Finishing the Garden Gates


Its been a long time in the making! There seemed to be almost constant interruptions for other things that had to be done, so the gates took a long time. That’s just part of the growing season. With all of the joinery cut and the final finish planing done, the gates are ready for assembly.


For the top and bottom rails I haunched the tenon to give it extra resistance to twisting. With this gate exposed to the harsh extremes of Colorado winter and summer, it needs all of the strength it can get.


I started by assembling my kumiko. Normally the kumiko notches alternate, at least on the horizontal, but having all of the vertical kumiko with an uninterrupted line is a cleaner look. If the lap cuts for your kumiko joinery are too tight the lattice will bow. Thankfully my practice has paid off and I’m getting pretty consistent with good fits.


Darla thinks my pile of finish shavings are a nice place to bed down. Adorable!


With the kumiko jigumi glued together I started assembling the frame elements. No glue on these mortises. I’ve been wondering about that of late, not using glue here. My primary source for skipping the glue is Toshio Odate. Both Jay Van Arsdale’s book and Des King’s book have you cutting kumiko mortises that are less deep, and King seems to glue absolutely everything that touches together. You’ll notice with Odate’s kumiko mortises that they are always as deep as humanly possible. Like, you know you’re deep enough when you’re ready to throw your mortising chisel across the room because the mortises are so difficult to clean out. That’s what I went for here. The kumiko mitsuke (thin edge) is 1/2″, the mortise 1/4″ wide, and 13/16″ deep.  Though, this is the first time I’ve used a brace and bit to speed my mortising up. A single hole down to depth in the middle of these small mortises makes things much faster. My bit was slightly undersized so that the chisel always sizes the width of the mortise.

Maybe I’m crazy for putting so much work into something that will be outside. How else do you know how your work will hold up than to see it weather? After all, these gates are just made from 2×4 construction lumber.


With eight foot ceilings in my shop I was cutting it pretty close to get the top rail and kumiko attached.


I set my planing beam down on some low saw horses to get the stiles on. I decided not to glue the mortises on the stiles either. If this piece holds together, it will be because of solid joinery.


With the doubled tenons I used, and two wedges per tenon, I had to cut a lot of wedges. These are red oak, about 1/2″ longer than the depth they will be driven to.


Okay, so finally I used a little glue. I don’t know if its technically necessary, but it can’t hurt. If I get six years from these gates and the wedges are still holding strong I’ll be happy. When you’re using a wedge either side of the tenon make sure to start and drive both wedges at the same time. If you only drive one it will make it really hard to get the other wedge started in the kerf.


After the glue dried I trimmed the wedges and tenon flush with the stile. The stiles were made slightly thicker than the rails so that the edges could be chamfered, necessitating a shallow recess be cut in the stile to have the hinge lay flat. Its still feels funny for me to take a finished set of shoji and start drilling holes for hardware.



I used some pretty standard stock gate hinges, with threaded bolts. You always wonder if you took your measurements correctly when dimensioning something to fit an opening. There’s added variables when you add space for hardware, and I wasn’t relaxed about the installation until both gates were up and hung plumb. I still need to buy a gate latch. For now, a piece of rope works a treat.

Well, how’d they turn out? The kumiko for the asanoha pattern was cut from redwood to set it off a bit from the rest of the kumiko. Hand plane finish, all natural. Finally, some gates that are tall enough to keep the deer from jumping!

Cutting Koshikake Kamatsugi (Stepped Goose Neck Splice)


I thought after last week that somehow this joint would be easy. Its certainly straight forward in cutting, but there were some new challenges that made executing this joint much more than a walk in the park. Time from layout to a joint that actually fit properly, 4hrs.  Can you see that I added a taper to the back shoulder of the goose neck? I thought it would make things easier and pull the whole assembly tighter.


Here, the layout for the female side of the joint. Both sides of the joint were marked out before beginning, no transferring of marks. All of the marking was done with a fine tipped marker (.3mm) and sashigane.


The layout for the male side of the joint. I started my cutting with removing the bottom half of the lap.


Once again, I did all of my cutting with the piece horizontal on the bench. For this joint I tried cutting horizontally, which I’ve never done before. Definitely felt a little strange, but I had already cut both sides of the line down to the shoulder with the saw vertically, so the saw was just finishing the cut. I should have stopped right here and sharpened the rip side of my ryoba, I knew it was feeling a little dull and floating around in the cut, but…but what, I don’t know. Sharpen your tools when they get dull and sawing is a pleasure of a challenge, not an exercise in frustration.


With the waste removed and pared flat I finished the layout on the bottom. With the timber in the same orientation I cut down one side of the cheek of the goose head, flipped the timber over and finished the cuts all the way to the haunch shoulder line. Then I made the cross cuts into the neck.


A chisel pared the waste along the neck line.


As you can see, a bit more wood has to be mortised out to define the horizontal and vertical haunches. I also chamfered a lot of the outside corners on the bottom of the goose neck.  It  was right about at two hours in when I finished this side. I thought The other side would be faster, but mortising with a chisel always takes me longer than I think it will.


