I’m in the midst of a spate of Gragg postings, mostly because that’s what I’m doing a lot of and I don’t think posting a steady diet about working my way through a book manuscript is all that interesting.
In building a Gragg chair there are two steps that are immensely time consuming. The first is fitting the rear seat rail into the side units, which serves to unify and distribute all the weight stresses. This takes about 4-8 hours for me to do this one element, it depends on how much magic is in my hands and how well my good eye is working that day. The other is cutting 17 open dovetail pocket joints, two on each of the six short splats and the front rail fittings for the five continuous splats. No way to put lipstick on that pig, they are nothing short of tedious. It generally takes me up to two days to finish them.
Cutting the insides of a dovetailed open mortise is just a pain, or at least it’s a pain by the eighth or ninth one. I’ve used a variety of tools and approaches — a saw, bench chisels, a skew chisel, utility knife, and especially a fish-tail flat gouge. The last was especially helpful but store-bought versions were just not quite right. The splay of the tool tip was too modest to really get into the sharp inside corners. I made a custom one for this purpose a couple chairs ago but for the life of me cannot find it now. Instead I decided to make another.
Looking through my box of derelict or disposable tools I found a 3/4″ Stanley “Handyman” chisel. It was a cheap almost cheezy tool of unknown provenance that still had the factory bevel and back. I recall modifying two of its siblings into 1) a dovetail chisel for small corner dovetails and 2) a small, short 3/16″ mortising chisel (this was before I fashioned my set of mini-mortise chisels out of plow plane irons). The thickness of the 3/4″ chisel steel is laughably thin for typical woodworking, but the thickness was precisely what I was looking for.
With my stationary grinder I created an extreme splay on the fishtail end of the chisel. Being careful and cutting in from the side I managed to avoid de-tempering the bevel. After running it through my usual sharpening routine I gave it a try.
Absolutely splendid!
It is now an indispensable part of my Gragg tool kit, executing the task perfectly. Even better it reduces the time for cutting one of the dovetailed pockets from an hour to 45 minutes or less. Admittedly, since all of Gragg’s chairs are painted thus obscuring the exact nature of this joint I had to reverse-engineer something that I thought seemed right.
When you have to do 17 of them for each chair, ease and speed count for a lot.
Making this chisel was an extremely well-spent 90 minutes.
When it comes to sharpening plane irons I am a sidewinder. I taught myself how to sharpen and this was the way that made the most sense to me as it exploits the natural kinesiology (the study of human motion) of the hand and arm. Straight-ahead sharpening fights the natural motions of the body as you have to constantly adjust the angles of the shoulder, elbow, wrist, and hand. Sidewinding takes advantage of the swinging motion of the arm. All you have to do is establish the bevel angle of the iron and everything else takes care of itself with minimal attention.
The pendulum motion of the swinging arm makes it almost effortless to shape a camber on a plane iron. This one was a new one for a Bailey #5.
My initial shaping of the bevel contour is done with a 60-grit belt on my 50-cent granite backsplash block. My first step is to gauge the bevel angle, which I determine by setting the iron bevel-down on the abrasive surface until the tip and the heel of the bevel are both in contact with it.
Then I start to introduce the curve on the bevel tip by rocking my body forward and backward, and swinging my arm in a pendulum fashion, moving the iron back and forth on the abrasive belt while maintaining the bevel angle. As I am doing this I rotate my hand to push down with my thumb and grind away the trailing corner on the push stroke and press down with my index finger on the opposite corner on the return trip. It really is almost effortless and idiot-proof. For this part of the process I usually wear cloth work gloves as the iron gets really hot, really fast.
After I get the curved tip I want I move on to finish the iron on my 220 and 1200 diamond plates, then finish it off with the 8000 ceramic stone. Sorry for the fuzzy pic, it’s the curse of the “Automatic” setting on the camera. I use the exact same motion for the sharpening stones as I used for the rough shaping on the granite block.
Originally I titled this post “The 5-Minute Camber” but I timed myself as I created this one. I was off by 40%. It took me just over seven minutes to go from start to finish. Shoot, I took longer to do the photography. I’m trying to teach myself to make short “shop tip” videos, maybe this process would be a good candidate.
When we acquired Shangri-La in 2001 I was especially delighted that there were standing outbuildings suitable for a variety of purposes. The largest of these was a small-ish chestnut log barn that had previously housed livestock. For our use as a storage facility the original tiny doorway was a hurdle, it being only about 5-feet high. In short order I sawed the next higher log section to open the entrance to one that was not an undue burden to use, even though it was still not large enough to pass through without ducking.
After opening the doorway I had a piece of chestnut log about 3-1/2 feet long by a foot wide and perhaps eight inches thick. It had been notched with edge tools when originally built so that log scrap had a flat side from the git-go. Some time around then I saw The Woodwright’s Shop with guest chairmaker Don Weber who described what he called “a bodger’s shave horse.” John Alexander’s 1978 Make A Chair From A Tree, was already a long-time part of my personal library, and his infectious evangelism for primitive woodworking inspired me even though my own work was not centered in this direction. Those two paths intersected with the chestnut log chunk and it wasn’t very long before I took that and a bunch of other pieces from the scrap box and made a short length three-legged shaving horse. I used it for many years as a traveling demonstration tool.
During the ongoing video project to document making a Gragg Chair I was vexed at being unable to locate and use this small shaving horse and instead had to use my gigantic one (6′ long), another almost-castoff I bought cheap at auction since the pivot rod housing was busted. I fixed it in an hour. It is a fine tool but not my favorite.
Recently I was rooting around the basement of the barn and came across the remains of the scrap chestnut horse. I do not have any memory of it making the trip from Maryland eight years ago, so I wonder of someone else tossed it into the big rental truck. I say “tossed” because all three legs were missing, and two of them had broken off inside the tapered mortises.
I cleaned out the detritus in the mortises then re-reamed the holes with this sweet tapered spiral reamer I got at a tool flea market for $12. It is beefy, precision-made, and still sharper than all get out. Two or three complete rotations is all it took to get the mortises perfect for new leg tenons.
Using the reamer as the pattern I shaved three new legs from waste stock up in the Gragg workshop/video studio up in the fourth floor attic.
In about an hour of total time I had the new legs cut, tapered, installed and trimmed to length. I flipped it over and it was ready to be put to work.
It is good sometimes to have scraps just laying around.
Well, maybe a concrete block rather than an actual rock.
While recently at our daughter’s house I had a trimwork task that required a nice wide, razor-sharp chisel. Unfortunately the only wide chisel I had was this Stanley one I bought from Big Blue for some demo work, and the dentated beveled edge confirms that usage. (This picture was after a minute of working on it, so it was a lot worse when I started out) Were I back in the mountains I would have had no trouble rehabilitating the edge but at that house I only had a fairly fine sharpening stone. So, I remembered a legend from my early days in the trade, when Pop Schindler had to re-cut some joinery while on-site for a project and simply picked a screwdriver from the installation tool kit and sharpened it on the curb to make it into a chisel.
I had the advantage of starting with a chisel, albeit a really beat-up one. Instead of the curb I carried a concrete block into the basement workshop and started working on the edge.
This picture was after about three minutes; after another three or four minutes with the concrete block I had re-established the bevel edge. I did take another few seconds to flatten the back. Like I said, it was a demolition chisel before this.
Another minute or two on the water stones and I was up the stairs to trim the door. The chisel notched the vintage cherry trim like it was a hot knife going through butter. My only hurdle was the contortions required to get the tool in the right place to work.
Last month my friend Ripplin’ John and I spent a week in the barn working on and brainstorming about our respective ripple molding machines, trying to get a model ready for the show-n-tell of Handworks 2020 (this is before we knew Handworks 2020 was being postponed by the Wuhan Virus). I had made a little progress on my machine since Ripplemania II but he had made great strides with his. During the week his main emphasis was on the lateral “wave” cutting function of his elegant machine while I was simply trying to get my newly designed cutting head to work properly.
Prior to our most recent week together I had also been working on the notion of improving the method and form of the ripple patterns themselves. I tried a number of different methods and jigs but wound up realizing that precision layout and careful workmanship was the key to producing a crisp, precise concave pattern. Throughout our week we discussed this issue and I am now thinking that concave is not the way to move forward, convex patterns may be the future.
As John was assembling his machine, a fairly lengthy process since it was totally disassembled to fit into hi vehicle, I was puttering on mine.
My new cutterhead was now configured with the cutter being positioned at the end of a long weighted swing arm rather than inside a spring-loaded modulating frame structure.
I reasoned that the swing arm was a simpler approach and determined to give it a try. Unlike John I narrowed the scope of my machine to do only one thing, namely cut ripple moldings of approximate 1-1/2″ stock width and 1/2″ thickness. No wider, no thicker. Thus my machine structure was much more restricted than his with approximately zero adaptability.
John had been working on two important evolutionary steps. First, the machine could cut but ripple moldings and wave moldings within the same overall machine structure. Second, that the machine could be mechanized and automated thus making it a more practical device for producing large quantities of moldings. As he recited a truth to me, “Turning the platen drive handle gets old after a surprisingly short time!”
I’ll let you know how we progressed next time.
The conceptual foundation of this saw is the ability to rotate the saw blade within the saw frame. When noodling the design and fabricating this hardware element I considered many options before settling on something nearly identical to Roubo’s.
I ordered a variety of of sizes and configurations of eyebolts from McMaster-Carr to play with.
The eye large enough to accommodate or incorporate the 1/2″ cylindrical rod for affixing the blade was simply too huge for the proportion to work properly. The eye of the correct proportion was made for a threaded rod way too small for the 1/2″ wide blade. In the end I took that smaller eye and drilled and tapped it for the cylinder rod of the right size.
I split the eye bolt with my jeweler’s saw and a #10 blade. I then drilled the cross-holes through which the blade could be attached to this split rod.
One of the last things I did before assembling the whole saw was to de-zinc the hardware parts by soaking them in a citric acid solution (about 2 tsp/pint of water). As soon as the mild acid began to work on the zinc the solution got cloudy. I had to monitor the progress and removed the hardware when I had gone far enough.
I assembled everything to make sure it fit but I did not give the saw a test drive since I did not have the saw blade that would be used for it.
I disassembled it and shipped it off to Wisconsin. I cannot wait to get a production prototype back to give it workout.
With the Roubo Joinery Saw prototype winging its way to Wisconsin I turned my attention to, well, Roubo’s “turning” saw from Plate 12, Figure 5. I used the identical design for the overall bow for the Joinery Saw, but there were more than enough twists and turns along the road to make it interesting.
The term “turning saw” has two particular meanings. First, the saw plate is narrow enough so that it could be turned as it worked its way through the workpiece. Second, the plate itself could be turned in the bow, enhancing the capabilities of the tool.
I began from the premise used in the Joinery Saw, namely that I would be trying to replicate Roubo as closely as possible based on his illustrations and verbal descriptions. As before I used white oak for the wooden elements. For the saw plate I just used a broken band saw blade since Bad Axe has not yet begun to develop theirs.
Thus the two crucial differences were the housing at the end of the bow arms through which the round spilt post passed, and the split post itself.
The bulbous end block was the simpler of the two problems to solve. I glued scrap blocks in place, laid out the bulb, and shaped it to the design. Drilling the hole for the post was simple enough.
Next time – the split post and related fittings.
I think that when creating a production-worthy Plate 12, Figure 3 prototype at the request of Mark Harrell of Bad Axe Tools, I spent the most time and creative energy in working out the problem of retaining the saw plate in the bow saw arms.
My first effort two years ago worked well enough for one guy making one saw, but it was IMHO insufficient for any kind of production run. The amount of work necessary to excavate the base of the wooden arm in order to receive this particular two-piece stirrup configuration made this option a non-starter.
Plus, like the configuration of my original Art Nouveau-ish bow saw frame it bore little resemblance to Roubo’s illustration. His description and illustrations clearly represent a folded “T” shaped fitting-with-pin for holding the saw plate in a slot at the bottom end of the arm. So, that is the direction I headed towards. At this point my only fundamental deviation from Roubo was that Mark’s saw plate had two holes rather than Roubo’s one.
I ordered some 1/32″ steel sheet, unfortunately it did not come in narrow configurations so I wound up sawing it by hand until I had pieces that would fit into my little shear/brake. Once I had the strips in-hand I started the process of figuring out how to accomplish this tricky series of bends. Fortunately I had my sweet little shear/brake from Micro-Mark to help me. Sort of.
I found that the brake’s bending was so crisp that the steel snapped off at that point. Actually, what happened is that I could bend crisp 90-degree corners then had to spread them in order to bend the next set of corners, then bend the original corners back. That’s when they snapped. I eventually did figure out how to get it done (see below).
After the steel sheet gave me trouble, before I figured out the solution I tried some .030 copper flashing I had sitting under the stairs in my hardware store. It cut like butter with my engraver’s hook and folded in the brake without a hitch. (I swear I do not know why Photoshop and WordPress cannot play nice with each other!)
Since copper worked so nicely but was pretty soft I decided to try something halfway in between, .030 brass. Unlike the copper it was really stiff so I annealed it on my hotplate before trying to bend it.
By this time I had hit on the solution to the earlier problem that caused the first steel stirrup to snap. Instead of making the first bends to 90 degrees then bending them open, I bent them only enough to establish the corner.
This allowed me to use the brake for the second bend. Then by hand I could fold the unit closed with a hammer and cold chisel until it was in the “T” configuration.
I worked this system in all three materials and found the solution to be darned near perfect.
I drilled out the stirrup and arm to match the holes in the saw plate and held the entire unit together with binding posts. I apologize for not taking a picture of this assembly all by itself.
With the entire stirrup now passing through the bottom of the saw arm there was virtually no joinery involved; just make the slot for the stirrup and plate, widen it a bit, and slip it all together. Inserting the binding posts completed the assembly of the saw and all it needed for working was the windlass and cord. I used a much smaller windlass bar than my earlier one, about 1/3 of the weight of that one.
I took pictures, dismantled it and shipped it off to Wisconsin.
Now, on to Roubo’s Turning Saw.
Aside from the fact that my first effort in producing a joinery bowsaw based on the Bad Axe saw plate didn’t look like Roubo’s, only one of several features rendering this prototype useless to Mark in designing a production run, it was way too heavy at over six pounds. That extra weight was fine for when the saw was engaged but sure was a nuisance when trying to maneuver it in between cuts.
So, I went back to the starting point and recut all of the wooden elements to both be slimmer and look like the saw Roubo illustrated in Plate 12, Figure 3. In the end the only dimension that remained unchanged as the length of the saw plate. Everything else was thinned, widened, thickened, or narrowed as I felt necessary.
The resulting weight reduction of almost 25% was noticeable immediately.
I literally enlarged this image to almost life size when creating the template. Again I used some of my stash of clear white oak left over from the fabrication of the Studley bench top for that exhibit five (!) years ago.
The cutting and configuration was straightforward in the gross sense, but I spent a lot of time getting the template for the end pieces just so, reflecting the illustration Roubo created for the tool. This is a tricky proposition, as the proportions and details of tools are not always perfect (see Plate !! for confirmation) but it had to be closer than the Art Nouveau-ish version I created for the first one. Once I was happy with the shape and proportions of the arms I made a solid template from which I could make as many tracings as I wanted.
Of particular interest to me was the juxtaposition of the stretcher to the end arms; the only way the saw made sense in the flesh was to have the stretcher be notably thicker than the arms. So that is how I made those three pieces. In the end I made the stretcher about 1-1/16″ and the arms 3/4″. I fussed over this detail for a couple weeks.
Notching the arms to receive the stretcher was also an issue as I had to make sure the strength of the arms was not compromised by too large a mortise. I accomplished this by making sure the loose mortise-and-tenon assembly was quite thin, just a bit over a quarter inch and shallow at a half inch. Since the joint’s only purpose was to keep the pieces in alignment that worked out just fine. All the stress was in compression so the tighter the windlass was turned the more solid the bow structure became. Unless, of course, it broke. Hence my fussing.
But that was not the worst of it as noodling and fabricating the stirrups for the end of the saw plate was way more problematic.
Stay tuned.
Some time ago I was asked by Mark Harrell of Bad Axe to build production-able prototype versions of the bow saws Roubo featured in L’Art du Menuisier, Plate 12, Figures 3&5. He was considering adding them to their product line and I was delighted to collaborate with him as I would be with any tool producer.
My first effort two years ago resulted in some success but some failure, so in recent weeks I have been trying to extract more of the former and expunge the latter. My first attempt yielded a prototype that was unsatisfactory in the important aspects of the tool; 1) it was way too heavy, 2) did not look anything like what Roubo described and illustrated, and 3) had plate fittings wholely unsuited for manufacturing (yes, I know I spell wholely different that the dictionary; the dictionary is wrong). So, I went back to the bench and started all over again.
In the coming days I’ll walk you through my successful attempt at Version 2 of the Joinery Saw..
Recent Comments