A Japanese timber framed garden shed

[GRadice]

Ditto!

I was wondering if these complex joints were developed to erect wooden structures without the need for fasteners and to be flexible enough in the event of an earthquake for the joints to interlock in "all" directions to prevent the structures to fall apart as a result of the shaking & quaking?
And another question is which joint to use for what... !!

Cheers
Yvan

From my amateur reading, I think you are exactly right. Historically what we think of today as Japanese carpentry started with Buddhist monks who learned their trade in China. The Chinese methods and techniques gradually became modified in Japan (some might say "refined"). There are also different styles and traditions within regions of Japan. And use of metal fasteners in Japan was restricted for a long time by "sumptuary laws" which restricted who could use metal in construction. To me the variations and subtleties of Japanese joinery are endlessly fascinating.

And after all this time, here is something new in pegged mortice and tenon joinery from Japan. Use a square peg but rotate it 45 degrees, and kerf the end of the tenon along the peg axis. Under tension, the tenon spreads and wedges tighter into its mortice.

As described in a blog post by Jon Billing, an American working for Somakosha, a small construction company in Japan who is trying to maintain old ways of building:

Maruta Bench - Build 3 - Big Sand Woodworking
杣耕社

I think this is brilliant and decided to use it for the pegged tenons on my shed. This will be the other end of the floor beam shown above.

6C064266-C269-4334-9930-6FE01857D0B1_1_105_c.jpeg

The stub tenons and haunch are there to resist twisting. That darned puppy cost me a couple of wonderful hours away from the shop today but I did get this far. I haven't cut the mortice for the peg yet since I need to make certain about the fit before I draw bore.
F70E4885-88D7-49E1-8A51-1FC7055D2D09_1_105_c.jpeg

[GRadice]

I finished the second of two "rod tenon mortices," (rough translation) and this went a little better than the first. A bit more refined.

C58263BA-CAE8-489F-A9F7-E9959ED24D33_1_105_c.jpeg

There is a nuance in cutting the half mortices for the two keys. They need to be tapered on one edge. I have been unable to find a good description in English for how to lay out and cut these little mortises but I do have a figure from book in Japanese by a master of joinery named Shinzo Togashi. The great thing about this book is that the layout dimensions are all proportions of the stock you have rather than fixed dimensions. I think the modern term is "scalable".

589A4E68-799A-44B4-A7A3-4871DBC4C79C_1_105_c.jpeg

Note that the mortices for the keys should be tapered by about 1/20 the width of the beam, not the depth of the mortice. In my case, the taper is about 5-6 mm.

Here is my layout for one side.

6E401CEC-84AD-4869-B5EE-7EB26F78F880_1_105_c.jpeg

And after chiseling and paring

B83F0BC2-8842-4AC4-AD27-C9E425400804_1_105_c.jpeg

[woodPixel]

That's the first time I've seen this measurement produced. Thank you.

I've two thoughts:

-- First is how these joints deal with the weather and rain build up inside the joints. Obviously it isn't too much of an issue.

-- Second, if you should ever sell the house, that garden shed may be proclaimed a Local Treasure and heritage listed!

[GRadice]

That's the first time I've seen this measurement produced. Thank you.

Here is the carpenter's joinery book I referred to:

B14C60DC-7ABA-4450-9365-726FD138E93B_1_105_c.jpeg76B0EB7C-E00B-4B75-A7A0-58013AA01B42_1_105_c.jpeg

Amazon.com

Togashi has other books on temple construction and gate construction if you really want to get in over your head!

I've two thoughts:

-- First is how these joints deal with the weather and rain build up inside the joints. Obviously it isn't too much of an issue.

-- Second, if you should ever sell the house, that garden shed may be proclaimed a Local Treasure and heritage listed!

Generally the joinery would be covered by wall infill and protected by flashing. Or the entire frame would be covered. On mine the frame will be infilled but the posts and beams outside surfaces will be exposed. I have a friend who does this and is advising me on flashing details.

It certainly will be a sturdy and unique place to house a lawn mower and pruning shears!

[GRadice]

Some WIP photos of the tenon end of the floor beams. I have done 2 of these and have 6 more to go. Here is how I cut the first two. I might change the order of cuts a bit later as I feel my way through this. Not too complicated but it does require being able to saw to a line. I knew this was going to happen on this project so for the last couple of years I set myself the goal of doing most of my saw cuts with hands saws even when I could have used a machine. That paid off, I think.

Cross cut to establish the end of the haunch. I used a rather coarse and aggressive 330 mm ryouba saw for this.

E5F6DC89-2E5B-4210-AD59-CE316F3AC3A1_1_105_c.jpeg
Then a rip cut with the same saw along one cheek of the tenon. That stray layout line to the left was a mistake. For this I knelt and cut down vertically Some stand on the bream and cut up. At my age I demurred.

F0D46824-9E4D-43C8-B5E7-187E92605FF9_1_105_c.jpeg
I try hard to not saw past my lines. Instead I leave a few feathers of connection and when the waste bit can be wiggled I knock it off with a hammer and then clean up the feathers with a chisel.
868F9258-9184-4C66-A2E0-FDF393C18FBC_1_105_c.jpeg

Then cut the tenon cheeks, flipping the beam as needed to get both sides even.
381E931D-DFFE-475F-882D-C9D33697E96D_1_105_c.jpeg79333DCC-94A5-4103-8B79-DBBA346F766E_1_105_c.jpeg

and then cross cut the shoulders, which in this case are also the ends of the stub tenons. For this I used a guide although it wasn't really necessary.

240807E3-DA14-463D-B738-1E1A43C32288_1_105_c.jpeg

and both cheeks done.

944CE9C7-1E82-4384-80AC-A0AC589DF9DA_1_105_c.jpeg

Then the rebates that establish the outsides of the stub tenons. I switched to a smaller, finer, 270 mm saw here since this will be a show joint. And a guide block checked for square.
617AD600-B8B9-4AEE-88BF-528F17E41BAB_1_105_c.jpeg

A4973045-44EC-486C-8D4C-76CC43E10608_1_105_c.jpeg

Then ripped the other side of the rebate.
5EA12633-01B5-4A8F-9E65-D72CB147D72F_1_105_c.jpeg

and both sides done.

A32CCC28-3B65-49C8-944E-231B2F398466_1_105_c.jpeg

Then reestablished the layout lines for the stub tenons and chiseled out the waste between them.

44CE0C3A-1281-489E-9984-92F0FA695C02_1_105_c.jpeg

Then chamfered the ends of the main and stub tenons.

0300CA4B-08A5-4F57-B57A-A01F861AE40A_1_105_c.jpeg

And finally checked the tenon width. Nominally 30 mm. From the saw it was about 31.5. After some planing and paring I got to within shouting distance. The softwood I am using is fairly compressible so tighter is better than looser. This was a good day.

AA663609-1B99-4F57-BB92-BB3C88328371_1_105_c.jpeg

[GRadice]

The adjacent floor beam has a long tenon to fit the long mortice I recently made. I ripped the long cheeks of the tenon on a bandsaw and did all the rest with hands saws and chisels as above.

185D7498-7BD7-401B-8234-46497E9B8B78_1_105_c.jpeg

On test assembly. The space left between is where the beams insert into a post.

0290640D-96FA-4309-9F82-508B9CC321E8_1_105_c.jpeg

How the half-mortises for the tapered wedges come together. Both halves of each are sloped as I described above, wider on the surface and narrow on the bottom.

1E9ADB99-90B8-469B-8A8A-E96719C7E7BC_1_105_c.jpeg

I made the tapered wedges parallelogram in section as you see. They can be made rectangular which are easier to make but that tends to spread the sides of the mortice when you tap them in. Maybe not a big deal on thick timbers but since I have only two of these joints to do I opted for the better engineered solution that concentrates forces along the line of the joint. Here are illustrations from Chris Hall's masterful monograph on splicing joints to compare.

Screen Shot 2022-09-11 at 5.40.38 PM.jpgScreen Shot 2022-09-11 at 5.40.49 PM.jpg

[GRadice]

Here is another nuance of making and fitting the tapered wedges (shachi sen). It is a little hard to illustrate. But if you taper the mortise from 36mm x 8mm at the surface to 31 mm at the bottom as I have done, then the mortise and wedge also have to taper from 8 mm to 6.89 mm at the wide end of the bottom of the mortice. Here is a view from above the wedge mortice toward the bottom of the mortice.

If you don't taper the wedge in both directions the wedge is 1.11 mm too wide at the bottom and pushes the mortise walls apart. It took me a while to wrap my head around the geometry but drawing it in Sketchup helped. Chris Hall has a fuller explanation in his monograph on splicing joints, which I highly recommend, again.

5BDEFEA9-755D-4209-AEED-196F87E5C748_1_105_c.jpeg

[GRadice]

Pecking away at it. I finished all of the tenons on the peripheral floor beams and now I'm on to the two sleepers that support the floor joists in the center of the shed. On of them is included in the three-beam-to-post joint I just described, and the other ties directly to peripheral floor beams. They are the two show here from below colored tan.34C0BD34-5AF8-4985-9EA5-F0AFE7B732ED_1_105_c.jpeg

One wrinkle that took me a while to design is that the tops of the two sleepers need to be 30 mm below the tops of the peripheral beams to accommodate the height of the joists and finished floor. That took some fiddling with the joinery layout.

On the left is end of the sleeper that meets the post. The notch in the tenon will capture the edge of the long tenon on a peripheral beam. The notch is cut out 1mm closer to the tenon shoulder so that the inserted long tenon acts as a draw bored peg and pulls the sleeper in tight to its post. Unfortunately this tenon had a bad crack and knot I had to cut out and scarf in a patch. None of it will be visible in the assembled joint. The other sleeper is joined with a housed double tenon, offset to accommodate that 30 mm step in beam heights. Both have a series of cogged lap joints for the joists that will sit on them.
FA6B51A7-0D62-4A08-A8E4-B9A0279132B9_1_105_c.jpeg

[GRadice]

The ends of the floor joists are supported by pockets in the side peripheral floor beams. I got those pockets chopped out today. Or so I thought. Here are the two peripheral beams and the two sleepers. Ooops.

63427E08-523B-4D69-924A-F738D40CC014_1_105_c.jpeg

But, as y'all say: no worries. I picked myself up, dusted myself off, and realized that all I have to do is adjust the width of those joints from 48 to 52 mm. I haven't milled my 60 mm joist stock yet so I can just make slightly wider joists. Catastrophe averted.

I have the revised joinery almost done. I would have finished today but at the end of the afternoon here my wife gently asked whether I could "STOP THAT #$@%&$ CHOPPING and please take the dogs out!" Which I did.

CC3D8E2F-6811-4E32-9FFF-2C67859D847B_1_105_c.jpeg

[GRadice]

I finished one of the two eave beams. It is about 4 m long.

4B272C44-F490-45E2-952C-42A733C36D95_1_105_c.jpeg

27885E8F-1E9F-47B7-BF37-EB421020E0F9_1_105_c.jpeg

Here is how an eave beam connects to a post and tie beam and supports a barge board at an end. The notches are for rafters on 1 foot/304.8 mm) centers.

B62FD009-BEDF-4C4F-B515-890F54F9FB08_1_105_c.jpeg

[GRadice]

A word about my double measurement system. While I was designing this thing my research told me that Japanese residential architecture was modular and based on the size of tatami mats. However, the size of tatami mats varied from region to region and were measured in the shaku system, now standardized at 303 mm. A tatami mat for construction purposes is today standardized at 303 x 909 mm but also by individual contractor preferences. Then you have to consider what standard size materials are available for sheathing and roofing and timbers. And tooling dimensions are almost all metric. It gets complicated. Buildings in Japan are still designed with the shaku system, and also in metric so carpenters need to be fluent in both.

Shaku (unit) - Wikipedia

In the USA of course, we insist on using the Imperial system so that adds another opportunity for "creativity." For example, roofing materials like shingles and roofing panels and timbers are almost exclusively in Imperial.However, a nominal 2 inch by 4 inch timber is actually 1.5 inches by 3.5 inches. Sheet goods are sold in Imperial width and length but most commonly metric thickness. It is a nightmare.

I decided the most important place to start was making sure the rafters lined up with the available roofing materials, which dictated a rafter spacing in feet and inches. I wanted the posts to line up exactly with the rafter spacing so that dictated that the post spacing be based on a 3 foot grid. So my plan spacing for the post centers is 6 ft x 9 ft, or 1828.8 mm by 2743.2 mm. Then I've dimensioned all of the of the posts and beams based on a post section of 120 mm x 120 mm. That is compatible with Japanese metric tooling and simplifies most of the mortise and tenon layout at 1/4 of beam dimensions/30 mm.

One of the USA guys who is advising me was trained in Japan in the shaku system and uses it for his designs, then translates all dimensions to Imperial to get plans approved by engineers and local building authorities. Of course computers do the math instantly but when you are in the shop it is a different game. Fun!

[woodPixel]

A missed opportunity to get the Shaku to 300mm in 1891, rather than 303.

So close!

[GRadice]

Yes! I should have mentioned that the shaku system, for those who don't know, is also decimal. 1/10th of a shaku is a sun, 1/10th of a sun is a bu. It could have been great medium between Imperial and metric.

I finished the joinery of the other, mirror image, eave beam.

2D573ED3-668D-421D-8B44-2C45A06DCF4C_1_105_c.jpeg

Next up is the ridge beam, sitting there patiently to the right. It bowed since I provisionally milled it a few weeks ago. I'll correct it tomorrow before I lay out and cut the joints.

[GRadice]

Mostly grunt work today with a hand plane and minor assist from the band saw to get the ridge beam squared and planed on three sides. A little more to go tomorrow.

7F0100F5-ED7D-4DD1-B88F-58BAF12995F2_1_105_c.jpeg

I had another load of Port Orford cedar milled last week. This for the roof rafters, floor joists, doors, windows, gable lattices, and some other parts. The mill kindly delivered it for free (because it was going with a much much larger load of their own) to a kiln just an hour from me. The kiln said they would dry it all for USD $125, a fair price to me, and sticker and wrap it in plastic when it was done. But that their kiln is down until they can get a computer board from Hong Kong which might take a couple of weeks before they can start. A global economy thing. But then it will take only about 10 days to dry my wood. So all together my wood might be available at the end of the month. Which is perfect timing for me and my budget.

[GRadice]

I have the ridge beam all squared and the top backed for the roof slope. I used a power plane for roughing and a hand plane for finishing.

762E5560-6AF2-46D8-9012-5B84C88770BC_1_105_c.jpeg

The English version of the Japanese carpentry book I'm using (Sato and Nakahara) has three common approaches for joining rafters to ridge. I'm using the center method since my roof will be exposed so I want a more finished look and I think the pockets will help the rafters resist twist.

F9D86971-ABF5-4356-8DAF-F037927A712F_1_105_c.jpeg

The ends of the ridge beam need to fasten the peak where the barge boards of the gable ends meet. That joinery is a little tricky since there are two relatively thin barge boards meeting at a miter with both stuck to the end of a beam. My meager collection of Japanese carpentry books are little help but I think I see a way forward. I'm looking at something like this with a dovetailed draw pin to hold the barge board horizontally and a stub tenon of some shape to support the barge board vertically. I know this will work for the ridge beam but I need to think on it a bit more to see how it will work with the joinery for the two barge boards.
DB09AA8A-92C6-47FF-83B7-7AFD354FC50B_1_105_c.jpeg