This is an account by Chris Algar describing his work to make a machine for grinding and polishing concave facets


The following notes are just a description of my attempt to build a faceting accessory. I have not included detailed building instructions or a constructional diagram because most people
who own a lathe will have more engineering experience than I have.

As if there were not enough difficulties and variables to contend with in normal faceting, I recently decided to see if I could cut a stone with one or more concave facets. Although 'concave faceting' is self-contradictory because facets are by definition flat, I can think of no better term to describe the process. Commercial machines are expensive and so I decided to adapt my Myford lathe for the initial testing.

As a 'toe in the water' experiment, I cut and polished a hemispherical dimple in the table of a round brilliant. I will not describe my method of doing this because the finished stone was not as good as one with a flat table. Although it had some oddity value, the stone convinced me to try a different approach. Instead of grinding against a spherical surface, I moved on to using a rod in the lathe chuck. This obviously produces facets that are concave in only one direction, i.e. part of a cylinder. The set up has to be rather different from that used when cutting flat facets because the stone must be held directly above the axis of the rod, if normal faceting is to be mimicked. Moving to either side of this position moves the facets around the stone in a similar way to the use of a cheater.

I made a slotted plate to attach the mast and faceting head of my Facetron to the carriage of my Myford lathe. The other part of the set up was locked down to the lathe bed and provided an adjustable stop against which a dop could be held (see photo). The grinding and polishing rods were held in a Jacobs chuck (safer than a 3-jaw chuck - I do not possess any collets). Using this set up, I cut a modified round brilliant CZ with concave facets replacing flats everywhere but on the table and girdle.

The stone was good enough to make me want to carry on, but the set up I was using had some drawbacks. Firstly, I like to sit down when faceting and this was not possible when working at the lathe. Secondly, unless I rocked the carriage back and forth while grinding the facets, score marks appeared and the profile of the lap was copied onto the surface of the stone. Such marks, which come from surface imperfections of the lap (grinding rod), can be avoided during normal faceting by sweeping the stone across the lap. This remedy is unavailable when cutting concave facets because, as already mentioned, the stone is held in one position above the lap. It is, however, permissible to move the stone parallel to the rod axis, and this can be done by rocking the carriage. The difficulties of using the set up just described convinced me that I needed to make a complete machine on which to use my Facetron head for concave faceting.

As usual, I made a series of compromises and guesses when planning the project. Planning may be too elegant a word for the process that in reality consists of think a bit, make a few mistakes and then think a bit more carefully. As I was building what would quite likely be my only concave faceting machine - not the first of a series of test models - I needed a usable end product. With this in mind, I tested the cam drive for the spindle by making a simple prototype in wood. Surprisingly, it actually worked and gave me an idea of what would be a reasonable speed at which to drive it. To keep down the cost of the project, I used motors and electronics from a surplus store, and as many items from my shed as possible. By deciding that I would not use water as a lubricant, I avoided the necessity for stainless materials and splash shielding/drainage. I also took into account that I could do pre-forming on my flat laps and so would not need to remove a lot of material with my cylindrical laps. It is difficult to give the total cost of building the machine because I already had some of the materials used, and did not use up all of those that I bought for the project. As a rough estimate, I would say that I spent £100 - £150. It took me a lot of time, of course, partly due to the fact that I was working out the design as I went. I should say, however, that I enjoy making the tools with which I am going to work. There is extra satisfaction involved in the same way as there is when working on stones that I have collected myself.

Although the angle of a flat facet should not be greatly affected by sweeping it across a lap, changing the angle at which a stone is presented to a cylindrical lap will make a marked difference. If the spindle is swivelled, it is possible to cut concave facets at an angle to the axis of the stone. The down side is that having a swivelling instead of fixed angle spindle means mounting the motors with the spindle on a plate above the base plate, instead of out of the way underneath.

I decided to have a reciprocating spindle to avoid the need to move the faceting head back and forth by hand. Adding this feature complicated the design considerably but I could see no other way around the problem of score marks. My usual method of trying to ensure that something that I build is robust enough is to use larger components/dimensions than seem necessary. I could not use this approach when building the reciprocating spindle, however, because I did not want to drive a heavy component back and forth. I finally settled for a 3/8in inner (rotating) shaft with a 1in spindle housing to hold the 7/8in bearings.

Before starting work on the concave faceting machine, I had a small amount of experience using normal ball bearings for rotating shafts, but none at all with linear bearings. After a bit of research, I decided to support the 1in spindle housing in bearings made from acetal plastic. Whether this will be a satisfactory solution in terms of component wear in the long term remains to be seen. I made the bearings by pressing acetal rod into the outer casing (stationary part of the unit) and then boring and reaming them to size. I found it necessary to bore an undersize hole in the acetal before pressing it into the housing, as solid rod loosened when bored in situ. To prevent the spindle housing rotating, I cut two slots in the outer casing and screwed pins through into tapped holes in the spindle housing. These pins also provide anchor points for the return springs. The ball races used are the sealed type and I have provided shielding for them. I could think of no way of sealing the acetal bearings but have provided some shielding.

A small spindle and cam provide the reciprocating movement. The camshaft rotates at 250 rpm (fixed) and gives a total linear movement of approximately 0.3in. The power comes from a small capacitor run mains motor, via pulleys of the appropriate ratio. The return springs have to be powerful enough to keep the cam against the spindle even with pressure applied to the lap during faceting. A 1/10 hp 12vdc motor is used to rotate the spindle. Pulleys and a dc speed controller provide a speed range of 0-1400, although the very low speeds are unusable due to lack of torque.

The spindle unit swivels on a metal tube through which pass the supply cables for the motors. The upper part of this tube is threaded so that the unit can be locked at a particular angle. I fixed part of a protractor to the unit (with its centre at the centre of the tube), so that swivelling angles could be measured. The plate and tube have a thin rubber covering in the appropriate places to minimise noise transmission to the base plate and plinth.

I made a 1/2in collet chuck on the end of the spindle to hold the laps. The design of the chuck is a compromise between minimum bearing overhang and its ability to secure the laps. The reciprocating action means that care has to be taken not to have the stone too close to the chuck, otherwise the two may collide when the stone is lifted for inspection (don't ask how I know). When cutting concave facets, lap diameter obviously controls the diameter of the arc cut into the stone. Eventually, I may end up with several sets of laps of different diameters, but decided to start with a set measuring just over 1/2in. This measurement came about because, although I am using 1/2in shafts on my laps, I used 15mm copper water pipe as the cutting lap surface. To hold the steel shaft and copper sheath concentrically in place while soldering them together, I bored an aluminium rod to the respective diameters. A press fit between the rod and copper tube was not possible because of the excessive clearance. The soldered laps were then turned down in my lathe. I made pre-polishing laps from resin/copper, drilled them and then glued them onto a turned down section of metal rod. The resin/copper was cast in a slightly tapered tube that was tapped and sealed at the bottom with a bolt. When the resin was set, a spacer washer was removed and the bolt was then used to push out the resin/copper rod. Unfortunately, these laps did not work as well (on CZ) as the flat variety and so I have been experimenting with phenolic resin rod. It may to take me some time to develop a system that will produce as good a polish as I can achieve with normal faceting. My cylindrical laps have cost less than £1-50 each and so are cheap enough to experiment with.


The support against which the dop is rested when faceting consists of an old micrometer barrel supported on a couple of linked 1/2in rods. This arrangement is very solid and provides adjustment in all directions.

So much for the methods of construction, but was it worth it? Well, I have only cut a few stones so far but the machine works as expected (hoped?) and the results are very promising. I cannot make comparisons between my DIY machine and the commercial variety because I have never used the latter, or even seen one 'in the flesh'. Adding concave facets to a stone gives effects that I have not seen on conventionally cut gems. Simply replacing flat facets on a round brilliant with concave ones is quite straightforward, but when the lap is swivelled matters become complicated. It seems worth persevering, however, because the results can be spectacular. The unpredictability and multiple variables of concave faceting have come as rather a shock because I am used to using Gemcad to pre-plan cutting procedures.

Anyone interested in the suppliers of the parts used in this project can reach me on: