I commend David Ellsworth’s article in the December 2019 American Woodturner on bowl turning. In it he advocates using a single bowl gouge sharpened as shown in Photo. 1 in different presentations. His method works. But is it the optimum method? I admit that I don’t know what the optimum method for turning bowls is, or even if there is one. But I’m certain about the following:

  • Bowls are best turned outboard or using an equivalent. This allows maximum freedom of body positioning and movement, and tool presentation and traversing.
  • The outboard facility must be absolutely rigid. Photo. 2 shows one I’ve recently constructed. Any flexing of the toolrest will be transmitted along the tool blade to the bowl surface.
  • If you’re right-handed you should turn bowl outsides with the workpiece running anticlockwise, and should turn bowl insides with the workpiece rotating clockwise. If you’re left-handed, use the reverse rotations.
  • Turning in both directions brings with it the problem of the chuck holding the workpiece unscrewing. You should use a lathe which has the facility to lock the chuck onto the spindle nose. One is shown in Photo. 3.
  • When turning the outside of a bowl, the shavings’ momentum means they tear themselves free. When turning a bowl’s inside, the shavings’ momentum means they don’t tear themselves free. Therefore tear-out is more likely when turning bowl outsides.
  • The smaller the sharpening angle of the active edge (that part of the edge actually cutting), the smaller the effort required to cut at a specified rate.
  • By far the most important variable in active edge presentation is side rake (Photo. 4). For finishing cuts on bowls an active edge’s side rake should be 70° or more. There are two reasons for this: the edge’s effective sharpness increases (Fig. 1), and the mode of cutting becomes more a sawing one. The reason that the swept-back sharpening shown in figure 1 can be used for finishing cuts is that when traversed with the gouge rolled on its side the furthest part of the active edge is cutting with high side rake.
  • A selection of curved toolrests enable tool overhangs to be reduced (Photo. 5).
  • You should traverse much more slowly during finishing cuts.
  • As cuts approach the intended finished surface they should be thinner. A thin shaving is weaker, and therefore less able to tear wood from below the intended surface.
  • When an active edge’s path of traverse is straight or convex, a hollow ground or flat active bevel is fine. When the active edge’s traverse is concave the active bevel should be convex.

Photo. 1 A bowl gouge ground using a pivoting jig to the shape advocated by David Ellsworth.

Photo. 2 An outboard turning facility. The steel plate supporting the tall banjo is 350 x 300 x 19 mm. The plate is braced with two lengths of angle, and a length of wood.

Photo. 3 The split collar chuck locking facility on Vicmarc lathes.

Photo. 5 Curved toolrests welded to 32-mm-diameter stems.

Photo. 4 Defining side rake The circular pencil line with the double arrow represents the velocity of the wood approaching the active edge.  If the tool’s active edge is at 90° to the velocity of the approaching wood, the side rake is zero.  How much that edge is then rotated is the side rake of the active edge’s presentation.  Here the side rake is about 50°. About 70° is desirable for finishing cuts on bowls.

Fig. 1 Showing how the effective sharpening angle the approaching wood perceives is the active edge’s sharpening angle multiplied by the cosine of the side rake. The first discussion of this appeared on pages 3 to 6 of Michael O’Donnell’s 1988 book Woodturning.

There are many other variables in bowl-turning, some are determined by your particular bowl-turning objectives. Mine are:

  • I use only gouges. I don’t therefore use scrapers.
  • I don’t sand. (After seasoning and during finishing the grain may raise, and I do then wipe the surfaces by hand with 320-grit abrasive paper.) Because I refuse to sand or shear-scrape away imperfections, I ensure that the finished turned surfaces are free from any tear-out and crushing. I also seek to eliminate any rippling, but am often not fully successful. The resulting bowl surfaces are then a memorial to my turning competence or incompetence. Interestingly, wheel thrown pottery with a rippled surface seems to be well accepted by the market. Why not therefore unsanded woodturnings with a rippled surface (Photo. 6)?
  • I typically seek to achieve a constant, and usually thin, rim, wall and base thickness (Photo. 7).

