Hacker News new | past | comments | ask | show | jobs | submit login
Five Hundred and Seven Mechanical Movements (507movements.com)
370 points by mrmaddog on Aug 17, 2013 | hide | past | favorite | 66 comments

In a similar vein, here is a video from the US Navy in the 1950s which describes how mechanical computers work (for calculating trajectories of missiles being shot from ships.)

Utterly fascinating to see how very simple mechanical devices, like differing gear ratios, can be used to calculate things like logarithms.


Fascinating, an analog computer. Reading about it on wikipedia, it seems analog computers have been almost completely forgotten in the digital age. It makes me wonder what kind of analog computer would be possible now with 21st century technology, especially because of discoveries in condensed matter.

For example, could an electrical analog system be a way of carrying out large-time molecular dynamics simulations? With proper tuning of the material, might you be able to create arbitrary potentials between the electrons within it so that they would interact roughly like atoms in an MD simulation? Or, perhaps you could create an analog chip where the electrons would 'naturally' solve various NP-hard problems, effectively by brute force?

Based on your comment, I did a little searching for the terms (single electron circuit) and found this interesting document. Maybe it's the kind of thing you're talking about:



For a completely different portrayal of futuristic mechanical computers, check out Neal Stephenson's story The Diamond Age. It's set in an era of nanotech, and the computers are nanoscale clockwork computers.

Actually, people are all the time now going the other direction (or proposing to): they use cold atoms to simulate setups from condensed matter. The general idea is that you can use lasers to site the atoms on a lattice, and then tune the interactions between the atoms to get all sorts of physics. A quick Google Scholar search turns up http://www.nature.com/nature/journal/v415/n6867/abs/415039a.... , wherein Greiner et al. simulate the Hubbard model (it would appear---I don't have access to the paper at the moment.

This is cool, because the Hubbard model is a simple and displays interesting phenomena, but understanding it is a hard problem. The Hamiltonian (that is to say, the energy) consists only of a kinetic energy plus an interaction between the spin up and spin down particles on the same site (e.g. if I have two spin up bosons and four spin down bosons, the interaction contribution to the energy is 8 U, where U is a constant parameter---the strength of the interaction.) Depending on this interaction strength U, the system might behave either like a conductor or an insulator.

The problem is hard to deal with analytically (for reasons I can't say I understand) and, as I understand it, the space of possible states is so huge that the numerics become computationally intractible at about a lattice 5 sites x 5 sites x 5 sites. So being able to see the phase transition happen is very neat, and exactly what you expect from an "analog quantum computer": simulating with cold atoms a system that we can't really simulate with ordinary computers.

You might also enjoy reading about this experiment: http://en.wikipedia.org/wiki/MONIAC_Computer

Reading about it on wikipedia, it seems analog computers have been almost completely forgotten in the digital age.

Musicians are still using them. Modular synthesizers are the bomb.

The concept of a really advanced mechanical/analogue computer is covered really well in "The Diamond Age".


> In a similar vein, here is a video from the US Navy in the 1950s which describes how mechanical computers work (for calculating trajectories of missiles being shot from ships.)

Guns, not missiles. At 0:50, the narrator refers to "initial shell velocity."

They were last used in combat in the 1990s, when the Iowa-class battleships fired their 16-inch guns at Iraqi positions in Kuwait. The Navy never bothered to replace them with digital computers, because they were already accurate enough for the guns they controlled.

That was a great video, thank you! Knowing basic gear mechanics helps a lot in understanding how cars function.

This 1937 video about differentials is also enlightening:


Why is it that old tutorial videos are so thorough and informative?

There was no Internet to fill in the details?

It was also a big deal to make a film like this. Today you can just prop up your phone on something and go to town, but these videos would have been a huge amount of work. Once you're putting in that much work, you had better make sure the result is good.

For Bay Area folk (or visitors), there's a restored mechanical targeting computer at the SF88 Nike Missile base in the Marin Headlands


Highly worth a visit if you're interested in this sort of mechanical computing stuff…

That was, as you said, utterly fascinating. It's amazing to see mathematical problems being solved with mechanical solutions. To think the leaps it took to translate these problems to digital solutions (e.g. what was [technically] an infinite set of inputs now recorded as a 32-bit floating point representation).

I don't know if any of these[1] ~50s videos are still shown in school but I found them amazingly efficient at conveying understanding.

[1] So far I saw : analog target computations, wheel differentials, wave diffusion

Another vintage series, about Fluid Mechanics by John Lumley (. Penn St Univ)


Would you mind sharing the link for the "analog target computations" video? I couldn't find it.


