and mechanical computers...
Ngl Rinoa, your username got me curious about your channel. Your pfp made me think it might have been related to gaming, as you well know I was incorrect lol (mostly). You come off as a relatively smart person in your videos, so I decided to subscribe. Have a nice day!
Same
The speed of the computer is suddenly shifted from GHz to Horsepower. Turbo boosting the CPU takes on it's original meaning!
It would be fun to have him build an actual windmill to power the CPU he makes. Or maybe to use mirrors and a small homemade stirling engine instead.
@@SapioiT I'm thinking more like 5.7 Hemi.
I need to upgrade my PC's drive shaft to play call of duty.
I think horsepower will just increase the "transistor count limit". To increase frequency, you would need a gearbox to drive the "clock" at a higher RPM.
@@odw32 You need horsepower for both. Torque and RPM.
"Is a mechanical computer possible?"
Alan Turning would like a word.
As would Babbage.
@@grimftl I was about to say Babbage.
Konrad Zuse as well
Jacquard wants to say a-hon!-hon!-hon!
No, It is not possible for two reasons (even if his idea is beautiful):
1) The more number of mechanical pieces in use, the more likely to crash. We already have lots of crashes, not want more of them.
2) The latency and timing that sucks!!! Even the electrons and light emissions have unacceptable delays in some complicated circuits.
Not even 3 rules - all computation can be reduced to a single rule: NAND. You can build all gates using NANDs, and you can build FFs using NANDs too. It's feasible to build an entire Turing complete computer using NANDs only.
NOR is universal too
@@prototypeinheritance515 Yes but NAND is more widely used.
I remember this is how most is manufactured
And nands can be has been built of stone(chrystal in nature) man mad structures duh!!!
@antonchannel And it's pretty cool now that i tried it.
Fascinating thought experiment.
Thank you! I'm still tinkering with the 2nd part of this, but I'm happy with the direction so far. Love your channel!
@@WesleyKagan looking forward to part 2!
Yea, a bit like Babbage's Analytical engine - except this will be binary so some of the 'mill' parts will be simpler although the analytical engine would have been able to calculate with 20+ decimal place accuracy.
The 'store' part is still going to be massive though. :-)
Destin is probably the reason I'm here.
@@eliljeho yup, algorithm is doin its job.
Coming soon: solid state mechanical engine management.
I very much hope that’s sarcasm, mechanical fuel injection is a thing and it is currently a step up from modern ECU tech
@@dylanzrim3635 It pains me to explain the joke, but I'm pretty sure the point was "a modern ECU, implemented as a mechanical computer instead of an electrical computer"
@@dylanzrim3635 Yeesh, humorless. Nothing about "engine management" implies fuel injection. And yes, MFI has been around since at least WW2.
Engine management via a difference engine.....
@@dylanzrim3635 step down*
interent: "i wanna put freevalve on my miata already"
wesley: "i wonder if i can do differential equations with switches"
Charles Babbage: "I beat you to it by more than a century"
Big brain reply
@@deusexaethera Allan Turing: "let's put input and output in the same place"
9:03 although its worth noting that nasa has actually been considering giving it a look for venus-based rovers, due to the fact that electronic circuits dont hold up well at all under the venusian atmosphere, whereas mechanical devices should hold up far better and remain working for longer. we may actually see mechanical computers make a come back, at least for venus missions
while that might be vaguely true, the amount of processing that has to be done on an modern rover is large enough that rover might have to be much bigger than anything else built before.
@@psd993 true, and the idea of mechanical circuits being used is still only a consideration for now, rather than something that's been decided on. If we can make electrical circuits work, then that's obviously a better choice, I mean there's a reason we moved away from mechanical processing for the most part. It's just that NASA is at least deeming it a worthwhile consideration incase electrical circuits aren't a viable option. Doesn't hurt to have a backup plan after all.
No, it's for a Mercury mission, not Venus.
@@revampedharpy09 perhaps mechanical sensors and instrumentation outside of a thermally insulated electronic processing unit will be the compromise they need.
You would still need to convert radio to mechanical energy, and you would need to make it REALLY compact. It would be better if they used the heat from the atmosphere to power a powerful cpu cooler.
It may not be efficient to a electrical CPU, but it gives way to looking at a CPU's in a different light. I think that is valuable. Being able to see things from different angles, always proceeds with progression.
I can't imagine how much friction will exist in the final system.
I was just thinking that.
