Joint design and wood movement...analysis and new thoughts

[Mark Singer]

Chosing the correct joint that will endure, not twist, and give the best strucural support is an essential aspect of furniture construction. Wood movement in joints is also critical to consider. In joining 2 boards of substantial width, say 8" or more , where the grain of one is perpendicular to the other, there are sometimes problems.. Where we often like to think that "True" mortise and tenon is the best choice, loose tenon has definite advantages in my opinion....I haven't read this in woodworking books, however looking at a free body diagram of a frame joint....my analysis seems reasonable.
Wood moves across the grain to a much greater extent than with the grain, and this fact is the source of the problem. It is the same reason why we often resaw our own veneer and glue it to both faces and both edges of a panel to be used as a cabinet door, or component. We see this construction in Krenovs work and others....he calls it "Real Veneer" an suggests a suitable core. When loose tenon joinery is used in a frame joint, the stabilty is similar to a real vennered panel and will eliminate most of the lateral stress common to joints in wider boards. The core is the tenon and should be selected for stability...quartersawn , vertical grain is the best choice. When the mortise is cut the cross section becomes a panel similar to veneer and a suitable core. The tenon is then coated with glue on all surfaces and becomes inert and very stable since the Tightbond II or epoxy is waterproof....it is unlikely the tenon will expand since the moisture content will not change very much. The mortised board is reduced to almost a veneer in thickness...say 1/4" or so....not much material to expand and cause problems. At the bottom of the mortise the wood is solid and the stress is ,as usuall, very high and will transtition to very low at the point of contact with the other board. This is exactly what is desired in a joint! If the tenon is small or it is "true " mortice and tenon design you are back to normal stress and potential problems.
Another overlooked point is that where the loose tenon grain is running the same as the mortise.....a much smaller embedment is required...since the long grain to long grain strength is occuring....in the other board the penetration depends on surface area and glue strength....much more depth is necessary.
In designing furniture these simple ideas are helpful in joint selection...try to think what is going on, rather than just selecting a joint at random...:confused:

[Steve Wargo]

Nice write up Mark, I've used the joint you describe above a few times and the "real veneered" type joint described by Krenov quite a bit. Never had any problems with it. I do a gree that some of the greatest battles in building furniture are trying to determine the proper joint, this is probably one of the reasons many furniture makers view the chair as voodoo. Great little article.

[Mark Singer]

I bet this Thread gets a lot fewer hits than my "Relabeling Job or What " thread! People don't care what holds something together....they would rather tear it apart....:rolleyes:

[Roy Wall]

Mark - I think I have a silly question.......

For glue ups....is it wrong to glue a loose tenon to only one mortise?- say the night before - so that the next day one could finish up the "GLUE UP" more like a traditional M&T joint. I usually get a little anxious.....so instead of glueing 4 parts at one time for one joint - I'd have only the one M&T joint per corner, etc... to deal with....getting the clamps on quicker, etc....

[Mark Singer]

Mark - I think I have a silly question.......

For glue ups....is it wrong to glue a loose tenon to only one mortise?- say the night before - so that the next day one could finish up the "GLUE UP" more like a traditional M&T joint. I usually get a little anxious.....so instead of glueing 4 parts at one time for one joint - I'd have only the one M&T joint per corner, etc... to deal with....getting the clamps on quicker, etc....

Do the long grain to long grain first...the faces of the joint must be completly free of glue so it really goes tight together...I reccommend all at once!

[Andy Fox]

Very interesting post. I would like to see more "why" posts like this.

[John Timberlake]

Seems to me that this joint would work best if the long grain to long grain part is slightly cross grain. That way, any splits that might occur in the loose tenon in the cross grain section would be less likely to transfer to the show surface of the joint. If the grain is perfectly aligned, seems not much different than a normal M/T joint.

[Gary McKown]

Mark said, "The core is the tenon and should be selected for stability...quartersawn , vertical grain is the best choice."

So would a tenon made from plywood (good cabinet grade or maybe your own laminate) be even better?

[Martin Lutz]

Mark, Thanks for the post. Its nice actually see a technique discussion. We need more of them with all of the talent we have available to access. You have really got the wheels turning in my brain with this. It makes sense.

