This unpretentious tool, for about six bucks, is surprisingly useful to modify concave curves on fairly narrow work such as table legs. I use it for fast, corrective takedown if my bandsawing has wandered off the layout line, or if I’ve changed my mind about the curve after having sawn it.

Its molded plastic handle and snap-in cutter certainly do not exude cool-tool cachet, but the varying curve of its sole, flatter toward the toe, steeper toward the handle, is quite effective. It cuts on the pull stroke. However, it tends to tear the wood and leave a surface too ripped up for efficiently transitioning to refinement with finer tools.

To remedy this problem, I hone the cutting face with a fine diamond stone. While this sharpens the cutting teeth, it has the more significant effect of limiting their depth of cut. This does make it a somewhat slower tool, but the resulting surface is considerably improved, so the whole process of refining the curve is actually faster.

The macro photo below shows the honed teeth, which have been lowered relative to the peaks of the “waves” on the cutting surface. (The cutting edges are facing upward. The honed area is the silvery line visible on each of the two teeth near the center of the photo.) This is similar to the working of an “anti-kickback” router bit, in that it limits the depth of bite of the cutter. If you go too far with the honing, the teeth will have too little bite or won’t work at all, so proceed gradually with the modification and test the tool as you go.

The tooth lines are angled to the length of the tool, pre-skewed, in effect, so I find it works best after this modification by pulling it with little or no additional skew.

A great tool it is not, but it does the job decently well. I wish Stanley (or Microplane) would make a longer version, say four or five inches, retaining the varying-radius curve, with room to place a second hand on the front of the tool. Not currently made, but perhaps a Shinto rasp in such a profile would be useful, or larger rasps in the style of curved ironing rasps, both with a knob at the toe for greater control and power.

Other options for a tool that is curved along its length and flat across its width include: metal and wooden compass planes, Auriou and Liogier curved ironing rasps, shop-made curved sanding blocks, and a new flexible rasp by Liogier that they call “The Bastard,” which I have not tried.

The Stanley Surform Shaver now comes with a bright yellow handle. Nuance that.

As I said, I’ll use whatever tool it takes to get the desired result for a particular curve in a particular wood. So let’s take a look at the available players and which make the cut (pun intended). Most of the game is won or lost on concave (inside) curves; the outside curves are easy.

Spokeshaves perform well on relatively narrow work with cooperative grain, but they can disappoint on highly figured woods, even using a skewed attack. The round shave sees almost all of the action, while the flat shave spends most of the time on the bench because I generally don’t find it has much advantage over a block plane or other flat tools on gradual outside (convex) curves. It’s all in the wrists.

The convex side of a half-round rasp is a good workhorse but has some weaknesses. If it is held at an acute skew, such as for steep inside curves, the teeth start to function ineffectively as tiny knives slitting along the grain, but if the tool is pushed more across the work, tearout results at the far side. Also, the tool is really working the curve at different points, and thus possibly different radii, at once.

So, for more control on gradual curves, I call up the Auriou curved ironing rasps (fourth and fifth from the left in the photo). They have an excellent reliable feel on the curve but lack speed, so they are not for hogging off a lot of wood.

The compass plane, which is a shaping plane in my view, was covered in two earlier posts, but a different twist on curved soled planes deserves mention. As discussed in the previous post, the sole must be set to accommodate the steepest portion of a chosen length of inside curve, so a given setting is approximate at best. Thus, a reasonable alternative to an adjustable compass plane is a set of a two or three wooden fixed curve sole planes, vintage or shop-made.

Hunting on vintage tool sites will turn up a few wooden curved sole planes with an adjustable toe piece to accommodate different inside curves. I have not tried one but I wonder if any readers have.

The little Lie-Nielsen spoon bottom plane is a different type of player but performs well despite its lack of size.

A card scraper is a good player if used in the proper role – great for smoothing curves but poor for shaping them because it simply rides whatever curve it encounters.

Underestimated but well within anyone’s salary cap is the curved sanding block. Customized in length, width, and curve, they can smooth curves but also can be designed as pretty fair shaping tools using coarse grit paper.

Speaking of sanding, the Ridgid oscillating spindle sander is very handy because it can be set up as a sideways belt sander or as a simple spindle sander in a range of sleeve sizes.

Back to the opening photo, the humble Stanley Surform shaver, a product of a program whose glory days are past, is an effective rough hogger. Surprisingly, it can be tuned to perform with a bit more finesse, as will be discussed in a future post.

The Allongee style gouge, #5 sweep, 38 mm, is a good reserve player for cross grain hogging in wide curved work, much like a freehand scrub plane for curves.

A couple of other tools are not in the TFC Team photo because, though they arrived with promise, were cut after tryouts. I found the flexible curved float file to be slow and awkward, and did not live up to the reputation of its flat cousins. The same was so for the Microplane flexible insert for a hacksaw frame. These are just this coach’s calls; you might like them.

When there is a simple curve in one plane, as shown below, to be made in multiples, I go to the pattern routing game plan, as represented by the pattern/flush cut bits on the right in the top photo.

Of course, just about all curves start with good, accurate sawing, which usually means a well-tuned bandsaw. That’s the fan base behind the whole team.

Next: nuancing the Surform Shaver.

The compass plane is an effective tool when thought of as a jack plane for curves. It is mostly a shaping plane, where the shape is a curve, not a flat surface as for a regular jack plane. It is mostly fantasy to think of the compass plane sleigh riding over varying-radius curves spilling out long silky shavings.

The most important step in efficiently forming true curves in solid wood is to saw consistently to a good layout line. However, there will inevitably be some lumps and bumps in the sawn surface, so the curve must be “faired” to make it pleasing.

Key to using the compass plane is that the sole must be set a bit steeper than the work piece for concave (inside) curves (see photos above and below), and a bit shallower than the work piece for convex (outside) curves. Furthermore, the planing should proceed into downhill grain, that is, with the grain, which means you may have to turn around often. Outside curves are generally easier to negotiate, and shallow ones can often be worked well with a flat-sole plane such as a block plane.

This all sounds good except that most of the interesting curves in woodwork have a varying radius (i.e. are not circular) and some reverse from inside to outside. So that means a single sole setting is ideal for only a relatively short length of curve. As a practical matter therefore, for inside curves, the sole is set to accommodate the steepest portion of a length of curve that you choose to work in which the radius does not vary too much. It is a matter of feel and judgment. Which is to say that these planes are not very practical for abruptly changing curves.

Because we want the plane to remove lumps and bumps, the shavings, especially early on, will mostly be short, and the cutting edge will engage and disengage the wood as you take fairly short strokes. Then as the fairing proceeds, the shavings will lengthen; that is, if the planets are aligned.

The compass plane is capable of fairing a nice gradual curve in the right circumstances and wood. Remember too, it can handle wide surfaces that are difficult to manage with spokeshaves and rasps.

Also, it is often helpful to initially remove some of the roughness of the sawn surface with a rasp (not a sanding block) to avoid a very rocky ride in the early stages of planing.

The anatomy of the compass plane does not permit it to transmit the wood-hugging stability that we expect from a good bench plane. I like to make the ride firmer and improve my feel of the plane’s interaction with the wood by placing my right hand as low as possible at the heel, sometimes with my fingers touching the top of the sole plate. Meanwhile, the palm of my left hand hugs down on the nose as my thumb reaches down onto the sole plate.

Ultimately, it’s not about the tool, it’s making the product come out the way you want it that counts. I’ll use whatever tool it takes to produce the desired curve in a particular wood. Sometimes, that’s the compass plane.

Next: scouting reports on each player on the tools for curves team.

