First Light XXIX

This long series of posts has been a great opportunity for me to review the use of Roman numeral in counting, that’s for sure. Yesterday’s post was written with a certain amount of time pressure, so I just went back and tidied up some of the more egregious typos and clumsy phrases – my apologies for butchery of the English language, not to mention the Japanese.

Today is part 29, and I will continue on with the description of the making of the hanging fish, termed ‘ge-gyo‘ in Japanese. If you’re new here, and wish to get up to speed on the ‘what’ ‘why’ and ‘where’ of gegyo take a look back to the previous day’s posting.

The Bloodwood centers for these gegyo were put together with an asymmetrical offset that left a small pentagonal hole in the center. I was playing around with the idea of leaving these openings as they were, and using the potentially intriguing negative space/porthole as a feature, however after thinking it over a while I decided to put centers in. These I cut from my chunk of Gabon Ebony using a plug cutter:

Here’s the two gegyo then after carving and installation of the Bloodwood petals into the Ebony pentagon ring and fitting of the center buttons:

Now here’s the perplexing question: how to attach the gegyo to the ridge and barge boards (hafū)? I mulled this over for many days indeed.

The Japanese use three different methods:

1) sliding dovetails on top of the gegyo fit into dovetail mortises on the lower edge of the hafū – clean and simple, though the gegyo cannot be readily removed if it, or the end of the timber behind it, needs to be repaired.

2) a mortise and tenon between gegyo and the hafū – one or two tenons each side of the miter. This solution results in exposed pegs on the face of the barge boards (a situation not to my liking, and one which would require extremely tiny mortises which were problematic to cut and make pegs for).

3) using dovetailed battens and dovetailed wedges to form a stiff-back method of union spanning the three pieces and helping lock them all together. This is the strongest method and the best way to go about the connection, it seems to me, however it wouldn’t work with my one-piece hafū.

My ‘simplification’ in terms of a roof with boards and no rafters and a hafū that was a one-piece version of something that would normally be two or three parts, along with the peculiarities of assembly (and the small size of the parts) – led to the result that none of the traditional attachment methods would work all that well for me. To top this off, unlike a gegyo on a regular building, tucked up high and which only needs to be rigid enough to withstand the wind, in the case of this lantern, about 6′ (1.8 m.) tall, the gegyo is well within grasp of the hand and, like candy, may induce someone to grab it and maybe even try to remove it. So, I need a strong connection that works with the particulars of the situation I had created with this ‘simplified’ roof. This proved to be a challenging issue, design-wise.

Here’s the solution I came up with, after much rumination. First I routed a piece of MDF to serve as a holder for the two gegyo, making them easier to work upon:

Here they are mounted in place, the MDF opening slightly tight to make the pieces hold in place by interference-fit friction:

I then laid out and began processing a pair of sliding half-dovetail mortises:

I was able to use a router to remove most of the material, but given the small size of these pieces and my lack of a small-enough dovetail bit, I hand-chiseled the dovetail slots:

A closer view from another angle – that Bloodwood and Gabon Ebony combo is hard on the chisel edges:

You can see in the above picture that the dovetails will help to also anchor some of the internal parts of the petal and pentagon rings to the backboard, thus, my comment yesterday about the glue being mostly a placeholder in this case.

Then I made up a pair of spandrels (brackets), which would triangulate the connection from the bottom of the ridgepole to the backside of the gegyo. First I cut the sliding half-dovetails:

On the underside of the ridge, I cut a 14˚ dovetail slot mortise:

Here’s the spandrel-gegyo connection:

The completed joint, drawn up:

A view from the frontside:

Notice that the spandrel remains to be cut yet from the blank.

I then severed the head of the spandrel (hah! -remember the discussion of the kanji ‘県’ yesterday?), followed by some chisel work, and then I processed the male 14˚ dovetails. Here’s how it turned out:

From the other side:

Later I cut some scooped facets along the curved arrises to give the spandrels an edge treatment.

Test fitting the spandrel into the end of the ridge:

They proved to be a very tight fit, requiring a mallet to tap them out – just what I was looking for in this case. The 14˚ bit is much better for sliding dovetails I think that the typically shallow angled, 6~9˚ dovetail bits used for carcase dovetailing. There are some 18˚ and 25˚ bits available from Japan that I would very much like to try out. I don’t worry about short grain issues too much with sliding dovetails.

Here’s the pair of near-complete assemblies:

I say near-complete, because there is yet more joinery work to be done on these little gippers yet.

Stay tuned – next is post 30.