How tight should a mortise and tenon joint be

[Phillip Ngan]

I'm making a mission style bed in which there are mortise and tenon joints between the corner posts and top and bottom rails. I'm trimming the cheeks to fit and was wondering how tight the joint should be for a dry fit, i.e. how do I know when to stop trimming the cheeks.

A finishing carpenter (who does very high end work for places like museums) told me that when gluing joints to make sure that there is a 1/16 inch glue line. That is, to have ample amounts of glue. Any tighter and the joints will fail after a while.

I would love to hear your comments.

[Tom Hintz]

I shoot for mortise and tenon joints that I can just push together by hand, perhaps with a little persuasion just to finish closing them up. I was taught that fit by a retired master cabinetmaker that was an absolute magician with wood and in everything I have done with that joint, he seems to have been right. (Again)

[Van Huskey]

I like them tight enough to ALMOST seat with hand pressure, I want to be able to fully seat with just a very light tap.

[Mike Cutler]

Phillip

A dry M&T fitup should be pretty tight. The wood shouldn't be deformed, but it should be fairly tight. Maybe a light tap with a dead blow mallet at most.
Once glue is applied it will act as a lubricant and you'd be surprised how easily they will slide together then.

A 1/16" thick glue line sounds a little excessive too me. I would have to hear the context in which the statment was made though to pass judgement.
Generally a 1/16th to an 1/8th gap is left at the bottom of the joint for excess glue to squeeze into.

[Charles Lent]

I was taught that they should fit so they can easily be pushed together by hand, but gravity should not pull apart when they are turned on end. In other words a very close and accurate, but not a tight fit. The tenon should also be slightly like 1/16 to 1/8" shorter than the mortise to allow for a glue reservoir and to guarantee that the shoulders fit tightly when the joint is closed. The only tool that I've found that lets me produce them with this accuracy in a limited production type environment is the Leigh FMT. It was just not possible to make them with this kind of fit repeatability until I bought the FMT. When they fit this way you only need to place glue on the sides of the mortise. With the final assembly they will require clamp pressure to force them together, as the glue will tighten the fit, but it will go all the way together easily and reliably. A little glue will squeeze up into the shoulder area, gluing it together, but very little will squeeze out of the joint.

[Mike Cruz]

I have to agree that 1/16" sounds way too big. Maybe a pocket for excess glue but not in the overall dimension. Wood glue needs the surfaces to touch (and be fairly tight) to work. Look how you clamp everything else together. You wouldn't expect to be able to lightly lay two boards next to each other and expect them to hold. :confused:

Obviously, if you have to beat them together or apply a pipe clamp to get them together, that might be a little extreme.

A tip I had learned was to apply glue to both surfaces and let glue soak in as much as possible before assembly. If you apply and quickly assemble, you might end up wiping off most of the glue. Of course this applies to situations where you have a choice. When doing multiples at one time, you do what you can. ;)

[George Beck]

I like the tenon to fit snug by hand with maybe a small tap to seat the tenon in the mortice. The modern glues hold so well I don't find you need lots to get a good joint. I measure tightness by making sure the piece doesn't move after seating, such as in an apron to a table leg. Usually I am shooting for being able to push the joint home by hand, have no wobble after seating and may have to tap a bit to disassemble the joint. Water based glues will cause the wood the swell a bit so I think a hand push compensates. A little trick I learned years ago, was after the joint is fitted as described and jus before glue up, I will take a sharp chisel and pare a little of the sides and cheeks of the tenon, forming a very slight wedge shape. This is no way close to 1/16". Perhaps only a few thousandths. The joint starts off loose and tightens as you drive it home and gives the glue a place to go.

[Jacob Mac]

I agree with everything above. Just be sure to put glue in the mortise before assembly. Otherwise, if you just put it on the tenon, a lot of it will scrape off and you will have a lot of squeeze out and not a lot of glue in the joint.

[Tony Bilello]

As others have stated, when dry fitting, they should fit with moderate pressure and not fall apart and yet be pulled apart relatively easily. Anything more than a few thousandths of an inch is too loose.
I think you may have misunderstood the 1/16" thing also. Usually you make the depth of the mortise 1/16" deeper than the tenon. This allows some room for excess glue and trapped air.

[Ken Fitzgerald]

If they are too tight, glue by hydraulic pressure will push the joints apart when glue is applied.

[Phillip Ngan]

Thanks everyone for the replies. Seems like there is a consensus that the tenon should fit with either a firm hand pressure or a light tap of the mallet. BTW the wood is maple and the glue is Elmers carpenter glue.

About the 1/6 inch glue line. My understanding about his point was that the glue line should be substantial since a thin layer of glue is prone to separate over time. It wasn't anything specific to the mortise and tenon joint because we weren't talking about those at the time. He mentioned that people tend to over tighten their clamps when edge gluing boards together. His advice went against my preconceptions so that is why I raised it here on the Creek.