Sorry for the confusion, but by "analog target computations" I meant the videos about trajectories mentioned in the post above https://news.ycombinator.com/item?id=6231166

Okay, stupid questions. - To what extent are mechanical engineers using these sorts of things? - Are there texts discussing the design / selection / integration of movements to perform particular tasks? - Are there are texts like that targeting application to Arduino driven robotics? Mindstorms?

I'm a trained watchmaker. I use, fix and sometimes build many of these as components of watch movements.

[1] Is what I know as a Geneva Drive. Back before we had fancy alloy springs and were forced to use Steel as the material for mainsprings because that's all we knew, watches had problems where a freshly wound watch would run fast and a watch that hasn't been wound for a day or so would start to run slow, as the strength of the spring tapered off. The Geneva Drive was a solution, though it's more of a hack, to only let the spring release power inside the middle of it's power arc, by preventing the watch from unwinding past a certain low point and preventing the user from winding the spring up to it's strongest point.

[2] This is a simple Heart Cam. Mechanical Chronographs (Stopwatches) use these to reset the chronograph runner to zero. A hammer, represented as the horizontal pin, is released from a caught position with spring tension on it, which slams into the heart cam and forces it to reset to a predetermined position. Incredibly simple design. No matter where the hammer slams down on the cam, it is guaranteed to reset to the same place. (It's also a golden ratio)

[3] This is basically the design of a modern instantaneous date wheel. The snail-cam ("D" looking thing) is generally affixed loosely on the gear. The worm gear drives the flat gear, naturally, which catches the snail-cam and drives it forward. As the cam rotates, it slowly raises up the hammer, which has tension provided by the spring. At a defined point, the hammer reaches the apex of the snail-cam and (since the cam is affixed loosely and is allowed some circular freedom) slams forward "instantaneously". The cam will usually have a finger that then flicks forward a date wheel.

[4] Fusee Chain. Similarly to the reasons behind the Geneva cross (loss of power as a spring unwinds), we developed the Fusee chain to compensate for the loss of power by acting like a transmission in a car. As the spring unwinds, it uncoils the chain from the spiral wheel, which in turn will have a high-torque output at first (~ 1:1 ratio), and slowly increase revolution while decreasing power. These were highly present in English watches, but are no longer produced (as we have fancy alloys that alleviate all of the necessity of these things)

[5] A Verge Escapement. Another long-forgotten mechanism. Escapements are the things that regulate the output of circular motion. It's basically a ratchet-and-pawl mechanism. You'll generally have an oscillator (like a pendulum) attached to the lateral verge which rocks the teeth back and forth, the drive train then tries to move the escape-wheel forward, but the verge only allows one tooth to pass per vibration (a vibration is one half oscillation). The "Swiss Escapement" has largely replaced this mechanism. [6] This is essentially the same thing.

[7] The Cylinder Escapement. In the 60's the Swiss freaked out because the Japanese started producing cheap, disposable watch movements. The result of this was a huge loss of Swiss watch companies as they struggled to compete. One of the ideas they came up with was to produce a large number of cheap watches, but they couldn't just drop prices, they also had to retool and drop quality, substantially. The cylinder escapement was not a new invention in the 60's, but it started to get a lot of use around then. On a personal note, these are terrible things. They were designed to be disposable, and subsequently were marketed to kids quite often. This is where a number of the Ingersoll watches came from (like the old Mickey Mouse watches). The original ones that are worth more than most startups in Mountain View are cylinder watches, and finding one that runs is the equivalent of founding a Facebook for watch collectors. [8] This is how it works when looking at it interacting with it's escape wheel.

[9] The Swiss Lever Escapement. B is the pallet, D is the oscillator (Balance Wheel). Escapements essentially work all the same. They're mostly just renditions of one idea. The swiss lever escapement was up until the mid 2000's practically the only escapement produced in wristwatches. The pallet stones have since been upgraded to be synthetic ruby, along with the majority of the other bearing surfaces in watches, since the introduction of this book, however.

[10] This is what I'd call a Daniels Escapement, or Co-Axial Escapement. If you've looked at an Omega watch since the mid 2000's (as referenced above), you've probably heard about the Co-Axial Escapement. This is what it looks like. It's much more complicated than any of the other escapements, and has whole books written about it, and yet is the simplest mechanism once you figure out how it ticks. It quite literally only touches the oscillator once an oscillation. All the previous escapements are required to touch the oscillator twice (once going up, once coming back down). This is literally the difference between a tick and a tock. This only ticks, it has no tock. The less interaction the pallet has with the oscillator means that there is less energy lost in it's oscillation, and helps with both accuracy and longevity (they wear out less, and can run longer per wind).