[Chris Padilla]

Seems to me that this joint would work best if the long grain to long grain part is slightly cross grain. That way, any splits that might occur in the loose tenon in the cross grain section would be less likely to transfer to the show surface of the joint. If the grain is perfectly aligned, seems not much different than a normal M/T joint.

Interesting point, John. However, it is well-known and accepted that long-grain to long-grain glue joints are the strongest...usually even stronger than the wood itself using modern glues. I think if you get a split anywhere along this joint, it is bound to fail regardless.

[Mac Cambra]

It is an interesting take on the technical reasons why one would make this choice over traditional a MT. I have noticed that David Marks uses MLT almost exclusively however he never has offered an explanation why.

I know Krenov (sp?) is one of his mentors.

Anyway thanks for the explanation, I think it makes good sense.

Sounds like you have an mechanical engineering background?

[Mark Singer]

It is an interesting take on the technical reasons why one would make this choice over traditional a MT. I have noticed that David Marks uses MLT almost exclusively however he never has offered an explanation why.

I know Krenov (sp?) is one of his mentors.

Anyway thanks for the explanation, I think it makes good sense.

Sounds like you have an mechanical engineering background?

I have a degree in engineering .....I do not attempt my oun sructural engineering ....just Architecture:rolleyes:

[Joe Blankshain]

Thanks Mark.
This is one of the many quality posts that you and others have provided as an effort to improve the techniques of all of the "Creekers". Keep up the great work.
I look forward to having a "cold one" with you when I travel to CA in the fall so that we can discuss the merits of other structural attributes in our hobby/obsession.

BTW: The Relabeling post was most interesting/entertaining

[Dave Ray]

Mark, thanks again for another good learning experience

[Don Baer]

Mark,
As usual you are showing that not only are are you a talented Architech/woodworker but also a fine teacher. This one on the many reasons why the creek the best woodworking site on the internet. I have used traditional MT joining for a while but never used the loose tenion method except as dowels. Now I have a good reason to use this method and know why. Thanks

[Kyle Stiefel]

Mark,

I enjoyed the thought process. As I slowly shuffle through my developing repetoire of techniques it is nice to really think about what is happening.

Thanks.

[Bob Childress]

Mark,

Your usual brilliant display of knowledge! :) Many thanks and a quick (and possibly dumb) question.

Given that your argument for loose tenons is valid, are there really any situations, IYO, in which a traditional M/T would be demonstrably superior to a loose tenon construction?:confused: (Aesthetics of through tenons, etc. aside.)

[Mark Singer]

Mark,

Your usual brilliant display of knowledge! :) Many thanks and a quick (and possibly dumb) question.

Given that your argument for loose tenons is valid, are there really any situations, IYO, in which a traditional M/T would be demonstrably superior to a loose tenon construction?:confused: (Aesthetics of through tenons, etc. aside.)

I kind of like making traditional mortise and tenon joints because it is more challenging....the loose tenons solve another problem if you can "read" wood...That is that for years on lpng tenons we made haunches at the top or top and bottom...sometimes with a "slip" or angle and this was to eliminate twisting...because the tenon had the same character and behavior as the stretcher or component....ie, if the strecher was prone to twist, the tenon would really twist! When you introduce a loose tenon...the character of the stretcher is relived and the loose tenon will exhibit its own behavior...meaning the joint will not twist or fail do to torsion...Expasion is relived since the glue imersed tenon has become neutral and inert...so to speak and the stresses of the strecher are greatly reduced since it has been hollowed to an almost venner like character....answer: Its a better joint IMHO:rolleyes:

[Tom Jones III]

Great thread! Thanks and lets continue to have threads like this.

* Is it a valid summary to say that you are suggesting making the loose tenon very large relative to the stretcher?
* What happens when the stretcher expands and the tenon does not?
* In your diagram, what happens if you place a large downward force on the right end of the stretcher? How would this be different than a "normal" loose tenon joint?
* You don't give any sizes, are there some sizes of parts where this technique is more or less appropriate?

* Wouldn't the joint be stronger if you cut out the middle third of the loose tenon making two separate tenons but keeping everything else the same?

[glenn bradley]

I love this forum.

[Bob Childress]

Thanks Mark. Your explanation is perfectly logical.