*L U B E I T*
grease cooling
actual adding machines from the period would basically be BATHED in lubricating oil at least daily if not more frequently.
@@IONATVS Imagine having to change your computer's oil every other month 🤣
*There's a GPU shortage
Wesley: alright what do I have in the scrap pile I can put together.
@gfgffgfg fgf Graphics Processor Unit.
OMG - he can use it to mine for bitcoin.
@@NotSomeone68 yes. However, be careful saying GPU because you would typically need a whole graphics card rathar than just the GPU.
@@NotSomeone68 also, there’s a silicon/chip shortage in general, not just GPU/CPUs so this comment still technically makes sense.
@@ShyDigi Likewise with a motherboard and CPU, though.
That logic gate looks like it was designed to survuve a neuclear blast, Skookum asf
Skookum As frig is acceptable
"Skookum" - must be in BC, Canada. That's not a thing anywhere else. :D
AvE lol 😂
@@Watchmedome3017 yup here's here too.
Dewclaw will be mortified that no ones, zeroes or angry pixies will be required
This is awesome!! I think the final design will be excellent for classrooms in order to teach kids how logic gates work, especially if it can be made in such a way that the gates snap together and be 3D printed.
completely agree I'm in high school and this the best explanation on how computers work I've ever found
@lucaraza5613 There are tons of explanations, Ben eater, Sebastian lague, etc., you probably just didn’t find any of them until now. In fact, I think Sebastian Lague’s is better, because you can follow along for free (you can follow along with Ben Eater’s but most of the time he just explains the gates and uses chips because real life is limiting. Also, real objects are expensive, so you also need to purchase a kit to follow alongs)
I haven't heard the joke about the "spherical cow in a vacuum" in ages. Talk about specialist humor!
1) Yes, it is possible to build a mechanical logic processor.
2) The frictional losses in a mechanical logic processor would make the entire thing melt into a pile of slag if it ran at anything close to a usable speed.
3) If you're really committed to doing this, you should use pneumatic gates or hydraulic gates instead of purely mechanical gates.
but what if he wants purely mechanical gates?
I don't think a "usable" speed is required for this kind of project. It's not really going to accomplish any significant work, just getting it working is an amazing feat
I think that was called Fluidics.
What if this was working in the vacuum of space?
@@sixhundredandfive7123 Heating, cooling, and different lubrication issues are very different. There is no "air" to carry away any heat. Parts exposed to extremes of temperature.
Main problem i see is friction. With enough gates in series you won't be able to move the inputs, since a single input will move the entire machine simultaneously.
What i think would be better is to have a mechanical clock input. This input would be what drives the outputs so that all the inputs do is decide whether the "clock" makes the output move or not.
For example the AND gate:
set A -> reset B -> pull clock -> nothing changes
set A -> set B-> pull clock -> output changes
I'm bad at explaining my ideas but i hope you get what i'm saying. The only drawback with the mechanical clock input that i see is that you would have so sequence it.
For example with your current design as soon as one input is changed the entire machine moves and the result is near instant, but with this mechanical "clock" you would have to strobe each section step by step, sort of like a wave traveling through the gates.
But then if you connect each section to the clock but with an inverter in between each section your clock it should be enough to sequence your clock in a "tick-tock" kind of way.
And thinking about it you only need two clocks i total, one inverted, the other normal and just connect every second gate to one and the rest to the other clock.
Anyways do what you do, will be interesting to see what you come up with.
Damn, you never fail to exude sheer genius.
Truly
Man, pls continue this project, it is amazing
Back when I was in Highschool I had a similar idea for a clockwork computer, using the direction of spin as bits. I drew up the plans for the gates and a potential processor, unfortunately, 3D printers didn't exist quite yet back then.
"Beauty in engineering doesn't necessarily have to be efficiency" - Absolutely agree!!!!
Sometimes the "inefficient" way is objectively better suited to a particular task. For example reading in a settings file "inefficiently" allows for much greater flexibility which ultimately ends up saving time, but it still takes more clock cycles to complete.
I was about to disagree with that statement, mainly as a knee-jerk reaction, because efficiency in design inherently has a beauty on its own. It may not produce rhythmic sounds, or have good-looking patterns, or even look interesting to the naked eye, but a well-designed integrated circuit is beautiful, as well as a plane that flies well, or any other simple and elegant solution to a problem. May be my inner German speaking or something, but there is an undeniable beauty in efficiency, a great satisfaction in seeing the best possible solution to a problem...
Why then I do not disagree at all?
Because there is one human activity strongly associated with beauty that is, speaking on a purely utilitarian term, an utter waste of resources: That activity is PRODUCING ART!
Art is probably the opposite of efficiency. It has zero utilitarian value, and still, it has existed longer than engineering... or probably art and engineering have coexisted forever, because the beauty of engineering comes from its likeness to art:
Rhythm, for example, is something humans really like. The basis of music is rhythm, and a pleasant and beautiful thing machines tend to do is producing rhythmic sounds... In rhythm we can find the beauty of a well crafted machine, because an engine with a problem will not produce rhythmic sounds. Electrical equipment on the other hand produces monotonous hums and is not pleasant to listen to. Nobody finds beauty on the sound of electricity.
I consider the best testament to human ingenuity the ability to take the beauty of efficiency and the beauty of art, and combine them into one single design or machine.
As you can imagine, I really, really like watches, and clocks, and engines, and trains, and railways, and propeller engines...
@@renard6012 Exactly! Although electrical waves seen on an oscilloscope are beautiful, so perhaps it's all about how you experience something?
I would actually disagree with electronic hums and droning being disliked. That is after all the basis of the Hurdy Gurdy instrument, a thousand years ago it was an instrument in nearly every church and was used everywhere because many societies at the time found beauty and perfection in an uninterrupted melody or note, and this produced what is basically a a wheeled violin that can be cranked to create droning near violin noises and the notes played on the strings are played by keys on a box over the fretboard.
Electrical hums can harken back to that view on perfection and can even be seen in modern EDM music with droning undertones that act like bass and texture to the notes on top of it.
Nevermind the fact that people play music with pure electricity with Tesla coils so even the sharp unpleasant noise of electricity can make music as well.
@@ComotoseOnAnime I absolutely love the sound of a hurdy gurdy - it also has some low-pitch clicking noises as the fret keys are operated. Very rich cello sort of sound, with a continuous drone in the background. Beautiful.
By the way, there is a practical application for this. On the surface of Venus, electronics don't work, everything needs to be mechanical. NASA is/was running a contest for a venus rover that uses no electronics and communicates by positioning pointers to be read from orbit with a camera.
That is interesting to hear about. I think micro-scale selective compliance mechanisms is probably the best way to go about it, and with metals they could withstand VERY high temperatures.
www.nasa.gov/feature/jpl/nasas-venus-rover-challenge-winners-announced
its probably going to have to be light enough to be shot in to space, unless it only needs to be 4 bit, he almost has a finished product
who the heck would want to go to Venus anyway
more info
initial contest announcement: www.nasa.gov/feature/jpl/nasa-wants-your-help-designing-a-venus-rover-concept
winners announced: www.nasa.gov/feature/jpl/nasas-venus-rover-challenge-winners-announced
Scott Manley video: th-cam.com/video/1DEvcJgBy0c/w-d-xo.html
Or maybe the proper question is, how can we get electronics working on the surface of Venus?
@@fluffmiceter1846 active refrigeration, vaccum tubes _may_ work, but I wouldn't want to run them through the vibration of launch and landing, let alone the thermal stress of code of deep space followed rapidly by the heat of reentry and the surface of venus (and there is a good reason we don't use them for computers any longer, their reliability is poor)
And since we don't manufacture large numbers of tubes any more, setting up a production line for them would not necessarily be any cheaper than developing mechanical computers, and the result may not be any smaller.
Isn’t this basically an irl redstone computer? Still awesome
Yep. I think the ultimate goal of any computer is; can it run crisis? And can it run Minecraft.
@@WesleyKagan Have you seen @Foone on twitter? Their test, for both salvaged retro computers and hacked up modern embedded devices, is "will it run Doom?".
Can we not physically do this with nano tech yet?
@@WesleyKagan I'm still not convinced Crysis wasn't some *really* elaborate HCF command with an emphasis on the F.
@@WesleyKagan bruh i build a mc turing machine and steve mould uploaded water computer basically same day. its been a few weeks and i sucessfully built 8 bit by 8 bit long multiplication in mc. i think computers in diffrent mediums might be a meme on the rise
As a recent comp sci grad, I found this fascinating. We had to build a simulated four-bit computer from basic logic gates, and we had to use Solidworks to model mechanical things, but we didn't put them both together! LOL! Also, nice (and very fitting) xkcd reference.
Guy is basically starting over with computing and going through the same minimization issues that original larger computer such as tube based computers ran into. I'm sure there were quite a few mechanical computing concepts back in the day. I mean the fact that certain fancy typewriters had caps lock, key memory, and more - its arguable this was all done before.
BUT!!!
I think its cool to start over like this as just cause it was done before by people in the past doesnt mean every permutation of possible solutions was experimented with, leaving some more interesting inventions on the table perhaps. We too often find what works through trial and error, move on, and rarely come back to that part of the problem unless the solution we found has a low efficiency or durability.
Here I think computers have very much moved on from larger mechanical components however I do still think certain insights can be gathered by revisiting this whole field. Glad to see people are still experimenting and thinking about this stuff.
Every other youtuber: I cant think of a video idea so I'll do a vlog.
Wesley: I cant think of a video idea so I'll reinvent the Babbage engine
"roll that intro.....I don't..I don't have an intro" is one hell of an intro
Came for the racecar, staying for the physics... yes yes yes, more content like this
There will be racecar videos soon too! Thanks!
Becauseracecar!
Here for race car stayed for physics
Yes, the same.
If the cow is in a strong vacuum and the vacuum container is spherical, then there would indeed be a spherical cow in a vacuum.
As an electrical engineer and a person who loves creative DIY projects I really like this video. I did want to point out that building gates for computers and such, they're actually designed as NAND gates and NOR gates as the simplest form. This has to do with the logic of how transistors work whether in series (for NANDs) or in parallel (for NORs). To create either an AND gate or an OR gate requires adding a NOT gate (which is just another transistor). Which is kind of counter intuitive when you're first introduced to gate logic in an intro to digital logic both because you're taught AND and OR gates first and because the symbols seem to indicate that you're building from AND and OR gates.
At any rate, I'll have to check back in and see if there any updates on this computer project. It'll be interesting to see what kind of optimizations you can make to your gates and see if it actually works.
I would love to see more of this project. Did you ever do anything more with this?
interesting thing I can already see is that the gates aren't exactly binary - since you can push them the gate inputs to different degrees. You've created some weird binary-analog abomination and that's pretty cool.
See fuzzy logic: en.wikipedia.org/wiki/Fuzzy_logic
As one of my EE professors used to say, 'All circuits are analog if you try hard enough.'
Electronic circuits have the same issue - a typical 5V logic chip will always see an input under 0.8V as a '0' and an input over 2.0V as an '1'. This leaves a range (0.8V - 2.0V) where the behavior is 'undefined'. (It may be off, on, or even weakly on or weakly off)
Some amount of effort is put into avoiding this undefined range.
@@cctmsp13 Detents.
As someone who looked into GTL systems I know the issue. At least in my use case I had some degree of control over the high and low thresholds.
@@cctmsp13 I once got a fortune cookie while on a business meal. It said "Even digital circuits are made of analog parts." It was like destiny.
Someone builds a CPU in minecraft
Wesley: So can we build a ryzen out of spare parts in the garage?
Technically yes, but the energy needed to drive it likely would render it useless. It's much easier to move a few electrons for a few nanometers than a few metal rods for a few millimeters, I think.
@@FlameRat_YehLon How cool would it be though.
@@FlameRat_YehLon OH REALLY!? DO YA THINK!?
@@OnFight1997 Computation is basically just wasting a whole bunch of energy to gain the tiny amount of information entropy. I mean, you wouldn't want your computer to just yeet away while doing stuffs so... Everything would turn into heat eventually.
And this kind of mechanical logic gate is huge (a few million nanometers vs a few nanometers in process level), which means you need loads more energy to drive it in order to achieve the same computation power.
@@FlameRat_YehLon when he said "cool" he probably meant "amazing to look at while performing some calculation", I guess he wasn't referring to temperature.
I don't know really how or why I got this channel recommended to me, but I am sure glad I saw this video! I think this is a real neat project and I can't wait for more videos!
ive wanted this for so long. thanks so much for working on this
Would love to see a video of your take on neural nets, would be fascinating to alot of us caveman mechanical engineers
I’ll put one together, it’s a super cool part of computer science
@@WesleyKagan I rang the bell just to see you explain NN
"Art itself is in nature, and technology just reveals where it's hidden." That's a very profound statement.
11011
I was enjoying the vid up until that point. To me it seemed a shallow pretentious statement, and completely misunderstands what art is and why it exists. Fortunately that really doesn't matter in a vid about making mechanical computers :)
It's a truism to say that art is in nature - after all everything is in nature - nothing profound there. The idea that technology somehow reveals 'where' art is is just nonsense - completely meaningless... unless you are shortsighted and need optical technology, or its nightime and need artificial illumination, or you are lost on the way to the museum, and need some cartographical tech... but those are just practical details.
