Joint No 1: Sumitome hozo sashi

This is joint number one in Sebastian’s PROJECT MAYHEM. Sumitome hozo sashi is the connection made between the corners of the sill plates of a timber frame.

I just finished! Rushed to the computer! Here is my best effort to cut this joint. From layout to assembly it took five hours. My stock was nicely squared 3×3″. I wouldn’t want to have cut this on anything smaller, and you’ll see why when it comes time to make the haunched tenon.


The tenon I left a little long, partly as a result of layout, and partly because it can then be sawn flush after wedges are driven in. I didn’t make the cuts for the tenon wedges because I’d like to be able to take this apart.


I used my sashigane for almost all of the marking and measuring, including the miter angle. You could even mark the tapered tongue with a direct 1:12 or 1:16 ratio using the sashigane, maintaining the overall proportion of the joint regardless of the size of the stock. I found a really very fine tipped marker and it worked very nicely, always providing a consistent line width.


Here is the layout for the mortised sill plate.


Layout for the tenon. There’s actually one line missing where I extended the tapered tongue cut line on to the end grain.


Layout for the dovetailed tenon post.


Because these are timber framing joints one of the things I decided to try as part of the exercise was keeping the timber horizontal on the bench. No vertical clamping in a bench vise.  If this were an actual 20′ sill plate you’d have no choice but to make your cuts like this.


My paring slick. The chisel for this was a Harbor Freight buy, literally worth less than a dollar a chisel. Turned the handle from an old broom handle on the lathe. The chisel is only hardened at the tip! CHEAP! But it works, you know?


This is one of two Japanese chisels I own. The other is my 6mm mortise chisel. It really helps to have good steel when making wide paring cuts on end grain. The quality of the cut surface is a direct reflection of the steel.


The post tenon, cut and chamfered.


Now for the mortised sill plate. I started by chopping the through mortise from both sides. If this were a real timber frame I’d be hogging out the mortise waste with a brace and bit. I didn’t have the right size bit for this tenon though, so it was chopped with a 3/4″ bench chisel. The haunches will be cut later with a saw.


The dovetailed mortise for the post was cut to a depth such that 1/4″ of wood was left between the two mortises. Sokozarai-nomi for cleaning the bottom to depth (a bottom so clean you can show it to you grandma). I also made a small depth gauge for this.


Then I cut and pared the vertical haunch, as well as the tapered tongue and miter. All my cuts were made with a 210mm ryoba saw. If I was cutting 6×6 or 8×8 beams I think a 270mm ryoba would be quite perfect. I then added the layout for the haunch depth. It is marked out to be 1/16″ deeper than the haunch on the tenon, so marking on the angled tongue won’t make a huge difference.


The sides of the horizontal haunch could then be sawn with the ryoba. This is one of those places where I’m glad I’m not using small timber.


The finished mortise.


The sill plate with the tenon was more challenging. I cut it in a bunch of spots, starting with the tapered tongue and the inside edge of the tenon, as well as one of the tenon shoulder lines that form the haunch.


I don’t have a frame saw so I just chopped down from both sides to remove the waste.


With the internal surfaces pared flat and clean I extended the tenon cheek lines on the back surface with a marking gauge. I tried to mark over the knife lines with pen, but there’s very little room to get in there.


BAM! Here is why this joint is easier in big timber. Now my saw has enough room to finish the tenon cheek and shoulder cuts.


The layout could then be finished and the tenon chamfered.


The rest of the joint was mortised in. Seriously, I’m glad to have this chisel.


The finished tenon with two dimensional haunches.


This could be a beautiful corner to a timber frame work shop!


And if you want me to pick next week’s joint, how about this nifty scarf joint? It uses two wedges, one from either side, as well as pegs to lock together.