I'll edge glue some test boards together with various clamp pressures, and test the joint strength in 5 years time, and report back :)

[Frank Drew]

Phillip,

Some glues have strength across a measurable (e.g. 1/16") glue line, epoxy for instance. But white and yellow glues really don't, the joint needs an intimate fit between parts for those glues to work best.

Another way to visualize it -- a 1/16" glue line on each side of a tenon would be a 1/8" gap if one cheek is tight against its mating mortise cheek, clearly too loose a fit.

[Neal Clayton]

from a mesurement perspective, since i'm meticulous about mortise fits for windows and doors, and check my cope height with a micrometer to ensure a proper tenon thickness, about 5 thousandths difference between the mortise and the tenon makes for a just about perfect fit.

[Phillip Ngan]

I'm going to send this conversation in a slightly different direction.

I was watching Frank Klaus' video on making Mortise and Tenon joints, and he stated that the critical part of the joint to get tight was not the cheek-to-cheek width but the other width (head-to-toe?). Now this makes intuitive sense to me. This fit he was referring to arrests the primary racking stress of a mortise and tenon joint. Do people agree with this? I guess the MT joint should be tight in both directions.

[Tom Hammond]

A finishing carpenter (who does very high end work for places like museums) told me that when gluing joints to make sure that there is a 1/16 inch glue line.


WHAT?!?!?!?!?!?!?! 1/16 ???????

That is CRAZY TALK.

[Kevin Begos]

Does any of the above advice change if using something like Gorilla Glue, which expands? I'm making some M&T window sash and had been thinking of using the Gorilla, instead of Titebond 3.

[Neil Brooks]

Here's all the science, data, measurement, and quantitative analysis you could possibly ever need to ... NOT answer your question ;)

http://woodgears.ca/mortise/accuracy.html

I've actually tried to keep his numbers in mind, and have, lately, made my M&Ts fit much more tightly than the first ones I made.

[Kevin Begos]

Thanks, that's a good site.

[Neal Clayton]

Does any of the above advice change if using something like Gorilla Glue, which expands? I'm making some M&T window sash and had been thinking of using the Gorilla, instead of Titebond 3.

i build traditional windows as well and have used gorilla glue and PL adhesive on a few. it will expand out of the joint, is the short answer. you'll find a bubbly mess on your copes and sticks the next day, as well as on the outside edges where the joint comes through, but the expansion isn't enough to defeat clamp pressure.

if you use bridle joints rather than true mortises you'll want to put a clamp from face to face to hold them shut, though.

all that said as long as your window design channels water away, as it should, i question the necessity for poly glues. it's a protected environment since there's no direct sun or water contact on the joint, only passive moisture contact, so i find TB3 to be enough.

[David Cefai]

Fine Woodworking had an article about clamping pressures.

Essentially their conclusion was that you could not achieve ideal pressures with normal clamps when using white or yellow glue. To me this implies a very thin actual glue line.

Note that this does not hold for epoxy. it is very easy to starve an epoxy joint.

[Kevin Begos]

Thanks, Neal. That's exactly what I wondered.
I'm going to stick with TB3.

[Gary Curtis]

A few technical reviews of glue joints written during the '90s threw out a figure of .04" -.06" on the cheek tolerance. The technical reports did destructive testing and are based on the best adhesion of long grain to long grain wood.

I use a machine called the WoodRat (a Router device) and can dial in that level of precision. I use Domino stock for the actual tenons, and those measurements result in a snug fit when assembling by hand.

I shoot for the same tolerances in Dovetails and Finger Joints.

Gary Curtis

[Chris Friesen]

The joint should go together with just hand pressure, but shouldn't come apart easily.

Also, you really do want to spread a thin layer of glue on both the tenon and the mortise--the latest issue of popular woodworking shows results from both that and mortise-only glue and there is a noticeable amount of difference in the final glue coverage. To avoid visible squeeze-out you can chamfer the edges of the mortise to make a small pocket for excess glue.

[Phillip Ngan]

Dave, thanks for the pointer to the FWW article on clamping. The main message of these articles is that woodworker tend to underclamp their joints. The article also mentions the recommended maximum glue line is 0.004", which is thinner than the cover of the FWW magazine at 0.005".

An interesting aside on these articles is that parallel bar clamps are the weakest of the long clamps, eclipsed even by the Quick Clamps. I found that interesting, but can see why because the round cylindrical handles.

[Leigh Betsch]

A few technical reviews of glue joints written during the '90s threw out a figure of .04" -.06" on the cheek tolerance.
Gary Curtis

Gary you sure that you don't mean .004 -.006?;)

[Gary Curtis]

Right. .004"

I don't write down these woodworking numbers much anymore. Need practice.