Welcome! Log In Create A New Profile


Mechanical Computing Working Group

Posted by sheep 
Mechanical Computing Working Group
February 22, 2010 08:49PM
From another thread.

Julie, let me know if you want to start up a Mechanical Computing Working Group. I'd suggest a thread to begin with, and then grow that into its own forum.

I have a hidden agenda: I want multiple gear families in the wiki as part of RBS. Also, mechanical computing is cool.

Would you like to upload your 3D models of Adding Mechanisms here
or here

This may supplement or supplant blogging.

-Sebastien, RepRap.org librarian.

Thanks for suggesting this. I am probably some time off having specific examples to post. Much of the recent research is somewhat copyrighted in academic journals and books, so I am not sure what drawings can be posted. Obviously my own models taken from the 19th century plans can be shared at will. The papers of dead professors I am not so sure about.

Borrowing examples from other threads. I define mechanical computing as cam and gear based examples. Historically these have been implemented in the following well known implementations.

The oldest of these devices is the Antikythera device. In the late 1950s, 1959 to be exact the late Derek de Solla price wrote a paper, which was summarized in Scientific American. He worked in this through the 1970s.

The late Alan G Bromley wrote a number of articles in both early computing publications as well as the British Horological Journal. Much of my work has been to independently follow on with some of Bromley's research.

Bromley's academic successors have managed some astounding breakthroughs using modern image processing and tomographic techniques. The result of this research can be found here.[1]

Ideally one should be able to take the tomographic slices and use reprap to create a study model, which shows the internal state of the gearing. Like most mechanical computing devices, the functions are based on a clockwork model of the celestial heavens.

In the following examples, this large central gear known as the "prime mover" will be seen as a common driving force. Where the reprap and other 3D printing technologies come in play relates to the need for gear teeth ratios containing prime numbers. These prime numbered gears are required to handle the non-linear functions mechanical computers are best at solving.

Technically the Antikythera device is not a computer. It is a calculator. A complex one used for determining eclipses and when the Olympic games are to be held.

Often dismissed by literature scholars as fantasy, there are quite a few indications that such devices, while rare were not uncommon. It would be the sort of thing a Roman senator would own. More on the techincal side we have the writings of Hero of Alexander and Philo of Byzantium. These detail technology from puppet shows (my favorite) to cranes and steam turbines. Why it took over 1900 years to combine the latter two concepts is one of the great mysteries of the world.

In digital logic, the elements are based on logic gates and oscillators. The fundamentals of mechanical computers are clocks,gears,cams, escapements and pinned barrels.

Clocks are either oscillators or pulse shapers. Pulses are shaped by a mechanical device called an escapement. Escapements come in two forms intermittent, or continuous. As found the Antikythera device has not escapement. It does however contain a program input section in the form a a grooved disk, which can react to programmed pins and markers.

There are many examples of clockwork displays in public squares throughout the world. While interesting and complex these devices fall out of the scope of this summery.

Most improvements relating to clockwork oscillators did not happen until the 16th and 17th centuries. These were further improved in the 18th century, where mechanical computing starts to separate from pure timekeeping. The most practical development was a new form of continuous escapement, which made musical boxes possible. This was the fly governor.

Forms of this governor had been used in automata since the times of the Greeks. The Franco-Swiss engineers in the Jura mountains worked out how to use a gear driven lead-screw to precisely control the speeds of rotating cams.

This combined with the concepts popularized by Cartesian mathematics made possible, program storage devices, such as music boxes, which could play interchangeable songs.

The Jaquet droz dolls, Detailed on my website represent the first example of a true stored program computing device. A large wheel on the writer doll can be programmed with tabs representing 40 characters of the roman alphabet. How these tabs are programmed determines what the doll can write. A stack of cams relates to the memory unit of a Harvard architecture processing unit. These cams represent in 3 dimensions plus time the letter to be written in Cartesian coordinates. All combinations of the 40 characters are possible.

The Jaquet-Droz automatons use all the concepts of mechanical computing. These are the ancestors of automatic lathes and riviting machines, Babbages calculating and computing devices, and most importantly the fabric weaving industry.

Jaquet-Droz followed on the work of Vacuanson, who created the automatic looms, which set off the industrial revolution. Vacuanson' looms used a single prime mover or barrel.

After Jaquet-Droz came Jaqard who made a change to Vacuanson loom, by adding a chain of cards to the prime mover. This makes the prime mover almost infinitely long.

In the 1830s and 1840s there exist two more examples, perfecting mechanical computing. The first of these is the clock of Strasbourg Cathedral. Strictly speaking these remain clock calendar devices which solve specific astronomical problems which are non linear. This clock is in effect made of many sub processors, some which move automata others calculate time. Mechanically solved are the lunar equation, Date of easter and the leap year century rule. The later is most ingenious inserting a separate card into a wheel every 4 or 400 years.

The second example are the computing devices of Charles Babbage and his son H.P. Babbage. This work was forgotten between 1870 and 1969 until some patent cases came forward. Since 1969 there has been a lot written.

Most of the foundational work on these archives was done by the late Alan G. Bromley, who was mentioned in connection with the early Greeks. The archives, left to the British public in the 1870s are now controlled by the for profit Science Museum of London. Through corporate sponsorship and private grants, examples of some of the simpler calculating engines have been built.

Simpler versions of these calculating engines were built by others during the 1850s and 1860s. Most scholars feel this work was a dead in and had little impact, on the developments 100 years later.

These mechanical calculators and computing concepts are a complex subject. Please feel free to open threads or discuss some of the details and myths of what was done, what could have been done and what can be done.

Returning to Swilgue's Strasbourg clock and the lunar equation calculator. Analog computers baased on this technology were to play a major roll in the WWII era time-frame. Much of this relates to the work of Vannever Bush and his differential analyzer.

This is a short summery of mechanical computing from the Greeks to the modern age. Please feel free to add to this summery or ask questions.

Re: Mechanical Computing Working Group
March 05, 2010 11:05AM
Julie (and others reading),

Although I think all-mechanical computing isn't really a core part of reprap, it might be useful -- if nothing else, it's pretty resistant to EMP. (Although we'd still have to replace the motors with something mechanical, as well.

One aspect of mechanical computing I don't see in your summary (did I miss it?) are the mechanical analog computers, notably the ones used to solve differential equations by means of drum and disc integrators. If you Google the phrase "differential analyzer" you'll find lots of links, including a pretty good article on wikipedia: [en.wikipedia.org]

One of the key developers of these devices was Vannevar Bush; here's a link with some pictures: [web.mit.edu]

Side note 1: My father worked for a short while on one of the MIT/Bush differential analyzers, but it was soon evident that electronic computers (whether analog or digital) would surpass the mechanical ones. I wonder when electronics' day in the sun will end. Optical computing anybody?

Side note 2: I mentioned the Jaquet-Droz automata in the first chapter of my dissertation -- somewhat over the objection of my first reader. He didn't think the history was needed, nor my linking of experimental robotics research to origins in science fiction. Luckily, my second and third readers weighed in on my side.

Larry Pfeffer,

My blog about building repstrap Cerberus:
Re: Mechanical Computing Working Group
March 05, 2010 12:33PM

I mentioned the connection between Swilguie's lunar calendar and the Vannevar Bush differential analyzer in passing in the second to last paragraph.

Lucky you to be able to write a dissertation on this subject. Something I would really like to do, but have not found the right history department to do the degree in "History of technology." that I would like. In the good ol U. S. of A, there is a strong dislike for the "not invented here and patented" attitude.

There is some really interesting observations on the real world in fiction too. I am still amazed at Dickens's Little Dorrit. This novel is undoubtedly Science fiction. Babbage's mechanical computer is the Macguffin and is even shown in one of the illustrations. As soon as I figure how to upload an image to the reprap wiki, I intend to upload that one as it may be the only documented evidence of what Babbage/Clement's workshops actually looked like. Note the Fictional names are Doyce and Clenham.

Also one can look at Hard Times for these Times as a novel set 100 years in the future from it's writing. The title is actually a pun on multiplication, the working title was Two times two is four. I like to think of The three novels, Tale of two Cities, Great Expectations, and Hard times as Dickens take on Past-present-future. A theme he is well known for.

What is not well known is the book Dickens was planning after he finised Edwin Drood, was to use the telegraph and Atlantic cable as the macguffin. He even wrote the opening paragraphs, envisioning the reader a the electrical signal traveling through the wire. This written years before Einstein's though experiment relating to riding on a beam of light.

Dickens work is socialist and often seen as a backlash to the industrial revolution. He was an observer, and got the details right. Especially when using nautical terms and describing clocks and watches. Even to the point, probably for research while writing Dealings with the firm of Dombey and son to join Lord Grimthorp's British horological society. His letters to that journal are hilarious.

Lord Grimthorp by the way designed the clock that rings the Big Ben bell. And to bring us somewhat back on topic, he used sewer gas pipe as shafting for the fly governors that regulate the striking speed. This was made by forming the pipe around a mandrel and seaming the join. After about a hundred years of operation this weld failed. Causing the clock to explode.

How this is slightly on topic, is that I have seen similar seams on thin walled conduit, which some have used as rails for table axis. Not that any of us expect our machines to last 100 years.

Re: Mechanical Computing Working Group
March 06, 2010 05:09PM

First off, my apologies for not reading your history of mechanical computing carefully enough. I should have caught your mention of V. Bush and his work.
Much though I like the subject, history of science/technology wasn't my dissertation topic. I just wanted to add some historical and literary context to set the stage -- my dissertation research was on the design and control of cooperating robotic manipulator arms.

I haven't read much Dickens, and that includes both _Little Dorrit_ and _Edwin Drood_. However, I saved a full (almost full?) set of my grandmother's, so one of these days....

Dicken's association with horology is likewise news to me. Ditto the use of sewer-gas pipe, and the pipe's failure as the culprit in Old Tom's demise. (If I recall correctly, the tower and/or the bells are named "Big Ben" but the clock itself is called "Old Tom."

Since you like history of science/tech, I hope you've read Neil Stephenson's _System of the World_ trilogy. If not, clear your calendar; you're in for a *treat*! They're long books, but I read fast and just loved all the discoursive details. (How can a techie *not* love books in which the Newton/Liebnitz feud over credit for inventing calculus is one axis about which the skeins of story are woven?)

One of the nice aspects of reprap is that, if the core ideals are realized, any given reprap machine won't need to last 100 years. If it breaks, one of its siblings can be used to make replacements for the broken parts. I know we're a long way away from that, but the idea of being able to do that (or even come reasonably close) is quite satisfying.

Your organ-building work sounds interesting. How to you like the sound and action of the one you built?

Larry Pfeffer,

My blog about building repstrap Cerberus:
Sorry, only registered users may post in this forum.

Click here to login