Shaping Yasuri


I’ve been working slowly towards the making of yasuri. Here is my first attempt at the feather edge file shape after forging, annealing, grinding, and filing.  I measured the included angle from one side of the file to the other at 15 degrees on the files that I’ve bought. That said, for this first attempt I didn’t measure any angles. Just how often is a yasuri used with both faces cutting? Not much.

Jason of brought forth a great link to a Japanese company making yasuri:

The pictures of their manufacturing process are like gold if you’re trying to backwards engineer a tool.

I used a dial micrometer to insure parallelism of width and thickness. The hardest part would seem to be keeping the thickness at the edge consistent. To be clear, the photo below is from the link above. I wish I had the time to make something like this. It may come to that.


In theory, if the thickness and width of the file were accurate as well as the angles your were filing at and the flatness of the angles faces, the edge would be of a consistent width. One of the easiest ways I can think of to check this without a special jig and dial micrometer would be to use an existing file to cut a profile into a thin piece of metal plate. That could act as a visual gauge to both thickness at the thin edge and angle between the faces. For that matter, lacking a micrometer, you could make thickness and width gauges as well, and file to fit. These files would have been made by hand before the use of dial micrometers, so I know there is probably a simpler way to approach the problem.


Using the knowledge I gained in my previous post, I started by cutting the teeth on the edges. Cutting the teeth pushes a little bit of metal out to both sides, so the file blank then has to be smoothed off on the remaining faces so that the file cutting chisel can slide smoothly. I thought my file blank was looking pretty clean and accurate. It wasn’t until I put the chisel to the first face that I noticed how rounded my crappy drawfiling technique had left them.


I recently had the same crowning problem from draw-filing a saw plate. My solution was to scrape the surface with my trusty old file scraper. I made sure the cutting edge was nice and straight, and painted on some layout dye to show me exactly where I was removing metal. A nice low angle and gentle pressure kept the surface finish smooth and free of chatter. This is the surface I left for cutting the teeth. Unfortunately, my technique was still too poor and it showed through in the teeth-cutting.


Here is my single cut on one side. Lacking a lead form to protect the teeth, I didn’t cut the other side. Not that it matters…this was forged from a piece of rebar. Short of grinding up some exotic animal parts and case hardening it will never be a file. Well, maybe a nail file.  Lots of the teeth in the above photo are not full width because I didn’t get the faces flat enough. I should have flattened it further on a sharpening stone. I may be able to scrape it flat enough with the use of a sen that has proper draw-knife style handles.

The other problem that is real clear from the photo is how quickly my chisel dulled. I sharpened the chisel out to 6000 grit on my Shapton stones before cutting  each face and the teeth at the tip of the file are quite nice for the first inch. As the chisel dulled I started swinging the hammer harder and my TPI increased as well.  Compared to the mild steel I practiced on in my last post the metal felt a little harder. Obviously my forge annealing was less than stellar.


From Wataoka yasuri company, sen dai for yasuri. As far as I can tell file blanks are simply draw filed before cutting, no polishing on a stone for flatness.  Their manufacturing process leaves the blanks very close to tolerance after forging, so no need to use scrapers to remove large amounts of metal.  To repeat a blacksmiths saying, “Five minutes at the forge is worth thirty minutes at the grinding wheel.” The sen dai is dead simple in construction, it probably aids greatly in accurate filing, and is now “on the list” of things to make. Right after a fuego…

File Cutting Demystified


I actually felt a little remorseful last night after putting out my last post on file cutting. Kind of jumped right into things without even a cursory glance at what other information was available on the subject. I stayed up too late last night reading about this, and woke up knowing what to do. Above is a great illustration of file cutting chisels, the cutting angles, and the resulting tooth shape. These are western patterns, but I think the principle is so simple it translates to what was seen in the video yesterday. The chisel on the left has a 50 degree cutting edge, with one bevel longer than the other on the back.  The smaller chisel, Fig. 809, is for cutting finer teeth, and has a cutting edge of, remarkably, 35 degrees.


