Gateway (Twenty)

With the post title I decided to jump away from the Roman numerals at this juncture, as I think after a while a lot of folks are going to possibly find the numbering more confusing than helpful. If not, and you are choked that the Roman numerals are no more, you have my deepest apologies. Onward we march….


Work continued today on the kiosk post splice repairs. First job was to buck the posts pieces down closer to finished size, which removed a good 20″ from each piece. That made them easier to handle.

After a bit of faffing about with fitting the posts in a horizontal position, I decided that it would probably be easier to tackle the fitting with the posts vertically-oriented:

In most cases, when post rot splice repairs are undertaken, either you are working on a complete structure, which has been jacked up and you fit the repair post in from below, or the structure has been completely disassembled and you are working on the sticks individually. Here, the structure is all together yet it is sufficiently small and awkward to handle that I ended up working on it in this manner. Like I said, splicing on new to old material has its challenges.

I started getting the splice on one post engaged:

At this point, about 1/2″ (13mm) shy of closure:

Another view:

Since the main portion of the joint has sloped surfaces, as the joint goes together it tightens up. It is important to judge the fit carefully, as it is inevitable that when the surfaces finally come all together you will need to make some final tweaks. If the joint was put together from too far apart with too much force, then it can be kinda hard to separate again. Conversely, you don’t want to overcompensate in the direction of easing the fit either as the joint will then be a bit loose when fully up.

Second post now fitted to a partial engagement:

A closer look – about 1cm apart at this point:

Getting closer with additional fitting work:

A bit of trimming and uttering and muttering later, almost there:

Another view after another whomp with the mallet:

Purty dang close….

I chose to drive them all the way together, and then found some areas that needed further fettling, and then pulled the joints apart. A while later, adjustments complete, the connections were reassembled:

Let’s see how things look:


Front face:

Shop is looking a bit like a disaster zone at the moment, but never mind that.

The other front face:

Looking passable for sure:

Closer in yet:



You will notice that the new wood piece is a bit larger than the old. Does this imply that I will be be planing the surfaces of the new piece down until they are flush with the old one? No.

One of the lessons of Japanese temple carpentry that I am working to understand better and incorporate into my work is the attention paid in the framing process to how things will be 10 or 20 years down the line. Wood and wooden assemblies, shrink, settle, weather and distort over time. Making things when new to fit crisply is all very well, but if 10 years later the settling and shrinkage of the material causes large gaps to form in those formerly tight joints then the carpenter has failed to anticipate. So, the trick is to understand what is likely to happen (and this will vary in different parts of a structure depending upon exposure to the elements, local climate, and how much live/dead weight is involved) and design the connections so that after the settling and movement has concluded the fit is good. This may mean that new pieces will be fitted to their neighbors with slight gaps in certain areas. Looks like a mistake was made, but it isn’t the case.

The old wood posts in this kiosk have shrunk and distorted and settled after 26 years outdoors. That wood has done most of its moving and weathering by now. The new piece needs to be left oversize as it is likely after 10~15 years of weathering and adjustment it will become more nearly similar to the piece to which it connects.

Further, the post will join to a metal shoe down lower, and needs to be square in section at that location. If I made the new post identical in shape to the old one, I would have to make it non-square in the spliced area as well, which would then leave me with the task of feathering out the faces from non-square to square in a run of about 20 inches. I’d rather not.

Finally, if one planed the new material down to match the old it would be difficult to do so without kissing the old wood surface with the plane somewhere. Then one would be left with applying some sort of stain to try and blend the parts, never mind the change in surface from weathered to planed, and it would be tricky to pull off the transition cleanly. As it is, I’ll be looking to stain the new wood to look similar to the weathered material. I’m hoping my contacts in the finishing industry can give me some good advice as to how to accomplish that.

Funny enough, the splicing repair was likely the trickiest joinery work on the entire project, so I’m glad to have gotten through cleanly. Everything from here on out is joining new material to new material, everything straight and square, which is so much more straightforward. For the kiosk, I’ve marked out for the cross member location and milled up the stock for that, so tomorrow I plan to complete that portion, which will wrap up the kiosk repair work for the most part. When the metal shoe fabrication is complete in a few weeks, I will return to this portion of the job and fit the shoes, re-establish the electrical chase up the back of one post, do final planing, chamfering and so forth.

Thanks for coming by the Carpentry Way. On to post 21.

12 Replies to “Gateway (Twenty)”

  1. Anton,

    thanks for the question. In answer, no, not worried about that at all. We're talking about a 'ledge' all of 1/16″ wide, so it is hardly much of a surface for water to sit upon. The splice itself could get wet from a heavy rain with wind, but that's not much of an issue really. The splice is well off the ground, so soil bacteria will not tend to get splashed onto the connection. That's one thing. If the splice gets wet the wood swells and tends to seal it up. Finally, there is a roof overhead keeping the vast majority of the rain off of the joint. The joint is not hollow inside so there is nowhere for water to get 'trapped' inside. Moisture may penetrate in certain circumstances, but the joint is fully exposed to the air and will dry evenly afterwards.


  2. Two questions:
    1. To what extent do Japanese Carpenters give to grain orientation of connecting members, for instance joining a tangential grain to a radial or rift type grain?

