Roubo’s Panel Clamps

When asked how many clamps they have, any woodworker worth their salt usually has two connected answers “1) A lot, and 2) not enough.”  Given the expense of manufactured clamps in our age, consider the relative cost 250 years ago when everything was made by hand.  I would imagine forging a single functioning iron clamp was the better part of a day’s work.

In part for this reason, Roubo and his contemporaries devised inexpensive, high-performance and practical solutions to the problems of clamping, especially for clamping up panels.  I too have followed their lead, and despite having many bar and pipe clamps I find these to be a terrific addition to the workshop.  The engravings are petty self explanatory, straightforward enough that I could crank them out in minutes.

My base stock for these is scrap 2×4, with clear grain if possible.  Laying out a series of square holes, off-set from the center line (almost certainly overkill, but then I tend toward overbuilding everything), I punched the 1/2″ square holes through the 2×4 with my mortiser.

That done I just re-saw the 2×4 on the table saw, yielding two identical halves of one clamp bar set.

Add a group of squared pins to connect the clamp bars and some wedges to tighten on the panel being glued and you are pretty much done.  NB: Before using for actual gluing all the surfaces of all the components should be coated with wax or grease to prevent sticking once the glue dries.

In use I just place half of the bar pair on the bench at each end of the panel assembly and insert the cross pins into the square holes, followed by placing the gluing subjects in place.

The other halves of the clamping bars go on top and once everything is squared the wedges are driven in to squeeze things together.

Finito Mussolini.

Desk Structure – Base Joinery

Cutting and assembling the joinery for the desk base was a chicken-and-egg sorta thing.  Since they were so forcefully integrated, did I cut the moldings into the desk frame first, then cut the joinery, or do it the other way around?

In the end I decided to cut some of the joinery first, like the mortises in the feet, but then mostly did the moldings first and added the upper joinery after that.

For the  tenons I marked and cut the saw-line first with a knife, excavated a bit with a chisel to provide a clean, precise shoulder, then sawed as usual.

The only real interesting exercise for the joinery was the main cross beam at the top of the base, immediately underneath the writing box.  Originally I had considered a simple mortise with screws from the outer face since that would be covered completely by veneers (I think the Senate replica I worked on was made this way), but then decided a better way was to make a haunched blind sliding dovetail.

That sliding dovetail tenon was matched with a dovetailed mortise.  I used a mitered block as a sawing guide and things proceeded smoothly in cutting the dovetail shoulders, followed by a series of cuts to make the waste removal easier.

Chopping out the waste with a chisel followed by a router left a perfect mortise.

The dry fit made me smile.


PATINA Toolfest

Fortunately for all of us afflicted with terminal toolaholism we are not the first ones down this path of compulsion; we stand on the shoulders of giants who got there first and established well-oiled mechanisms to feed their “needs.”  EAIA, M-WTCA, MJD, Superior, all there to provide you with a focus for spending to satisfy the urge (if you sign up for the latter two they will deposit tool listing directly into your email in-box).

Another such group is the Potomac Antique Tools and Industries Association that meets monthly, and once a year holds its mondo crack-house flea market, dealer sales and auction event.  Since this often coincides with the Highland Maple Festival back home I can attend only occasionally.

This year was one such occasion, as our guests for that weekend of the Maple Festival canceled and I was free.  So off to Damascus MD I went with a little money and a shopping list.

The tool flea market in the parking lot begins around sunrise, or so I am told, I generally arrive about 7.30 and find the festivities well underway.

At 9.00 the inside dealer sale opens, and the morning is spent fondling, testing, purchasing, and yakking about tools.

I almost pulled the trigger on this one, but the quality/price point just wasn’t good enough.  But next time I will recount the great deals I did make.

Roubo Joinery Bowsaw Prototype – Frame

With the stirrup system finalized for anchoring the saw plate it was time to move on to the frame.  Armed with some extra-dense 5/4 white oak I dove in.   I wanted to make sure the frame was both simple in construction and beefy enough to withstand the stresses of tensioning such a robust plate.

The overall structure couldn’t be much simpler — two vertical arms connected by a crossbar that was inserted as an unpinned mortise-and-tenon into the arms.  Once I had the dimensions and proportions where I wanted them I used my mortiser to cut the pockets in the arms and sawed the tenons on the crossbar.


Then I moved on to the housings for the stirrups.  It was a simple matter of laying them out against the base of the arm, removing the material so that the stirrup fit neatly, then sawing a slot for the plate to go through.

With the arms and crossbar cut to length and fitted together, and the stirrup housing made, I sawed the curved shape of the arms on the bandsaw.

I assembled everything together just to make sure the parts all worked together before moving on.  I really was pleased with the manner in which it all fit together.  It seemed a little beefy, but I had not put it to work yet.  Besides, it is considerably easier to make elements smaller ex poste than to make them larger.  It made me recall on of my Dad’s favorite quips in the shop, “I just don’t understand this.  I’ve cut it twice and it is still too short.”

With rasps and spokeshaves I shaped the arms to be more congenial to being hand held.  Once it was far enough along to give it a test drive I assembled it completely and strung the top with multiple strands of linen cord for tensioning, found scrap stick (a practice spindle from the writing desk) to act as the windlass paddle, and it was ready for the race track.  I’d added a small vanity flourish at the top of the arm so I just knew it would saw like a banshee.  I cranked up the tension until the plate twanged like one of Stevie Ray’s guitar strings (before he broke it) and lit into a scrap of wood.


And it did saw like a banshee.  Made from concrete.  It was so heavy I actually grunted when picking it up to use the first time.  Somehow I had to hog off a gob of mass or otherwise it was a two-handed-only tool, and I wanted something that could be used with one hand.

Roubo Joinery Bowsaw Prototype – Plate Anchors

Before I got too deep into making the “bow” of the bowsaw I realized I needed to work out the details of how exactly the saw plate was to be anchored to the bow frame.  Given the robustness of the saw plate from Bad Axe and the illustrations and commentary from Roubo I knew this was not a casual thing.  The amount of tension required to make the saw plate perform well was considerable given the dimensions of the plate, so the anchors for the plate had to be able to withstand the force requisite for making it function well.  And, given the likelihood that any user of such a saw as this might well want to swap out the plate from one utility to another, taking advantage of varying plates that could be available.  So, the fitting of the plate to the frame needed to be not only exceedingly stout but also easily reversed or swapped-out.

The clue to the preferred manner of fixing the plate within the bow frame was pretty clearly described by Roubo in his commentary to Plate 12.  While the simplest method would be to simply drive a pin through the foot of the bow frame and the end of the plate, there was and is a better way.  And he tells us how to do it.

There is still another way to attach the blade in the saw, which is to use stirrups, which are pieces of sheet or flat iron that you fold into the form of a “t”, and that you attach to the two ends [of the blade into the stirrup] with a single nail the same way as above. These are then inserted into grooves in each of the arms, which you take care to fasten tight enough to hold them. This method is very good because the stirrups holding the arms from their back sides make full use of the arms’ strength, without which they might split, and you only use one nail to hold the blade at each end because if there were two it would prevent even blade tension, figures 11 & 12.

I took this information/description and started running with a couple of changes.  First, the folding of bar stock into the “T” was not possible with the material I had on hand, although it might be fine with lighter steel flat stock like 1/32″ or 1/16″.  I didn’t have that (nor did the local fabrication shop) and I was too impatient to wait for some to be shipped to me.  Second, the plates supplied by Bad Axe have two bolt holes, which I believe are necessary to house fittings strong enough to tension the plate.  Still, I really liked the concept of a “T” shaped stirrup to affix the plate in the frame.

So instead of folding thin flat stock for this purpose (although I am certainly likely to try it in the future with either thin soft steel flat stock or brass, although I like the solution I came up with for other reasons) I sawed some 1/8″ x 1″ x 1″ angle stock I had on hand.  I cut the length such that the body of the plate would be fully housed but the teeth were exposed and unfettered.

Then I cut off most of one side of the angle stock to reduce the arm of the stock from 1″ to 1/2″ so that when two pieces of the cut angle stock were placed together the configuration would be a “T”.  I used the finished plate itself to provide the layout holes into the stirrup “T” plates.  One half of each pair of the plates was drilled and tapped, the other was drilled and countersunk to fit machine bolts I had in my hardware stash.

Placing the two halves of the “T” stirrup over the pair of bolt holes in the plates, and screwing in the machine bolts, the task of mounting the plate to a stirrup was finished and it was time to move on to the bow frame (this picture is a bit out of sequence and was intended for another purpose but you get the idea).

Roubo Joinery Bowsaw Prototype


One of the many peculiarities of mid-18th Century Parisian workshops and their accouterments as documented by Roubo revolves around the absence of backsaws and panel saws in the tool kit.   These were bow saw and frame saw folks, and they used bow saws for any number of functions at the bench including stock prep and dimensioning to delicate dovetail work.  As for the panel saw, it was only mentioned in one brief sentence and then only in the context of building carpentry.  For whatever reason the French furniture and joinery shops were committed to open frame saws, whether bow saws or sash saws, as their workhorse tools.

Subsequent to my visit to Bad Axe Tools in November and the conversation it sparked about saws in mid-18th Century Parisian workshops, both Mark Harrell and I were noodling the concept of interpreting those saws for contemporary craftsmen.  I think we arrived at the same point more-or-less simultaneously, and I suggested the need for exploring bow saw prototypes based on the saw plate for the Bad Axe one-man Roubo-esque frame saws.  Mark bit and expressed immediate support.  By “support” I mean that he would provide me with some saw plates for me to try out the concept, and would fabricate custom specified plates for me to experiment with.

***A note about such a collaboration: I gladly share my knowledge and curiosities with anyone interested in pursuing the same paths I am on.  I’ve made available all the observations, opinions, and documentation I have, including detailed photographs of my own vintage saws, to anyone who asked.  In that regard I am not showing any preference for Bad Axe, they were simply the folks who asked me to engage in this exploration.***

Since I already had a Bad Axe frame saw and thus the dimensions of their standard saw plate I was able to begin fabricating a bow saw as close to Roubo’s as possible, or at least as close as I could guess.  I grabbed the saw plate, some of the ultra dense 5/4 white oak from my stash (left over from making the Studley bench top replica for the exhibit), and got to work.

The results were instructive.

Next 2018 Barn Workshop Calendar Reminder

The complete 2018 Barn workshop schedule, which I will post every couple of weeks to help folks remember the schedule.


Historic Finishing  April 26-28, $375

Making A Petite Dovetail Saw June 8-10, $400

Boullework Marquetry  July 13-15, $375

Knotwork Banding Inlay  August 10-12, $375

Build A Classic Workbench  September 3-7, $950

contact me here if you are interested in any of these workshops.

Leafy Friends Find A New Home

A few years ago my friend DrDan gifted me with four tropical hardwood tree-lings that he had grown from seed; mahogany, Chinese rosewood, sandalwood, and another that I cannot recall at the moment (padauk?).  Of those four, two remain alive but laboring in our alpine climate in the Virginia Highlands.  Despite moving them indoors in front of a large south-facing window during the winter, they struggled no matter how hard I tried caring for them.  The days here are just too short in the winter since the mountain blocks much of the sun by mid-afternoon, and the cabin too chilly.

That led me to seek out a new home for them, and my longtime friend Tred volunteered his yard in South Florida.  On a recent trip we swung by Tred’s place where these two tree-lings will be planted to mature and flourish.  I’d bet that in a hundred years we could go there and see a fine specimen of a Cuban Mahogany (sweitenia mahoganii) tree on the right alongside a Fragrant/Chinese Rosewood (“huanghuali” or dalbergia odorifera) tree on the left.  I hope they do well in the climate of south Florida, and will go back to visit them periodically.

While visiting we got a nice view of Tred’s shop, which survived the recent hurricane just fine (an ancient live oak in the yard did not fare so well).  I think the footprint is about 75 x 40 feet, and the full bank of north windows provides spectacular interior lighting.

Out front is a cool gnarly chair made from a mahogany stump.  Aesthetically intriguing but lacking in the comfort department.

Gluing Curved Panels With Roubo’s Coopering Cradles

When constructing a bowed or undulating panal or structure, one of the challenges is to decide the manner of achieving the shape.  Commonly there are main four methods employed: sculpting/carving from solid slabs, a la bombe’ chests; glued laminations (home made curved plywood); contour- sawn layers or segments glued together in a stack, sometimes called “brickwork”; or coopering, which involves the assembling of long sticks or boards that are isosceles trapezoids in cross-section (“staves”) with the legs of the trapezoid angled such that the assemblage follows a desired curve.

If you consider the forms of furniture common in Roubo’s time, it is clear that there was a need to produce large numbers of curved panels reliably and quickly.  This calls for coopering.  And, for each individual curvature/radius a unique set-up was required.  Roubo presents two distinct solutions to the problem.   The first, marked “a” in his drawing, shows a pincer/bar clamp and a group of spacer shims to align the individual staves to assure the proper configuration while the glue sets.  This approach strikes me as very finicky, and thus time consuming, to achieve standardized results.

Roubo’s preference was stated unequivocally in his commentary on Plate 102.

Before speaking of gluing curved wood, it is good to enter into the details of gluing those pieces which, although straight along their length, are only curved on their width, like panels that are curved in plane, columns, etc.

For curved panels [in plane], they hardly differ from straight ones. As to the manner of joining them and gluing them, it is only a matter of using a clamp to bring together the joints because when these same panels are a bit curved, the clamps always make them a bit more or less curved than is necessary. You remedy this by putting cauls between the panel and the clamps [bar], which is always placed at the side of the bulge, as you can see in figure 5, side a. No matter what precautions you take, the cauls that you are required to tighten or loosen twist the joints and prevent the glue from working properly. Even when the panels are thin, the clamps bend them and even break them. That is why it is much better to make cradles that you [cut] hollow in the same shape as the panel, which you join and hold in the cradle by means of a wedge, see figure 5, side b.

There must always be at least two of these cradles for gluing a panel and even three for one a bit larger. One should also observe that the angle of the hook “holding the wedges” of these cradles be a bit sharp [acute] so that the panel cannot shift when closing it, see figure 6.

I do not know how to conceal the fact that this method takes much longer and consequently is more costly than the first one because it is necessary to make as many cradles as one has of panels of different curves. These considerations should diminish the advantages that result from using these cradles. These same cradles also serve to fasten [peg] the curved work, which is always better than the clamps, which distort the joints and sometimes break curved crosspieces.


The second approach, marked “b” in the top drawing and important enough to make another detail drawing to communicate the essence of it, makes more sense to me.  In fact I have used this approach countless times beginning back in the pattern shop where we produced dozens/hundreds(?) of core boxes for creating cores for casting pipes.  Although our patterns were permanently assembled, the process is conceptually identical to using a coopering cradle.

The general process is pretty fool proof.  First, lay out the curve with a compass or trammel on a pair (or more) of wooden timbers.  Cut out the concave form.  Since these would be used for gluing it is almost certainly true that these would be slathered with wax or tallow to prevent the curved panel from being glued to the form as it is being built.

Saw as many staves as are needed for the panel being built.  On the edges of the staves plane or saw a slight chamfer so that the staves fit nicely into the curved form, with intimate gluing surfaces aligning between each stave.  This angle can be determined during the layout of the panel, and transferred with a bevel gauge.

Place the beveled staves into the form to confirm the fit.  Make sure the forms are square to each other.

Fill any extra space with loose boards, and pinch the curved panel staves together with individual or compound wedges.    If there is glue between the coopered staves, once that dries the task is finished and after knocking out the wedges the curved panel can be removed to work further.



Measurementally Challenged

It was supposed to be somewhere between “occasional flurries” up to a maximum of 1-3 inches.

Hmmm.  We ended up with nearly a foot.  It’s been a pretty light winter for snowfall, but March is definitely coming in like a lion.