As you also point out, M/T is more challenging, which for a duffer like me translates to L/T being less challenging and thus less prone to operator error. You can cut boards to exact length and have much less fiddling about to get the depths etc. just right.:D

Of course, the obvious conclusion from this argument (and I have thus far carefully avoided alluding to it in this thread) will be played out next April when a certain tool will be available in the US.:eek:

No, I'm not looking to start that discussion yet. :rolleyes: This thread is about tenon joinery and Mark's well-reasoned arguments. I always learn when Mark posts.

[Mark Singer]

Great thread! Thanks and lets continue to have threads like this.

* Is it a valid summary to say that you are suggesting making the loose tenon very large relative to the stretcher?
* What happens when the stretcher expands and the tenon does not?
* In your diagram, what happens if you place a large downward force on the right end of the stretcher? How would this be different than a "normal" loose tenon joint?
* You don't give any sizes, are there some sizes of parts where this technique is more or less appropriate?

* Wouldn't the joint be stronger if you cut out the middle third of the loose tenon making two separate tenons but keeping everything else the same?

Tom,
Really good questions!
The joining of the tenon to the stretcher forms a "composite" kind of like plywood, it becoms a sandwich,,,almost a torsion box that is balanced since the skin material....the walls of the mortise are equal...they move together for the most part. The downward force is resisted by the composite and it should develop almost the full strength of the weaker of the members. The sizes are conceptual and not dimensioned....1/4" min on top and bottom 3/8" on the ends for mortice walls. Small tenons will not solve the expansion of cross grain to long grain...they are strong in groups. This design is for the big boys...the 8" and bigger width joints...like the rear shelf of the teak sofa recently posted....That was the inspiration for this thread and the thought and design process of this joint... Many of the woodworking joint designs are tradditional ...hand me downs and from our own work and experience we can question, re-think and improve on the traditions...Not too many years ago...foundations were not bolted down...shear walls did not exist ans soil analysis was investigated with a steel rod pushed into the soil....someone asked the question:confused:

[Tom Jones III]

Tom,
Small tenons will not solve the expansion of cross grain to long grain...they are strong in groups. This design is for the big boys...the 8" and bigger width joints...like the rear shelf of the teak sofa recently posted....

Thanks for both the thread and the response to my questions. I think I have a better understanding of your original post now.

My question about making 2 loose tenons was not so much to solve the expansion problem rather 2 tenons would add significantly more surface area to the joint. For that matter, 2 side by side tenons would accomplish the same thing depending on the dimensions of the pieces being joined.

[Gary Curtis]

I am waiting in the wings to buy the new Festool loose tenon machine. They call it the Domino.

Their International (Germany) website has some interesting graphics showing stress tests on various wood jointing methods, including their own. On a stile-to-rail joint they measure the applied torque needed to make the joint fail. The force is applied at 90 degrees to the joint.

As I recall, a loose tenon is about 3 times as strong as a biscuit. Pocket screws are shown, as well.

That's why I'm holding off on buying a biscuit joiner until the Festool is availablel in January. sorry I didn't want to hijack this thread into the pros and cons of biscuits. Just wanted to stick with the engineering aspects of the discussion.

Gary Curtis

[Chris Mann]

Originally Posted by Tom Jones III
>My question about making 2 loose tenons was not so much to solve the
>expansion problem rather 2 tenons would add significantly more surface area
>to the joint. For that matter, 2 side by side tenons would accomplish the
>same thing depending on the dimensions of the pieces being joined.

Depends on the joint. In Mark's example, by splitting it up into two tenons, you'd actually be decreasing the glue surface area. Only two sides of each tenons are long grain to long grain. Of course, maybe you mean making the tenon thinner and adding two of them, but then at that point you might as well be using biscuits.

One other benefit of loose tenons is that you cut your pieces to actually size, you don't have to remeber to add on that extra length for the tenon.

[Chris Padilla]

One other benefit of loose tenons is that you cut your pieces to actually size, you don't have to remeber to add on that extra length for the tenon.

