how did they do it?

fredag den 10. august 2018 kl. 00.32.24 UTC+2 skrev RichD:

a guy on youtube is making a replica mostly using tools and layout that would have available at the time

wheels laid out with dividers and made with files

his channel is worth a look, he did a clock before

The man who replicated it was on a Nova show about the Antikythera mechanism on PBS a few years ago. It explained how he determined it was 53 teeth and all...

lution debate -

Once Darwin had proposed evolution by selective adaption, that argument was dead. The Antikythera mechanism isn't all that close to anything it might have "descended" from.

Cast them? Filed them out of circular blanks?

Euclid (323?283 BC) preceeds the Antikythera mechanism by about a c entury.

Carefully. A 53 tooth wheel would be a bit of a challenge, but Archimedes h ad worked out an algorithm for calculating pi in 250BC, so it would be poss ible to work out how long each of the 53 segments around the rim of the whe el would have to be. Messy arithmetic, but even using 22/7 as an approximat ion for pi would get you quite close enough.

--
Bill Sloman, Sydney

Compass and straightedge, classically. If you can get a flat surface to mark on (stretched parchment?) and inscribe a circle, you can subdivide it. Powers of two are easy, so you could use a binary fraction rendition of 1/53, 2/53, etc. to get to any degree of accuracy you want.

That's why 'trisecting the angle' is an old well-studied problem.

Lens design used large sheets of vellum, H6 (very hard) sharp pencils, and ratio-of-sines graphic calculation, up 'til the 1970s. And a gizmo called a 'polar planimeter' integrated areas, and an exceptionally clever device existed to do harmonic analysis (for tide calculations, largely); it computed a Fourier transform, mechanically.

I'm more impressed with the ability to produce flat bronze stock; someone with a hammer is no match for centerless-ground rollers.

As I vaguely recall, in the Nova "Ancient Computer" video, Michael Wright is shown marking off the gears using Dykem Blue and dividers. The teeth are cut with a triangular hand file. That's insane.

"Ancient Computer NOVA HD" The rather skechy part on cutting the gear teeth starts at 23:40.

Much easier is to start with a metal disk that is the outside diameter of the gear, and wrap a piece of paper around the circumference of the uncut gear. Then divide the piece of paper into whatever number of teeth is needed. Some trial and error was probably involved since all the gears are prime numbers and therefore can't be neatly divided into smaller sections. Once the paper is properly divided in the proper number of sections (teeth), the paper is re-wrapped around the circumference, and the gear is scribed between the teeth, where the triangular file cuts.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

And of course: first make your file! ...

Mike.

exactly, it's an almost trivial problem. Just slow to do. Thread works if no paper.

NT

Am 10.08.2018 um 05:52 schrieb Jeff Liebermann:

It is even easier when you nail the disk into the middle of a large table. Then you can draw a large circle around it and sketch the ticks on the table in large scale. Then you need only a needle & a ruler to transfer the marks to the disk.

Cheers, Gerhard

fredag den 10. august 2018 kl. 08.44.32 UTC+2 skrev Mike Coon:

It might have been easier to make a model of the gear wheel in wood, which is easier to work, then use the wooden model to make a mould in sand, take out the wooden master and pour molten bronze into the mould.

When I need to replace a shattered cast iron grate in our house in Brighton , England, that's exactly what I did - carved up a chunk of plywood so that it fitted into the fireplace where the original grate had gone, cut out a bunch of slots to provide the ventilation, and took it over to the foundry where they used it to make a sand mould, which the filled with molten iron.

Worked a treat, and didn't cost much, if you neglect the time I spent cutti ng a dozen slots and smoothing off the surfaces.

--
Bill Sloman, Sydney

If you are careful about your choice of units, you can get very accurate. Draw your circle 135 units, and your 53 teeth will be 8.002 units around the circumference.

They probably did that by old mirror grinding techniques to get the final truly flat material by working three pieces against each other the same way optical flats and snooker tables are still made today. On the plus side they wouldn't need to polish it as carefully as for a mirror.

As Bill said they knew pi well enough to compute the right length for a known diameter of wheel and would doubtless have made a jig to allow precise construction of the gear wheels.

They could also have done it by successive approximation given enough time and patience. Prime numbers of teeth are more tedious to do.

53 & 127 are both inaccessible to the Neusis construction known to the ancient Greeks so they would have had to do a bit of a fiddle.

Engineers of the day would have been more likely to use this sort of method to get practical results for N-gons - straight edge and compasses geometrical purism came a bit later.

--
Regards, 
Martin Brown

Nice, thanks. GH

Take a piece of string, tape, whatever. Mark out equal ticks, say 1 cm apart. Make a loop with 53 (or 127, whatever) of these marks, spread it out into a circle, maybe around a board with lots of concentric circles drawn on it. Draw lines from the marks to the middle.

Cheers

--
Clive

I beg to differ. It's very difficult to get a smooth surface and impossible to get much detail because of the size of the gains of sand. It's also tricky to obtain thin sections, such as the gears. Sand casting is great for large chunky objects with rough surface finish, such as bathroom plumbing fixtures and manhole covers. See any picture here of anything that resembles the smaller gears in the mechanism? Sand casting are also full of inclusions (holes) where something turned to a gas and created a bubble. If I sand cast a gear with 233 teeth the same size as the mechanism, methinks half the teeth would be missing, and the other half would soon fall apart. If the Antikythera mechanism had been cast, it would probably have been the size of a refrigerator and weigh a 1000 lbs.

I had a backyard foundry running at my house for a while. Propane forge, green sand, 50 gallons of water in case something caught fire, etc. Mostly, I worked in aluminum because of the low melting temperature. I sold everything to a jewelry maker about 20 years ago, who lets me use it when I need it.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

ch

I don't know so much about bronze, but aluminum is easy to cast. I did som e casting and made a car door handle which was less than a quarter inch thi ck. We used a flux to remove dissolved gasses and my casting had zero hole s. The only problem I had was not enough risers and the metal shrank away from the mold on the side with the decorative pattern which was crossed gro oves in the original plastic.

For a gear the only parts you care about being smooth are the shaft hole an d the touching surfaces of the teeth. They will need to be properly drille d and filled no matter what.

Rick C.

Loads of really ingenious methods are outlined in this book:

You can make a *perfectly* flat surface (terribly important to have one of these in every serious engineering shop) by placing two slabs of granite on top of one another, separated only by grinding paste and agitating one against the other. Eventually they rub off each other's imperfections and you get two perfect surfaces for reference purposes from which you can then make other highly accurate tools and so on. Clever stuff!

--
This message may be freely reproduced without limit or charge only via  
the Usenet protocol. Reproduction in whole or part through other  
protocols, whether for profit or not, is conditional upon a charge of  
GBP10.00 per reproduction. Publication in this manner via non-Usenet  
protocols constitutes acceptance of this condition.

Three. Two will give equal positive and negative spherical sections (or paraboloids if you grind them right). Only in the case when three faces are ground against each other, will that spherical section have zero curvature, a plane.

Tim

--
Seven Transistor Labs, LLC 
Electrical Engineering Consultation and Design 
Website: https://www.seventransistorlabs.com/

Same way they made suits of armor. Cold beat it flat with a hammer. Then make it reasonably smooth and flat by grinding between two flat stones with some sand as an abrasive. It takes a long time to do it this way, but in 250BC, the time clock hadn't been invented yet.

Drivel: My litmus test for reverse engineering technology is to check what method was used preceding it. If there's no technology preceding the idea, it's probably wrong. The problem with the Antikythera mechanism is that much of the technology doesn't fit the time estimates. For example, everyone seems to assume that steel files were used to shape the gears. Yet, the ability to make such fine and precise files is far into the future. However, what did proceed gear the cutting technology of 250BC was knife, weapon, and agricultural technology. Much of this was based on the use of abrasives instead of cutting tools. The teeth could have been made by notching a knife blade, and using it like a chisel to notch the teeth. Or, the knife blade could have been ground flat on a stone, and the teeth ground using sand as an abrasive. It didn't even need a knife as a piece of hard wood with abrasive grit pounded into the surface would have acted as a tolerable file. Etc...

Of course, it could have been made by aliens from outer space.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558