File talk:AntikytheraMechanismSchematic-Freeth12.png

Incorrect placement for lun1-4.

As per Freeth et Al. 2012 which this image cites as a source, this image has the 20 tooth lun1 and lun2 reversed with the 27 tooth lun3 and lun4 (the 27 tooth pair should be on the bottom). R0uge (talk) 19:21, 4 August 2014 (UTC)

The updated version of this image still has an issue. To clarify, lun1-4 are currently in the correct positions, but lun1 and 2 should have 27 teeth, while lun3 and 4 should have 20 teeth. R0uge (talk) 18:58, 12 January 2015 (UTC)

Zoom in to the top of http://dlib.nyu.edu/awdl/isaw/isaw-papers/4/images/figure06.jpg from Freeth's paper and you'll see what I mean. The tooth counts are simply mislabeled. As-is, the lunar phase indicator will rotate at the wrong rate with respect to the moon's position. R0uge (talk) 15:43, 13 January 2015 (UTC)

I'll see about the tooth count.

An interesting thing. M. T. Wright figured out the phase moon display from the x-rays he took of the wreckage. The wreckage showed the lun4 and lun1 gears, the spindle of the sphere, and the spherical hole it left. He wondered about that lun2-3 gear; it doesn't seem to do anything important. By turning the lun1 gear around, it could directly engage the lun4, with the correct gear ratios. The measurement of the collar attached to the ball spaced the ball from the gear just enough that the ball fit its socket and lun1 and 4 meshed. Wright concluded that the someone took the mechanism apart and re-assembled the the ball-shaft-lun4 gear backwards onto the shaft, and that lun2/3 is a figment. I was tempted to depict it that way. See http://fsoso.online.fr/antikythera/DOCS/The%20Antikythera%20Mechanism%20and%20the%20Early%20History%20of%20the%20Moon%20.pdf

That is interesting! And makes of sense given the economy of the rest of the mechanism. I quite like how you've updated the image. R0uge (talk) 07:11, 20 January 2015 (UTC)

Updated the file again. Added a little cutout to illustrate Wright's design for the lunar phase display. Note the vertical order of the N gears - sometimes the horse can learn to sing. SkoreKeep (talk) 03:03, 18 January 2015 (UTC)

I had a few days of bother. It seem to me that the difference between Wright and Freeth's lunar gears had implications which no one had considered.

The Wright model is straight-forward. The moon and the sun positions get subtracted; in logic terms the sun is subtracted from the moon (M-S), or perhaps (S-M) = -(M-S) which results in the ball moving against the the motion of the lunar pointer rather than with it. That's no biggy; the visual result will be essentially the same.

But what happens when Freeth adds in the extra gear? My first analysis was that it was in effect negating the the sun input, to get (-S-M) = -(S+M). The sidereal lunar period would be 32 days (more or less) rather than 27. Had no one notice this?!!? I had no mechanism or simulation which could tell me if that was how it worked.

After a better look it now appears that the gear negates the output of the difference, so apart that the ball would turn the other way, here is no difference. The key thing for me was figuring out how the extra gear fitted in. It has to be axled upon the rotating lunar pointer because it won't maintain contact with the other gears unless it is; it appears that the physical point of difference is where lun1 and lun2 mesh, and the reversal takes place where lun3 and lun4 meet. Any comment from the experts here?

I guess that's what I get for being an electrical, not mechanical, engineer. :) SkoreKeep (talk) 02:07, 25 January 2015 (UTC)

Incorrect placement for O and P gears

As per Freeth et Al. 2012 (same as above), O1 should mesh with N2, and P1 should mesh with N3 (right now, these are reversed). This then shifts the vertical positions of the O, P, and Q gears. See Fig. 6 in section 2.4.1 here: http://dlib.nyu.edu/awdl/isaw/isaw-papers/4/#section2 R0uge (talk) 20:14, 30 December 2014 (UTC)

In other words, N2 and N3 need to have their names reversed in order to match the naming Freeth arbitrarily used, and to stay with the convention of numbering the gears top-down they need to be redrawn a bit; otherwise there is no significant difference. They are hypothetical gears (N2, N3, all of O and P did not survive), so whether N2 or N3 is next to N1 in the AM is not known. The placement of the gear axes (left to right) is also not important, since this is a schematic of a 3D structure. I am under no obligation to make my drawing match Freeth's exactly, as long as the teeth numbers and notations agree with his tables; however, I will endeavor to please. SkoreKeep (talk) 01:07, 9 January 2015 (UTC)

Looks good! I agree that the horizontal placement of the gears is unimportant, though I would argue that the vertical placement is, as this was specifically laid out in Freeth's paper, and this image is stated as a representation of that model. It's of little consequence however, since the most important change was the correct gear ratios down to the calendar pointers. R0uge (talk) 18:58, 12 January 2015 (UTC)

Uploaded new version with both of the above problems fixed. In this case, N3 is above N2 (as I suggested above) so that the names of the gears match Freeth's. Many small cosmetic changes made while I was at it. SkoreKeep (talk) 22:16, 9 January 2015 (UTC)

Fixed the order of the N gears. SkoreKeep (talk) 03:04, 18 January 2015 (UTC)