    2. How important is consideration for Compression Set?


  3. Jack,

    thanks for the questions.

    I think that in japanese carpentry every effort is made to match the grain orientation of connecting timbers in a splice. You can imagine that two pieces connected so that each had growth rings at 90˚ opposition to its partner would mean irregular steps in the mating surfaces for much of the year. In this case, the original timber I am splicing to is boxed heart, and there is no way I will be using boxed heart. The timber I am splicing with has a rift grain orientation, more or less, so it is the most 'neutral' sort of orientation for a square timber, its tendency being to diamond when swelling or shrinking. Rift grain gives a fairly seamless visual match to the boxed heart, however this point is a bit moot since the grainy-ness of POC is very much subdued anyhow.

    Compression set. Well, not much of an issue with splices like these, as the long grain abutments are relatively gentle in slope. When the komi-sen (assembly pin) is driven in, the end grain abutments are driven tightly against one another and that is the primary means by which the joint transmits load. While swelling of the timbers when wet can cause a certain amount of compression set, the only place this is really a concern is in the stub tenon portions. Looking at old splicing joints of this variety, I've never noticed them to have significant gaps from compression set, so it is not something I am worrying about. Compression set could also happen if the joint were subject to high bending loads, but I really can't imagine this happening in this situation. Again, the key aspect is that it is the end grain abutments of both joint halves which undergo high compression when the peg is driven in, and compression set issues are not much of a factor.


  4. Was there ever any consideration for replacing the original posts with all-new ones? Or is this kind of joint typical in Japan when similar circumstances arise?


  5. Darrell,

    thanks for the questions. Yes, there was consideration given to replacing the posts entirely. That would have however necessitated complete disassembly of the kiosk. At this point, it is hard to say which approach would have been the quicker one! Not knowing how extensively the rot had penetrated up the post, my plan was to cut the wood for the splice away, and if the wood was sound, continue with the splice repair. If the rot had been more extensive, then plan B would have been to remove the post entirely, which would have meant taking the structure completely apart.

    In Japan it is more typical to replace only the portions of a post (or other framing element) which have rotted out, most especially when dealing with important structures of historic significance. This kiosk is not such an important structure in the grand scheme of things, however the repair I am doing is a typical one.

    Given my druthers, I would have built an entirely new kiosk which stands on 4 posts, so the metal shoe foundation system could have been done away with altogether, however that was not in the scope of work.

  6. Chris,

    Think the approach taken is great. I very much respect the way your work employs intelligently chosen jointry to address structure and accomodate the behavior of wood. An aesthetic is expresseed that is artistic not for the sake of embellishment but one of honest care-full practicallity.
    Keep on keep'n on bros. Your on a tear lately.
    Been a delight. Tico has asked before…You work hard…Do you sleep ?
    Happy Holidays.

    S. Henderson

  7. S. Henderson,

    appreciate the comment and happy to learn you have been enjoying your visits here.

    As for work, I am struggling to stay away from the shop right now as the work is so enjoyable. Sunday is a nice day to work – the building is really quiet, and there isn't much traffic to deal with coming and going. That said, will be taking at least one day off per week from now on if at all possible.


  8. Hello Chris,

    I've been reading your blog for years and really appreciate it. All those difficult Japanese joints that fit so extremely well, are nice to see 🙂

    But how do you fit these large joints so well? How do you know where it is too tight, and how much material needs to be removed? You can not look inside the joints, so how do you do it?

    Thanks for the lovely insights you give us with this blog,


  9. Jeroen,

    thanks for the comment and the questions. I'm sure my joinery fitting is marginal at best, however I can share a few observations. Larger material presents advantages and disadvantages as compared to smaller. On larger pieces, the surfaces are comparatively easier to check for squareness and flatness than when working with smaller joints and their surfaces. On the flip side, joints cannot be as readily cut in as large a variety of techniques as for smaller joints, due to tooling limitations, however most of the techniques transfer over from smaller to larger. If a larger timber is not properly seasoned or has internal stresses, then cutting things like scarf joints can cause the parts to deform appreciably, complicating the joinery process.

    Cut one half of the joint to your marks and measures, and then cut the other half to conform to it – that's about it. Based on the type of wood and size of connection and amount of interface, you may wish to have a little play in some portions or a bit of interference, depending.

    The wood I'm working with now compresses fairly readily, so a slight interference fit is fine. Other woods, like ebony, tolerate almost no interference fit.

    'Too tight' can be judged by feel and sound during assembly, however if the tenon, say, will start into the mortise, and the surfaces of both mortise and tenon are flat and square, then the joint will generally go together. It's important to check all surfaces prior to assembly, looking for non-flat areas in particular. Shoulders need to be checked meticulously for squareness. With large timbers it is usually easy to get a square and/or straightedge onto the surface and you can check in many locations. If your layout is correct and the cutout produces clean and flat surfaces, then things generally go together well.

    Preceding all of the joinery work, it must be emphasized, is taking care to obtain stock that is dry, straight, free of wind, and square. The joinery will generally go smoothly if those initial conditions are met.


Anything to add?

error: Content is protected !!
%d bloggers like this: