Category Archives: Design

Experiments In Ammonia Bending Wood

Back in the 1960’s, researchers discovered that wood could be plasticized by use of treatment with ammonia gas in a pressure chamber. The ammonia would dissolve the hydrogen bonds between the lignin and the cellulose fibers, allowing the fibers to slip past one another, and enabling extreme bends to be made. When the ammonia evaporated out of the wood, fresh hydrogen bonds would form, setting the bend permanently in place as if the wood had grown that way. In the 1970’s through the present, some experiments have been done regarding the process, but it has yet to meet with any commercial application, mostly due to the difficulty of safely obtaining, storing, handling, and using deadly, corrosive anhydrous ammonia gas.

Meanwhile, woodworkers agree that kiln dried wood is difficult to bend, and that air-dried is the way to go if attempting to make steam-bent wood products. However, many parts of the world are unable to reasonably source air-dried stock and are forced to make do with kiln-dried. Many have discovered that adding a little household grade aqueous ammonia solution (1-3% concentration ammonium hydroxide) to their steam generator helps with bending kiln-dried wood. Ammonia fuming is a common woodworking technique to darken high-tannin woods like oak, and while stronger concentrations work faster, low concentrations have been found to still produce equal results if exposure time is lengthened. This, combined with stumbling across this 2015 study by an Iraqi university, combined to indicate to us that aqueous ammonia could be used as an effective wood plasticizer, significantly reducing the risks involved with working with the ammonia.

We have begun to experiment with treating wood with a 29% ammonium hydroxide solution, and the results are quite promising. Historically snath manufacture has been a costly process, with the steam bending resulting in a very high breakage rate. With good quality ash wood in scarcer supply than ever before thanks to the emerald ash borer, methods for reliably producing the complex bends of a scythe snath are the needed if a traditional or semi-traditional wooden American snath is to remain in production. Now that we’ve developed a soak tank and functioning method for producing a bend that forces compression, we will need to manufacture more clamps to produce the full 3D curvature needed for a finished snath and refine our drying methods to reduce checking due to the rapid evaporation rate of the ammonia.

Forged Traveler’s Celt Mockup

A forged rough mockup of our Traveler’s Celt wilderness tool design. Depending on how it’s rigged up, the tool may be used as an ulu knife, chisel, splitting wedge, bark spud, carving spade, axe, adze, and more. This piece does not represent the final intended form of the piece, but rather an initial testing sample prior to having finished pieces produced. The handle here is made from a black locust branch.



The Balanced Axe

This is a living document and will see continued updates and adjustments as time permits.
Of all aspects of axes, one of the most-touted but least-understood concepts is how they balance. One often hears that an axe has “good balance” or “bad balance”, but what does that actually mean? Ultimately, all axes are balanced, but the question lies in how that balance impacts the orientation of the handle and the bit in use.
Shown here is a comparative analysis of a classic American style axe (a Council Tools Velvicut 4lb American Felling Axe) alongside two 1300g (2.89lb) Italian axes by Rinaldi: a Calabria and Trento pattern respectively. The grip point on the handle is represented by a blue dot, the center of gravity by a green dot, and the axle as the red line running through them.
The axle is the axis through which the tool will naturally balance and rotate from, and is always represented as a direct line running from the grip point (wherever that may be) through the center of gravity. When the center of gravity lies external to the body of the object, it can be easily found by suspending the object by two different points along with a plumb line. The object will automatically pivot at the grip point to bring the tool into balance. Make mental note of the position of the plumb line, and grip the tool from another point and take a reading of the plumb line again. Where the two plumb lines would intersect in space is the location of the center of gravity. Knowing these dynamics, it is possible to easily analyze any axe without the need to know specific weights or using formulas or computation of any kind.
When people talk about long bits having head wobble (see Cook et al.) they usually make the mistake of measuring from the eye, when really they should be measuring from the axle. As a result, the influence of bit length is often greatly exaggerated, especially when combined with misunderstandings of the effects of curved vs. straight handles.

An overlay of the Velvicut (red), Calabria (blue) and Trento (green) axe heads, aligned at the edge apex.

Having a deep bit allows for deep notches without glancing your cheeks and minimizes risk of barking the neck of your handle against the mouth of the notch. However, it tends to move the handle off a single axis, and so one must be more mindful of which hand is delivering force and how when grasping under the head before a swing, since it creates a plane (triangulation) rather than a simple line. This can be easily compensated for in technique, but one must be aware of it, and ideally one would make a handle with an offset neck to bring as much of the handle as possible along a single axis. Many shipwright’s adzes exhibit just such a neck offset because their manner of use (a third class lever) made it necessary for clean results. This requires starting with a large piece of wood with good grain orientation to be sufficiently strong, and so more commonly one finds axes with relatively straight handles, even with curved American handles.
Note that many American axes could also do with a little more offset in the neck, and that with off-axis handles a change in the grip point will change the presentation of the bit relative to the stroke. This can actually be used to advantage in certain situations. Gripping higher on the handle will yield an axle that presents the bit more open during the stroke, while gripping it lower will present the bit at a more closed angle.
offsetcalabriahandle            offsetvelvicuthandle

A Calabria and Velvicut axe with the handles altered for ideal offset.

As another way of phrasing the dynamic, imagine you lopped off the rear face of a sledge hammer to reduce its weight, which leaves the eye now at the rear of the head. It would then need the forward face’s angle adjusted slightly to close its presentation to the target and an offset handle to give it the same handling characteristics it had before (though now with less weight.)
Now that we’ve seen how that dynamic plays out, let’s examine how changing different variables impacts the axle, location of the center of gravity, and orientation of the bit.
In the upper left we have an axe or hatchet of fairly conventional orientation and balance, with a short handle. As before, the grip point is shown in blue, the center of gravity in green, and the axle in red. However, we’ve now added the line running between the heel and toe of the bit in fuchsia to help with visualizing the orientation of the bit with the axle.
In the top center, the bit orientation and grip point are held constant, but the bit has been extended. This shifts the center of gravity forward, which pivots the axle forward along with it. In the top right we have the same extended bit, but the grip point has been changed. Because the extension of the bit took the handle more off-axis, the new axle leads to a more open presentation of the bit. The axe has been rotated to put the axle in vertical orientation to highlight this. If, when the bit was lengthened, the handle had been offset to lay along a single axle, the change in grip point would not have resulted in any change in presentation.
In the bottom left, the axe is identical to the one in the top left, but the handle has been extended, effectively closing the presentation of the bit in use despite the head not having been altered in any way. Lastly, in the bottom center and right the heel of the blade has been moved out and then inward respectively.
 Lastly, to illustrate that axes with offset necks are not just some theoretical armchair speculation, here are some actual examples of American axe heads fitted to properly offset handles, image courtesy of Axes by G-pig, and used with permission:
Manufacturing such handles on a mass scale, however,  would require much larger pieces of wood for blanks, and most axes that have handmade handles running off-axis are generally made that way because it is more expedient and convenient, or possibly because the maker did not even understand these principles and so did not even realize that an offset would be of benefit.
Fortunately, most axes with off-axis handles have a wide enough neck to provide counter-leverage during the two handed span hold. This counter leverage is only needed for a brief moment at the start of the stroke, and is thereafter unnecessary as the hands converge to a practically singular grip point and a natural balance is almost immediately restored. An offset handle, in most instances, merely allows good technique to not require so much mindfulness and is more automatic or “fool-proof”.
The above show the axle of the tool (red line) and corresponding lever arms (blue lines in center figure) of the tool when used with a sliding upper hand (numbered green dots) and fixed lower hand (blue dot.) As previously, the red dot shows the center of gravity, and when the axle and handle diverge the axis along which the hand is sliding is shown by a green line.
On the left you can see that the axe head is balanced by its large poll, and so a straight handle is an appropriate match, with no imbalances imparting torque on the hand in horizontal blows.
The middle axe has the same profile, but the eye has been shifted to the far rear, causing the center of gravity to shift forward in response. This now causes the handle to run off-axis and we can see the lever arms imparting torque on the axle at different points along its length as the upper hand position changes during a sliding stroke. As the hands converge we can see that the lever arm gets shorter and shorter until it becomes essentially insignificant. As such, the infamous “wobble” of a poll-less axe is mostly imparted at the beginning of the stroke, and–while not the ideal–if bearing this in mind it can be compensated for in technique by applying appropriate counter-torque at the start of the stroke and making the slide as early in the stroke as possible.
The third axe now shows the poll-less head with the handle corrected to bring the main length back along a unified axle. This axe will afford the bit size-to-head-weight advantages of a poll-less axe with mostly equal balance to the polled version. One will note that while the axe will now balance properly, the upper hand cannot go as high on the handle as the straight one without running off-axis again. The handle also is trickier to make than in the case of the other two examples and requires better grain alignment to minimize runout.
In case the previous diagrams have been a little difficult to visualize, this diagram simplifies the relationship by eliminating complicating factors. Rather than an “axe” shaped head, we have a simple long, eyeless bar as if the handle were welded to the solid head. The top view shows us the forces at work when the axe is held horizontal. The intersection of the handle’s trajectory and the centerline of the head is shown by the red circle, and the handle treated as massless. A triangle is placed at the point of rotation to indicate the fulcrum forced by the two-point grip. The two sides of this “teeter-totter” are colored to assist in seeing their relative length, and the lines are copied and shown below the head for a clearer comparison.
In the first figure we see a balanced “T” shape, with mass being equally distributed to either side of the center of gravity, and the handle running directly towards it.  This tool is in perfect rotational balance.
In the second figure, the handle has been shifted to one side and the lever arms are now imbalanced, causing the longer end to want to drop. The hollow magenta circle marks the center of gravity and the dotted line indicates where the axe would be rotating from if held by the bottom hand only. With the second hand in play, the forced axle of the red line is where the axe will rotate when held/suspended loosely. However, torque applied along the red line will cause the tool to attempt to rotate around the natural axle of the dotted line.
In the third figure, the handle is now offset to align the handle with the natural axle. The red dotted line shows where the handle had previously run in the second figure. The lever arms are now brought back into balance and the “teeter-totter” is now equalled out again.


Path & Presentation: Understanding The Cutting Stroke

Note: This work is a living document and will continue to see updates as we have the opportunity to write them.

In cutting tools today the most common topics stem around steels, heat treatment, and (in folding knives) locking mechanisms. If you’re lucky, you might see some discussion around cross-sectional geometry and its impact on cutting performance. However, one aspect of edged tool design that seems to almost never be discussed is the impact of the profile of the tool on its optimum stroke pattern, or even how strokes themselves behave. This is a fundamental and profoundly important aspect of edged tool design, and culturing a deep understanding of it can greatly assist in matching the correct tools to their best functional contexts.

Any stroke of a rigid object consists of two variables: the path and the presentation. In the following diagrams, the path is shown as a red line, and for clarity’s sake the heel of the blade is bound to it, riding along it as if affixed to a track.


Presentation is the orientation of the blade relative to the path. A green line is used to represent the path traced by the toe of the blade and the depth of the swath made by the total stroke, though the red line is considered the dominant path of the two. In this case there is no path (just a single point at the heel) and the presentation of the blade is being altered by pivoting it at that point.KnifePresentation

Here the two actions are combined. The heel rides the path while the presentation shifts through rotation relative to the path.KnifeRotationAlongStraightPath

You can see how an object presented as a target to the blade would only be cut by this motion if it existed in the space between the first and second frames of the animation, after which the spine begins to precede the edge, and the edge is pulled away from the target instead of moving into it. This brings us to the subject of edge engagement and stroke optimization.

To begin, let’s demonstrate using this straight-edged knife cutting a target against a flat anvil surface. As before, the red line represents the path of the stroke, while the green line described by the toe helps visualize how the presentation of the blade is affecting the depth of the swath (the area the edge actively passes through during the stroke.) The act of cutting consists of a combination of pushing and sliding forces, in varied degrees. Here we see an isolation of sliding force, without any pushing.

A fully neutral slice. The edge glides along the target medium without any downward force to give depth to the cut.

As you can see, no green line is visible because the edge is running perfectly on top of the path itself, and as a result, there is no depth to the swath. In order for the knife to cut the target, the path and presentation need to be altered to add depth to the swath and place the target within its boundaries.  However, the edge can be considered as fully engaged because its full length is sliding along the surface of the target, albeit with no penetration at all. This unusual situation will become important later as we delve into more complicated aspects of cutting strokes, and will be referred to as a “neutral slice” from this point onward.

Shifting from a neutral slice, let’s switch to the opposite extreme by rotating the path 90°. This happens to switch this blade to what is, from here on out, referred to as a “fully open” presentation, in that the depth of the swath cannot be increased any further. Rotating the blade in either direction would result in the depth of the swath narrowing, but would cause either the toe of the blade or the heel of the blade to be leading the stroke depending on the direction of rotation. Regardless of the shape of the edge itself, the fully open presentation will always create a swath as deep as the straight-line distance between the heel and the most distal point of the edge.

A fully pushing cut. The blade passes fully through the material, but only a small region of the edge is engaged in the cut.

The problem here is that only a small part of the blade is doing all of the work, which–in addition to causing more wear on one region of the blade in repeated cuts–is less efficient than spreading out the work over more edge length. As an edge is effectively a slope, this is much like how climbers tackle otherwise unscalable inclines by zig-zagging up them. It stretches elevation over a longer distance, effectively making it like climbing a longer ramp to the same elevation. So let’s see what happens by altering the presentation of the blade to narrow the swath, bringing more edge to bear on the target.

An angled push cut. More edge length is used in the cut, but the presentation means that the target must be in open space for the cut to be performed, like hanging off the edge of a table.

This angled pushing cut is the same principle employed by the infamous guillotine, spreading the cutting force required over a longer length of edge than possible in a perpendicular cut. However, it obviously produces a notable limitation: you need to have empty space for the tool to pass into. This cut works fine if a target were hanging off the edge of a table, but if cutting in the middle of a broad, flat surface like a cutting board, you cannot force the handle through the board. A different approach would have to be used.

Let’s try “opening” the presentation of the blade relative to the path and trying a pure slice again.

A fully engaged lateral slice. A change in the presentation of the blade and the run of the path allow the whole edge to be put to work. However, the hand has little clearance of the anvil surface and the anvil surface itself is not being used to best advantage.

Now we’re getting somewhere. We’re now able to engage the full length of the edge in the stroke. However, you may notice that there’s now little room for the hand, and if the edge didn’t sit so far forward of the handle we would have to lift the heel of the blade instead of the toe and make a drawing cut to provide this effect. Additionally, the anvil surface is no longer opposing the direction of force, and so isn’t lending a helping hand in immobilizing the target as we cut into it. Let’s try a combination of slicing and pushing forces instead.

A fully engaged cut that combines slicing and pushing actions to give the hand good clearance and support of the target from the anvil surface.

The edge is now fully engaged with good clearance for the hand and the anvil surface is opposing the applied force from the cut, helping to immobilize the target as we cut into it. Chances are that this resembles the action of how you already use a knife in the kitchen, because it’s what you’ve found to provide the best results. Now you know why!

[To Be Continued]

North Star Snath–Early Prototype

An initial proof of concept of the North Star snath. The snath is produced in two parts, and joined by an aluminum elbow. The halves in this case were both the same, but were technically both the upper half, as that was the component I received samples for. This resulted in too strong of a bend in the neck of the lower end, but the production version will have less severe of a curve.

The halves come overly long on purpose, allowing the user to trim them down to desired length. They can then be rotated in the aluminum coupling, allowing the snath to “shapeshift” to best adapt to the user’s preference before being drilled and bolted in its final position. This has the benefit of allowing for a truly one-size-fits-all scalable stemless snath, and allows the snath to pack down for transport or shipping. Note the strong lateral bend of the upper half. This both places the hand in a very ergonomic position, but the end can be used as a grip in its own right when lifting the lay of the blade while mowing, as circumstances sometimes dictate.











Assessing Form Factor In Cutting Tools: The Line Test Method

When analyzing knife and tool designs there are a wide range of approaches that can be used to develop an understanding of a particular tool’s ideal applications. One of these methods that I’m fond of using when initially sizing up a tool is the line test method. Imagining a superimposed straight line over various points of the tool’s outline is a quick and easy method for establishing rough concepts of tool clearance in use. That is to say, it helps you get an idea of how much space your hand will have in use, what regions of the blade will be making contact at what orientations relative to the target, and if any regions of the blade would be prevented from cutting against a broad flat surface. For instance, if you were cutting atop a chopping block of some kind, many forward curving blades would need to be chopping on a block of a certain height and width in order to deliver a blow along the interior of the blade’s arch without the hand striking the ground.  To demonstrate this method, observe the differences between the following lineup when the test is applied.


To begin with, we’ll start the the most basic test–seeing what a line looks like describing a “table” surface, and what the tool would look like laying against it with one point of contact somewhere on the blade and one somewhere on the handle. This is the same as placing the tip of the blade on a table surface and rolling it back until the handle contacted it.


RinaldiZoccaStraightLine RinaldiZoccaStraightLine2 ColdSteelKukriMacheteStraightLine ImacasaPangaStraightLine TramontinaStraightLine ProDynamicStraightLine CondorHudsonBayStraightLine BeckerBK2StraightLine CondorBushloreStraightLine MoraHDCompanionStraightLine SvordPeasantStraightLine OpinelStraightLine

The line test can also be used to assess things like what part of the blade will be in contact with a surface when held at a given angle to it. This is often useful when considering specific task applications where the target will have a certain spacial relationship to the user. I often think of it in terms of if the target will be sitting above or below the elbow, and by how much. The following images show one example of the line test being used to approximate the angle at which the tip contacts the plane surface. However, if you have a particular set of tasks in mind for a knife, imagine the plane formed by a “line of best fit” by your targets and try using the line test at those angles to see if an appropriate region of the blade is being contacted.

RinaldiZoccaTipLine RinaldiZoccaTipLine2 ColdSteelKukriMacheteTipLine ImacasaPangaTipLine TramontinaTipLine ProDynamicTipLine CondorHudsonBayTipLine BeckerBK2TipLine CondorBushloreTipLine MoraHDCompanionTipLine SvordPeasantTipLine OpinelTipLine