* If you have linked to this page from another site and want to see the most recent post on the Carpentry Way, click here.
I've been working on the design for the mizuya and considering various options. With so many options, sometimes it can take a while to get mental traction with a direction in which to head. I took a look again at my Chinese Classical furniture texts and magazine and found several excellent examples of cabinet forms that would work, with some tweaking. They gave me what you might call hunches about hutches. I've also looked at Korean traditional chests, which share attributes with both Japanese and Chinese examples.
At this point I am leaning towards frame and panel construction for this piece, if for no other reason than it allows for a more pure form of joinery, as, unlike joined carcase construction, the connections need not be reliant upon glue. In fact, insofar as possible, I will build this piece without any glue. Regular readers of this blog will I'm sure have a good handle on some of my reasons for minimizing the use of glue, so I'm not going to take another stroll down that lane at the moment. I do however have more to say on the matter and I was feeling a certain sense of validation though during a recent re-read of an article Bruce Hoadley wrote for Fine Woodworking, back in the day when it was worth buying. In issue #21 (March/April 1980), Hoadley wrote an article on why dowel joints tend to fail and proposed a solution involving elastomers. While I have zero interest in dowels or elastomers, I noted the following comment from Hoadley when he compared the performance of oak dowel joints held together with PVA (white) glue:
Predictably, with fairly thick elastomer layers...the joints are able to withstand severe moisture cycles (6%-24%-6% MC) without losing withdrawal strength. The same cycle destroys a standard PVA joint. For example, in oak joints with white glue, it took an average of 1,100 lb. to pull apart un-cycled joints. But after a 6%-24%-6%moisture cycle the average withdrawal resistance was only 41 lb.
It's a shocking difference in performance to see how much weaker a glued joint can be after being subjected one round of moisture cycling - at least when using regular woodworker's PVA adhesive. A change from 1100 lbs. to 41 lbs. is a 96.3% decrease.
Hoadley's empirical observation suggests that one thing missing from the 'tests' one sees every few years in the woodwork magazines, where they compare the strength of various forms of joinery, is that they do not examine the performance of the joints after moisture cycling. Therefore these tests appear to be of dubious real-world value. Unless you live in one of those places where the humidity level doesn't change much during the year, and are sure that all the furniture you make will remain in such an area, the issue of moisture cycling and its effect upon the material and the connections is something to which you need to pay attention and an in-service factor which your designs ought to reflect. Yes, that was a value judgment!
Now it is true that many modern adhesives, like epoxy, form bonds stronger than the wood itself, and that joints with decent amounts of glue surface area are typically strongest of all joints. Examining the results of destruction testing though, in most cases we see that the glue bond holds well and that the surrounding wood shreds and blows apart. However, it begs the question: if the connection is to fail, do you want it to fail in such a manner? Once the joint has failed, what about repairing it?
It's like the power take-off (PTO) winch on my Toyota LandCruiser. Driven by the engine, the winch can pull some serious weight, however the engineers who designed it thought through what usage the system might see, and wisely placed a shear pin in the system that will give way if the load becomes too high while winching. This allows a $5 part to fail instead of something more catastrophic occurring. The system is only as strong as the weakest link, and ideally the weakest link is adequate for most tasks and is inexpensive to repair and easy to repair. This aspect attributes to mortise and tenon joinery, and makes it durable, resilient, and more easily repaired down the line. And if glue is needed, as it surely is when edge joining boards to make a panel etc., it makes more sense to me to use a glue that is reversible, like hide glue or fish glue, etc. If you're designing with repair-ability in mind, you're not designing to make pieces last, and that means you're not respecting the material you work with. Quality construction generally starts with quality materials, and quality materials are generally less abundant and took longer for nature to produce, so they must be used wisely and in a sustainable manner. 'Sustainable' is just another marketing term these days it would seem, however for me it means that if the tree grew for 150 years, than anything made from that tree should have a designed lifespan of at least 150 years, and be recyclable/re-purposable if at all possible.
Another reason I'm trying to design glue-less, bolt-less connections for this piece is that I am leaning towards using Cocobolo (dalbergia retusa) for the frame components. The finest classical Chinese Ming furniture was typically made of one of two very hard highly-polishable hardwoods, huanghuali or Zitan. Neither of these woods is available, practically speaking, however huanghuali is a rosewood, dalbergia odorifera, and it seems to me that another rosewood would be a fair substitution. Generally the rosewood available today is on the small size, and most species, due to over-harvesting, are now banned for trade altogether. One rosewood still available, though not in what you might call abundance, is Cocobolo. I managed to locate a stash of it several months back in Tulsa, Oklahoma of all places - this material that was imported into the US in the 1960's and has been drying for 50 years now. So I have stumbled into a modest pile of it, some sticks of which are quite large chunks indeed. Cocobolo is an oily wood which is tricky to glue and finish, and that suggests to me that the 'best' way to use it would be without glue. It's best to avoid the conflict and downstream problems of failed glue joints if at all possible.
'Wide panels' and 'cocobolo' are two things that don't really go together. While one could glue up narrower boards into wide panels, there is the aforementioned glue problem. I know there are ways to overcome that challenge, however I would prefer if at all possible to use wood of sufficient width to make a panel without recourse to glue-up, or making an overly-busy frame with many dividing bars to accommodate individual slender panels. So, at this point I'm thinking that the panels in this piece will be bubinga, which is readily available in wide pieces.
At this point the mizuya is being designed around a cocobolo frame and bubinga panel system, however this is early days yet and the deisgn is by no means hardened into final form. This is the sketch as it sits now:
The drawers are about the only storage arrangement I have settled on so far. The very top section will have sliding doors.
Here's a view of the back, where I have positioned a wide board, double-tenoned into the post at each end:
The drawer dividers connect to the back board with multiple mortise and tenon joints and should impart decent shear resistance to the entire cabinet. I anticipate taking further steps in that direction of structural bracing yet. At this point the rear panels are obviously too wide, and will be vertically divided. I haven't set the location for those vertical divisions yet, as they relate to the arrangement of sliding doors and drawers on the front. The rear frame and panel backpieces will be made so as to be demountable, just as I did in a bookcase build several months back.
All for today - the design work rolls onward into places unknown. Thanks for your visit. On to post 2