I enjoy incorporating curves in my work and so have explored lots of different tools and methods for shaping, refining, and smoothing them. Years ago I used a new Record #20 compass plane but then got rid of it. The problem, however, was mostly in my approach to the tool. I’ve harbored mixed feelings about the metal compass plane since, but have finally come to peace with the beast since owning this vintage Stanley #20 for the past year.

I’ll get into the function and handling of the tool in the next post, but here I will detail its tuning and modification.

This #20 was manufactured sometime in the years 1933-1941, as best I can tell. It arrived from the seller fundamentally sound – no cracks in the main casting, working sole adjustment, and japanning in excellent shape.

These planes need all the help they can get with chatter dampening so I replaced the thin Stanley blade and chipbreaker with a hefty Hock A2 cryo blade (#BPA175) and chipbreaker (#BK175), 1 3/4″ wide. I prefer the durability of A2 for the way I employ the #20, which I’ll discuss in the next post.

Patrick Leach notes that the #20 (and #113) have unique chipbreakers so I carefully checked the diagram on Ron Hock’s site. The critical parameters are the chipbreaker’s slot-to-edge distance and the length (the short dimension) of the slot. These worked out beautifully. The #20’s advancing fork engaged the chipbreaker slot very well despite the increased thickness of the blade-breaker set. Also, the disc in the lateral adjusting mechanism nicely engaged the blade slot.

Unfortunately, the thicker blade-breaker set caused severe pleating of shavings, and bad clogging. To remedy this, I disassembled the sole by knocking out the pin at each end of the sole and freeing the dovetailed connection between the sole and the body, then filed the forward side of the mouth to widen it (barely advancing into the row of pins that bind the flexible portion to the dovetail block), and added a slight forward angle to the throat, all to make more room for shavings to escape. It also proved necessary to round over the crisp bevel on the back of the chipbreaker.

This solved the clogging problem very nicely, and the beefy A2 Hock set outperforms the Stanley set! Suprisingly, I have not found the wider mouth to be a problem for planing curves. 

The frog needed minor truing. I reattached it as deep as it would go, then, after reassembling the sole, filed the landing below the frog to be mostly level with the frog to increase support for the blade.

I flattened the sole around the mouth with a diamond stone. There is no point in flattening beyond the vicinity of the mouth in a compass plane with its flexible sole. A general clean and lube, and touch ups with a file here and there, finished the job.

Consistent with the purpose that I assign to this plane, I sharpened the blade with a medium camber and made sure the corners would not catch the work piece.

There are other options in metal compass planes including a Record #20, Stanley #113, other variants of the #113 style, and current versions of the #113 by Kunz and Anant.

The metal compass plane is a bit of an odd animal and one must come to terms with it, as will be discussed in the next post.

Is a thinner kerf saw more accurate? Does that make a skinny saw better? After all, we associate thin with accurate, such as thin pencil lines or thin gradations on a rule.

Accurate sawing means a clean, neat kerf that consistently splits the layout line, with the kerf in the waste wood. This comes from teeth of appropriate design and pitch for the task that have a small, consistent amount of set. Further, the saw plate must be produced straight and stay straight throughout cutting. The sides of the teeth should also be cleanly free of burr.

The sawyer must employ good mechanics, aided by good tooth geometry, saw balance, hang angle, and other mechanical factors. With all that on your side, you can physically sense true cutting, split the layout line, and visually monitor the progress with accuracy.

But is thinner kerf width, per se, more accurate? I don’t find this to be so. As an example, my .012″ plate Japanese rip dozuki holds no advantage in accuracy by virtue of its thinner plate over my .018″ plate Western dovetail saw. In fact, because of other factors, I find the latter is more accurate. Yes, this is an apples-to-oranges comparison but my eyes and hands can tell that factors other than plate thickness are the deciding ones in determining relative cutting accuracy between these saws.

Similarly, my carbide tip bandsaw blade makes a considerably wider kerf than my steel blades but it cuts more accurately. We also don’t think of thin kerf table saw blades, whatever their other advantages, as being more accurate than standard kerf blades.

