Please help me repair this design fault

[Dell Littlefield]

I have a teak danish modern chair that I am trying to match with a bench/coffee table. I am using mesquite as it closely matches the teak. I did not realize the tremendous strain/twisting action the cord would put on the side pieces. Driving the L nails in after the weaving was completed caused the rails to twist enough to break the glue joint. I would like to repair rather than start over. This was a very tedious project. The bottom of the legs splayed out about 3/4 inch. I think I can get enough glue in the split to pull it back together but obviously need more strength.

I now realize that I needed stretchers across the legs. Any suggestions on how to best accomplish this? Through dowels? Rabbeted 1/2 inch stretcher? External stretcher glued
and maybe screwed?177144177145177146

[Steve Schoene]

Screws are seldom a good solution since wood moves as humidity changes and screws don't. Perhaps the dowel concept, but to be through dowels they would have to be just simple cylindrical dowels, not adding much to the design. I would use two cross members on the ends, one a short distance below the seat/top and another closer to floor level. I would even tie that lower stretcher end to end.

As I see it, you have a couple of sources of weakness. First the bridle joint is considerably more subject to failure as you showed than a mortise and tenon would have been. And, the choice of the leg timber may also have contributed since it appears to me, though it is a bit hard to see so I might have it wrong, that you have grain runout just below the bridle making that a particularly weak spot. I would also ask how thick those legs are--I don't really have a scale to go by but they look a bit delicate.

I don't understand L nails--where were they driven, and how heavy are they? Could they have acted as a wedge?

[Dell Littlefield]

I was under the impression that a bridle joint was stronger than a mortise and tenon and chose it for that reason. The factory made chair I am matching has a bridle joint at that location on the front legs. It has caused no trouble in the forty years I have owned it. (Albeit a double bridle joint) The legs are one inch thick. I agree on the grain runout, that one slipped by me. The L nails are at one inch intervals into the rails on the underside. The cord is woven and hooked on the L nails. After weaving, the nails are driven further in to better secure the cord. I under estimated the twisting force that would be applied to the rails or I would have incorporated stretchers. I do have two stretchers between the long rails under the top to prevent bowing caused by the weave.

[Jay Allen]

It won't help you fix what you already have, but I would say that at least part of the problem is that the tennon portion of the bridle joint is too thick. That makes the outer "cheeks" too thin, whis is further weakened by the shorter stretchers being cut into that surface. How are they attached? mortise and tennon? dowels?

[Dell Littlefield]

I agree with you about the thick tenon. I didn't consider the twisting motion when designing. The cross rails are mortise and tenons. The unseen stretchers are held by dowels. I figured the dominant force would be to cause bow to the long rails, apparently they are thick enough to withstand the twisting motion. I am thinking about adding stretchers with screws. Dowels partially through the stretcher up from the bottom would provide cross grain strength. I'd then plug the holes or add raised buttons. Comments?

[Peter Quinn]

It won't help you fix what you already have, but I would say that at least part of the problem is that the tennon portion of the bridle joint is too thick. That makes the outer "cheeks" too thin, whis is further weakened by the shorter stretchers being cut into that surface. How are they attached? mortise and tennon? dowels?

I'm thinking the same thing. The tenon part of that joint is too thick, or the female part of things is left simply too thin, and the grain you have at that break is pretty short, sort of turning the corner in a weak way. This is more of an autopsy than a remedy of course. A fix might be some extra rails just below the existing ones, with loose through tenons driven through the outside of the legs into the new rail. And maybe some glue and clamps to bring the split back together. I see why you would hate to loose all that beautiful cord work, I would too. Perhaps a brass rod as a rail epoxied in and possibly held in place with a hidden machine screw to the inside of the legs would work as well for both strength and visual interest?

[Lee Schierer]

I'm not sure the design is repairable so that it is safe. Your bench appears to be about 40" long. Using the Sagulator (http://www.woodbin.com/calcs/sagulator.htm) it appears the sag will be about 1.5" at the center when a 200# (not counting the dynamic loading) person sits on the bench. The bowing of the long side rails will also splay out the legs putting significant torsional stress on the bridle joints, which I agree are mis-proportioned. This will be very bad if the bench is placed on a slick floor surface like tile or wood. It might fare better on a carpeted floor where the feet won't slide. Tying the legs together with stretchers near the floor would certainly help it's ability to carry weight. I would do some load testing in the shop before placing this in the house where a guest might sit upon it.

[Greg Hawthorne]

I think Lee's observation is correct - the seat as it is at the moment is probably too fine for today's well-proportioned frames. Stretchers across the shorter sides would suppport the joints, but the greater strain will be along the long side, so support is needed here also. The simplest way to attach the short stretchers is with through-dowells, possibly plugged to improve the aesthetics, closely followed by a skein nut (threaded insert nut) if you don't mind visible hardware. The longer side is more problematic as it would be more difficult to blend stretchers in with the existing design, so here I would have a central long stretcher attached to the stretchers on the shorter side, as in traditional chair design.