Oh, and don't our memories and tape measures constantly fail us. I've gone through 20-30 tape measures and still have measuring problems.... :confused: :p ...not sure what to do about the memory problems! :eek:

[Jeff Farris]

Mark, is there a point of diminishing returns on this method? Is it as effective in 1-1/2 to 2 inch members as it is in your 8 inch example? Obviously, in the smaller members the degree of movement between the two is less of an issue, but the rest of the argument would suggest that the floating tenon would be a better choice -- certainly easier to construct.

On your sofa, how did you machine the thin walled mortises?

Do you have a rule of thumb for the appropriate depth of mortise / length of tenon?

It sort of sounds like unless you are going for the aesthetics of an open or through MT, a floating tenon would be the better choice.

[Al Willits]

Ah...at the risk of bringing this informative post down a notch or two, I have a few questions and considering I have maybe 20 M/T's to my short career, please humor the old guy, thanks...

I understand you say that the loose M/T is stronger, so would I be better to use this instead of the standard M/T when ever possible?
This would be related mostly to furniture making btw.

Also, what's a "strecher"?

And...when making M/T, I find I can either make the tenon match the rounded edges the router leaves, or chisel the mortise to match the square corners of the tenon.
Any prefered way?

Al great post btw....thanks

[Mark Singer]

Mark, is there a point of diminishing returns on this method? Is it as effective in 1-1/2 to 2 inch members as it is in your 8 inch example? Obviously, in the smaller members the degree of movement between the two is less of an issue, but the rest of the argument would suggest that the floating tenon would be a better choice -- certainly easier to construct.

On your sofa, how did you machine the thin walled mortises?

Do you have a rule of thumb for the appropriate depth of mortise / length of tenon?

It sort of sounds like unless you are going for the aesthetics of an open or through MT, a floating tenon would be the better choice.

I used a slot mortiser which is on my MM Elite S and it worked very well.

Members over 5" should get addressed with this joint design concept. In other cases of shorter cross grain interface....normal design with smaller tenons or tenon pairs is better IMHO

Proportins are a matter of experience...reducing the walls too thin is not a good idea....they need to engage the tenon and act as a resitin couple for torsional stress... Look at a steel wide flange beam or truss and you will get some idea of the relationships of flange to web proportions

[Chris Padilla]

Ah...at the risk of bringing this informative post down a notch or two, I have a few questions and considering I have maybe 20 M/T's to my short career, please humor the old guy, thanks...

I understand you say that the loose M/T is stronger, so would I be better to use this instead of the standard M/T when ever possible?
This would be related mostly to furniture making btw.

I think Mark answered this above for you. :)


Also, what's a "strecher"? A stretcher is normally a stiffener used on legs supports of a desk, table or bench to minimze racking or sideways movement of the bench (think about planing hard on workbench and having it wobble as you plane)


And...when making M/T, I find I can either make the tenon match the rounded edges the router leaves, or chisel the mortise to match the square corners of the tenon.
Any prefered way? Pure personal preference. It is akin to asking what should be made first: the mortise or the tenon?

[John Stevens]

As I recall, a loose tenon is about 3 times as strong as a biscuit. Pocket screws are shown, as well. [] That's why I'm holding off on buying a biscuit joiner until the Festool is availablel in January. sorry I didn't want to hijack this thread into the pros and cons of biscuits. Just wanted to stick with the engineering aspects of the discussion.

Gary, I'd been thinking about bringing up the subject of biscuits myself, because I think it's fair game in any discussion of alternatives available in M&T joinery...especially when discussing floating tenon joints.

The question most often asked on discussion forums is “which joint is strongest?” But unless maximum strength is the most important quality of the joint, then a better question is “which joints are strong enough?” Once you answer that question, then you can make your final choice based upon other considerations--appearance, ease, cost, tools, materials, and subjective feelings of "craftsmanship."

I’ve found that biscuit joints are tremendously strong as long as I follow Lamello’s recommendations. We can all easily imagine any number of situations in which biscuits will not be strong enough for a given design...but so far, with a small amount of imagination, I’ve always been able to imagine an alternative design suitable for biscuit joinery. I can’t say what it’s like to use other biscuit joiners, but with the Lamello, strong, accurate joints are very easy and very fast to make as long as you design your piece with biscuits in mind.


I own many Festool products and I’m looking forward to buying the Domino. But I’m glad to have my Lamello biscuit joiner, and I’ll be surprised if the Domino completely replaces it.


Of course, maybe you mean making the tenon thinner and adding two of them, but then at that point you might as well be using biscuits.

I’ve found it so fast and easy to make a strong joint with biscuits that I look at it from the opposite direction--under what conditions would I be better off using a floating tenon instead of biscuits?

[Mark Singer]

I do not reccomend bisquits for frame construction at all! They are only to be used for panel construction IMHO. At a recent Maloof seminar atended by several SMC guys, Maloof said he has had lots of problems with them and does not use them. Everything has its place and bisquits are fine for joining panels of sheet goods or solid wood edging etc... Not good for joints subject to shear and torsional forces like tables and chairs...I don't care what Lamello says! The bisquits fracture easily from shear forces sometimes even during dry assembly...
Pocket screws are fine and I just used them for a simple table I posted on SMC...I used them because I could hide the screws and the componets were wide in cross section allowing the forces to distribute well and create torsional resistance.. Not good for frame joinery either!

http://sawmillcreek.org/attachment.php?attachmentid=36177&d=1144637557

[John Stevens]

At a recent Maloof seminar atended by several SMC guys, Maloof said he has had lots of problems with them and does not use them. Everything has its place and bisquits are fine for joining panels of sheet goods or solid wood edging etc... Not good for joints subject to shear and torsional forces like tables and chairs...

We've been here before.

http://www.sawmillcreek.org/showthread.php?t=31971

Thomas Stender writes that he has used biscuits for making chairs for years and has never had a failure. No disrespect to Sam Maloof, who I greatly admire, but when it comes to biscuits, I'll trust the advice of the person who has more experience with biscuits in chairs and tables instead of the one who, by his own admission, has almost none.


The bisquits fracture easily from shear forces sometimes even during dry assembly...

That has definitely not been my experience. However, I've only used Lamello and Porter Cable biscuits...can't say whether others are prone to breakage caused by the application of shearing force. But anyone who is wondering what to believe in this regard can try this: make yourself a good, tight butt joint and glue it together with a single biscuit. Then try to break the biscuit with shear force alone, no wracking or torsional force. Should be a real eye-opener.

As for torsional forces, it depends on the plane(s) in which the torsion is applied...but this is a design challenge, not a problem that is insurmountable with biscuits. If a floating-tenon is suitable for a particular joint, then biscuits will often be, too. After all, they're a type of floating tenon.

[Mike Cutler]

I'll throw in my .02 on the subject.

I am a traditional M&T guy. It's what I learned, so it's what I do. However, I have been playing around with the "Singer" style of loose tenons that Mark presented us with a few years back. I like 'em. Fast clean and easy.

There are some "thumb" rules for traditional M&T joints That I learned. They're not etched in stone, just a place to begin planning a joint.

1. The tenon is 1/3 the thickness of the mortised piece.Most folks default to 3/8" on 4/4 stock though
2. The width of the tenon is 5 times the thickness of the tenon, not to exceed 4" or the tenon is then divided. Dividing tenons does give more glue area,but it also relieves the amount of stress that a single long tenon could place on a board. Something to do with the leverage principle here.
3. The length of the tenon is 2/3 the width of the mortised piece. The longer the tenon, the stronger the joint .
3. Mortises are started at least two board widths away from the end of the mortised piece.

I see one advantage to an integral tenon in the application of the haunch to increase strength. I'm not sure of a floating tenon can perform the same mechanical function as an integral haunched tenon if it was cut the same way. Mark???

I always make the Mortise first. It's much easier to cut a tenon to size than change the dimensions of a mortise. Trust me on this point.

A few points in favor of the loose tenon that I didn't see.

1. Alignment. As long as all the material is mortised to the same reference the joint will go together cleanly even if the mortises aren't perfectly centered.

2. Save $$$$. let's say you have some,oh.... $15.00 bd/ft of Wenge, and you make a traditional, integrated M&T joint. If a table stretcher needs to be, say 10' long, and then mortised. You would need a 10' 6" long board to start with. With loose tenons you just need 10'. Less waste.

I don't think that I am in favor of having a different material for the loose tenon for the following reason;
On the day of the glue up the moisture contents of all the pieces of the joint need to be at the same moisture content or difference in expansion and contraction can occur which could effect glue strength, and joint integrity. not to discount fit up problems. I would use the same wood, or a very similar wood for the tenon.

