A few nights ago, [Chris Fenton] was hanging out at NYC Resistor putting in some time on his electromechanical computer project. You might remember [Chris] from his tiny Cray that he’s putting an OS on. It seems [Chris] is going back in time about 150 years and has set his sights on a 3D printed version of [Babbage]’s Analytical Engine.
The Analytical Engine
was is a remarkable feat of engineering and machining. It was the first programmable computer. Shame, then, that it was never built in the 1800s. [Chris] isn’t building a glorified calculator like [Babbage]’s polynomial-computing Difference Engine – he’s going all out and building something with conditional looping.
[Chris] calls his device an electromechanical computer, so we’re assuming it won’t be crank driven like the version in the British Science Museum. Right now, he’s constructed the decade-counting gears that are vitally important for the ALU of his design. All the parts were printed on a Thingomatic, so we’re betting [Chris] is going to be relying heavily on the MakerBot automated build platform for the thousands of parts he’ll have to fabricate.
Check out the video from NYC Resistor after the break.
23 thoughts on “3D Printed Electromechanical Computer”
Maybe I’m missing something, but nowhere in the video do I hear mention of him planning to build either of Babbage’s machines. He just aays he’s building an “electromechanical computer”.
Feel free to point out where it mentions otherwise.
Whoa – I’m definitely not building an analytical engine! My current design is based on lots of stepper motors and mechanical switches. I’ve been trying to teach myself some mechanical engineering skills, and really wanted to build something with tons of moving parts, so this seemed like a fun way to do it. The thing I’m fiddling with in the video is a 3-digit base-10 register (when appropriately connected to a stepper motor and some zero-detect sensors).
I actually just read the description on this – no idea how HaD inferred all of that from the video (although I’m flattered they think I’m that good at mechanical engineering! I think this is the first thing I’ve built with gears that hasn’t torn itself apart).
What book did you find this design in?
I’d be interested to see if 3d printed plastic components would have the right combination of precision and strength to actually be able to make an Analytical Engine. It certainly should be much cheaper simply in terms of material cost.
I think it might be a worthwhile endeavor to perhaps build part of the ALU just to see if you can get it to work as a sort of proof of concept. Coming at it from an electrical analogue standpoint might give you a leg up. This site: fourmilab.ch/babbage/applet.html has a good simulation as well as some definitive articles that might help you should you be interested.
Also if you ever have some questions on the mechanical side of things feel free to send me an email. userjjbATgmailDOTcom
Another example of hackaday’s wonderfully accurate reporting…
“I’d be interested to see if 3d printed plastic components would have the right combination of precision and strength to actually be able to make an Analytical Engine.”
I don’t think so- pretty sure that in the past someone tried to build a replica out of Meccano which didn’t work because there was too much flex in the (steel) parts.
The reason “Because I can” in my book has to be the best responce to the question of why your trying to do something (:
To build an entire computer from printed parts not only going to take a while to print but also to design.
Can’t wait to see it working (:
exactly the same thing I thought.
the guy’s got talent, let’s see what he comes up to.
if this machine is Turing complete he could simulate a more modern computer and play doom at one frame per decade.
I have a hard time understanding how a set of gears can do math. Can someone explain or provide a link about how it works?
Prepare to waste your life: fire control computer.
Just finished watching that series of videos. I have so many ideas. Just using that video in a trig or calculus class to teach ‘car jocks’ about how integrals and differentials can be represented in things like integrator and some combination of differentials (combinations for finding something like sin(x)dx); instead of just hammering on “you’ll need to know this later” It’s easy for some geeks to grok it without holding a set of gears, but a printable set of teaching aids for each major principle… off to work.
Totally agree with you on that one Quin, ‘Tis a much needed teaching resource!
Awesome! I can’t wait to see what derivatives and improvements get made on the design.
I can’t wait to see it do derivatives, or integrals, or Taylor series!
and sorry mods, hit report on Jim’s comment when I went for reply.
In our computer history class the professor mentioned that babbage’s analytical engine would be impossible as the frictional heat would melt some of the parts, if the machine as a whole could actually last that long whithout falling apart. The calculations assumed that bronze gears were used. He didn’t give us any reference tough. I wonder if it’s true, as I could not find it…
Based on nothing than a combination of common sense and engineering guesswork I think your professor is talking out of his ass.
First, it’s not like it NEEDS to operate at some speed. You could run it arbitrarily slow and have no worry of melting anything.
Second, things would break well before they melted. Thermal expansion, softening of components, etc.
Third, the amount of frictional force necessary to be dissipated to MELT a metal components would be prodigious. The whole things is so intimately connected that it just wouldn’t be possible to produce enough frictional heat to overcome conduction to surrounding parts. Hell at melting temperatures the natural convection and radiation would even be significant.
tl;dr Slap your professor in the face, give him a thermodynamics textbook, and the back of an envelope.
LOL, yes, apparently his professor is dumb as dumptruck.
He’s not a physicist and I was pretty sceptical as he didn’t give any reference but he is otherwise well informed so I just couldn’t believe he’d be lying – but yeah you’re right
Then again a lubricant can be used to both reduce friction, and cool the components. The problem of heat generation, and problems related to heat generation, aren’t insurmountable problems, even back in the day.
That looks exactly like a scaled up mechanical odometer to me. If you pull one apart, that’s pretty much how they are arranged with 20 teeth per drum and a partially toothed gear engaging a layshaft.
Did Babbage really invent that? I think it would have been pretty time consuming to fabricate those in metal as the missing teeth would have to be carefully milled off after the full set on the other side had been hobbed.
The required specialized machining was one of the downfalls of Babbage’s engines. In order to machine the necessary parts he and his machinist Clement invented several new techniques/tools to do so.
However an old artisanal law dictated that any tools made in the course of a job were automatically property of the artisan (which up until the industrial age was reasonable because the most complex thing an artisan might use was a simple lathe etc).
So when Babbage pissed off the already uneasy relationship with Clement, many of the tools were unavailable as well.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)