Why have I adopted these three objectives? Because my equipment and techniques allow me to, because sanding is boring and dusty, and because I enjoy the challenge. I don’t ask that you adopt these objectives. They may be inappropriate; for example if, say, you want to minimize production durations. In the remainder of this article I’ll focus on the equipment and techniques which I use to enable me to achieve my three objectives.

Photo. 6 The rippled surfaces of a Chinese elm bowl and a wheel-thrown jug.

Photo. 7 A 185-mm-diameter bowl with a constant wall thickness of 5 mm. An early example of a bowl cut in half to show wall thickness is on page 125 of Richard Raffan’s Turning Wood published in 1985.

Achieving thin rims, walls and bases

The conventional wisdom is that you hollow bowls using a series of arc-like cuts as shown in the top half of Fig. 2. However if you seek thin rims and walls (and ideally bases) you’ll need to hollow in layers typically about 20 mm thick as shown in the bottom half of Fig. 2, and advocated on page 172 in my 1997 book The Fundamentals of Woodturning and on page 65 Ernie Conover’s 2000 book Turn a Bowl. To monitor the wall thickness I use homemade calipers (Photo. 8). Shining a light through the wall is also useful

To turn the bottom thin, if the rim is full the usual method is to grip the bowl by its rim. I usually use the chuck plate system which I invented in about 1980 (Photo. 9). This equipment grips the whole rim, and allows you to compress it, thus imparting increased rigidity to the bowl’s base , and allowing you to turn even deep footrims safely (Photo. 10). Before chucking the bowl I take certain measurements (Fig. 3).

Fig. 2 Tool traverses for bowl hollowing: top, suitable for thick walls; bottom, for thin walls.

To produce thin walls, first cut away the bulk of the waste leaving a still stiff wall, then cut away the remaining waste in a series of layers (shown green). Each of these layers is cut away using of a series of progressively thinner cuts (shown blue).

Photo. 8 Plywood calipers for monitoring bowl wall thickness. The darker one has an adjustable flexible plastic tongue cut from a milk container, and can be used with the lathe running.

Photo. 9 About to chuck a bowl by its rim to turn off the waste from the bowl’s base.

Photo. 10 Left, a bowl on a tall footrim; right, a stem cup. Both pieces are turned from the wattle  Acacia decurrens.

Eliminate tear-out

Not often, but most commonly with green sapwood, if the flute is presented in line with the approaching wood’s velocity, tear-out of parallel grain may be very difficult to avoid. A more horizontal flute presentation probably allows the lower, upward-facing flute surface to act like a hand plane’s chip breaker.

Minimise rippling

If you want to turn the truest possible cylindrical spindle, and avoiding tear-out was of no importance, you should use a wide skew presented at zero side rake (Photo. 11). Similarly, to minimise rippling, the component of the shaving width parallel to the direction of tool traverse should be as long as possible.(Photo. 12) This of course conflicts with the desirability of at least a 70° side rake for the active edge for finishing cuts which greatly shortens the horizontal component. How to achieve an optimum compromise? Increase the flute radius of the active edge. I therefore use big gouges.

Eliminate crushing

If you use a hollow ground or effectively straight in longitudinal section bevel, when traversing along a concave path, to achieve a desirable near-zero clearance angle, the bevel heel must act like a fulcrum (Fig. 4). The likely result is a ring of paler crushed cells. In some woods this damage can descend to a considerable depth. By using a convex bevel whose convexity is tighter than that of the concave path, crushing is avoided. Also when hollowing in layers where the traverses of finishing cuts are short and overlap, it is easier to feather the start of a cut.

I first showed how to grind convex bevels on page 293 of the 1995 edition of The Practice of Woodturning, and later on page 50 of The Fundamentals of Woodturning. To date I am not aware that anyone else has adopted convex bevels despite their simple and obvious logic, although I note the increasing promotion of compromise two- and three-facetted bevels.

Grinding convex bevels is easy. Dress the periphery of a grinding wheel concave with a dressing stick (Photo. 13). I adopt a 35-mm-diameter recess because it is the tightest concave curvature I turn in bowl surfaces. You can vary the convexity of a bevel by the tilt of the gouge blade which has to be manipulated by hand (Photos. 14 and 15). I invariably hone all the tools I use for spindle turning, but because in bowl turning I present active edges with high side rake and the edges are therefore cutting largely by sawing, I don’t bother to hone.

A concave grinding wheel periphery can also be used to grind hollow-ground bevels on gouges (Photo. 16).

Fig. 3 Taking measurements in order to determine other measurements which enable the bottom surface of the bowl to be turned accurately and the intended base thickness to be achieved.

To finish-turn the base with the bowl reverse-chucked, I first turn off a thickness i equal to c – (e +d). I then mark diameter h, blend from that circle to the finish-turned part of the wall, and then turn the area under the dimple concave to a depth of e-f.

Photo. 11 Turning a true cylinder with a skew traversed with a zero side rake presentation.

Photo. 12 The length of the horizontal pencil line is the length of the horizontal component of the shaving width. Here the side rake is about 60°. To increase the length of the horizontal component (desirable to lessen surface rippling) one can thicken the shaving (undesirable), lessen the side rake (undesirable) or decrease the curvature of the gouge flute.

Fig. 4 Archimedes of Syracuse is supposed to have said “Give me a place to stand and with a lever I will move the whole world”. In doing so the vertical force on the fulcrum equals the force exerted by Archimedes plus the weight of the world. Similarly levering an active edge around a bowl surface using the bevel heel as a fulcrum exerts sufficient pressure on the wood’s surface to cause noticeable cell crushing. Engraving from Mechanics Magazine 1824 courtesy Wikimedia Commons.

Photo. 13 About to re-dress a grinding-wheel periphery with a dressing stick. The wooden template has a diameter of 35 mm.


Photo. 14 To grind a bevel with the same convexity as the concavity in the wheel periphery, manipulate the gouge within a plane running through the grinder’s spindle.

Photo. 15 Grinding a less convex bevel by manipulating a gouge in a more vertical presentation.

Photo 16 You can also hollow-grind bevels on (spindle) detail gouges in a concave periphery.

Deep bowls

The sharpening angle of the gouge advocated by Ellsworth is 55° to 60°. It cannot therefore be used in a zero-clearance-angle presentation in the bottom of some bowls with steep walls. I also mentioned earlier in this article that the smaller the sharpening angle, the easier the edge cuts. I therefore use several different bowl gouges for turning bowls (Photos. 17 to 19). I also sometimes use detail gouges. Others do too. I remember in about 1980 watching Tasmanian turner Adrian Hunt who used about four difference gouges.

I also sometimes use a ring tool (Photo. 20) in bowl bottoms, especially with soft woods. A hook or ring tool is merely a gouge with its shaft perpendicular to the usual direction. Because it cannot be resharpened many times, I use it only for finishing cuts. And because it will tend to be used on bowl insides and when traversing the active edge in a concave path on bowl outside, it too should have a convex bevel which in most brands you will have to carefully grind yourself.


Much of the bowl-turning advice one reads focusses on the tool(s) used. I believe that the focus should instead  be on the active edge and its preferred geometry and presentation. After all, you can achieve similar active edge geometries and presentations with different tools.

The equipment and techniques that I currently use (because I continue to experiment I’m likely to modify them in the future) are somewhat more demanding than the norm. Even so I hope that this article will cause you to reassess and perhaps modify some of your current objectives, equipment and techniques.

This article was published in the May/June issue of The Australian Woodworker.

Photo. 17 The gouge I use for roughing bowl insides. It has a 16.5-mm-diameter shaft. I use a similar gouge but with a hollow-ground bevel for roughing bowl outsides.

Photo. 18 The 16-mm-diameter detail gouge ground square across with a sharpening angle of 25° which I use for finishing cuts on bowl outsides.

Photo. 19 Three gouges with 22-mm-diameter shafts with progressively coarser sharpening angles and convex bevels that I use for finishing cuts on bowl insides.

Photo. 20 A ring tool, a HSS version of the carbon tool steel hook tool which has been used for centuries.

Hook and ring tools are gouges with their shafts perpendicular to their flutes unlike normal gouges which have their shafts and flutes in line. A ring tool’s cutting edge must be presented at high side rake. Therefore only the short length of cutting edge opposite the shaft is used.