accuracy / precision of measurements

[joeq]

I have been looking closely at the proportions of some diamonds and I''ve noticed that they don''t always add up.

For example, it is physically impossible to have the combination of crown pavilion angles, table size, total height, and pointed culet that the certificate states. The difference is small, but definitely there. I would think that this is due to imprecisions in the measurement of the diamond.

So, which measurements are typically more accurate than others? I would guess that carat weight, diameter, and total height would be the easiest to measure and therefore the most accurate. I''m guessing that the specific angles are the most difficult. Nowadays, jewelers just drop it in a Sarin or OGI to get the numbers, so I guess I''m asking what level of accuracy to expect from those machines.

It occurred to me that the difference could also be due to using different orientations for each measurement in a non-symmetrical diamond. Are the measurements averages across the diamond?

I would guess that on a Sarin machine, all of the numbers on the report are from actual measurements. On an AGS cert, are ALL the numbers there from actual measurements or are any of them extrapolated? Does AGS use Sarins to get their numbers? Or does it depend on the specific AGS lab?

I also wonder about the precision of the measuring process. If you run two Sarins on a stone, how much variation will you see? How about running the same stone on two different machines? With something like crown angle, when I see an angle of 34.63, how many of those digits are significant?

(Of course, .01 crown angle isn''t going to make a bit of difference. I''m just curious as I try to model these diamonds, because you need to extrapolate the measurements that you don''t have.)

I''m guessing Garry has lots of experience in this area

. Garry, what do you think?

 

[derekinla]

The reason why certain #'s (i.e. depth percentage) don't always add up is secondary to how the girdle is measured and reported. As for variability in the reported #'s, there is definitely variance from machine to machine and even for an individual machine. You will also get different measurements for different parts of the same stone so reported #'s are often averages.

 

[dimonbob]

Also some of the numbers are averages. Throw one average in there and you can't get the number you are looking for.
Diamonds are not measured out to the 4th decimal like some machines are. They may be in the future but some engineer will have to invent the machine first. We have enough engineers buying these super ideal cut diamonds, why don't you guys get your heads together and come up with a better Sarin machine?

 

[Garry H (Cut Nut)]

Given that 15 years ago getting to within a degree and 1% would take 10 minutes, we have come a long way with the development of Sarin and the me too companies.

Sarin spend more effort on assessing rough diamonds and what to cut from them - that is 80% of their business.

Before HCA, there was no great need for super dooper accuracy. Most labs simply quote table to rounded percentages. The rest of the stone is quoted in height and depth 0.5%'s. Only AGS gives angles and more accurate data on reports.

The good news - ther are companies developing new and better systems.
One will solve the biggest problem - one you guys are completely unaware of - azimuths.
We all see angle and % variations - but what matters far more is the direction a facet is facing.
Say the pavilion main is 40.6 or 40.8 depending on the accuraacy of the gadget. What we do not know is if the facet is actually in the same plane as the one on the other side of the stone. i.e. is its azimuth 180 degrees opposing. Existing systems make certain alogorithm assumptions. Remember the facets are measured in sillohette.
If you can calculate the azimuth of each facet - then you can actually create a deadly accurate model of facet meet points and all sorts of other symmetry related issues.

In the future this technology will make H&A's accuracy images of diamonds.

There are some very clever folk

As for angles, well.... we wont need them because the whole cut grade process will be automated into the measurement process. Minor facets, major facets, any and every shape, stones with extra facets - symmetry deviations - the whole shootin match.

 

[Richard Sherwood]

Hey Garry, is this advance in equipment about to happy anytime soon, or is it three years or more down the road?

 

[Garry H (Cut Nut)]

some of it is really high high end Rich - GIA, HRD, EGL and IGI type clients.

But a better mouse trap is not real far off.