Now, I’m not saying get a dovetail saw with a .042″ plate, nor that thin plate saws are necessarily bad choices. I do think confusion arises in assessing and choosing saws because thinner plates are sometimes associated with other factors that promote accuracy such as nicely set fine teeth, or comparing a good quality thin Japanese saw with a poorly made thicker Western saw.

Within limits, however, one ought not assume that, all else being about equal, a thinner plate is more accurate. In some cases, contrary to the assertions of some vendors, it may be less accurate.

There are many factors that produce an effective, accurate saw. You may, for various reasons, prefer a thinner plate saw. But I suggest don’t get charmed by skinny saws. Rather, consider the whole picture, I’d say, and see how the saw really saws.

In saw descriptions and discussions, there is often the implicit assumption that a thinner saw cuts proportionately faster than a thicker saw. At the risk of setting this up as a straw man case, the assumption goes that, as an example, with all else being theoretically equal, a .012″ plate will cut twice as fast as a .024″ plate. Further, it follows that a thin-kerf saw has, within limits, this distinct advantage, assuming that its other sawing parameters can be controlled to maintain good function.

This would be analogous to a 24″ wide swath of snow being twice as hard to push as a 12″ swath, all else being equal. However, I don’t think saws work like that!

Let’s think about what a saw tooth does. A rip tooth cuts and plows the wood at the bottom of the kerf. In this, kerf width is probably roughly proportional to the effort, and thus inversely proportional to speed. At the sides of the kerf, the tooth shears the wood, and there the task is approximately the same regardless of kerf width.

The crosscut tooth severs the wood fibers at the sides of the kerf where, again, the task is approximately irrespective of kerf width. At the bottom of the kerf, where it is a lesser task of shuffling away the broken wood, the work is probably about proportionate to the kerf width.

Thus, in both cases, especially crosscutting, this simple idealized analysis suggests that, all else being equal, twice the kerf width does not mean half the sawing speed. It is not like pushing snow.

In reality, all else is never equal, of course, and the dynamics are surely more complicated than described here. Nonetheless, this way of looking at it at least gives some basis to explain my real world observations using many saws that, within limits, thinner kerf saws do not seem to give a proportionate advantage in cutting speed over thicker kerf saws.

Again, my argument is against this as an assumption that may be made by some when comparing saws. This is applicable in comparing among Western saws, and generally comparing Western with Japanese saws.

Further, as plate thickness is reduced too much, especially in Western saws, disadvantages ensue. Among these, depending on other design parameters, is a tendency to distort in the heat and action of sawing. Also, energy intended for cutting seems to get wasted in vibrating the skinny saw plate, somewhat akin to action of a thin or poorly supported plane blade.

In summary, skinnier is not as attractive as it might seem. It is important to look at the whole picture when choosing saws.

Next: Ah, but is the thinner kerf saw inherently more accurate, all else being equal? Does that make skinny better?

Sometimes in woodworking, especially in unusual constructions, there comes a deflating realization that things would be a lot easier now if a different turn was made several steps earlier in the process. It’s not an abject mistake but it is time for the fix-it crew.

Here are several notable tools among the many whose modest bearing belies their performance in the clutch.

The pair of left and right-handed crank-neck 3/8″ skew chisels get into vertically and horizontally restricted areas to remove small amounts of wood that are preventing parts from fitting well. Widely available and inexpensive, I find them more useful than straight skew chisels.

For precise paring in more accessible locations, long paring chisels allow much finer control than a regular bench chisel. Using the dominant hand at the back end, the tool’s length allows fine control of the attack angle of the edge to produce clean cuts.

The Japanese azebiki saw has short, curved rip and crosscut edges. It’s great for starting cuts on a flat surface, and for getting into restricted areas. I adjust the set to a bare minimum on my saw.