A selection of file cutting hammers. The weight of the hammer blow directly translates to the size of the tooth you are cutting. These are western Sheffield pattern hammers, wrought iron with welded steel faces. They look rather Japanese, do they not?


Workholding for file cutting is ancient and simple. A thong, strap, loop of cord, with your feet holding things. I still need a little more work on this, the file wants to move in the direction of the chisel blow, but I know the concept is sound. If I actually want to cut a double sided file I’ll also need a sheet of lead to protect the cut face while cutting the opposing face.


I cut my cold chisel down to size and reground the cutting edges to form a 50 degree angle. These cold chisels are tempered to be tough, not so much for edge quality, so I’m not jumping right to 35 degrees.


As with any cold cutting of iron, a light coat of oil on the surface will give cleaner cuts. Teeth are cut from the tip of the file down. After each light hammer strike you slide the chisel up until you feel the bulge of the previous tooth. I tried my best to hold the chisel so that the front edge was at 90 degrees to the file face. Light blows! We’re trying to cut files for saw sharpening here. Yesterday I was hitting way too hard because of my obtuse cutting angle. Today it was easy. Like, dumb easy. So easy this used to be the occupation of young boys earning piece work. So easy you have no excuse whatsoever not to try this.


Here is my single cut, first try. Looks good, feels good on my finger. This would cut if it was hardened.


For double cut files you smooth the tops of the single cut off with a file so that the chisel can slide smoothly. There is a great deal of variation in the angle that the teeth can be cut at with respect to the long axis of the file.

Today is a beautiful day, the stars must have aligned last night.  I’m finally going to buy a bag of lump charcoal, dig a hole in the ground, and hook up my shop vac to a steel pipe.

The forge will be born!

Cutting File Teeth By Hand, Sort Of

In the comments from yesterday’s post Jason brought forth a great  video link: of a Japanese gentleman cutting the teeth of a file with chisel and hammer.  This is exactly the kind of thing I can’t resist trying my hand at, so I sharpened up a cold chisel and found some mild steel to cut on.


The cold chisel is a 3/8″ Dasco brand, with cutting edge at 80 degrees. I made the assumption that a finely polished cutting edge would give a better polish to the cut. I used my smallest ball-peen hammer. Based on the video it looks to be about twice the weight I need. The little chunk of mild steel I have clamped in the vise is 5/16″ wide and polished.


The concept is simple enough, but the execution is everything. This was my first attempt. Every variation in the angle of the chisel, the force of the hammer, everything must be perfect. I was trying to move the chisel forward about 1mm per cut.


Attempt number two. In the video it appears as though he is holding the chisel leaning away from him. This would mean the edge of the teeth he was cutting would also be pointing away from him. The thing that confuses me is that he appears to be cutting the teeth from the top of the file downward. I tried that and only ended up flattening the tooth I had previously cut. Which is to say, I’m cutting starting close to me and moving out in the direction of the teeth. There were a few moments where I worked into a good rhythm with speed and the cuts actually became more even. One blow per cut, no double strikes, that only made things worse.


From there I tried the same technique on a piece of 1/4″ square bar.  I could feel that the forming process had work-hardened the surface a bit.


I don’t suppose someone has a camera that could take good macro shots of file teeth? Even better would be a microscope with camera. This is a shot with my cellphone through a magnifying glass…

Its still not quite the magnification to show the teeth in good detail. Suffice it to say, the teeth are crisp and clearly defined by their two meeting edges.


By comparison the teeth I cut with a chisel are all over the place. To little force on the hammer and you end up with only a little hump of metal. To much force and the metal curls back over in the opposite direction of the tooth.


Obviously its would take a fair amount of practice to get a usable file. But think! All of those worn out yasuri could be annealed and re-cut, each time becoming slightly smaller. There’s a lot to be figured out yet: the cutting angle of the chisel and the angle that it is used to the work. Work holding the chisel on the anvil while it is being cut. What kind of heat treatment is needed for yasuri steel? Is it an oil quench?

