Grading the Princess Cut by Bruce Harding

[pricescope]

Many thanks to Bruce Harding (aka beryl, the author of Faceting Limits and "Diamond Design" Revisited) for his new article Grading the Princess Cut.

 

[oldminer]

Thank you for posting the article and thank you to the author for allowing us to see yet another part of the unfolding tale of diamond cut improvements. It does make it sound as if cutters are going to have many difficult lessons to learn as well as the possibility of failed best efforts on some rough. This target of perfection may be difficult to obtain with regularity......but I am not a cutter. Personally, I'd prefer to see nearly all princess cuts perform well. Maybe, over time, this will become the norm.

 

[JohnQuixote]

(Touche' Dave)


Bruce, thank you for sharing your observations - and for the incredible illustrations, as usual. You have a gift for that.

Comments...

In designing our forthcoming 0 princess Brian has walked this path as well. Much connecting of dots must be done using the guidelines. Bruce hits several nails on the head. I humbly suggest that observations such as these be considered as an addendum to the existing guidelines. If not in whole, perhaps Peter will consider separate input from Bruce, Paul, Brian and others surfing the front edge of the new princess wave… Paul’s already hanging 10, Brian’s in the water & other swimmers are applying wax to board (or facet to wheel if you prefer).

This also seems an opportunity to give due praise. DiamCalc is often seen as a simulation/assessment tool ‘round these parts. People should know that behind the scenes it has unsurpassed value as a design tool for new proprietary cuts & development of existing ones. From personal experience, Brian’s use of DC with ACA and now our 0 princess design has saved a lot of weight in R&D.

Of course, you have to know what you’re doing...As a theorist I come up with some pretty wicked ideas, only to have them sidelined by the cutter due to this thing called ‘reality.’ But gee…it was so easy to do in wireframe.

For those who don’t know Bruce’s pedigree, he has written pieces that literally changed the way we judge diamond beauty. He was the first person to notice and write about the effects of an observer on a gemstone’s appearance in 1975. Obstruction research ultimately led to new quality terms within our paradigms such as AGS’ inclusion of ‘contrast’ and GIA’s definition of ‘pattern’ within their definition of scintillation. He has been of invaluable assistance with Brian’s discovery & explanation of facet yaw. He is also a prince of a guy.

 

[Garry H (Cut Nut)]

Thanks Bruce for sharing this little known issue.

I learned a way to overcome this using DiamCalc last year.

I took a 3D scan (Thx Rhino) and opened it in Diamcalc. When you imnport a real scan into DiamCalc there is a little known option called ADVANCED, which is normally grayed out. the tab is in the middle of the right side area just above MAP.

If you open that tab you can cut or uncut facets.
I took a stone and (cant remeber - did i add or remove) recut the chevrons.

Now there is a way to do this with the standard DiamCalc 2 chevron model - you have to save it as a .dmc and then re-open it as an imported model i think. Someone might find out how ???

Anyway if you look at this screen shot of the before and after you will see what I mean.

Also on the advanced page you have a list of every facets slope and azimuth - hugely detailed.

Of course modelling in GemCad is easier - but this allows you to quickly make a small change in DiamCalc and see the results of your ''virtual repolishing''

 

[Garry H (Cut Nut)]

Oops
Forgot the screen shot.
The 2 DC windows are open on top of the advanced page

 

[belle]

thank you for sharing this information. i am but a cheerleader on the sidelines watching the evolution of the princess cut performance, and it''s great to know that there are real improvements in the works.

 

[Garry H (Cut Nut)]

Here i have butchered a chevron facet

 

[JohnQuixote]

"I saw a broker looking at FireScope or IdealScope image simulations by software to see what these cuts would look like. However, the image he saw will not be what he is getting if the secondary facets are not cut the same way as they are represented in the computer model that is making the image." -BH

Date: 8/13/2005 6:08:02 PM
Author: Garry H (Cut Nut)

...When you imnport a real scan into DiamCalc there is a little known option called ADVANCED, which is normally grayed out. the tab is in the middle of the right side area just above MAP. If you open that tab you can cut or uncut facets. I took a stone and (cant remeber - did i add or remove) recut the chevrons.

Probably a good catch, Garry. I didn't think of that as the 'representation' issue.

 

[strmrdr]

Thank you Bruce interesting article.
Its nice to read some hard core diamond cutting articles from time to time.
I have to admit that princess cuts are my least favorite diamonds but I have to admire what AGS has done for them.

 

[beryl]

Thanks, Leonid, for publishing this. Someday perhaps I will learn how.
Thanks to others for kind comments.

I think that some missed the point: the computer definitions in different software (such as OGI and DiamCalc) may not match each other, they may not match what AGS tested, and they may not match what the cutter does. So, if you''re trying to grade a Princess cut by computer-simulated images, don''t put excess faith in light-leak patterns or the color grade of the AGS grading charts.

