the Carpentry Way: Wood Moves                                                          

Wood Moves

Today I want to start on the topic of accuracy in work, a facet of woodworking that is of particular interest to me. Whenever a person undertakes to cut pieces of wood so as to form joints, and then fit them together, the accuracy of the fit is a major concern. It is a critical concern actually, because the closeness of the fit determines in large measure the structural integrity of the joint. Joints that fail in loading are dangerous in buildings for obvious reasons.

Any joint, like a splice or scarf will be weaker than an uninterrupted section of timber, so the joined timber as a whole is only as strong as its weakest point - namely the joint. A fit of a mortise and tenon that has too much internal slop will allow for the pieces to move when loaded and strength will be greatly diminished as a result. If the fit is too tight, then the risk is that the receiving piece will split when the tenon is forced in, rendering the joint greatly weakened. This is less a concern in working with softwoods, as the wood inherently can be compressed more - nevertheless, I would say that a lot of timber frame joinery, as practiced, relies unduly on the power of multi-ton come-alongs and massive sledges to bring the joinery together, which is often a bit too tight to begin with.

In working with hardwoods, there is far less tolerance for error in the fit, and from working hardwoods in the process of furniture-making, I have learned much about achieving good fits, and these lessons have transferred well to timber work I think. In denser hardwoods, the difference between a joint that fits just right, and if glued has just the right amount of space for the glue, is on the order of 0.005". Thus, if you want a perfect fit, you need to be able to cut the joints with something close to that level of accuracy, more times than not.

When I returned from Japan to Canada in the summer of 1999, I managed to gain employment a day or two after touching down with a company on Vancouver Island called Daizen Log Tech. It was run by a Japanese owner and shared lot space with B. Allan Mackie's famous Log Building School. Several employees were Japanese, and the bulk of the production was log home shells made for export to the Japanese market. They also had a 'timber frame' division (I use that term loosely), and that is where I, er, slotted in.

I was told early on by my supervisor that their tolerance of fit for a mortise and tenon was +/- 0.25", which I thought was a lot. Their rationale for such slop in fit was that the parts were easier to trial assemble and take apart in the yard (which was true to an extent), and that after a one-month sea journey in a container, they didn't want their Japanese crew wrestling over joints that had swelled and become too tight to fit. I knew this was plain wrong, but I was the new guy so I said nothing. I was told that under no circumstances did they want me fussing over the inside of joints in places no one would ever see.

I was also told by a veteran worker there, after I asked him why there wasn't a jointer ( a woodworking machine for creating flat surfaces) in the timber shop, that the owner had injured his fingers using it, and that besides, "what did you need a jointer for? The tablesaw cuts a straight line". My eyes widened, and I knew this to be completely fallacious. While the saw can and does cut a straight line, the behavior of wood, its movement, after rip cutting is another matter. If the surface of the piece of wood if twisted or bowed, etc, then the process of cutting 90˚ (or whatever the angle) in relation to that at best will produce an uneven result, and at worst is dangerous to the machine operator. However, he was the veteran and I said nothing. I've since learned that this is a relatively common perception. Most shops do not have jointers.

A little while later, when I was discussing with another worker my aim to get a good fit with joinery, the fellow turned to me, and with a definite tinge of scorn in his voice said, "what are you worrying about that for - are you crazy? Wood moves!" I stared, and gulped.

Daizen was quite a scene. They had a crew of 20-odd people in the yard, along half a dozen in a window and door plant. At any given moment there were typically 3 log shells in production, along with one 'timber frame', and the cacophony of 10 chainsaws was omnipresent throughout the working day. New employees were expected to have their own tools, including chainsaw, and starting pay was a less-than generous $8.00/hour. I lasted 6 months, and moved on to working for a real timber frame company on another part of the island. Eventually Daizen closed down, laid off all its workers, and the owner moved, acquired a Hundegger automated timber cutting machine (thus solving the 'employee problem') and now, last I heard, produces timber components with that set up.

Anyhow, Daizen was an interesting place to start out my woodworking career, as I got to see that 'professionals' are often not all that knowledgeable or professional in their work, regardless of how long they had been doing it. I was surprised to find that people with 20 years experience, though accomplished in their work, often had a narrow grasp of wood and woodwork outside of the specifics of their jobs. I didn't realize then that for most woodworkers, the job ended at 5:00, and they didn't give it further thought until the following morning, thus they didn't learn much beyond the day-to-day concerns of their job-specific tasks.

Since the Daizen days, I have worked in a variety of shops and situations, along with having a long period of self-employment. My thoughts and desires to achieve fine tolerances for the fit of joints has often been greeted, I have found, with either dismissal, disbelief, or scorn by other woodworkers. I find these attitudes curious.

Another aspect of tight-fitting joinery is simply the appearance of tight-fitting joinery. When any joint is drawn up tight, most of the internal mechanism is concealed from view, and thus what is available to the eye are the intersections on the surface. One way to get tighter, gap-free intersections is the process of undercutting, as seen in the picture to the left. This picture I grabbed off a web site page explaining 'how to undercut', as a path to joinery success, however I would say the picture is better described as how NOT to undercut. To my eyes, the undercutting has been taken too far. If the tenoned vertical piece in the assembly were loaded downwards, then instead of having good bearing surface at the shoulder, it instead bears along a narrow edge at each side, and the tenon would probably bottom out to boot. This makes for a weaker joint, and one vulnerable to splitting if loaded. If the tenoned piece were loaded side to side, the knife-edge at the outside of the shoulder would dig into the surface (end grain being much harder then edge or face grain) and crush the fibers down, leaving a gap later and a loose fit. Additionally, if the lower receiving piece shrinks appreciably, the undercut will become quickly apparent, and in the effort to hide a gap by mean of undercutting, the result will be to show an even larger gap. If the receiving piece swelled, the grain on its face gets crushed again, and when it shrinks back, there will be a gap and thus a loose fit.

I'm impressed when I come across crisp-fitting, 'light-tight' joinery, however if I later learn that the joint was heavily undercut to achieve that, my esteem disappears. The mechanical connections achieved by joinery are only as good as the integrity of the fit. Placing too high an emphasis on the external visual aspect at the cost of a quality fit internally is a poor decision in my view. Taken too far, undercutting a joint results in little more than a candy shell of a connection, and this defeats the purpose of the joinery in the first place.

Undercutting in fact, is often a 'blunt instrument approach', a means of quickly resolving issues in a joint fit that is not closing cleanly. While working faster is, economically-speaking, of interest, this shortcut does nothing to improve the craftsman's ability to actually achieve a crisp fit, thus undercutting can become a bit of a crutch. More often than not, it seems to me, undercutting produces inferior joinery.

Undercutting can play a role in circumstances where relieving a surface slightly will help in dealing with seasonal wood movement or shrinkage between pieces in a joint. An example of this is the practice of gluing boards edge-to-edge, where a slight amount of material is planed out of the middle of the contact zone along the edges, thus the boards abut at their ends but have a slight opening between them at the middle. This is termed 'springing the joint' and it allows the construction to suffer moisture loss at the end grain zones of the board (where moisture is most rapidly lost or gained) without the joint splitting open.

I have studied joinery from both Western (English, German, French and Hungarian) and Asian (Japanese and Chinese) perspectives, and in general, the western joinery methods are relatively undeveloped in comparison to the joinery elaboration seen in Japan. Not in every case, but generally speaking. English scarf joints, for example, reached what might be argued as a higher pinnacle of development, obviously due to a chronic lack of long straight timber in that country from the 1400's onward. Necessity is the mother of invention, as they say.

A really excellent resource for the study of English joinery methods and developments are the works of the deceased Cecil A. Hewett, especially in his works "English Cathedral and Monastic Carpentry" and "English Historic Carpentry", which is linked at the right of this page in the book list.

Hewett undertook the study of old buildings in the Essex region (primarily) of England, and by showing the development patterns of joinery was often able to come to more refined understanding of the historic process of building. The interesting thing he notes, to my view, is that for any given type of joint, the development process is very similar: early crude forms are gradually improved upon, time and consideration, sometimes simple insights of brilliance, proving the superior forms, until after often hundreds of years the zenith in development of the joint form is reached. Often, it is the case that the 'best' joint from a structural perspective is not such a good joint from the economic perspective, as complex joints take more time to cut. Thus, after a zenith of development is reached, there typically follows a drop off where inferior imitations are made, either by subsequent carpenters who did not fully understand the brilliance of the penultimate form, or who were, surprise surprise, trying to cut corners with a quickly-cut imitation. Thus, from perfection follows degrade, technically-speaking.

The point of interest here is that a carpenter of today, ignorant of such matters, may well see an old timber building and seek to emulate the joinery and structural systems, when in fact he is copying something that was poor in the first place. An example of this are pockets for joists cut into the Summerbeam simply by chopping out a series of rectangular pockets along the arris of the summer beam to house the floor joist. The pockets cut into the upper surface of the timber and thus weaken it unduly. The carpenter who studied the mechanical concerns and history of development in forms of joinery for floor joist attachment to carry beam would learn in time that the surfaces to leave alone are the ones loaded in compression or tension (usu. top or bottom surface in a beam), and mortises are best placed in the neutral axis of the receiving timber. Thus, it can be seen that the tusk tenon, illustrated on the left, is the superior form of joint for this application, in most cases.

Another example of poor joinery in timber work are single housed dovetails. While these facilitate easy assembly in that the joists (or other similar part) can be simply dropped down into the receiving pockets on the beams, this joint, especially when cut out of green timber, has really poor performance characteristics. When loaded in tension (which is what the dovetail is supposed to resist), the long grain side of the dovetail is opposed by end grain in the dovetail mortise, and thus when the surfaces meet the side grain of the male dovetail is readily compressed. Thus, the joint can withdraw from its housing quite easily. Factor shrinkage into that when using green material, and what results is a near-useless connection, albeit one that is easy to assemble.

There's a book out there, the '"Craft" of Modular Post and Beam', which shows a housed dovetail proudly on the cover photo, as an example of 'craftsmanship' or something 'cool' I imagine. The joint is 'reinforced' with a pair of dowels driven down from the top, as if the designer of the joint intuited that it wouldn't actually resist withdrawal very well, so he put a couple of dowels in there to 'beef it up'. The dowels render the logic of the dovetail, weak to begin with, largely moot, and in fact under tension I doubt the dowels would add significantly to the load resistance of the joint anyway. Love of form has trumped functionality - sadly this is true in many places. The housed dovetail illustrated on the cover of that book is only a hint of the gems of information you will discover inside, if you know what I mean.

As I mentioned in an earlier thread, I avoid single dovetail construction in timber work, and try to limit them in furniture as well, particularly in places where loading on the joint might be higher - like a table or chair leg for instance. The dovetail performs best when used in multiples, like in carcase construction (see "Steps Along the Way" on this blog), or when used with a long siding abutment surface, illustrated to the left, such as the Chinese use for battens to reinforce the underside of a table top, or the battens medieval English carpenters used to reinforce planked doors (and so forth).

Anyhow, those are just some preliminary considerations about joinery - there's much that can be said on the topic. Let's not forget though that an accurately cut joint, if a poor design, results in poor construction.

In the next post I will look more at the topic of accuracy, and how I strive to achieve that in my joinery work.

Labels: , ,