For the female side, removing the step that houses the horizontal haunch of the male side, and finishing the layout.


I used a saw to cut for the vertical haunch , as well as cut as deeply as possible into the goose neck before chopping the waste with a 3/4″ chisel.


These interior surfaces are what make this joint so challenging. There are so many of them that need to be perfectly vertical, or the joint will never go all of the way together. I used a small bevel gage set to square to check. All of this paring and checking, it really slowed me down. I also used this bevel gauge as a depth gauge when paring the bottom of the mortise. It was nice being able to run my chisel, bevel down, to clean the waste from the bottom.


I grabbed the taper angle from the side of the layout with my bevel gauge and marked the interior of the mortise wall. I thought about just eyeballing this for the cut using the markings on the side of the beam, but that would have been a mistake. Getting this taper correct was the key to having this joint fit properly.


The finished joint. Well, I forgot one thing, chamfering the outside edge where the bottom of the goose neck meets the vertical haunch. That was mistake number one that kept the joint from going together all of the way.


For the longest time the tip of the goose neck fit flush, but the step was still 1/16″ high. I took it apart, checked my surfaces for square and flat, pared a little here and there. I must have taken it apart a half dozen times and each time could not figure out what was preventing the step from sitting all the way down. Finally I realized that the taper at the back of the goose neck was too acute on the female side of the joint. It was keeping the back of the goose neck from seating all of the way, and I thought it was a problem at the step. I was wacking it with a heavy hammer, thinking, surely it can squeeze in there a bit more. Nope, it needs to be almost perfect. The joint fit properly after paring. The taper became snug the last 1/16″-1/32″ before the two joints were flush. With a good fit the two pieces of the splice lay in the same horizontal plane, flat on the bench. I did manage to introduce a little error in my marking, you can see a small gap where the shoulders meet over the step. I need to get faster! Way too slow, and I was trying to move with a purpose.

I’m ready for more, what do you think for next weeks joint?

Wedged Mortise and Tenon

I realized that I’m getting ahead of myself with some of this beautiful Japanese joinery I’ve been practicing. I skipped over the wedging of the tenon for Joint No. 1 and really missed out on picking up some important details. Namely, paring a taper on the outside of the mortise such that the wedge truly locks the tenon in the mortise. Having planned to use this wedged tenon in my shoji garden gates I’m currently working on, I figured a practice try was in order.


This is a shoji stile, 1-1/8″x1-3/16″, with the back taper on the mortise 1/8″ wider at either end than the tenon. This puts the wedge taper at a 1:5 ratio, quite obtuse by most measures. I figured if I could get away with this heavy of a wedge, I could comfortably relax the taper ratio for my garden gates and not worry about splitting the tenons when driving the wedges. My reference for the proportions was Jay Van Arsdale’s book on shoji making.  He describes how the back taper extends about half way down the depth of the mortise, and the cuts in the tenon for the wedge are made at a diagonal. This was news to me, even having read this book…

I thought tenon wedging was done with a wide kerf saw to make the cut for the wedge in the tenon, no back taper, and wedges with very little taper. It hurts to be wrong, but the time to learn is now before I produce a joint that lacks strength and longevity.


I’m used to tapering in the sides of my blind mortises so that the tenon is compressed as it is driven together with the mortise. Toshio Odate, in his book “Making Shoji”, describes the shape of tategu through mortise as an hour glass, pinching the tenon in the middle. He mentions nothing about flaring the outside of the mortise wider than the tenon, but perhaps there is a middle road. This picture isn’t representative of the taper I pared in the end grain. In general, I’m trying to taper the mortise in so that the middle is about 1mm narrower for softwoods. I figure compressing the tenon at the point that it most wants to split when the wedges are driven will go a long way to keeping the joint together. If I was to cut the end grain to a barrel shape the tenon would have room to split.


Can you see the saw kerf for the wedges in the tenon? They were cut with a very fine saw starting towards the outside edge of the tenon and leaning inward. I’ve always seen these cuts made straight down the tenon, but I imagine this diagonal cut lowers the chance of a split running out from the tenon, as well as allowing more flexibility at the tip of the tenon where it needs to move over the  most. I should have made the wedges about a centimeter longer than needed and cut off a bit of the sharp point on the wedge, but this still worked out.


Look at how much room there is! This is the sweating moment, when you wonder if wood really is this flexible of a material (and I’m a bowyer).


Success! One of the wedges needed to be driven a little deeper than the other necessitating the use of a nail set to drive it down, but it doesn’t show much after planing. I feel like the back taper on the mortise was more than needed, but it comforts me to see just how much the wood will move over without splitting. The compression at the middle of the mortise was definitely a good idea.


And now in practice, using a small sliding bevel to gauge the back taper of the mortises for my garden gates. I used 1/8″ more width either side for these mortises as well, but the stiles are 2-1/2″ wide, making a taper ratio 1:10, half of the previous example. Screw the glue, wood is good.

Joint No. 2: Koshikake Kamatsugi (Stepped Gooseneck Splice)

Were back at it again! This is joint no. 2 for our series Project Mayhem. As always, if you’re reading this feel free to join in and we can learn together. The rules are here, but to recap a la Sebastian:

No electric tools.

Document the process.

Document the tools used.

Time it.

Discuss it.

Blog it.

I knew I wanted to do a scarf joint, I just didn’t realize how awesomely complicated they can get in execution. I thought the gooseneck splice was complicated, but its unfortunately one of the simpler ones, and builds on the layout and execution skills developed in the previous exercise. So here we go…


I have tried to provide some simple ratios to help proportion the joint.


This is a great joint for connecting groundsill beams together when you’re building your dreamy timber frame shop.

Gentlemen, there are seven days from today.

Finish by June 30, 2015


Making and Using Aburatsubo (Oil Pot) for Mortising


I’m working on all of the mortising for my shoji garden gates. Because I love my 6mm mortising chisel I’m using it for both the kumiko mortises and the rail/stile connections, which are twin 6mm mortises. But Japanese mortise chisels have very little taper tip to handle for clearance when cutting deep mortises.

In comes the aburatsubo, a small vessel with silk cotton or a rolled cloth soaked with oil. The difference that a little oil makes on the chisel while you work is amazing, especially the deeper you are cutting. Not only does it lubricate the sides of the chisel, the swarf seems to come free more easily as well. I had my reservations about getting a little oil on the important glue surfaces of the mortise, but I’m going to follow in the wisdom of Japanese tategu shokunin of the past and give this a try. Smooth is fast, and oil makes things much, much smoother.


The tools from left to right: 1/4″ dovetail chisel for cutting down the sides of the mortise,  sokozarai-nomi for cleaning the bottom of mortise, flat-tipped screwdriver for loosening chips in mortise, 6mm Japanese mortise chisel, hammer.

I find with very small and deep shoji mortises that the chips come free much more easily if I cut down the cheek of the mortise. When I started with chisel mortising I used various wood mallets, which are great, but I’ve come to prefer the density of a steel hammer when chopping. I’ve become extremely fastidious about the placement of my tools when working. Lets face it, mortising by hand is slow compared to a power hollow chisel mortiser or a spiral up-cut bit on a plunge router. I own neither, so I seek to work as quickly as possible, and every little repeated movement adds up.


Part of my quest for faster mortising involves the final trimming of the end grain surface. Tategu use a slight inward taper to the mortise to compresss the tenon as it is driven together. Its great because shoji can assembled without the use of any clamps, but judging the cut can be difficult. Depending on the size of the mortise I try to tilt the cuts in for a total of 1mm to .5mm less width at the bottom. My method at the moment uses a small wooden square that I watch as I cut to the line. This way I cut it once right as I want it and don’t fuss around wasting time carefully paring the endgrain. It teaches you how important hitting the chisel in the right spot is to cutting a straight line. If the cut is too heavy the chisel wants to travel back in under your cut line, making for a weaker assembly.


Here are the twin 6mm mortises for a rail/stile connection. I mortise in the traditional Japanese manner. Starting in the middle cuts are made inward, producing a ‘V’ notch as close to your final depth as you can manage. Chips are continuously pushed up and out as each side of the ‘V’ is cut down. Then vertical cuts are made back to the line. Its really important in terms of speed to cut as close to your depth with the ‘V’ as possible.  The aburatsubo made these deep mortises possible, and kept me from stabbing myself in the thigh while trying to yank a stuck chisel from the mortise…


The aburatsubo is traditionally made with a small section of bamboo, 1-1/2″ – 2″ diameter, but anything will work really. A small glass pint size mason jar with a bit of rolled up cotton t-shirt would be easy enough. Or maybe you live in Hawaii where bamboo abounds.  I have a lathe, so a nicely turned object is what I gravitate towards. I started by boring out my square blank 1-1/4″ wide by about 2″ deep.


I then turned down a lightly tapered mandrel in my lathe chuck to fit the bored blank.


If the taper is too tight it can split the blank, so I leave things barely snug on the mandrel and bring the tail stock up to keep some pressure on the blank while it is roughed to round. Once the major shaping is done the tailstock can be backed off and the bottom finished slightly concave so that it sits properly on a flat surface.


I used a couple of coats of paste wax to seal the grain and keep the oil from soaking through the bottom. The silk I used for the ‘cotton’ filler is prepared sliver for spinning into yarn.  I’m not convinced that silk is somehow better than some cotton balls or a bit of old t-shirt, but I have it so I’ll try and find out.

This is the first time I’ve used aburatsubo for mortising.  To quote Toshio Odate, “When mortising, the tategu-shi continually stabs his mortise chisel into an aburatsubo filled with cotton and clear vegetable oil to lubricate the chisel.”  When I try this the silk ends up packing down into the oil pot, only lubricating the first half inch. Maybe I need a bigger pot that I can form more of a “nest”. For now I just pulled a bit of silk out the top and use the aburatsubo to wipe the sides down periodically. The amount of oil it uses is miniscule, so I’m using camellia oil, though expensive, because I love the smell.

Does anyone else use an oil pot while mortising? This is one more thing I’m surprised not to have tried sooner. Its already made my mortising twenty-five percent faster, at least.