I have TONS of examples, data and illustrations for anyone who is interested. Much of this will be very useful to cutters and to manufacturers of cutting equipment who will make the new click wheels; I hope this reaches them.

 

[Garry H (Cut Nut)]

It would be interesting to have a 3D example or two Beryl, to see if we can see the differences, and find solutions to rectify the problems.

 

[beryl]

Garry:
. I have made dozens of 3D illustrations and they appear visually identical graphically, but I am sure their FireScope or IdealScope images would be different.
. As you know, I have great difficulty posting illustrations here, but Leonid has promised to help, so I will try.
. With the help of a computer guru, I just recenlly got DiamCalc working again and see the latest one for the first time. It is 2.3.0 HASP, which Sergey gave me in Moscow, May 2004. I am sure it is no longer the latest. Can I do what you say with it?
. I am afraid of GemCAD since finding errors in it. Perhaps Vladimir can help me to import AutoCAD dawings into it, or to create dawings in it. I will check with Sergey after your reply.

 

[beryl]

. Here is the initial version of 2-chevron design, with both side and corner slopes specified, radial locations of meets PQR as specified by AGS, and also using AGS' original definition of height Q.
. Note the odd azimuths. This, in itself, is not a problem because the equipment manufacturer can make the click wheel at any angles (I asked ADRI recently); the problem is that this type of definition requires different azimuths for each Pav1/Pav2 combination.

 

[beryl]

. Fig. 31 is basically the same stone as Fig.30 except I specified azimuths which are typical indexes approximating those of Fig.30. I kept the same radial locations of meets PQR but, after specifying side slope of 60°, their heights were automatically defined by that and the chosen azimuths. The result was a corner slope of 42.21° instead of 42°. I think this in an acceptable approximation to Fig.30.

 

[beryl]

. Fig.32 is also similar to Fig''s 30 and 31 but this time I specified the corner slope at 42° and used the same indexes as azimuths, getting 59.81° side slope - also an acceptable approximation.

. Much more later; must go, briefly.

 

[beryl]

. In Fig.34 I changed the azimuths to even divisions of 45° = 15° & 30° = indexes 1 & 2 of a 24-tooth wheel (if there were one). Specifying the corner slope at 42°, the side slope becomes 58.93° - still not a bad approximation of the 60° in the original cut.
. Note the 'angles between facets' indicated at right: - the changes of 16.10°, 11.76° and 10.25° are much more uniform than in the previous 3 examples and, I think, would therefore give a more distinct scintillation pattern. I have never seen these angles indicated in any other gem design analysis* but I think this relationship is significant. I will be anxious to put these into DiamCalc to compare the results.
. Each of the insets at upper right is an exact scaled-down version of the larger layout. Put side-by-side, I would defy anyone to discern the difference - or also in 3-view drawings - but I bet they will look different in DiamCalc simulations - both moving and IdealScope.
. Edit Sep.04: I have just purchased GemCAD software and find that it shows these data in 'Point on Edge Dialog Box' as "Angle between facet normals". In accompanying text it describes it as "the dihedral angle between the two facets that share the edge." I had forgotten this proper term from crystallography.

 

[beryl]

. Attached are slope data for 3 different combinations of indexes: 25°/5°, 30°/10°, and 30°/15°. You will find that one of these will give you close to the Pav1/Pav2 slopes used in the AGS grade chart.
. Note also, however, that it is not necessry to achieve these slopes precisely to read a grade; think of these as coordinates on a chart and find the combination you have in the large pixels between them.
. 3 more combinations on next page - forum limitations would not take it all at once.

 

[beryl]

... and here's the other half of the 6 azimuth combinations I chose to approximate various Pav1/Pav2* combinations cited by AGS on their grading chart.

. Perhaps the most important thing to notice here, is that if you use the same click wheel for all values of Pav1 or Pav2 the difference between Pav1 and Pav2 values changes. That is why you need 6 wheels to approximate all the Pav1/Pav2 combinbations cited by AGS as coordinates of the grade chart.
. 35°/20° matches 6 of the AGS combinations; if i had to use only one click wheel, that's the one I would have made.

* Sorry about that: if you are not familiar with the AGS charts, Pav1 is the slope of the main facets and Pav2 is the slope of the corner facets.

Much much much more if anyone is interested. I can send an MS/EXCEL spreadsheet if you want to plug in your own azimuths and Pav1 or Pav2. Radial distances to meets PQR stay the same at .05, .15, .25.

