Mizuya (2)

Design work has been proceeding along with the kitchen storage tansu, or mizuya. I’ve fleshed in some more off the piece and thought I’d share some pics of the piece as it sits now in concept-land.

At this point, I’m thinking that the uppermost level will have 4 sliding doors, the level below that 3 or 4 sliding doors, then the 4 drawers (with a possible horizontal splitting to happen with one of them so as to have 5 drawers), and on the bottom level a pair of sliding doors on the left and a hinged door on the right.

An elevation view for the front:

Elevation view, side:

Perspective view, rear:

One more perspective view from up on high:

There will be a few more interior shelves to come yet.

Corner Conundrum.

The design of the corner joints took a lot of time to solve to my satisfaction. In the end I remembered a form of Japanese timber framing joint that I had used on a well pump shed project many years ago that might work for this application:

This locking joint, which shows a miter on the exposed face, is termed dodai sumi dome hozo sashi shiguchi. The word dodai refers to the mudsill, however a form of this joint is also used to connect a type of eave fascia, urago, around a corner and up the gable end of the structure, where it gets a different, and equally long name.

I chose this form of joint because it mechanically locks together using wooden pins, shows a miter on the exposed faces, and allows adequate toom for a tenon to pass through. I haven’t seen this joint used on furniture before, however adapting timber frame structural joints to furniture is how it was done at the dawn of antiquity, so to speak, so nothing unusual here.

I made some modifications to the standard form, and here is what I came up with:

The post receives a haunched, double wedged tenon.

From above you can get a good idea as to the configuration of the parts:

First off, one of the joint halves slides over to partially interlock with its neighbor:

Then the joint slides 90˚ the other way to fully interlock with its partner, and the 2 fixing pegs, which are slightly tapered, are brought into the party:

Once the two sills/rails are connected up, the parts are slid down onto the post tenon:

Then the two wedges are driven in, completing the connection:

From the inside, one can see some of the joint mechanism, however in the cabinet I am making that mechanism will largely be hidden from view:

On my cabinet, here is an outside view of the same sort of corner connection:

If you were on a chair, you’d be able to see some of the joint on the top surface:

The usual view however, where you would be looking up at the joined area, nothing is given away:

I felt pleased to find a good solution to the corner join problem and have applied the same joint to all 8 corners of the cabinet.

The cabinet components are quite rectilinear, and that could look a bit severe, so I’m planning to soften the aesthetic a bit by using some molded curves and possibly beading on the doors, and have an idea to include a bit of delicate latticework as well. I’m still nowhere in the quest to figure out the hardware, though at this point it is mostly going to be a pair of hinges and possibly some drawer pulls.

Thanks for coming by the Carpentry Way. Comments always welcome.  On to post 3

14 thoughts on “Mizuya (2)

  1. Chris

    The design is looking good (except for the paucity of drawers) and your corner joint looks like a winner. But please explain how one arranges three sliding doors. Are there three separate tracks? Or does one only access 1/3rd of the cupboard?


  2. Tom,

    well, to have three doors and convenient access, one needs three tracks, which would entail slimming the doors somewhat. I kinda threw out the “3 or 4” doors quote without thinking too much about the track and door issue, so upon reflection, I will probably not be going with the three doors. I am leaning towards the hinged door on the lower right, but haven't ruled out drawers either! Thanks for your comment and question.


  3. That corner joint is _very_ nice. What kind of tolerances would one need to keep a joint like this looking nice? Are we talking in the order of 1/100 of an inch or even less?

    In a previous house I had a kitchen with two solid sliding doors per segment under the counter and in the cabinets, but I found them rather impractical. You can only see into half of the cabinet, and what you can see is pretty dark. And if the handles are worked into the doors as opposed to protruding, it is quite easy to get your fingers caught when you're opening the doors, especially if the doors are big.

  4. Roland,

    thanks for your comment. I think any joint should be cut with the closest tolerances one can achieve (of course, with some balance given to the time involved…). 1/100th of an inch is, to me, a fairly large gap in a hardwood furniture joint, and if cut out went awry to the degree that such a large gap was produced, I would either repair the joint or cut a new piece.

    As for the issue of sliding doors blocking some portion of the opening, that is certainly the case, and a drawback to using sliding doors. With more tracks, one can reduce the amount blocked down to 1/3rd or 1/4 of the opening. There are solutions to that problem of light passage though, and you will be seeing them here in the next post or two in this series. As for getting one's fingers caught in recessed pulls, I can't recall ever having had that happen to me personally, but I imagine it could be painful. I'm not sufficiently apprehensive about something like that to go to the trouble of ditching the sliding doors, which have their advantages.


  5. I would guess that on a joint like this one would test fit and work on the parts until you've achieved a good fit. So what would you say is an achievable tolerance in this case? Are we talking in the order of 1/1000″?

    The use of the different slots and keys to keep the pieces aligned is ingeneous. Also the use of two fixing pins to clamp the joint in two directions is also very interesting.

    On a CNC milling machine it would be relatively easy to make these parts accurately, but by hand it looks like a tough job!

  6. Hello again Roland.

    a tolerance of 0.001″ is achievable with a router, however not without taking some fairly thorough measures, including aluminum jigging. I haven't quite gone to the extent of aluminum jigs – yet.

    I find in practice, so long as I am paying attention, being present, etc., I can achieve a tolerance on the order of +/-0.005″ most of the time. That's what I shoot for. A slightly dull cutter on the router will cause some of the harder woods to deflect during cutting, or compress and case harden, and this is one factor affecting accurate repeatability. One can arrange to use a series of router bits to achieve a decent cut, with the idea being that the final pass is taken with the sharpest bit. It's not always convenient to do so of course, and sometimes one is under a bit of time pressure and finds that the ideal really sharp bit is not on hand. So, you do the best you can.

    I think the important thing is not letting a substandard result head out the door, and thinking ahead to prepare a few extra pieces in case something goes awry during cut out.

    As far as the ingenuity of the joint, we can thank some anonymous Japanese carpenter way back in time for that. I've added an internal dado and stub dovetail because the joint is less well supported than it would be on a concrete foundation, but the form of joint mechanism is a Japanese invention. I haven't come across the same connection in historic Chinese joinery, not yet at least.


  7. Before we got our CNC milling machine, we did a lot of routing with air-powered routers and aluminum jigs. I think it's impressive you can reach +/-0.005″! I don't think we ever got that tight in composite materials. But wood is probably easier to cut; In 4 mm thick carbon/epoxy it took us about 20 minutes to route 3 meters (10 feet).

    WRT sharp bits, composites eat carbide for breakfast, so we tend to use PCD. They give a nice clean cut, and last for a long time. Some manufacturers claim that PCD lasts 200 times longer than carbide when cutting wood.

Anything to add?