>I am making a clear pine table about 5' 3" long by 3' wide, with legs 4" square in section - the design calls for a rather narrow apron, just under 3" wide, and this will be the only bracing for the legs.
Instead of a conventional tenon, I am considering recessing the whole of the apron section into the leg, to a depth of about an inch. Would this be better than a narrower conventional tenon? Or perhaps should I recess the full width into the leg, leaving about a half to 3/4" shoulder at the bottom? I would greatly appreciate some advice.
>A full stock width tenon would be slightly stronger, but you would be setting yourself up for assembly problems. All of your mortises would have to be the same depth exactly and all your apron pieces would have to be exact in length and thickness as well. Any flaws in the mortise would show in the final piece. A slight shoulder 1/8"-3/16" would greatly help in assembly and would hide a mortise that's just a skosh too big. Lots of table are made with just an apron for bracing and they last a long time.
>If the apron stock is 3/4" I would make the tenon 3/8 and full width.I would not bother with the shoulder at the bottom. It is just another operation. Just cut your mortises accurately. If you want the shoulder make it 1/8 or so. I would use a haunched tenon also. Cut the blanks for the legs one inch longer than needed trimming the extra material off plush to the top of the apron after glue up. This extra material will keep the blank from splitting while machining the mortise and gluing up.
I would think that some sort of corner bracing under the table connecting two corner rails would also help stabilize the legs and wouldn't be seen from above. I've used this approach on smaller tables and it really helps. I agree with Ed and Virgils advice regarding the rail tenon.
>I posted the same question a few months ago and got similar advice. Ellis (as in Wallentine) advised me to make the mortises 1/2" wide and at least two inches deep. I had suggested that I was gong to make them 3/8 and In fact had already cut them. On his advice, I went back and recut them and cut the tenon's to fit. I also had legs almost four inches square.
The purpose of this joint is to transfer all the stress on the legs to the rails and have the strength of the rails resist that stress. Smaller shorter tenons can't do that. Either the tenons will flex or break or the glue joint will fail. If the tenons are too large, the mortises will weaken the top of the leg and the leg will fail. I suppose if you tested a perfectly built table to destruction you would have the top on one leg fail, one rail break, one glue joint fail and one tenon fail.
It occurred to me that the "weak link" in my design was the long side rails. They were about six feet long and only 3/4" thick. I had no thicker stock. After I made the tenons and glued up the table, I laminated another piece of stock (also walnut, but not very pretty) to the inside of the long rails and tightly against the legs on the inside. Mine were four inches wide. The people that I made the table for commented that the legs were very sturdy, as if the table were made of a single piece of wood.
I also agree that cutting the shoulders on the tenons is very important. For assembly, they are your "control surface". You pull the legs onto the tendons until the gaps close and you are "there". They cover any sloppiness that might have crept into the M&T's. I think they also add to the strength, because the change the characteristic of the stress on the glue. The shoulders act as fulcrums so that the stress is a lever action trying to withdraw the tenon from the joint rather than just trying to rotate it causing the glue to fail.
I did not cut my mortises deep enough to cause intersecting tenons. My thinking was that must weaken the top of the leg. With a nice fat leg, you can still have plenty of depth on the tenon with out intersecting the other one. I cannot imagine that two tenons that are mitered on the end and touch each other develop any strength from the end of the miter.
I think it would be interesting for someone to do testing on this and make up joints of numerous designs and test them to destruction. But to do that well would require making dozens or hundreds of joints of different types of wood, with different types of glue, using different tolerances for the joints and different techniques for cutting them and so on to cover all the variable. OTOH, we can sort of reverse the process and there are those among us who have the privilege of working on or owning the pieces that that survived for centuries, and a little forensics will tell us what worked for them. The fact that we can sort out stylistic requirements and then find that big deep M&T's with a haunch on at least one side has survived thousands of times for many years and we see almost none with rails just butted to the legs tells us a lot.
>Actually, Barry, I think "Woodworker's Journal" did tests over a couple of back to back issues about a year or so ago. They tested not only different types of joinery but types of glue, if my memory serves me. Pretty interesting.
>I would be interested in the article on tenons and their testing. Often those tests are not exhaustive and generate a lot of flack from all of us who think we know better. Some of it well intentioned and some of it just bluster. I would like to see a test where they broke the joints under steady pressure and then some kind of impact to simulate someone falling against a table or someone sliding a table across the floor with the legs chattereing. I would also like to see them store some of the samples for a year and five years and ten years and then break them to tell us how well the glue holds up over time.
Rob, the "one third rule" mentioned does not exactly apply to your situation. Its generally applied to things like cabinet doors where the rail and stile are the same thickness and you are trade the strength of the tenon off against the strength of the mortise which serves as a housing. The tenon can be thinner sense the mortise has something of a bracing or reinforcing affect on it. OTOH, the larger tenon tends to weaken the mortise or at least the housing or cheeks. Again, a balancing act.
On a table leg, the leg is usually much thicker then the rail, so the tenon can be much thicker than one third of the thickness of the rail. What you trying to find is that magical combination that is as strong as possible.
One of the things I thought of was adding a third "rail" to each leg. This would be a piece running back under the top at a 45 degree angle. You could cut the inside corner of the top of the leg off at a forty five to make a flat area bout an inch wide. Then cut a mortise into it and add a block with a tenon running back under the table. I was considering running a rib under the top from side to side and joining this brace to the rib. It even occurred to me that the brace could be wider that the rails. Say, for your three inch rails, it could be curve on the bottom and be five or even six inches wide where it joined the leg. Since it curves up and recedes under the table, it would not be much of a knee bumper and not very noticeable. I guess "knee brace" might be a name for it. Never seen a table built that way and I wonder why? Is that it is unnecessary or that no one thought of it? I was going to do it on the table I built, but it turned out to be so sturdy it seemed unnecessary. Maybe I answered my own question.
>You should use at least 1" stock for the apron, especially so in something weak like pine. Shoulders on the the tenon as described above are essential for strength.
>Rob, I wouldn't use 3" wide pine for the rails. They might be sturdy enough to support weight on the top without flexing --though I have my doubts -- but they don't offer you enough shoulder height for rigidity. By all means, use shouldered joints because, as Barry said, the shoulders are critical for preventing racking. If you have enough glue surface in the mortise-and-tenon joint -- i.e., the tenon is wide and deep enough for the design strength of the joint in a given piece of furniture -- the shoulders will make it impossible (famous last words) to lever the rail up or down relative to the leg. The problem with your narrow pine rails is that the most shoulder you can get is 3". Also, since the table is made of pine, both the legs and rails are subject to compression when leverage is applied.
There are no hard-and-fast formulas for joint specifications, because there are too many variables (wood hardness, gluability and glue strength, machining tolerances, length of rail, height of shoulder, dimensions of tenon, etc.) that enter into it. Sometimes you have to go with your gut feeling, which hopefully is informed by experience or by talking to other craftsmen you trust. You also have to decide how much of a safety factor you want to build into the equation. In my case, I like to design for worst-case scenarios, like if Refrigerator Perry should trip and fall against one end of the table. That would amount to a torture test of the leg-and-rail joints, and I have my doubts as to whether 2" x 3" tenons in softwood would stand up to the test. Go as deep and wide as you can, and if you need even more glue surface, make double tenons.