Edit Aug 18 AM: ERROR: Heading for 2nd column should be 'AZIMUTHS 40-25°' (not 40-35°).
Edit Aug 18 PM: ERROR: Heading for 3rd column should be 'AZIMUTHS 37-25°' (not 40-30°).

 

[beryl]

. And here is a 3-chevron design with AGS-chosen radii to meets PQRS (.05-.30 by 3rds) but common-index azimuths. I like this one, using indexes 1,2,3/40, because the angles between faces are most evenly distributed at 11.4°, 8.0°, 6.5° and 11.4° (recall my comment about scintillation). The one with indexes 2,3,4/40 gives interfacial angles of 20.6°, 7.4°, 6.5° and 6.1°.

 

[strmrdr]

For the non-cutters on the board can someone define:

indexes (using toothed wheels)
and
click wheels

I think I know what both are but if one the experts can chime in with a good definition it would be better than what I could do.

 

[beryl]

. Of the infinite 4-chevron designs, here's one using AGS' chosen radii to meets PQRST (.05-.325 by 4ths). I prefer this one using indexes 1,2,3,4/48 because it has the best interfacial angle balance (10.2-10.2), as compared to indexes 1,2,3,4/40 (Fig.72) and 2,3,4,5/48 (Fig.74).

 

[beryl]

. For the daring soul - here's one with 4 chevrons and a chamfered corner. Try it in CZ first.
. My MS/EXCEL Spreadsheet and TI-83 Plus calculator program both include this option. Either is available on request.

 

[beryl]

Strmrdr: Sorry; this discussion does not belong here, but where else would I put it?

. I have photos of my machine, with index wheel, and a click wheel photo that Garry sent. At this moment I can't find either of them. They perform the same function.
. Both are on the rotating part of the faceting head which carries the dop with the stone on it; they ensure the accurate rotation from one rotary angle to the next - such as even increments of 45° to cut the 8 main facets on a round brilliant - and hold the stone firmly in that position as it is lowered onto the cutting lap. A little spring latch is lifted out of a notch so the wheel can be moved to the next similar position, then drops into the new notch to hold it steady in the new position.
. An 'index wheel' is a round disc with teeth, like a gear, and the latch is moved radially in and out of the teeth. 'Index wheels' are used by hobbyists and colored-stone cutters to make various cuts. They have teeth all around, and the cutter uses only those he wants for a given design. Mine has a 'mask' to cover the unused teeth.
. A 'click wheel' is a cylinder with notches on the end into which a latch is moved axially. They are used mostly by diamond cutters and have only the teeth that the cutter needs - to minimize costly mistakes. Sometimes they have teeth only for the mains or for the break facets (diamond cutters call 'halves') to further minimize the chance of errors. I will provide photos when I find them, if someone else doesn't do it first.

 

[strmrdr]

In my opinion this thread and article does belong here.
Although I dont get all of it I find it very very interesting and am learning more than I ever wanted to know about princess cuts.
Please consider doing more stuff like this and when I see something that when explained will help people understand it better I will ask.

Thank you for taking the time to post this here and discuss it.
If it leads to better cut diamonds im all for it and it belongs here!

My 2c anyway :}

 

[Garry H (Cut Nut)]

I see if I get some time tonight Beryl
Here is a picture of the type of critical gadget that is an index wheel for a round brilliant pav facets (I think)

 

[beryl]

Strmrdr:
. I found pics of my machine, with its ''index wheel'' but they have too many bytes and I don''t know how to change that (I am a computer klutz).
. However, I attach a poor copy of a photo that Garry sent me - of a ''click wheel''. By the spacing, I would assume that this is for brillianteering the crown of an SRB - the center one for the star and the two beside is for the breaks, or ''halves''. The mains would have been already cut by another person, so this skips over them.

 

[strmrdr]

kewl thanks

 

[adamasgem]

Date: 8/14/2005 9:57:19 PM
Author: Garry H (Cut Nut)
I see if I get some time tonight Beryl
Here is a picture of the type of critical gadget that is an index wheel for a round brilliant pav facets (I think)

Thats what it looks like gary, sometimes the index wheels are also color coded on the side, depending on who made them.

 

[beryl]

. Here is pic of faceter head on my machine (much-trimmed to fit forum limits).
. Platform at lower right moves vertically to lower stone onto lap and thus control depth of cut.
. This machine is unlike most hobby machines (which have a mast) but similar to most commercial ones.
. Diamond machines have an extra axis (vertical) to change direction of grain relative to lap motion.

 

[beryl]

... and here is close-up the index wheel and latch.
. The 80-tooth wheel shown is not commonly-used - that's why it looks so new.
. Most-used by hobbyists are 64-, 72- and 96-tooth wheels. I also have a 120-tooth wheel.
. For oval gems I make my own index wheels, with only the teeth I need (article in DiamondTalk 2002).