The very flexible .020″ hand scraper is easier to use than thicker scrapers to clean up localized surface defects that can arise in the late stages of building from planing tearout or handling dings.

I’m almost embarrassed to say how often I use the little 1″ x 2″ .016″ mini scraper for fix-ups. I keep some edges with a hook and some without, and use it pulling, pushing, angled, skewed, or even flat against a surface to solve all sorts of problems.

The low-profile ratcheting driver is another tool that I might be lost without. It accepts 1/4″ hex-shank bits and can be used for driving and, patiently, for light drilling when necessary. With this tool and with a right-angle attachment for a power drill, it is very handy to have shorty drill bits available.

Of course the most important tool in a jam is the one on your shoulders. Pause, step back, collect, think, and be optimistic – there’s probably a solution!

The Veritas Shooting Sander uses the principle of shooting – a guided vertical cutter is pushed to engage a work piece that is stably oriented by a surface and a fence – but uses sandpaper instead of a plane blade as the cutter. It’s simple and useful.

Though it certainly is not intended to replace shooting with a plane and a good shooting board, I’ve been so far finding it handy for odd-shaped parts that cannot be fully backed by a conventional shooting board fence, and for small parts.

As we would expect from Veritas, the tool is well made and thought out. The accurately made anodized aluminum extrusion body and the nifty adjustable wooden handle are good reasons to forego a shop-made attempt at this low-cost tool.

The shooting board I made for it is straightforward but there are a few fine points. The base is 3/4″ MDF, 23″ long. The work surface is 7 3/4″ wide with a nice straight edge against which the sander runs. The track for the sander is 2 1/8″ wide with a 1″-wide outer guide rail.

The work surface must be elevated at least 9/32″ above the track surface for the sandpaper to meet the lowest part of the work piece. I made the work surface from two pieces of MDF (just what was handy) for a total thickness of 11/32″, which gives a little margin for error when applying the sandpaper to the tool. That is, the bottom edge of the work piece is sure to be within the width of the sandpaper, even if I don’t apply the PSA paper to the tool perfectly accurately.

The fence is about 1 3/8″ high, screwed down 3 1/2″ from the end of the board with slightly oversized clearance holes that allow fine tuning for squareness.

Break in the shooting board just as you would for a plane shooting board by running the sander along the edge of the work surface so that a tiny width of sandpaper, say 1/16″, cuts a miniscule rabbet along the edge of the work surface. Then screw down the 1″-wide guide rail on the outside of the track so it is snug against the sander for the full length of the track.

A generous amount of oil-varnish finish toughens the MDF surfaces. Finally, I waxed the track. It all works well.

1 1/4″ wide adhesive-backed sandpaper strips are used for this tool. These are most economically made by slicing 2 1/2″ Klingspor PSA abrasive roll paper down the middle of its width. The paper strips that Lee Valley supplies are Klingspor’s.

After removing the first piece of sandpaper from the tool, I cleaned the residual adhesive off the tool with a citrus-based remover, but did not then clean off the slightly greasy residue of the remover. I found that subsequent sandpaper stuck plenty well enough and left hardly any residual adhesive when removed.

The tool is very easy to use but there are a few caveats. The sandpaper leaves grooves that are surprisingly deep for a given grit. That is simply because the tiny grits on the sandpaper are running in the same tracks over and over, unlike with regular hand sanding where the slight variations in movement erase most of the tiny grooves.

The work goes slower than shooting with a plane, especially since sandpaper seems to cut slowy on endgrain. Also, the thickness (height) of the work piece is limited to just under 1 1/4″.

The tool can be used ad lib to sand odd angles without using the fence by holding the work piece very firmly and offering its edge at the desired angle (such as indicated by a scribed line) to the sander running in the track.

All in all, this so far has been a worthwhile addition to the shop. My sense is that it will increasingly become a valuable quick “problem solver” tool that I’m very glad to have.