Automata: "Robots" designed hundreds of years ago [video]
youtube.com
youtube.com
I've seen the Jaquet-Droz automata[1] in Neuchatel on the one day a month they run them. They're demoed by a watchmaker who understands and maintains them.
The three automata are the Musician, the Artist, and the Writer. These were made between 1764 and 1778. The Musician and the Artist are just playing back pre-recorded motions from a set of cams. To increase the length of the recording up, there's a stack of cams, and after one turn, the stack moves vertically to play the next cams. So there are two clockwork trains taking turns - playout, and cam selection. It's a beautiful piece of work, especially when you realize someone made all those cams by hand, with a file.
The Writer, which writes text with a quill, is programmable. There's the stack of cams that move vertically to switch cams, as with the others. But with the Writer, the cam selection is programmable. There's a programming wheel made of little screw-on sections of different heights, and a supply of cam sections which indicate what letter to print next. It's an encoding with at least 26 different levels, probably more. I'm not sure if letter case is encoded on the main cam.
It's all very compact, fitting inside the bodies of the dolls. There's no huge mechanical box hidden away somewhere. Even today it would be tough to make that mechanism work, although there are still watchmaking companies that could do it.
Better video of the Writer.[2] You can see the cam stack and the programming wheel working.
I don't think he is trying to explain how they work, he's just trying to give an overview of the topic, with particular focus on how they influenced the society around them, and I think he did a good job.
This logic is derived from the way better striking clocks work. They have a stepped cam called a "snail", which is the program for how many times to strike vs the time of day.[2] It's just a table from 1 to 12. But there's no reason it couldn't be an arbitrary function.
It's beautifully made, but not mysterious.
vaucanson (and others) had previously devised a digitally controlled loom, though it didn't come into general use until jacquard's later improvements, and the pascaline and leibniz's stepped reckoner were controlled by digitally encoded numbers, but jaquet-droz's work may be the first machine controlled by a digital character encoding
even if it was just a plotter rather than a theorem prover or something
I wonder if one of the various YouTubers with a CNC mill might be able to build a replica --- that would be a video (series?) I'd certainly watch. The tolerances involved would be extremely difficult 250 years ago and probably involved much manual trial-and-error, but are likely not out of the ordinary for a modern machine shop.
The complexity isn't in the manufacturing of the gears. It was extremely difficult back then, easy now.
The challenge is in the design. Even with today's advanced CAD software I use and access to CNC machines, etc. It would be quite the project for me to design one of these. A lot of it is that much of the real in depth complicated mechanical engineering isn't super common anymore, much of it has been replaced with electronics and computers. I've made a good niche for myself kind of being the old school mechanical engineer who knows how to design those old school type mechanisms, but with modern CAD and design automation software tools.
My dad's a master jeweler, I've always wanted to make a fancy Orrey with him.
Having it do one thing after the other (by default) seems challenging but not impossible. What would a for loop look like?
I cant begin to imagine how to design such a device but if you provided less complicated building blocks it seems something one could learn pretty fast.
Well, this documentary answered that question for me: there used to be other applications! And if you imagine how you would recreate some of these automata yourself, you'll work out why we don't use these techniques today: it's not because we are under-utilising a valuable craft; it's because we have software.
It's not that we can't find other applications for tiny gears and springs, it's that we have better options. The only reason to create things out of gears and springs is if you don't have software and stepper motors. But now we do, and we don't even realise how great it is!
It used to be that if you wanted to do something complicated, you had to painstakingly make it all out of bespoke gears and springs, because there was literally no other way, but nowadays we use simple generic components for the mechanical parts, and we put the bespoke parts in software, and we can get so much further with so much less effort.
There will always be a place for great craftsmanship with tiny mechanical systems, the same way there will always be a place for bushcraft and there will always be a place for retrocomputing. But that the mainstream has moved on from these things is not a step back, it's for a very good reason.
none of these use simple generic components for the mechanical parts
it's true that there used to be more things built that way, like naval fire control computers, pid controllers, curta calculators, pascalines, cash registers, and teletypes; and watch gears and springs are smaller than most of these, because it was worth paying the huge premium for portability
i do agree that with newer technologies you often get much further more cheaply. for pid controllers or communications even analog electronics are much faster and cheaper. plastic snap-fit parts replace a lot of latches that used springs. hydraulics are often better for power transmission, gearbox applications, and especially static mechanical advantage than shafts, cables, chains, and belts. tiny electric motors replaced line shafts with thin copper wires. digital readouts with glass, magnetic, or capacitive scales are much better than handwheel graduations or dial indicator. nc and cnc lathes replaced cam-driven screw machines. fluidic analog amplifiers and digital logic gates have been totally replaced with smaller and faster electronics. cmos is faster, cheaper, and lower power than ttl. and yeah, now that we have 3¢ microcontrollers and subthreshold microcontrollers, almost anything that can be simulated with a computer program will probably cheaper and easier to debug that way
fuck steppers tho
And the same way they will always be a place for amazing works of art (which this sort of craftsmanship absolutely qualifies as). Some people will always create crazy incredible things purely for the personal joy of having done it, and bonus points if it brings joy to others as well. ;~)
Also, these are going to survive EMP very well.
I've made a little niche for myself as a mechanical and mechatronics design engineer who has one foot in the old world and knows how to design these types of complex old school mechanisms but with modern CAD and design automation tools. Much of the old world mechanical engineering stuff has been lost in the last few decades and digging it up and finding uses for it in my career is a passion of mine.
No designs, of course, but people’ve evidently thought about autonomous machines for a long time.
Of course there’s a living river in the same book. So.
Discourse on the Method[0] - René Descartes - 1637
A precursor to the Turing Test[1]
0. https://en.wikipedia.org/wiki/Discourse_on_the_Method
1. https://en.wikipedia.org/wiki/Turing_test#Philosophical_back...
The comparison between cams and vinyl records was very interesting, and it was all bordering on and some fully implementing programmability.