As far as the binary logic - that's cool. It would be neat if there was a simple mechanical implementation for a NAND gate, because that is all you need - every other type of logic can be build from NANDs
@@oldmossystone I completely disagree with your blanket, and quite frankly derogative, statement. A statement that is just as shallow as you claim that the original statement is.
Because:
There would be no flute without the technology to drill holes in bones or forming wood. Not to mention metalworking, but that is much later.
Body paint works as sunscreen, and extracting the paint and mixing it is also technology.
That statue that always was inside that block of marble?
It required many skilled hands to turn the marble rock into a block, and even more technology to create the tools used in all of the processes from extracting a slag of stone to polishing the final statue.
Making a strict divisions between art, science and technology is dangerous. Art, including stories, have always inspired scientists and engineers to create new things, while new technology and scientific breakthroughs made new kinds of art possible. The links between technology and science most definitely goes both ways as well.
@@57thorns The implication in your original statement is that art is hidden and needs technology to reveal it. If this were true, all art would require technology to reveal it. This would exclude e.g. singing from art. QED.
While various forms of technology can be exploited in the creation and dissemination of art, there is no essential need for technology in the pursuit of art.
There is no need for a 'division' in order to make a distinction between technology, art and science either. Division would be bad, distinction is necessary in order to understand what each of them is, and to help us understand their unique value in our culture(s).
Returning to your example of the Block of marble. The statue was not always inside the block. The statue is a result - a side effect - of an expression of the artists state of mind during that creative process. The important point here is that if there was no stone, the expression could still have occurred in some other way - the Block of marble is not the essence of the art. The statue itself isn't even the essence of the art - otherwise any statue would be art (clearly not the case, many are just ornaments with no greater value). The statue or song or painting is a conduit - a medium through which to communicate.
A book is a useful analogy here. A book is a physical thing with pages and words. But the book is not the story, and the words are not the story either the story is an abstract construction of ideas that the reader decodes by reading the book.
A statue is like a book, it is a medium through which an artist has encoded some abstract meaning. If the statue is art, it is because it contains some deeper meaning encoded by the artist well enough that it can be decoded by the 'consumer'.
In the same way a story doesn't have to be in a book, it could be spoken word, or a movie, the essential meaning encoded in a statue could potentially have been encoded in a painting, or some other art 'form'. Not all available encoding forms require technology.
@@oldmossystone And yet, you can't accurately retell a story over time without a book (or other media that conveys words, including, but not limited to wooden staves, loose sheets of paper, movies, computers, clay or stone tablets).
I appreciate the fact that you don't have an intro.
SAME BRO
The Bruce Lee intro - My style is no-style.
Awesome video! Once in a while, I daydream about alternate implementations of logic gates. One fun one would be water-driven gates next to a waterfall.
I love that you are trying this. I had one thought though. The problem with doing this mechanically is that unless each piece requires little or no force, as it adds up you will have to use springs, grease or put the whole thing in a vat of oil to keep the force levels down. Kind of in a similar way where they made the first computers, they had to deal with current losses as more chips are added. This was done by having each chip be self powered and specifying what the maximum current draw could be for outputs and minimum for inputs. So I think that might be a challenge that you would need to overcome. Mechanically I think grease/ oil / springs might be a good way to help, but im a computer engineer not a strong mechanist.
I’m here for the nerdery. Whether it’s car nerding, or mechanical engineering, or whatever.
Nvidia head quarters right now: He built it in a cave, with a box of scraps!
Underrated comment lolol
Hahahaha
Omfg 😭😂💀
Yes, but we aren't Wesley Kagan
@@francescoesposito9462 Lol
I'm a CS student and a petrol head so I think I just found my new role model
oh god. you gave me a challenge *to build the first gaming computer made of mechanical parts
Hey dude! Just wanted to say I think you’re doing a great job here! Spherical cow in a vacuum! LOL
Thanks man! Love what you’re doing too, we need to chat again soon!
Well this is getting interesting, I work on turbine engine fuel controls for a living, essentially a hydromechanical computer, although much more analog than binary. Its impressive what can be achieved without the use of electricity, these take into account the throttle position, barometric pressure, temperature, and current engine rpm along with a few other parameters to meter the correct amount of fuel at any given time. And all designed in the era of slide rules no less! A digital system would likely be better, but they still do the job. I look forward to seeing where you will take this.
My friend used to work on pneumatic logic for tube trains doors and aircraft wings or something like that. Again, no electricity, but I do thing it was binary decision logic stuff
That’s fascinating, I still am amazed at what engineers did before computers.
Do you watch AgentJZ?
@@WesleyKagan Check out the vacuum logic setup of an 80's era Honda and imagine having to troubleshoot it: www.autoweek.com/car-life/but-wait-theres-more/a1860746/run-away-screaming-1985-honda-cvcc-vacuum-hose-routing-diagram/
@@Default78334 is that why my '87 CRX never passed Smog?
Great video! I was surprised you didn't mention hydraulic computers. Lots of older automatic transmissions are, very basically speaking, just sets of logic gates that are controlled using hydraulic fluid and vacuum.
Have alwasy thaught about this. Compliant designs seem a realy good way of doing this as are far more simple. Would also allow for extreem minaturisation as could be manufactured using steriolithography similar to a silicon chip.
Fun fact is that one of my professors basically worked on a design kinda like that on the silicon level. There was a poster outside the door when I went to go have her give me one more submission time cause I submitted the wrong assignment. It used a relay design and the price bends with a compliant design on it if I remember correctly.
It sure seems like these things need to be 3d printed to ease the mamoth task of building a functioning computer.
Indeed, but this scale will have a certain pleasing clackety aesthetic quality to it.
Ya beat me to it :) I was going to suggest the same thing for the same reason. Honestly I'm surprised micromechanical computers aren't a niche market already. Seems like they'd be useful in nuclear disasters, to go into high radiation areas where electronics don't work.
This is the same mechanical systems that were used on mechanical train signalling systems. These were popular in the 1940's onwards. They are still used on some of the older london underground lines. They are called interlocking machine rooms
I randomly ran into your page and I am super stoked about watching your videos. Thanks a bunch!
Oh no they made Minecraft CPUs a real thing
I think shader packs will really bog it down though.
Here is a real world lightsaber :-)
th-cam.com/video/xC6J4T_hUKg/w-d-xo.html
6:59
ah yes, the exclusive AND gate : only turns on when the laws of physics are disobeyed and all logic is broken
My favorite
Hey, that looks like a really neat project!
Also looking forward to seeing more of the FreeValve-Miata. Keep up the good wörk!
Freevalve miata!
I subscribed because of this video. I did something similar when I was young using relays but it never occurred to me to try to make it use no electricity at all. This is a very cool project and I'm really excited to see it get done.
11:34 "Beauty in engineering doesn't necessarily *have* to be efficiency" couldn't agree more. That's why vintage technology still capture the hearts of many. Things like hourglasses, sundials, wind and water mills, those things are still aesthetically enjoyable despite being inefficient technology today.
"so we're going to assume there's a spherical cow in a vacuum"..
This cracked me up
Looks like a fun adventure! I imagine the resistances of the system are going to stack up pretty quickly though. Electronic circuits can be built with gain, to increase the strength of a weak signal. Can you achieve the same with your mechanical system? Where does the energy come from to do so?
the only solution I see is hydraulics but hen why not just make 100% hydraulics logical gates. propagation would be slower though
@@geemy9675 a hydraulic computer would be the easiest kind to build that I can think of, you could use air, water, mercury whatever lol most electrical circuits could be copied into a hydraulic circuit anyway. Would be slow but also would be cool, my optimized guess for core speed is like 100 hertz, possibly 1 Mhz lol but I think it'd be really hard to get that fluid to move any quicker and still retain accuracy in the system
Matt Parker built a calculator out of domino pieces. But it can only do one calculation. I guess the pieces hold potential energy which is good.
Other people built calculators with "marble gates". You could let marbles roll down and then lift them up on several places to add energy again.
Are marbles too slow?
You could power each mechanical gate and only let the connected gates direct the energy instead of providing it on their own. That's how real electronic computers work, don't they? Maybe push some rod upwards and another rod can connect it to another gate or disconnect it, like a clutch.
This is a "CMOS NAND Gate": en.wikipedia.org/wiki/CMOS#/media/File:CMOS_NAND.svg
It connects a high voltage / push to the output, if at least one input pulls and it connects a low voltage / pull to the output if both inputs push. So, in the end it "not-pushes" exactly if input a AND input b push, like a NAND gate should. (IDK if that helps anybody, sorry. I was pretty intrigued by that difference between the flow of the 1s and 0s and the flow of the voltage in computers when I first learned about it.)
