Trammel Points for Circular Work in Carpentry

Need to lay out an arch? You’ll need some trammel points and a beam of sufficient length. After a survey of what was available to buy I decided to make my own.

Why three? Most simple arches can be formed with just two. Approximations of ellipses, the so called “bricklayers arch” similarly use two. If you want to draw true ellipses though, you’ll need three. So go ahead and make three and thank me later, haha.

I recently had a chance to read both Hasluck’s “Carpentry & Joinery” and Collings’ “Circular work in Carpentry and Joinery”. They’re both excellent texts and Hasluck’s stuff is in the public domain, so go treat yourself to a bit of free knowledge if you care to understand a bit better what modern carpentry has been reduced to.

The wedge and point sit in a gentle sliding dovetail, which keeps things from popping out of place. Both are contained within the long grain of the vertical clamp side. I used a thread box to cut the threads for the screw and tap the threads. Really, they’re just made out of a little scrap I had lying around. The points are 1/4″ mild steel.

I was asked recently to cut an arch-top molding for the inside of an entryway door. Its no small feat, really the domain of specialized shaping machines in custom mill shops. I’ve seen a tilting router setup that can do some decent profiles in a home shop, but how did the carpenters of old make this stuff before shapers and profiled knives? Radiused hollows and rounds of course. A look at carriage makers planes is very instructive.

The layout for the molding I have to cut wasn’t one listed in the books I have, but follows a standard model that I was able to deduce once back home in my shop. Fun stuff!

As an aside, have you ever noticed how much Mentos “The Fresh Maker” candy look like bi-convex go stones?

Roubo Frame Saw Eat Your Heart Out

There is an alternative for woodworkers to the Roubo style frame saw for medium scale resawing and sawing of smaller boards. Introducing the Woodmizer -1, haha. Its a 450mm rip saw with about 3tpi, made by MIG welding a tang to a western panel saw.

I’ve felt the need for a smaller saw for ripping boards. Especially as the piece of timber that you’re milling becomes lighter, a full scale maebiki-oga becomes very hard to use, simply too much saw for the cut. While I would love to have a frame saw I feel that a pull saw offers several advantages in comparison.

First, a pull saw is much lighter and easier to use when the orientation of the cut is less than perfectly plumb. Second, the tooth line is shorter, faster to sharpen, and fits the natural length of stroke for a single sawyer. Frame saws have a definite advantage of allowing two person use, which creates good efficiencies, but most of us are working alone, and three feet of saw blade is a stretch to use evenly.  Third, there is no problem running a pull saw through a wide cant, where a frame saw can only cut material that fits between the frame. The wider the frame the more skill would be required to make sawing smooth and comfortable.

As it turns out I cut  ridiculously aggressive tooth geometry on this saw. Over the course of trying it out I’ve basically tuned it back to a neutral rake angle by back beveling the teeth with a micro bevel. Any saw will need tuning to the specific species and quality of wood you are sawing. In my case the limiting factor in milling pine is the knots. The fastest saw is not the most aggressive saw, its the smoothest saw that you can keep running with ease and rhythm in the cut. If you waste a lot of time hung up pulling through knots in tortured jerking fashion its not a fast saw. I had to push the tooth set way way out, but the kerf is still about 1/16″.

MIG welding butt edged sheet steel is not easy. My first try at the tang weld quickly failed. Forge welding would be much preferable, but using MIG meant that I didn’t spoil the saw plate temper and it didn’t need much correction of plate flatness.

Somebody asked about my preference for an ink line over a chalk line. An ink line serves as more than just a snap line, its the pot you dip sumisashi in for layout. Plus, its not that hard to make your own.

The ink throws off the line better than chalk, and sumi ink is a rich and satisfying black colour, without worry of it fading or rubbing off.

The sumitsubo is a fiddly tool, I’m still learning to keep the ink the right consistency and the wadding just the right degree of wet. Sometimes I find the sumisashi quite difficult to give a good line, especially when marking with the grain. In the end the ink pot, bamboo brush, and Japanese carpenters square have been designed to work really well together and its worth the difficulties to learn their use.

My latest Japanese franken-saw greatly exceeded expectations. I know just enough about saws now to be dangerous. The handle is a little low to the tooth line but I wanted to keep the tang weld as long as possible for strength. But if you’re looking for a cheap way to put a saw together that can handle material out to about 12″ (although fletch cut at 12″ might make a bitch out of your kerf, quartersawing is more stable), its a viable option. The perfect saw in this range for green timber in my mind is a bit larger with a thicker plate, but for such an ugly saw it does still work, and at a fraction of the cost for putting a Roubo style frame saw together. What do you think?


Heavy Trestles for Log and Beam Work

I felled a tree the other day and needed some low heavy trestles to bring the saw logs up to a height that’s comfortable to saw horizontally with maebiki-oga saw.

I didn’t want to worry about carefully preparing the lumber for joinery. Working from a center line allows you to use lumber that is twisted or bowed, but joints need to be housed.

How about a nice big log for the top of the trestle, plenty of space for driving dogs and not likely to walk around  when pulling on the saw.  I’m a fan of draw boring for pegging mortise and tenon joints.

Good work holding makes me a happy sawyer, and creates the most efficient conversion of human energy to the saw.

Height for the trestles are an inch or two below my knee, about 20″. Sawing being such a simple activity, one would assume that it is easier than in practice. In reality you will find that small details make a big difference, getting the log to a good height for sawing is one such detail.

Now if I could keep my ink line from freezing in this cold, haha.

Temoto Tansu that the Beetles Killed

The Carol Link’s dissertation on tansu has been a great source of design inspiration since I had the chance to read it. I’ve noticed over the years that western woodworkers making tansu tend to make joinery choices that they are familiar with, especially dovetailed drawers and casework, and I am no exception (Corey Smith out of Hilo  is the case in point: Ronin Daiku).

The form I patterned after is temoto tansu, and it seems to be typical to have the combination of drawers, hinged cabinet/drop in cabinet door, and sliding doors. This is also the first piece of furniture I’ve made from lumber sawn by hand. The carcass is quarter sawn beetle kill Ponderosa Pine, the cabinet front is in Siberian Elm, and the nailed back is T&G Western Red Cedar. The elm is considered a weed tree in my part of the world, and doesn’t typically have a straight growth habit suitable for conversion to lumber, but pieces can still be obtained long enough for drawers and small doors. The hardware is brass from Hida Tool and Hardware.

The elm does have wonderful color. Unfortunately the quality of the pine I used didn’t allow for long clear panels that could be cut to wrap the grain from side to top.

The cabinet door is a joined plank from vertical grain material. It almost looks like it came from one board, so close! I don’t think I’ve come across a good explanation on the internet of the joinery for mitered breadboard ends. At first I thought of using a slot cutting router bit and inserting a spline to join up. Thankfully Woodcraft didn’t have one in stock, and on my drive home it occurred to me that its not anything different than a very long mitered face blind tenon. Of course, blind tenons need to be deep as possible for strength, but it wouldn’t make sense to cut a groove nearly the full depth of the end pieces. So the tenon ends on the main panel are only full length near the miteres. That way for the breadboard ends the open mortices can be cut first, followed by plowing a groove about 1/2″ deep, then the miters can be cut and the fit snuck up on by shooting on a mitering board with a plane.

I swear the dry fit was perfect. Then I put the glue on, the joint was very tight, and slightly misaligned as I drove it on. It didn’t want to properly center as I tapped it down and split the panel at the arris of the miter. I was able to pull it into line with clamps. In the future I’ll have a clamp across the panel before the ends go on to prevent splitting. Live and learn, right? At least it doesn’t show on the front. A boxwood butterfly will keep the split from travelling further.

I dovetailed the drawer sides, but used tongue and groove boards for a nailed bottom, with glued runner strips. I don’t think I’ll make drawers like this again, it just rubs me the wrong way.

The elm molding around the bottom presented one of the most interesting challenges of the whole piece.  I’ve never snuck some secret mitered dovetails inside a molded profile, and it was a challenge to figure out the order of joinery. I designed a smooth profile for the molding that I could cut with a normal kanna, and tried it out with pine first.

The stock was dimensioned and marked directly from the planed carcass after assembly. Next, the dovetails were cut. On larger panel work there is a specialized Japanese plane and shooting board for trimming the miter, but I’ve seen this detail on federal period furniture and knew there had to be a simpler way to trim the joint.

