Thread: Kintaro Yazawa Box with Key Hole Joint

When questioned about the construction of that keyhole joint Yazawa simply said: "My secrets will go to my grave".

I have been fascinated by that joint and have been attempting solve it for almost 20 years. I think that I may have a workable solution in two parts:

1. Craft the pins and tails of the joint, then
2. Make an elegantly simple variation so that it may be assembled.

I do not want to reveal step 2 in case it is an intellectual dead end. However, I am stuck on Step 1. How do you actually cut that joint with the needed level of precision, and lack of tool room?

Drawing - Assembled.jpg



Drawing - Dissassembled.jpg

Drawing - Rear View.jpg

Remember, this joint cannot be assembled until I work some "impossible magic" on it. However, I cannot test my hypothesis until I make a pair of joints as above.

The joint must be cut to extremely fine tollerances - precision is essential. Some parts seem relatively straight forward:

• The round tenon is just a drilled hole,
• The flat tenon - continuation of round tenon - is also easy.

But I cannot think og a way to cut other parts, certainly not with the required level of precision:

• Keyhole tenon contains both round and flat elements,
• Round part must be truly round,
• Flat part must be flat and precise,
• Mortices either side of keyhole tenon are not uniform,
• Curved section must match perfectly with round part of tenon,
• Ditto - flat section.

How can I cut this? Yazawa has done it, so it is not impossible.

Is it possible that only the top fingers on the front and back panels are shown to be square, while the remaining fingers aren't square at all but cut at 45 degrees.
The joint could then be assembled by joining the pieces diagonally.
The whole box would have to be done at the one time, as the sides would pull together as assembly takes place.
A bit like the impossible dovetail
Tom

The keyhole tenon will look like this:

• Very high level of precision needed,
• Indicative thickness of flar part is 3 mm,
• Indicative diameter of round section is 6 mm.
• Transition between round & flat is highly visible.

Drawing - Keyhole Tenon.jpg


The mortices on either side of the keyhole tenon will look like this:

• Must accept the tenon perfectly.
• All comments re tenon apply'
• Indicative maximum width of mortice is 3 mm.
• (Two mirror images give variant perspectives.)

Drawing - Complex Mortice.jpg


The fact that most of the work has to be done inside rather restrictive mortices really complicates this. Maybe it is too hard?


A doctor friend defined lararoscopic surgery as being akin to painting the lounge room through the keyhole.

Comments, please.

Originally Posted by Scribbly Gum

Is it possible that only the top fingers on the front and back panels are shown to be square, while the remaining fingers aren't square at all but cut at 45 degrees.

That was my first reaction, too, Tom. I have tried drawing it multiple times with variant angles - 45°, 50°, 60°, etc, but none worked. Also tried it with "circular tracks" so you twist it in situ - another failure.

Yazawa has confirmed that it is not a varient of the angled dovetails, twisted dovetails, rising dovetails, etc. but has otherwise remained stum.

Ooh interesting! Could the keyhole hole be done with a forstner bit? And the key made with a tool similar to a dowel making block, only with a slot for the finger to pass. Perhaps even a plate with the full array of holes for the board so it could be registered with the edges for accuracy and perhaps pressed through with a press rather than hammered. I mean, the keys would have to be shaped roughly to size so the block just does the final cut, similar again to the dowel making block.

If you zoom in on the detail photo of the joint, it looks like there is a slight mismatch in the grain on the left hand keyhole tenon, right where the corner radius finishes. My guess is that they are cut off with a fine saw, the box front & back are thinned down by the thickness of the saw cut, then the keyhole tenon fitted as a loose piece after the box is assembled. Only the very outside of the joints have to be a neat fit, the rest could be made with clearance as it can't be seen.

The sides are laminated together on the left one in the photo, join is in line with top of 2nd box joint and top of and bottom of 4th box pin

Originally Posted by 62woollybugger

If you zoom in on the detail photo of the joint, it looks like there is a slight mismatch in the grain on the left hand keyhole tenon, right where the corner radius finishes. ...

I have enlarged the original photos of three boxes and there is absolute continuity of wood grain through the tenons - no hidden glue joints.

On an earlier version he had a hard right angle corner to the box - a similar join could have been hidden there - a mitre join can be almost invisible.

But in these versions Yazawa slightly rounded the corners, so there is nowhere to hide.

It is definitely all solid wood.

Originally Posted by rwbuild

The sides are laminated together ...

No laminations - look at grain continuity.

Originally Posted by mic-d

Ooh interesting!

I have been interested for almost 20 years. And frustrated for most of that time.