[11] This is a rotary Wankel engine, like the Mazda RX. Not a watch part, but still a "modern device"

[1] http://507movements.com/mm_212.html

[2] http://507movements.com/mm_097.html

[3] http://507movements.com/mm_064.html

[4] http://507movements.com/mm_046.html

[5] http://507movements.com/mm_234.html

[6] http://507movements.com/mm_302.html

[7] http://507movements.com/mm_294.html

[8] http://507movements.com/mm_295.html

[9] http://507movements.com/mm_296.html

[10] http://507movements.com/mm_313.html

[11] http://507movements.com/mm_425.html

Now, this is a real hn comment. Coming out from the shadows to provide further information on a topic that most of us don't know much about at all!

I always wondered how it was possible to build the small pieces that go into mechanical movements with a 200 years old technology. I guess they needed precision lathes or other tools to build them. How much was a trade secret and how much was known by everybody? What would it take for someone today to replicate what was possible then?

Actually, did you ever think about writing a book on mechanical watches? There's a need for a modern book. Most of the material out there is from the 50's.

Funny thing about mechanical watches, the technology hasn't changed in 150 years. The technology around tooling has changed substantially, which in turn allows us to produce more accurate and cheaper watch components and movements, but we're still producing the same old watch design that we (Americans) stole from the Swiss and started mass-production of.

The book I used in school was "Theorie de l'horologie" [1], which is a fairly modern book. I supplemented this with "The Bulova Watch Repair Training Manual" [2] and "Practical Benchwork for Horologists" [3]. I would argue that the most useful book for practice was "Practical Benchwork", which was originally released in 1938, it's latest edition being from 1988.

[1] http://www.amazon.com/The-Theory-Horology-Charles-Andre-Reym...

[2] http://www.esslinger.com/the-joseph-bulova-school-of-watchma...

[3] http://www.amazon.com/Practical-Benchwork-Horologists-Louis-...

And it sounds like new materials have obviated some of the designs too.

Still, I'd read about how these machinese were developed and then made obsolete by other technology.

Great comment, thanks.

I imagine this is blocked from being appearing higher up on the page by the parent not having more votes than the parent's siblings.

(Do votes for a child have an effect on the parent's sort order? A workaround would be to vote the parent up when voting for the child, but you shouldn't have to do that...)

Sussman (MIT,SICP,SICM) did a class reversing a watch, very interesting.


Here is a demo of the Daniels escapement [10], http://www.youtube.com/watch?v=PVhSQ_Azkr8

Wonderfully elegant solution. Oh, and great post, uxp.

Good post. I became very interested in mechanical watches a few years ago.

Great answer. Thank you.

As a former mechanical engineer, it depends on your chosen course curriculum. Standard courses that all M.E.'s take generally don't cover topics like this. Instead, you deal at a more fundamental level with courses such as physics, calculus, statics and dynamics, heat transfer, thermodynamics, material sciences, etc. When other majors were allowed to take pretty much any elective, we were required to take 400 level engineering electives. This is where you had the option of choosing courses that would teach this sort of stuff. For example, I took a course in gear design which covered topics related to this. Another engineering elective I took was vehicle dynamics. Robotics courses are often an option for students as well.

As far as texts go, I would just search Google or Amazon for "Machine design". I don't have any particular recommendations because at the time I viewed text books with the same level of excitement as most other students, which is to say none at all. The two books that I did keep after college are the following: 1) Product Design and Development: http://www.amazon.com/Product-Design-Development-Karl-Ulrich... 2) Machine Design: An Integrated Approach: http://www.amazon.com/Machine-Design-Edition-Robert-Norton/d...

Thumbs up on the Ulrich textbook. I took a class with him at Penn, and it was fantastic. Highly recommended if you ever get the chance.

Not a mechanical engineer, but there are mechanisms books that explain modern mechanisms and their uses and analyze them from an engineering perspective.

I love the 507 movements book (I've had the paper copy for years) but an important thing to remember is that many of these mechanisms have been replaced by modern electronics, motors and software. Mechanisms are cool to watch, but they often have to be aligned when initially set up, they wear in complex ways, may have to be taken apart to be lubricated and can be very difficult to fabricate. In many cases a mechatronic solution (combo of electronic motion control and a simpler mechanism) is better than a purely mechanical one.

A car differential that lets wheels rotate at different speeds as the car turns is interesting. A performance car differential supplanted with solenoid-driven clutch packs under software control gives you the ability to immediately give each wheel the maximum amount of power that it can transfer to the ground

Not much. Most of the time it's easier to just use many simpler mechanisms and time and actuate/control them singly instead of these mechanisms. At least for the more complicated ones. Most of these (that I've seen haven't gone through them all yet) deal with turning one form of motion, circular or linear, into another with varying alterations, speed, period, or direction change, which was really important when all you had access too was a spinning power source, say a water wheel, and you have to convert that energy into every type of motion for your machine.

