Re: Musings on wide tenons
Dave Mount
>To address Matt's original question, I wouldn't hesitate to make 4" wide tenons on your 5" wide rails.
On the overall issue of wide tenons, I got into a discussion about this on a board some time ago, I don't remember where. As far as split tenons versus single wide tenons, the consensus was that that practice, as Adam suggested, was originally intended to stabilize the mortise by bonding the sides of the mortise together in the middle.
I think a lot of M&T design depends on whether you consider glue to be a "permanent" (relative to the life of the piece) bond. If that bond is viewed as "permanent" for all intents and purposes (i.e., it is acknowledged that when the glue fails, the piece will fail), then the glue bond between the tenon and the mortise holds the mortise walls together, so that a divided tenon is unnecessary. I'm not advocating this perspective, I'm just pointing out it that affects engineering of the joint.
The expansion across the width of a rail (I'll use a mortised leg and a tenoned rail as a reference joint) is the same regardless of whether the tenon is split. So the only argument I can find for split tenons as a way of accommodating expansion is if the two smaller tenons actually flex. This seems unlikely to me, but I can't say it's not so. If a piece is built wet and later shrinks, it seems to me a split tenon would encourage splitting of the rail to a greater extent than a single tenon, by focusing pressure on the outer edges of the rail.
Another option is a single wide tenon for which the mortise is cut over width to allow for expansion, and the joint is only glued over a portion of its width. This keeps stress from expansion/contraction off the glue joint and allows other parts of the tenon to move unrestrained. However, the joint loses some strength in regard to racking because the physical restraint of a tight-fitting mortise is lost. Resistance to racking is now limited to levering the shoulder of the tenon against the leg, which is opposed by the glue (or pegs) resisting tension pulling the tenon out of the mortise. While this is less strength than you would get from a tightly fitted mortise, if it keeps expansion from damaging the joint, it may be a sacrifice that's worth it.
One could strengthen this latter arrangement by gluing the center of the tenon and pegging the sides. You could even go crazy and elongate the holes in the tenon to allow the pins to slide, kind of like a breadboard end.
It seems to me that many wide tenons are done not for strength per se, but because design aesthetics dictate that a wide piece join a perpendicular member, and we just reflexively make wide tenons. In the case of a wide rail, a wide tenon is doing two things. One is (potentially) strengthening the joint by providing a lot of glue surface and/or a wide lever arm (mechanical advantage) to resist racking. The other is to align the joint across its width to resist twisting. Where the wide rail is needed more for aesthetics than for mechanical strength, I think some alternate designs might be better.
For example, consider a 10" wide rail. Rather than making an 8" or so single tenon, or a split tenon of equal overall length, I think it would often be better to make a single, narrower deep mortise (say 3-4") and tight-fitting tenon, then a shallow groove extending out the full 8" or so with corresponding haunches on the tenon. One would then only glue the deep tenon and leave the haunces to float (no glue on the haunches), providing alignment only. The shallow groove housing the haunches can be cut oversize to accommodate expansion. This gives the strength of a tight fitting deep mortise, but allows cross grain movement without much stress, and still resists twisting of the rail.
In the case of a rail that receives racking stress mostly in one direction (like a rail in a door, one could offset the smaller deep mortise/tenon to the side of the joint that is in tension and create a greater mechanical advantage to resist racking.
Someone else mentioned wide tenons in timber frames. I think this is a different beast for a couple reasons. One, timber joints are not glued, which changes some things (as it would if furniture joints were not glued). Second, most timber frames are constructed so that there are a number of opposing forces that work together to hold the frame in place, frequently involving diagonal braces to resist racking. This is a very different situation from a purely rectalinear (sp?) construction like a door, where the resistance to racking has to come entirely from the M&T joint itself. I still would be concerned about the walls a 12" wide mortise blowing out if the joint were subject to much twisting and the mortise walls were not sufficiently thick. This would be a case by case thing.
My thoughts, probably not worth any more than you paid for them,
Dave