Could the keyhole hole be done with a forstner bit?

That would probably be the best way to cut the round part of the mortice. The flat section of that keyhole mortice could then probably be cut with two passes with a very fine saw such as a Japanese dozuki or dovetail saw.

And the key made with a tool similar to a dowel making block, only with a slot for the finger to pass. Perhaps even a plate with the full array of holes for the board so it could be registered with the edges for accuracy and perhaps pressed through with a press rather than hammered. I mean, the keys would have to be shaped roughly to size so the block just does the final cut, similar again to the dowel making block.

This really stumped me. Perhaps a very rounded gouge or perhaps a custom scratch block or .... ?? But how would the keys be "roughly shaped to size"? There is very little clearance for tools in the mortices that you are simultaneously cutting as you shape a side of the adjacent tenon.

Graeme, what do you think we can rule out? Because of who's done it, should we rule out lamination of wood? Is it safe to assume it's 'joinery' not face veneers for example, or something like this model I found, which seems quite unlikely due to fragility.
huh.jpeg

We can rule out a circular pivot into position because you can't get the last corner of a box together this way I think.

Can we rule out compression of the wood before assembly and boiling water after to swell the 'dowels'. Ive seen some pretty radical compression of pine in puzzles. The base of the box looks like Paulonia which would be a candidate, but the sides look more like NGR?

I am favouring an idea that there are voids in the joint and the dark shadow in the close up seems to support this. I have the beginnings of an idea for how the cuts could be made fairly simply using a dozuki and coping saw and one other hand tool ;-D I'd like to try a joint first though. If you'd like to know what I'm thinking I could take it off line by pm just until I try it out

Originally Posted by mic-d

Graeme, what do you think we can rule out? Because of who's done it, should we rule out lamination of wood? Is it safe to assume it's 'joinery' not face veneers for example, or something like this model I found, which seems quite unlikely due to fragility.

Kintaro Yazawa has a stella reputation and works in solid timber. He has stated explicitly that it is solid timber, absolutely no veneers, but has declined to go further - "My secrets go to my grave" - great marketing.

I have seen very high resolution photos of three of those boxes and have been unable to detect any irregularities in the grain pattern. This confirms to me that it is solid wood.

I first saw that "solution" that you quote some 10-12 years ago. It simply did not work because the short grain was too fragile.

huh.jpeg Short Grain.jpg

We can rule out a circular pivot into position because you can't get the last corner of a box together this way I think.

I could not make a circular pivot work, either.

Can we rule out compression of the wood before assembly and boiling water after to swell the 'dowels'. Ive seen some pretty radical compression of pine in puzzles.

Yes, on pragmatic grounds. There is very little space in the mortices. How could you get the jaws of any clamping device into that space and then apply pressure?

The base of the box looks like Paulonia which would be a candidate, but the sides look more like NGR?

I think that the timbers are Japanese white oak and Japanese red oak, but I profess zero expertise in identifying Japanese timbers. Paulownia is certainly popular in Japan, but ... ?

Yazawa uses traditional urushi lacquer, which may be very high gloss or matt, on much of his work. Presumably the boxes are finished with urushi which may impact on the colour.

I am favouring an idea that there are voids in the joint and the dark shadow in the close up seems to support this. I have the beginnings of an idea for how the cuts could be made fairly simply using a dozuki and coping saw and one other hand tool ;-D I'd like to try a joint first though. If you'd like to know what I'm thinking I could take it off line by pm just until I try it out

Mmmmm! If there is a void, how would you create it? There is so little tool room.

If you have a wrkable idea for cutting the basic shape, please try.

I am stumped on how to cut the rounded part of the keyhole tenon and the mirror image part of the mortice. Its all about tool room, I think, or some brilliant lateral thinking?

PS: The Yazawa keyhole joint has been discussed extensively on many forums starting about 15 years ago - including Woodworkform and FestoolOwnersGroup and others.

I think the round, dowel shaped bit will be at 45 degrees to the sides, you cannot see this as it’s completely hidden.
If they are all cut and drilled at 45, all 4 sides will be able to be pushed together simultaneously, creating the illusion that you can see from the outside.
This is the only way this joint can be accomplished using solid timber.

Originally Posted by Ironwood

I think the round, dowel shaped bit will be at 45 degrees to the sides, you cannot see this as it’s completely hidden.
If they are all cut and drilled at 45, all 4 sides will be able to be pushed together simultaneously, creating the illusion that you can see from the outside.
This is the only way this joint can be accomplished using solid timber.

A variant of the "impossible dovetail" technique??? Impossible Dovetails | Popular Woodworking