@@jomanout5866 I would use water and put it next to a dam.
Damn, you my friend have definitly an underrated channel.
"Charles Babbage. Paging Mr. Charles Babbage! Please come to the white courtesy phone. Somebody on TH-cam who obviously didn't take computer science has never heard of the Analytical Engine."
I'm in Computer Organization currently for my Software engineering degree currently and this video/concept is amazing. Keep up the amazing work!
Careful about a spring return, you might introduce too much impedance.
My man is an inventor proper!!!!!
Same thing was done with water on another Channel a while ago. "Intel wet inside" or something. Same thing though.
@@MaNNeRz91 It was done on Steve Mould's channel with help from Matt Parker (stand-up maths).
@@chitlitlah that's the one 👌
"I'm in a little uncharted territory here"
You and me both buddy.
I can't wait to see were this project goes. I really wanted someone to build a functioning mechanical computer that was programable. Thank you for giving me a view on what that is like, and hope to see more.
Can you imagine Babbage, as the world's first computer geek, coming up with these things all on his own? It's no wonder he never finished the analytical machine, kept thinking up better ways to do the calculations.
I guess that's why Microtech named a whole city after him. "New Babbage"
“Hey, how long do fresh eggs keep after you put ‘em in the fridge?”
“Uh~, here-let me look it up.”
*GOD-AWFUL CACOPHONOUS CLACKING*
Whether it's the sound of a mechanical computer tapping away, or the pony express delivering letters, both will be loud. Pony express for delivering bits might be faster though.
@@WesleyKagan Speaking of the pony express, cars and planes carrying hard drives and/or sd cards is still faster by total download time than anything else. And it may always be, those things keep storing more and more things in less space, even while Moore's Law slows down in other respects. (And hey, while I've got you, I highly recommend building out the full common logic gate set, not just using the minimal complete set (e.g. the AND, OR, and NOT gates you mention, or you could've just used NORs and nothing else, like Minecraft). That's all well and good for Minecraft (I made one, it was fun), but having XOR, NAND, etc, will let you have fewer components, lower friction, etc., all while keeping to the spirit of the project. May also want to make custom Muxers, Flip Flops, etc? E.g. google the wikipedia article "List of 7400-series integrated circuits". The low-complexity stuff on that list is probably good inspiration for what you might want to consider making specialized modules for.)
Why would you put eggs in the fridge??
@@WolfgangDoW Don’t get cocky with me just ‘cause you don’t know nothin’ fancy.
So I should be wearing gloves and PPE watching this video as well, yes?
Absolutely
Also applies to styropyro
And a safety mask :))
Looking forward to seeing where you go with this. If you want some inspiration, the Zuse Z1 is a general purpose mechanical computer, and lots of Zuse's notebooks are available online, although sometimes it's more fun to figure it out from first principles yourself.
Yes, Zuse built some cool mechanical computers.
11:18 In my opinion, art is anything that somebody keeps for reasons beyond practicality. That's exactly what your computer is, and we're better off for it.
Man I swear you and me are so much on the same wavelength it's scary. Keep it up I love your videos. All of them, every topic.
"A diary to my madness" is a frickin' mood if I've ever heard one.
the NAND gate alone provides a COMPLETE logical system, btw.
Don't even need to go that far. You can use a 3-state inverter with pull to get it all!
(or more easily with just a NOR gate which is effectively the same thing but with the pull resistors and buffering added in)
@@Fingerblasterstudios NOR is obviously also a complete logical system
I think a good way to make a NOT Gate could be to put a simple gear in the middle, that grabs both pistons on opposite sides, so when the INPUT piston is moved inwards, as a 1, so is the OUTPUT piston, as a 0, and vice-versa.
I've got this in my recommendation, man I wish that there were more videos you make like this.
Subscribed because this promises to be fascinating.
Three immediate thoughts: friction, tolerance stacking, hysteresis. They're all factors that are going to be working against each other, and you.
Whilst you said "I don't need to de-bounce the clock signal from a 555" my gut feel is that you are going to need some kind of mechanical buffer to sit between every few modules to ensure you keep a clean 'mechanical signal path'. Which is pretty much a mechanical analogue to debouncing an electrical signal. Now if I had any idea about what such a buffer would look like, I'd say...but I've no clue :)
"I dont have an intro" lol, immediately had to press the like button
Too bad I can not double like this comment.
I instantly hit the subscribe button when he said: "rolled that intro, wait i don't have an intro"
LMFAO
Make a never ending mechanical while- loop and you have free energy!
Just don't let ElectroBOOM see this comment 🤣
"And like the Meryl Streep in the Devil wears Prada".. you have carried this video wonderfully informative. Great video. Keep them coming.
First time running into your channel. I know nothing about cars nor do I wish to, but computing intrests me and this video was really interesting, can't wait for follow-up!
"Beauty in engineering doesn't necessarily have to be efficiency" - Well said.
So instead of talking about a 3 GHz processor being on the low end, we'll start talking about how a 3 Hz processor is a middle of the road one...
I've seen a fully automated, programable loom built in the 1500's on display in the Technology Museum in Milano.
These Jacquard looms are indeed programmable, however, they do not actually do calculations. It represents a mechanism to record a set of instructions and have those instructions performed reliably. But that must then be sent to a General Purpose Processor; a device that can perform calculations via an Arithmetic Logic Unit, and then, as directed, be able to save the results to referenceable locations in some form of referenceable storage.
It's a similar situation with player-pianos, which inspired paper-tape memory: they recorded instructions, but the piano does not do any calculation, nor saving, nor retrieving of such calculations in order to execute further instructions.
@@verdatum I believe what I saw predated Jacquard looms by at least 150 years. It was built somewhere in Northern Italy in the middle ages.
@@trespire Oh hey, that's right, 1500s does indeed predate jacquard by quite a bit...In that case, that's rather interesting. I know that jacquard didn't really invent programmable looms; he just put together some ideas of others before, but I thought the ideas he borrowed from where fairly recent to him. I'm under the impression that the London Science museum has an impressive collection of looms that demonstrate this progression; I've always wanted to check it out.
You have inspired me to do research! :D
@@verdatum It was a small loom, 400 to 500 mm across, was damaged and not in working order. It had very complex mechanisms all out of wood, forged iron and strings. I went back to look at it 3 times, still couldn't figure it all out. It was the most delapidated display in the museum, and yet impressed me the most.
Making a modern mechanical computer was actually an idea I was just thinking about recently. So you can imagine as soon as I saw this video I had to watch, I am happy to say I also definitely wasn't disappointed. Great content and I hope to see more!
Probably the most interesting channel on the ‘Tube at the moment. Thanks Wesley, keep it up.
I now know how Homer Simpson feels when Lisa explains something to him....
In real life, Mechanical Computers are better off as analog rather than digital because analog machines doesn't take as much space as digital machines.
“Playing doom on a mechanical cpu”
This is totally awesome! I know about Konrad Zuse's mechanical computer Z1, but I never seen befor explained this basic components. Thank you very much, it really helped me to understand it!
THANK YOU for this. In a sci-fi setting I roleplay in, I have a where a race of aliens get into space before they've mastered electronics. However, they are absolute masters of clockwork and so their ships computers are completely mechanical. And this showed me how this could work.
You were reading about Babbage's Analytical Engine recently weren't you?
I've built a calculator in Minecraft and just seeing this title made me worried about the size of computer this could result in. Definitely fascinating though, and I wouldn't mind having some mechanical logic gates as decorations.
"the function of art is to hold a mirror up to nature"
you can consider the mirror technology
DNA would not work if left handed.
Awesome! I was going to mention XOR gates, but you covered that! Well done!
man this is quite the amazing video. I havent gone through you channel but i would really be interested in hearing more about this project, and your ideas about computers in general. great content.
How is nobody talking about the "exclusive AND-gate" at 6:59?
Exactly. It feels to me that an AND already is exclusively AND. There's no ambiguity about the inputs when the output is 1.
@@AquilaSornoAranion Yeah there is the XNOR gate which is sometimes (falsely) called an XAND.
"assume we have a spherical cow in a vacuum" HA
Dude if you don’t have electricity check if you forgot to pay a bill first, rather than just build a mechanical computer straight away. But on a serious note: brilliant!
Buddy, I just subbed specifically for this project and the fact that you shouted out Ben Eater, right before citing Douglas Adams. This is an awesome project.
Hey man ! Just a tip that might be useful; I call this the "If it floats, it's where the mics goes", which basically means whatever you're wearing, you're mic goes on the most steady part of the cloth that's the most going to move.
Like, your shirt's neck not the t shirt in that case lol.
Anyway great video dude, subscribed !