Who out there with a forge can give this a try with steel that will harden, see if it will really cut? I tell you, every time I answer one question three take its place. Here is another video of file teeth being cut:


Homemade Osaehiki-Nokogiri


I need to start making dai, specifically a naga-dai-kanna (jack plane) and hikouki kanna (kumiko thickness plane). Inomoto sells one here for about $900 FRN, so yes, I’m making one.  The one hang-up is the small saw used to cut the shoulders on a dai for holding the blade.  Once again, Inomoto carries this saw here for $230.00 FRN. Its such a tiny saw I thought I’d give it a try to make one.


I started wondering if I could even cut such small teeth. Odate, in “Japanese Woodwoking Tools” says of the osaehiki-nokogiri, “The blade is about 13cm to 15cm long with 22 to 28 teeth per inch. Okay, I knew I wasn’t going to be cutting 28 teeth per inch with a 100mm yasuri, 16 teeth per inch seemed much more manageable and easy to lay out with an inch scale. The teeth for this saw are nezumi-ba, basically a Japanese rip tooth for hardwood filed with fleam like a cross-cut.


This is the throw-away saw I’m making it from. Up-cycle anyone? The plate is .020″, which still feels too thick for the TPI I am cutting, but its the best I have. Eventually I’d like to be able to thickness my own plate with sen, but I need this saw now.


I use machinist layout fluid to make it easy to see my cut lines, which were made with a carbide tipped scribe. I was able to incoroporate the full length of the plate, tang and all, within this 270mm saw.


I used an abrasive cut-off wheel on an angle grinder to cut the saw from the old plate. For the most part I didn’t worry about dis-tempering the plate, but along the tooth edge I liberally applied water to keep things cool.


The saw blank, tang and all. I deburred the edges and cleaned up to the line on a grinding wheel, and polished a bit by hand with 220 grit silicon-carbide sand paper. The little circular indentation at the heel of a saw was cut with a small circular file.


Layout for the tooth spacing. One tooth every 1/16″, with layout dye and carbide to make the mark. These marks shouldn’t be deep, just enough pressure to expose the metal under the blue layout fluid.


At this point I cut the handle recess and glued together the handle blank. By the time I had the teeth cut, the handle was dry enough to shape. Having a router plane makes it easy to cut even depth tang slots. I used my 1/8″ detail chisel to get into the tip of the tang recess. My detail chisel was re-ground from drill bit shafting. Works great!


With the handle glue drying I started cutting the teeth. I know from a great video that Sebastian linked to of a Japanese saw maker that cross-cut teeth are cut with fleam angle, all at once. It somehow seemed easier for me to first cut small rip teeth, and then convert them to cross-cut.  Its really hard to get good macro shots with a cell-phone camera, so bear with me.


Here is one side of the teeth filed to cross-cut, with about 25 degrees fleam.


Once the general shape of the teeth was cut, I jointed down the teeth and went back, evening up everything with the yasuri. Odate says, “Because plane bodies are generally made from oak, the osaehiki-nokogiri has deep, nezumi-ba style teeth.” I’m not sure if my teeth are “deep” or not, probably a bit on the stubby side if you look at the saw that Inomoto is selling, but the teeth on that saw look like Ikeda-me.


The finished tooth-line, looking a lot like western cross-cut teeth. There’s no set in the teeth of this saw (thank god) which makes things easier on my vision. Even at 16tpi these teeth are microscopic.


The handle is a bit fat looking for the blade, but I have large hands so this is comfortable for me. I’ve already put the edge through a bit of oak and maple and the cut surface it leaves behind is nice and smooth. Definitely not a fast cutting saw, but that’s not the point. Now I can make some dai! I may even use this saw to cut the dovetail slot for the bamboo spring on the sole of hikouki kanna because it is already set up for hardwood.

Tools for making other tools. I love it.