For those who like printed books, Dover has an inexpensive edition of the underlying book [1].

There are many similar books, such as "A Victorian Handbook of Mechanical Movements" [2], and "1800 Mechanical Movements, Devices and Appliances" [3].

[1] http://www.amazon.com/507-Mechanical-Movements-Mechanisms-De...

[2] http://www.amazon.com/dp/0486498123

[3] http://www.amazon.com/Mechanical-Movements-Devices-Appliance...

If you like this kind of thing and you haven't run across the Equation of Time cam from the clock of the long now, take a moment and check it out:


Really cute mechanism in http://507movements.com/mm_123.html

The little pins hit the V in the center gear and cause it to shift to the other gear. Neat!

That kind of stuff, is why I've always wanted to be an engineer. In my mind, I visualize arrangements of systems and networks exactly the same way as these.

Building stuff, pressing the "On' button, and seeing it all work is just the best feeling in the world.

FWIW, these are great fun to print out on a 3D printer and play with.

Is there a library of ready-to-print STLs for this? If not, I know my next Thingiverse upload! :-)

Here's a YouTube video that shows some of these: http://www.youtube.com/watch?v=mkQ2pXkYjRM

This reminded me of a recent paper on automatically designing animated mechanical characters. Some of these movements could be used to incorporate different kinds of animations into these characters.


Some of these I had to see animated to even grasp what it might do. Even then I couldn't figure out what use it might be.

A couple of them describe stuff that used it but most were just technical details. It would be nice if there was a link to something that used it so you could see why it was ever made (or maybe some of these are just for fun).

I couldn't get through them all but will come back because it's fun to watch. More fun than it should be really.

edit: example of one that has me guessing why it exists: http://507movements.com/mm_115.html

You probably have one of these if you have a printer. It's basically the same mechanism as on the output paper guides, that causes a movement of the guide on one side to match the other side, and for the paper path to therefore remain centred.

Normally you can take the output tray off and turn it over to see the mechanism.

For number 115: by moving the bigger piece side to side you spin the two wheels. Each wheel spins the same amount but in opposite directions. Shafts attached to those wheels can be used to drive components that need to move equally but in different directions.

Perfect! I couldn't have even visualized that. I just saw gears and didn't even think about a rod off that.

Another thing that would be interesting is to know when some of these came about (some are older than people might think [1]).

[1] http://en.wikipedia.org/wiki/Odometer#Ancient_China

Back in the early 1900s there were a handful of people making physical models of a number of these movements (and some that went beyond this incredible collection). They occasionally tour, Boston's Museum of Science had a big show of William M. Clark's collection back in 2006, I think. Anyway, Cornell has "The Kinematic Models for Design Digital Library" where they try to keep track of all of these physical models:


People are so damn clever.

I am always in lovee with inventions, especially ones where they are complex or extremely massive/heavy, where they were done without modern computers.

I find that the amount of large scale steel industrial works, done before our current era are incredible.

The fact that these movements, visualized in the mind - then described in 2D, drawn by hand, are fascinating.

This looks like a wonderful thing. But I can't see any animations! What am I missing?

Edit: ah, only the color diagrams are animated. The top of the page links to a blurb about that.

"Look for the color thumbnails. They identify the completed animations. Use the prev and next links (above right) to browse the thumbnail pages"


Thanks! Your comment and my edit must have been posted concurrently.

Is there any software that would allow one to easily build and simulate these kind of mechanical movements?

Yes: Algodoo (http://www.algodoo.com/), a 2D physics playground. It used to cost money, but now it’s free.

Just downloaded that and started playing around -- fantastic. I need to force myself to go to bed.

Wow, that's amazing. Thanks for sharing.

that is fantasic - thanks for the link!

I'm much more excited about watching all of these little mechanical animations on a Saturday night than I ought to be...

Oh damnit! Now I miss my lego technic sets!

The animation for http://507movements.com/mm_349.html is actually a bit misleading, since there are two degrees of freedom: the lateral movements of the lower and upper halves are independent.

19 - 22: Am I reading this right? Exponential reduction of the power required? I gotta try this.

It should be exponential reduction of force required, but the length is longer, so the same work.

Clicked thinking I was going to be introduced to a new startup with a pretty cool name ... actual mechanical movements not bad either, however.

this is beautiful; this will definitely come in handy when I start machining...

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact