Notes on Workshop Techniques



9. Spherical Turning

9.1 Methods

The term 'spherical turning' in this instance refers to all non-linear turning rather than just the generation of true spherical surfaces. Various techniques exist to impart a smooth curved surface to the workpiece, and some of these are outlined below:

9.2 Ball-turning attachments

Whilst it is just about possible to turn a ball shape using a combination of cross-slide and carriage feed screws, most would agree that it's far from easy. You can approximate a ball by working to a previously tabulated set of figures designed to place the tool point in the correct position in the x-y axes, intermittent cuts being taken leaving a 'stepped' ball shape which then requires finishing with a hand-graver or file. To make life much easier it's better to use a purpose built toolholder designed to swing the tool point around in a short arc. Many designs exist, and just one of them is that designed by George Thomas (example chosen because it's a particularly good design) and a kit of parts and plans are readily available from Neil Hemingway. The typical use for such a device is in the making of ball handles (you either love them or hate them!), dome-nuts or other decorative features. They are not designed to produce a complete sphere, though it comes fairly close. My own design makes use of a worm reduction gear and is a simple affair that nevertheless produces good results - if somewhat more slowly than the hand-operated swing tool holder.

Worm-drive ball turning tool holder.


With any of these attachments you cannot work very close to the chuck, and clearance for the tool holder needs to be allowed for by having the work some 2" or so out the chuck jaws. This is obviated to some extent by the design of the cutting tool which is cranked to best advantage. In use, it saves time to rough out the shape of the ball beforehand, just leaving final finishing for the swing tool. It is essential that the pivot point coincide with the center of the intended ball, otherwise you might end up with a mellon instead of a ball.

Geo H. Thomas' design for scalable ball handle dimensions.


GHT describes a method for producing ball handles, the dimensions of which are scaled according to the size of the balls. This gives a pleasing symmetry to handles of differing sizes. His technique for producing the handles involves cutting the large ball and shank from one piece of bar-stock, and the smaller ball from a separate piece, the two then either screwed, pressed or Loctited together. When I make mine I make them from one piece, the two balls being formed first on either end of the bar-stock, and a special fixture used to hold the handle for turning the taper on the shank. This fixture consists of a piece of 1" copper bar 1-1/4" long with a deep center one end (7/8" diameter) which fits onto the 60 degree point of my rotating center. The copper will not damage the surface of the finished turned ball, an equally good job could be done with a piece of annealed 1" brass. The other end of the fixture also has a deep center and the larger of the two balls sits in this, the smaller ball being held in a cup holder in the chuck. In this way the shank can be machined to a taper. It's a tight squeeze operating the top-slide handle at the required set-over to cut the taper so the larger end of the two balls is best held at the tailstock end.

9.3 Ball-generating cutters

I have never experimented with the rotating cutter technique myself, but have read about it in an article in Model Engineering some years ago (I'll look the reference up shortly...). Other than the fact that it is possible to restrict spherical machining to only a part of the workpiece circumference I see no advantage compared to other more simple methods. The added complication of using a substantial cutter demands a rigid milling spindle be used, with a more powerful motor than is normally catered for. If you have a proper milling attachment then perhaps the method would be useful to you. Having said that I seem to remember that the article referred to above used a hand-drill mounted on the cross-slide, but you still have the difficulty of making the mounting I guess.

9.4 Cross-slide Link

Deceptively simple at first sight, the practical problems of providing a pivot at the back of the lathe (which really needs to be adjustable in both x-y axes) may prove to be a stumbling block. In principle at least, the method is very workable and will give a good finish on the work. There are few occasions I can think of where it would be necessary to produce a radius on a turned bar of several inches (perhaps crowning pulley wheels). A similar method can be applied whereby a link extends from the tailstock end of the bed to the cross-slide (along the lathe axis instead of square to it), thus concave facing of large radii may be achieved on work mounted on the faceplate or held in the chuck. If you are into making spherical metal mirrors this might be good starting point.

9.4 Profiling Device

This is a very flexible system in that it can be used make a variety of otherwise difficult to machine shapes. All that is required is either a 3D solid model or a 2D outline cut from sheet metal and held in a suitable position such that a stylus mounted at the rear of the cross-slide can trace along the outline. The result will be an accurate copy in 3D solid form. Much information on the practical aspects of using such a device is given in the construction notes accompanying the project for making a Profile Copying Attachment. It is of course necessary to provide a suitable adjustable mounting for the models to be copied, and this can conveniently make use of a taper turning attachment if one is available. Complex shapes such as bells, domes and other non-spherical objects can be marked out on a piece of 1/16" brass sheet and cut to shape. Only a half- axial profile is required and this can often be copied from workshop drawings. The blank to be machined is brought to approximate rough profile using straight inwards cuts using the stylus to measure the depth limit of the cuts. The cross-slide feed screw is then disengaged and finishing cuts applied by setting the tool point at the widest part, pressing the cross-slide inwards with the hand and using the handwheel traverse to trace the outline. I have found this method quite successful.

9.5 Form Tools

An example of a form tool for making handrail stanchions is illustrated below. You can see that the cutting edge is in fact the entire outline of the stanchion, and this is where the problem lies. Large cutting edges lead to chatter and to be at all successful very low speeds are required (slowest backgear at about 35 rpm), and plenty of cutting oil. Even better, a lathe mandrel handle can be used to reduce speed even further. The finish achieved is a reflection of the finish on the cutting edge, and this is unlikely to be as good as can be achieved with a single point tool. To make the form tool shown it is necessary to use either a suitable piece of silver steel filed flat, or a piece of oil-hardening gauge plate. In either case, the hole (dotted outline) would be drilled first and countersunk to provide a cutting edge. The remainder of the outline would be filed or milled, again cutting back to provide about 7 degrees of front clearance. The tool would then be hardened and the final edge put on with a fine slip stone. The tool is best mounted upside-down in the rear toolpost which offers the greatest rigidity.

Schematic of a form tool for generating handrail stanchions.



Return to top

(c) Chris Heapy 1996.

Back to previous page