The miter is trimmed before the outside profile is shaped so that there is a flat face to register against a miter paring block (The above photo shows this incorrectly). If the miter is undercut even slightly you’ll never get a tight joint once the molding is profiled. It was counterintuitive to trim into rising grain, but sliding the chisel along the jig for a shearing cut leaves a nice clean surface.

But then how do you mark the profile for the molding with the joinery cut?

It turned out to be simple. The dovetails are kept below the level of the molding profile and a sample piece of the molding is cut to a matching miter for marking.

Last the rebate was cut which the carcass rests on, and a few screw blocks used to hold it to the carcass bottom.

I can tell that the quality of my work is improving, but there is still much to be improved on. For instance, the combination of plainsawn grain for the sliding door panels, vertical grain for the cabinet door and diagonal grain for the drawer fronts leaves the piece looking a bit muddled. The hardware I ordered is nice, but I didn’t make the effort to have it all match in finish. I would have preferred to use pine for the bottom molding, but worried about denting. It detracts from the visual framing of the cabinet fronts, but at least helps to balance the depth of the sliding doors with the drawers, keeping it from looking top heavy. I need to up my game and use a molded profile for the front carcass edges so that I have a decent excuse to use sword tip miters on the shelf joinery.

The work totaled out at 143 hours.

Its snowing and beautiful outside.

Sad Asanoha


About a year ago I made this gate with wanton abandon towards the fundamental principles that act on hinged frames, namely our old friend gravity. The hand plane finished surfaces look really good though, haha. So I keep saying, oh, better get on fixin the gate with some diagonal bracing.  But really, a wooden gate should have a roof, so… it doesn’t get fixed because its not worth fixing without a roof on there first. More on that in a bit.


I managed to get a 3D printer working (Flash Forge Creator Pro), finicky machines these printers. After a whole bunch of fuss calibrating the build table to the extruders I still couldn’t get parts to stick properly, hot plastic mess ensued. For such an expensive machine, they are still far from perfect. The fix was a four dollar bottle of max strength hair spray, figures.

I’ll leave the existential quandary of using a tool like this for later when I have less to get done. For right now there’s a satisfying feeling to have it merrily printing out useful parts while I’m in a pit digging a foundation.


I also bought a Rikon 12″ Planer/Joiner combination machine. When you get to the point of having rough sawn lumber for a full house to process you have to pick your battles. That or start taking apprentices.

Its a decent enough machine for the money, just barely there on quality. It would have helped to know that the chip extraction is a total joke without a power dust extraction system. Even with the jointer tables flipping up and down to use the planer it seems to hold alignment. After years of planing all my stock four square with hand planes its also a bit of welcome relief, allowing me to get to joinery much faster. Of course, long thin material like shoji stiles will still have to be jointed by hand, and I kind of laugh at the idea that the planer table could adjust with the tolerances of my kumiko thicknessing kanna.


I was left with the shipping crate that the thing came in, after an ordeal with a stupid truck driver and hiring a forklift to move the thing into my shop. Oddly enough I had just enough material from cutting up the crate a bit to sheath a 7.5/10 pitch roof on top. Providence wanted it to be a chicken coop, so I build a coop.


And with the new planer it was blow and go to get the siding material resawn from 2x dimensional lumber, down to 9/16″ board and batten.


The 3D printer came in handy for printing the hinges, and I used a simple post foundation. Didn’t have to buy anything for making it but the asphalt shingles, a whopping $10 from a building recycling store.

Of course chickens also need a fenced run to keep them from eating all your tasty perennial food forest crops, and a fence needs a gate to get in and out. Around here I guess that means I need to make a gate too.


To come full circle on today’s post, a proper hinged gate, courtesy of knowledge from a Chris Hall series A Bracing Situation.

I used mostly lap joinery to speed up construction. The rails all taper from the hinge stile to the latch stile, lightening things up where gravity has the most effect. The diagonal brace meets the upper rail with a butt joint, but with a particular shape of cut that allows it to bear against more end grain in compression. And, it being cheap lumber, I painted it to moderate movement from moisture cycles.


Hopefully my shoji inspired main gates can get some tender love and care, but there’s no easy way to retrofit wooden diagonal bracing.  And the roof over it will be mostly western red cedar, currently drying a bit in my shop awaiting resawing for decking and shingles. Now if I can just remember to close the window above my drafting board and keep the rain from ruining my plans. Wouldn’t a roof help with that? Haha.