un-roundness

[beryl]

An article by Sergey, Yuri and Garry in the Russian Journal, and in English* at the website www.cutstudy.com, mentions roundness of diamonds as a factor in cut grading. It describes how an object can be 3-sided but measure the same across the girdle in all directions. They show only a special case - 3 lobes with radii from opposite corners; however, there are an infinite number of situations which produce this effect.
The attached illustration shows a 5-sided case, and that the radii need not be from the opposite corners, but from corners of a smaller polygon within. This illustration was presented to the authors by e-mail but has not appeared in the forums.
The number of ''sides'' is always an odd number; this occurs commonly in industry in the process of ''centerless grinding'', used for ball-bearings, etc. In that industry we used a ''waveometer'' to measure the number of ''bumps'' per revolution and the amplitude of those ''bumps''. Surprisingly, perfect roundness (within a few millionths on an inch) is undesirable in ball or roller bearings!
This also occurs commonly when drilling holes, where 3- and 5-sided holes result when using 2-fluted drills. It is very noticeable when using a flat drill across-grain in wood.
My son, a fine jeweler (see www.hardings.net), used-to offer unround diamonds, at considerable savings, for earrings, where the ''error'' is not noticeable.
This will be a factor in the upcoming conference on diamond cut grading and may lead to more common use of unround diamonds at significant price reductions.
* The English is not an exact translation; some changes were introduced. You may wish to read both articles.

 

[strmrdr]

I havent had time to read the linked material but I will it looks interesting.

Kinda off topic but Im curious if anyone else knows the reason for not wanting the bearings perfectly round or perfectly smooth.
I had an interesting discussion about this with an aircraft parts machinist a while back.

hint: think lube :}

 

[Rank Amateur]

Are diamonds interntionally cut this way or is the un-roundness a result of bruting or some other process? I picture bruting as having only one radius center.

 

[beryl]

Strmrdr:
If ball/roller bearuings are too round they emit unpleasant noise. I once externally-honed some sets of cup/cone/rollers for a tapered roller bearing within millionths of an inch (I forget how few - it was 40+ years ago). The manufacturer put them through his Quality Control department with a lot of standard production, without them knowing they were different, and these were the only ones rejected - because of the sound.
I have heard it said, by a bearing manufacturer, that the number of 'bumps' should be between 17 and 100. Less than 17 produces too severe mechanical shock and reduces life. More than 100 results in audible noise (but speed and damping by lubricant should be factors too).
Too smooth is another problem. One manufacturer told me that they skid and burn if rotation starts too fast. I recall (again 40+ years later) that 18 minroinch finish was the low limit (I have seen 60 in ball bushings). Contact curvature and direction of finish are significant factors (more interesting stories lurking there).
Not related to bearings but to roundness and finish ...
When in the cotton yarn machinery business, we found that a spinning ring worked best if it was between .003" and .008" out-of-round. This has a cross-section like a railroad track and a little steel loop travels around it at great speed (about 3" diameter at about 19,000rpm). We theorized that with more than .008" it bounced too hard, while with less than .003" it didn't hop through air enough to cool it sufficiently; either way they burned-up faster than when within the .003-.008 range.
Unexpected things happen when things get smaller than usual. Two famous cases occurred with the surface finish on two aircraft engines in WW2. In the first case the pistons were too smooth and wouldn't hold oil, so they honed them to make them rougher! In the second case surface finish on a crankshaft was changed from 8 to 3 microinch and the engines burned-up prematurely; while friction coefficient decreases as surface roughness decreases - within then-known range - it increases again if the roughness goes below 5 or 6 - things tend to cling and weld together (I have also seen this happen with valve stems and packing). Wudjubleev that everything becomes black when it has a 1/10 microinch finish? (need interference microscope to measure this).
I can tell you more interesting and surprising things about surface finish and some unusual things I did.
==============================================================
R/A:
I am sure that it occurs in the bruting process.
In the past this was done mostly with one diamond rotating on a spindle and another - of junk grade - held against it on a stick. If you have ever used a wood lathe you know that your hand can vibrate and produce un-roundness this way.
Now, I am told, they mostly rotate two production diamonds side-by-side, bruting two at once. They would have to rotate in the same direction, which would reduce the tendency toward un-roundness, but some is still possible - note the changing billboards with triangular slats that rotate the same way.
In my observations, un-roundness in diamonds is generally two-sided = ovalish. I would guess that this is due mostly to the shape of the rough.
=============================================================
I am leaving on trips today and will be back about April 29, so I will be unavailable for comment until then.