Lost Art Press

Author: meghanlostartpress

The 18th-century Shoulder Knife

Posted on by meghanlostartpress
Roubo-on-Marquetry1
Fig.1 The shoulder knife was an important part of the ébéniste’s 18th-century tool kit – and yet this simple tool has all but disappeared.

This is an excerpt from “To Make as Perfectly as Possible: Roubo on Marquetry” by  André-Jacob Roubo; translation by Donald C. Williams, Michele Pietryka-Pagán & Philippe Lafargue.

One of the joys of researching the old ways of doing things is that every so often you encounter an amazing “new” way of accomplishing some task. Such is the case with the shoulder knife, an indispensable tool in the ateliers of Roubo’s world. The tool’s utility is remarkable, and I am still discovering new uses for it.

We all have our favorite shop knives; mine is a Swiss chip-carving tip that gets used in many ways. And – like you – I have tuned it exactly to my preferences. Yet, more and more I find myself reaching for the shoulder knife that I made at about the time this book project began.

One of the issues of knife work is balancing the power and control integral to its use. Typically one of the limitations is the amount of force you can bring to the cutting tip, and the precise control you can exert on it. The determining factor is often the amount of handle you can grab comfortably. In fact, that is why my favorite knife has a small blade but a comparatively large handle. Still, I am limited to having only one hand on the handle. A shoulder knife overcomes that because the handle extends all the way from the knife tip to, well, your shoulder. You can obtain great power and control because it allows you to grab its handle firmly with both hands and leverage it off your shoulder.

Roubo-on-Marquetry2
Fig. 2 The long handle of the shoulder knife allows immense control because you can grab it with both hands and brace its handle against your torso.

The shoulder knife has practically disappeared from the woodworker’s tool kit, and to my knowledge only one company supplies them commercially. Making one is fairly straightforward. Although it is a simple tool, mastering it is not so.

The first step in making a shoulder knife is to make a pattern so it fits exactly your upper body’s dimensions and posture relative to the work surface. You can make a template from something as simple and disposable as heavy cardboard. A good starting point is to simply grab a yardstick tip in your hands, drape the stick over your shoulder and make note of the measurement from the work surface to your shoulder. Mark this out on the cardboard, then draw an arc to mimic the curve of your shoulder. Cut this out and compare it to your own body. Revise it until the match is the one you want. I made perhaps a half-dozen patterns until I got what I wanted, and then I cut that pattern out of three or four layers of cardboard and bonded them together to make it sturdy enough so I could get a good feel for its shape and fit. Just to make sure, I made a final pattern out of a piece of 6/4 softwood.

I selected a piece of scrap walnut to make my first knife, and a slab of ancient oak for the second, which is a few inches longer than the first. I used disparate methods for building each.

I made the walnut knife from two pieces of 3/4″ stock laminated together to make setting the blade much easier – even though the final thickness was 1–1/8″. I traced my pattern on both pieces and cut out the shape of the handle. Using a knife and chisel, I excavated a void matching the shaft of a Swiss blade purchased at a woodworking store on the two inner faces that were to be glued together in the final assembly. When the fit was perfect, I glued the whole package together with hide glue, with the knife blade embedded in and protruding about 1″ from the long handle.

For the oak-handled knife, I started with a 6/4 slab, then traced and cut out the shape I wanted. When I was satisfied with the overall shape, I sliced it lengthwise on the band saw. Recycling an old chip-carving blade, I excavated a pocket for the knife haft, then temporarily tack-glued the two pieces back together to shape the handle. (This is unlike the first knife when I assembled the knife and then shaped it.)

Roubo-on-Marquetry3
Fig. 6 The completed shoulder knives. The walnut one is a little shorter than the oak. Each knife is designed to fit its user.

With spokeshaves and files I shaped the handle to my preference, inserted the blade and glued the whole thing back together with hot hide glue. After shaping the business ends and adding compression-fitting brass ferrules, I coated both handles with shellac and wax, made the leather blade guards and called them complete.

My skill at using the shoulder knife is growing, but it is not yet to the degree where it is second-nature. But classical marqueteurs probably used it about the way we would use a scalpel for cutting filigree in paper.

One of the main differences between the manner of creating marquetry between the way I did it for decades and the way that Roubo practiced the art has to do with the assembling the compositional elements into the background. I had previously always sawn them together in fairly typical tarsia a encastro technique, and frankly it is still the practice where I feel the most comfort. But for Roubo and his contemporaries, the elements were often set into the background by scribing the element’s outline into the background with a shoulder knife after the background had been glued to the substrate.

Roubo-on-Marquetry4
Fig. 9 While the shoulder knife allows great control, you first must develop the skill to harness that control in order to follow a line.

This is in great measure the definition of David Pye’s “workmanship of risk.” Careful examination of enough old pieces of marquetry will indeed reveal instances where the knife got away from the marqueter.

— MB

The Plate 11 Bench

Posted on by meghanlostartpress

Plate-11

This is an excerpt from “With All the Precision Possible” by André-Jacob Roubo, translation by Donald C. Williams, Michele Pietryka-Pagán & Philippe Lafargue.  The following text is part of an essay about Roubo’s workbenches written by Christopher Schwarz.

One of the biggest obstacles, downsides and joys to a French bench is the massive slabs used to construct it. Finding wood that is big enough to use without laminating thinner pieces together can be difficult. Laminating thin pieces together to make the thick pieces required for the top and legs is a lot of work without the help of machines.

French-Bench
With any slab workbench with through-joinery, you will experience some shrinkage of the top around the tenon and sliding dovetail. This subsides eventually until it is almost imperceptible.

If you do find stock that is 6″ thick and 22″ wide for your benchtop, it almost certainly will be wet in the middle and prone to distortion. The first French bench that I built used a 4-1/2″-thick cherry slab that had been seasoning in a lot for about five years. The first couple years with that bench were rough. The top shrank at least 1/16″, leaving the through-tenons and sliding dovetails proud of the benchtop.

After planing those flush, the top didn’t shrink much more, but it sagged a bit in the middle during the third year. And now the benchtop is quite stable – yearly humidity fluctuations have little effect on it. The tops of the legs and the benchtop are always in the same plane and the overall shape of the top is consistent.

The French oak that I used in 2013 was likely even wetter than the cherry. For starters, the oak was thicker. And thick material takes a lot longer to dry than thin material. When we first cut into the oak, we used a moisture meter on the wood and found its moisture content in a few places was off the charts. Most places on the bench were about 30 percent moisture content, which is quite wet by furniture standards.

Two months after completing the bench, the top was so wet that it would rust the surface of a holdfast left in a hole overnight.

Like the slab cherry workbench I’d built years before, the oak benchtop shrank around the tenons by more than 1/16″ during the first six months. And the middle of the benchtop began to sag. I flattened the oak top twice during the first nine months in order to be able to plane thin stock on my benchtop.

This begs the question: How flat does a bench need to be? The answer is: It depends on your work. If you plane woods that are less than 3/4″ thick, benchtop flatness is important. I shoot for getting the front 12″ of the benchtop so flat that I cannot get a .006″ feeler gauge under a straightedge anywhere in that area.

If you work with thick stuff or do mostly carpentry, you can be more cavalier.

So if thick slab workbenches are so difficult to find and fussy at first, why bother?

After they settle down, slab workbenches move very little. The same forces that make the top dry slowly also retard its ability to take on much moisture during the seasons (thanks to Steve Schafer for explaining this via Fick’s Second Law, a diffusion equation). After about five years in your shop, your benchtop should be well acclimated and monolithic.

— MB

Two-Foot Rules

Posted on by meghanlostartpress
Rules-1
One leg of this scale has been cleaned with lanolin. The other has been wiped with wood bleach, which lightened the boxwood but didn’t affect the markings.


This is an excerpt from “The Joiner and Cabinet Maker” by Anon, Christopher Schwarz and Joel Moskowitz.

The two-foot rule was the standard measuring device for woodworking for hundreds of years. The steel tape was likely invented in the 19th century. Its invention is sometimes credited to Alvin J. Fellows of New Haven, Conn., who patented his device in 1868, though the patent states that several kinds of tape measures were already on the market.

Tape measures didn’t become ubiquitous, however, until the 1930s or so. The tool production of Stanley Works points this out nicely. The company had made folding rules almost since the company’s inception in 1843. The company’s production of tape measures appears to have cranked up in the late 1920s, according to John Walter’s book “Stanley Tools” (Tool Merchant).

The disadvantage of steel tapes is also their prime advantage: They are flexible. So they sag and can be wildly inaccurate thanks to the sliding tab at the end, which is easily bent out of calibration.

What’s worse, steel tapes don’t lay flat on your work. They curl across their width enough to function a bit like a gutter. So you’re always pressing the tape flat to the work to make an accurate mark.

Folding two-foot rules are ideal for most cabinet-scale work. They are stiff. They lay flat. They fold up to take up little space. When you place them on edge on your work you can make an accurate mark.

They do have disadvantages. You have to switch to a different tool after you get to lengths that exceed 24″, which is a common occurrence in woodworking. Or you have to switch techniques. When I lay out joinery on a 30″-long leg with a 24″-long rule I’ll tick off most of the dimensions by aligning the rule to the top of the leg. Then – if I have to – I’ll shift the rule to the bottom of the leg and align off that. This technique allows me to work with stock 48″ long – which covers about 95 percent of the work.

Other disadvantages: The good folding rules are vintage and typically need some restoration. When I fixed up my grandfather’s folding rule, two of the rule’s three joints were loose – they flopped around like when my youngest sister broke her arm. To fix this, I put the rule on my shop’s concrete floor and tapped the pins in the ruler’s hinges using a nail set and a hammer. About six taps peened the steel pins a bit, spreading them out to tighten up the hinge.

Another problem with vintage folding rules is that the scales have become grimy or dark after years of use. You can clean the rules with a lanolin-based cleaner such as Boraxo. This helps. Or you can go whole hog and lighten the boxwood using oxalic acid (a mild acidic solution sold as “wood bleach” at every hardware store).

Rules-2
Here I’m using a zig-zag rule and a carpenter’s pencil to lay out the cuts on the pine stock for the Packing Box. I dislike zig-zags for this work because they don’t lay flat. They have the precision of a hand grenade.

Vintage folding rules are so common that there is no reason to purchase a bad one. Look for a folding rule where the wooden scales are entirely bound in brass. These, I have found, are less likely to have warped. A common version of this vintage rule is the Stanley No. 62, which shows up on eBay just about every day and typically sells for $20 or less.

The folding rule was Thomas’s first tool purchase as soon as Mr. Jackson started paying him. I think that says a lot about how important these tools were to hand work.

— MB

Classic Orders – An Exercise

Posted on by meghanlostartpress
Classic-Order-1
Fig. 2.4.1. The classic orders dominated pre-20th-century furniture-design books. Above is shown a Corinthian capital.


This is an excerpt from “By Hand and Eye”  by Geo. R. Walker and Jim Tolpin. 

The lifeblood of craft has always depended on knowledge passing from one generation to the next, and I struggle finding words to convey the importance that classic orders played. This is an opportunity to walk in the footsteps of thousands of artisans gone before you, a chance to learn things that cannot be put into words, because this leads into a room in your imagination. The classic orders aren’t about memorizing some nifty proportional recipes. In fact, it’s the furthest thing from recipes. It’s about learning to see. The physical act of drawing challenges the mind to reshuffle and see things anew. Try not to approach this like you’re learning a task or skill; instead just immerse yourself in this rite of passage. Have some fun with it, and let the ancients knock down the cobwebs and pry open some windows in some long-forgotten play space in your imagination.

Classic-Order-2
Fig. 2.4.2. You’ll need a sharp pencil, an eraser, several dividers, a straightedge and a couple fine-point markers to make your pencil lines permanent at the final step.

Grab a clean pine board about 8″ wide and 3′ long for a canvas. If (when) you botch the first attempt, simply plane or sand to reveal a new surface for another go. Pencil in all your lines then, after the entire drawing is complete, go back over your pencil lines with a marker. Think of it like a maze or a puzzle that will change the way you think and make new connections in your imagination. I encourage you, as always, to do this with pencil and not a computer to make sure you get the most direct connection between the portal of your hand and inner eye.

A word about scale. Because you will be drawing a relatively small image, some of the details will be too awkward to draw with a compass. For elements such as moulding profiles or the finer points on the capital, draw a separate detail sketch in larger format with a compass. Once you have completed the larger sketch, go back and hand sketch those details in. You’ll be pleasantly surprised by how well you can freehand sketch once you have the boundary of the form established and little practice on the larger detail drawings. This has real value in furniture design, also. For example, a volute is a delightful form to work into a design, yet because of scale, almost always requires drawing freehand. Generating a volute with a compass will inform your freehand attempts. Also because of scale, don’t attempt to use geometry to draw the entasis (slight convex bulging) on the upper two thirds of the column, just draw a straight taper.

In this drawing exercise you will render a Roman Doric order based on James Gibbs’ “Rules for Drawing the Several Parts of Architecture” (circa 1732). There are five orders – Tuscan, Doric, Ionic, Corinthian and Composite – that exist in an almost endless number of versions and varieties to draw and explore.

Classic-Order-3
Fig. 2.4.3. The moulding at the top of this base is a proportional extension of the base below it. The half-circle indicates that it’sone-third of the base’s height. The quarter-arc shows the linkage between the height of the moulding and the projection of the base.

A few points about communicating proportions using arcs. One common way to show how a proportion relates to another element is to use a half-circle or quarter-circle to indicate a connection. Typically, a half-circle extends a mirror image proportion along the same line. Conversely, a quarter-circle mirrors a proportion from one element to an adjacent element but from horizontal to vertical (or vice versa).

Start by organizing the form (Doric order) into its major vertical parts: the beginning, middle and ending, better known as pedestal, column and entablature. Draw a vertical centerline and establish the top and bottom of your drawing with a pair of horizontal lines, leaving yourself a few inches of margin above and below. Use dividers to step off these major elements and indicate their boundaries with horizontal lines. Once you establish the height of the middle (column) you can determine the module. In the case of the Doric, divide the column height into eight equal parts. That’s the diameter of the shaft near the base and also, therefore, your module. Now – and this is important – draw a small module key in the space below your drawing. Many of the elements that follow will be simple divisions of the module, for example, the column-base height is a one-half module, so having this key handy will speed up the drawing process. To create a key, draw a horizontal line and mark off two modules end-to-end using vertical hash marks to highlight them. Then use your dividers and, through trial and error, step off one module into halves, quarters and eighths. Then step off the second module into thirds, sixths and 12ths.

Classic-Order-4
Fig. 2.4.4. On this Doric order the diameter of the column at the base is the module. That is, one eigth the overall height of the column. Once you find the module, step off a key with simple divisions of the module. You can then use the key to quickly reset your dividers as the drawing progresses.

Start with the largest divisions and work down to the smaller details. Once you have established the overall column height and diameter of the shaft at the base, there are a couple reference lines to pencil in. Note that the column height is divided into thirds and that the lower third’s shaft diameter remains constant while the upper two-thirds curve in gradually – an effect the Greeks called entasis. (As I mentioned earlier, however, at this scale you may want to just render the entasis as a slight taper rather than as a curve.) Also note the use of reference lines: One extends the outside diameter of the shaft above the column while a second extends the outside of the column base below into the pedestal. These lines allow you to step off the horizontal projection of elements in the pedestal and entablature.

Once you’ve established the overall vertical organization, draw in the details of the pedestal. Start by stepping off the vertical organization and then establish the horizontal projection for each part. Most are a function of the module or pulled from an adjacent proportion. Move up to the column and then the entablature.

For certain, you will take a wrong turn or two and have to backtrack and rethink it. It’s all part of learning to see proportionally. When your drawing is completed, you’ll not only have some studies to hang on the shop wall, but you’ll also have created an important mile marker on your journey to becoming an artisan designer.

— MB

How Slow-Grown Oak is Different

Posted on by meghanlostartpress

Slow-Grown-Oak-1

This is an excerpt from “Make a Joint Stool from a Tree” by Jennie Alexander and Peter Follansbee. 

If you have a number of oak logs to choose from, then you can go through the checklist of factors that affect the work ahead. Once you find a straight grained log that’s nice and even with little or no taper, has a centered pith in a mostly round shape, and no twist in the bark – then you’re ready to work that log. But there’s one more thing. You can go even further and look at the rate of growth in the tree’s annular rings. Fast-growing oak has widely spaced annular rings; sometimes up to 1/4″ per year. This timber is exceedingly strong because it has fewer rings, which create a great concentration of the dense latewood that grows in the summer. But the resulting timber is visually distracting. Its radial face comes out looking heavily striped. It can also be difficult to work; it has an uneven texture resulting from the widely spaced transitions between the earlywood and latewood.

Slow-Grown-Oak-2
The most striking difference between the stripey-looking fast-grown oak and the even grain of the slow-grown example is the radial face. That this is also the principal “show” surface of joined furniture makes knowing the growth rate of your log helpful.

The slow-grown oak is more even textured, both visually and for working. While technically weaker than its fast-grown counterpart, slow-grown oak is still well-suited for joined work. This furniture is grossly overbuilt by stress standards, so the decrease in strength is not a factor. The benefit is the consistent texture, ease of working and a closer visual match to the timber used in 17th-century work done in New England. You can’t always get what you want, but if you are faced with two otherwise evenly matched logs, try the one that grew slower. The only thing better than riven radial oak is slow-grown riven radial oak.

— MB