If you really want to add some long term reliability to either an integrated, or loose tenon. Use Foxtail wedges. This is a wedged tenon inside of a blind joint. It will never come apart agian if done properly. There is no dry fit though. Once a foxtail is done it's permanenet. Glued or not.

John. Not to be contrary, but every article I've read that compared the relative strength of a biscuit joint to an M&T. The M&T joint size was made the same size as the biscuit, and not sized according to convention. IN my mind this serioiusly skews the results. The articles all pointed to material failure, prior to joint failure, but in a properly sized M&T joint the overall failure, of joint or material is diminished. Many years ago, in beginning woodshop, we always had to present our projects to the instructor.( Mr. Brower) on dry fit up. If a project could not hold it's shape with the mechanical properties alone of the selected joints. We didn't pass. I personally would not use biscuits for load bearing joints. They would not have met Mr Brower's criteria.

Just a few thoughts on the subject.

[Mark Singer]

Mike,
Great response and good general rules of thumb fo M &T joints. The same proportions work well for loose M&T joints, A couple of things to think about are that the long grain to long grain Mortise to tenon needs less penetration due to the bonding... The second considertion is that if you size the tenons traditonally you have the expansion problem,,, That is where this larger tenon seems to help. I like using the sme material, although I dont think it is that essential since we glue different woods and use dowels and still develop a good bond. The advantage of finding a better , more stable piece of the same wood is that the inherent character at the joint is neutralized...if it were to twist because of the wood character...it is now relieved since a large section was removed and a more stable piece inserted.
I must admit I am not an expert on bisquits, but I have used them for about 15 years in applications where they made sense. I would never use them in a chair unless it were panel construction as opposed to frame construction. It seems if you isolate the joint , it relieson the bonding of material on the face of the joint...the surfaces that touch to a great extent...if the grain of these pieces is prpendicular to each other,,,I think myself and most woodworkers wood question the strength of such a joint. A good mortice and tenon joint will fail only after the tenon fractures the mortise,,,,in a bisquit failure I would think that the surface bond and the bisquit would shear , never reaching the fiber strength of the resisting wood...Maybe I should make a sample and see.:rolleyes:

I forgot the question of the haunch....with a loose tenon it might not be necessary since you can increase the size with less fear of twisting....haunches are typically used to resist twists from the nature of the tenon acting over a wide distance....not as much a problem if it is cut from another piece.

[Michael Cody]

Very interesting thread -- I've long been a fan of loose tenons, even use Beadloc for a lot of stuff.. Now I mostly turn stuff, but I still do the occasional flat project so I think have a different question?

Since a normal loose tenon is basically a long<>long grain glue up on the one side (the one that would have tenon for M/T joint) and since today's glues should be stronger than the wood on such a joint, why would it be any less strong than a integral formed tenon? Assuming the joint were sized properly. I don't see how a loose tenon should be any weaker than a normal M/T and if the stock is of sufficient size to make a loose tenon joint, it would seem to be the best choice for any load-bearing joint - that is how I read Marc's info and I agree with it .. I see how on some thinner or smaller joints it's just not possible to form the joint with sufficient wall thickness so a normal M/T is the best choice. Am I misunderstanding something here. I don't want to get into a biscuit argument here, I've used them for lots of stuff, just normally not end <> long grain joints, there loose tenons are my preferred way to do it.

[John Stevens]

I must admit I am not an expert on bisquits, but I have used them for about 15 years in applications where they made sense. [snip] Maybe I should make a sample and see.

I'm not an expert on biscuits either...but I have done many test joints in maple, spruce, plywood and MDF. Because I have no training as an engineer, I don't claim that my test joints amount to anything more than anecdotal evidence. But the strength of the various biscuit joints surprised me. Especially the biscuits glued into slots running perpendicular to the grain of real wood. The results of that will really get you thinking about what happens to the glue and wood fibers as they form a bond in a biscuit joint. Perhaps it will expand for you the "applications where they make sense."

Also, no need to reinvent the wheel. There have been a number of tests published in magazines. Here are four that I know of:

John D. Wagner, "Choosing the Strongest Joinery for Doors," reprinted from Fine Woodworking in Practical Design, Solutions and Strategies (Taunton 2000). Article compared loads at failure for door-frame joints consisting of two- and three-biscuit joints, traditional mortise and tenon, loose tenon, dowel, lag bolt and tongue and groove joints. Thicknesses and lengths of mortise and tenon, loose tenon and dowel were larger than dimensions of biscuits.

Bruce Gray, "Testing Joints to the Breaking Point," Fine Woodworking, March-April 2001, pp 74-79. Article compared loads at failure for various joints consisting of traditional mortise-and-tenon, floating tenons (rounded and square-edged), two-biscuit joints, and four variations of dovetail. Dimensions of the traditional and floating tenons were not given, but were noted to have been thicker than the biscuits.

Robert W. Lang, "The Anvil Test," Popular Woodworking, December 2005, pp 76-79. No quantitative results measured; rather the author looked at the way various joints, including biscuit joints, failed when anvils were dropped on them.

Sven Hanson, Chuck Ring, Mark Stavig, David Zamora, Tommy Guess, and John Emerson, "Testing of Plate (Biscuit) Jointer and Adhesives for Applicability in Constructing Architectural Products," on the internet at http://www.woodworking.org/WC/GArchive98/Abstract/abstract1.html as of 5-6-2004. Article compared loads at failure for biscuit joints and M&T joints in simulated door frames.

All of the tests showed that biscuit joints were not as strong as traditional M&T and loose tenon joints. However, as I said in my previous post, maximum strength is not usually an important consideration. Usually, the most important consideration is adequate strength. Big difference. Let's take a practical example.

In Bruce Gray's article cited above, he tested a joint in which boards were attached perpendicularly, with their faces all in the same plane. (Flat, like a solid wood door.) The wood was 3/4" thick and 4" wide. The length was not given, but it apeared to have been about twelve inches. The average failure load for the traditional M&T was 6,000 lb, with one of the samples reaching 6,600 lb. The author noted, "That's like piling three cars on top of that test sample." A double-biscuit joint failed at about half the load of the traditional M&T. So in other words, it withstood the stress of having one and one half cars piled on it.

Now imagine a chair made with one such biscuit joint at each of the two side-rail-to-rear-leg joints. The two joints would fail when you piled three cars on top of the chair. To me, that seems like enough of a safety margin. (Enough to safely reduce the width of the pieces from 4" to something more in line with what you'd use in a chair.) But let's not stop there.

If you used proper design for a twin-biscuit joint, you'd increase the thickness of the side rails and rear leg to one inch, and the joint would be even stronger. If you added stretchers running between the front and rear legs, you would distribute the stress to not only the stretcher joints, but also the joints between the side rails and the front legs. Now you've got an even greater margin of safety.

This is why I don't think it's productive to question the strength of biscuit joints in the abstract. I think it only makes senst to look at the strength of a biscuit joint in the context of the design in which you plan to use it. If you can accommodate the design to suit the strength of the biscuits, then biscuit joints may be considered along with other joints, and the decision of which joint to use would depend on other factors, such as aesthetics, labor costs, equipment costs, etc.

If you can't accommodate the design to suit the strength of the biscuits, then by all means, use a different type of joint. However, I think most folks underestimate the strength of a biscuit joint, and overestimate the stress that any given joint will see in normal (i.e., not abusive) use. When folks advise others to dismiss biscuit joinery based on an underestimation of their strength and an overestimation of the strength needed, they do a real disservice to their fellow woodworkers who don't know any better. If a biscuit joint is adequately strong and is satisfactory in terms of costs, aesthetics and that subjective feeling of "craftsmanship," then it's a waste of time and money to use a floating tenon or traditional M&T joint instead. Time and money that could be spent in pursuit of more woodworking projects, or with family, friends, or on other fun pastimes.


I would never use them in a chair unless it were panel construction as opposed to frame construction. It seems if you isolate the joint , it relieson the bonding of material on the face of the joint...the surfaces that touch to a great extent...if the grain of these pieces is prpendicular to each other,,,I think myself and most woodworkers wood question the strength of such a joint.

Not sure if I understand what you mean. As I understand it, in both the traditional mortise and tenon joint and the floating mortise and tenon joint, there will be perpendicular grain in the mating pieces. Most woodworkers don't question the strength of those joints...until the width of them gets to be fairly large (hence the reason you started this thread). In a biscuit, the grain runs diagonally to the long axis of the ellipse. So in your example, the grain of the biscuit would run at 45 degrees to the grain of both pieces. Not sure why there would be a problem with that: the biscuit is small enough that differences in wood movement between the biscuit and the mortise would be negligible, especially in light of the elasticity of the PVA glue that is used with biscuits.

Moreover, if you do a test with a biscuit in which the slot runs perpendicular to the grain, you'll see just how much stronger the glue joint of a biscuit is when compared to glue joints in which the mating surfaces do not expand as much. My guess is that when the biscuit expands, it drives the glue into the adjoining surfaces in a way that simply can't be duplicated in a traditional M&T or floating M&T. (Maybe if you clamped perpendicular to the plane of those joints, but think of the extra time and labor compared to the biscuit joint!)

In addition to the strength of the glue joint, there is also some mechanical strength. When I dry-fit three or four biscuit joints, I have to use a rubber mallet to get them to fully seat. I have to use the mallet again to disassemble them. And when I insert the biscuits into the slots with glue, I usually have to tap them with a metal hammer to get them to seat fully in the slot. Sometimes when they appear to have hit the bottom of the slot, I give them one really good whack! And no, I have never experienced a biscuit shearing, even when being whacked with a metal hammer. :)

[Mark Singer]

John,
I am open to reason...I will make a few samples and let them dry and then test them...maybe I will make a M&T joint to compare....I feelings are based on the way I have been building furniture for many years and not tests. I must say though my intuition so far has resulted in good solid results...not to sy that the same coudn't have been achieved with bisquits...Even if they test good , I know I will always use more traditional methods for joinery....Its just what I am use to...Thanks for the links which have test values

[Gary Curtis]

Fine Woodworking magazine featured a detailed analysis of floating tenons in an early 1990's issue. Sorry, I'm away from home, so I can't reference the issue with an exact date.

But I just received Taunton Press' book "Joinery", and the same article was reprinted there. It gave a lot of information about grain direction and its effect on gluing.

I'm sure the article can be downloaded --- for a price --- from the Taunton Books website.

Gary Curtis

[jonathan snyder]

Hi Folks,

This is an excellent thread, very informative and timely for me since I just cut my first M&T joints last week.

Mikes rule of thumbs are just what I have been looking for

[quote=Mike Cutler]
1. The tenon is 1/3 the thickness of the mortised piece.Most folks default to 3/8" on 4/4 stock though

I have one question. I recently read that the mortise should be larger than a third of you stock thickness. I cant remember who wrote the article but they indicated that this would make for a considerably stronger joint. Is this what you are saying when you say most folks default to 3/8" on 4/4 stock, and is that 1" finished dimension or 3/4 finished dimension?

Another question, weather you are using a loose of fixed tenon, how tight should the tenon fit lenghtwise. If it is a fixed tenon, one would expect some movement. Should the mortise be cut a hair longer to allow for expansion.

As you can tell, I am quite new to all this. Thanks again, this is great stuff

Jonathan

[Mike Cutler]

Hi Folks,

I have one question. I recently read that the mortise should be larger than a third of you stock thickness. I cant remember who wrote the article but they indicated that this would make for a considerably stronger joint. Is this what you are saying when you say most folks default to 3/8" on 4/4 stock, and is that 1" finished dimension or 3/4 finished dimension?

Another question, weather you are using a loose of fixed tenon, how tight should the tenon fit lenghtwise. If it is a fixed tenon, one would expect some movement. Should the mortise be cut a hair longer to allow for expansion.

As you can tell, I am quite new to all this. Thanks again, this is great stuff

Jonathan

Jonathan.
Without reading the article, and seeing the example. I won't refute the larger joint. A few things to consider though. In an M&T Joint.

The tenon has the ability to act as a lever in the joint. If the remaining thickness of the walls of the mortised material become too thin. The tenon can actually blow the mortised section of the material out. A lot has to do with how a joint will see stress, and in what planes. If the mortised member is fixed, and all of the stress is along the edge than having a larger cross section of tenon would add considerable strength in that application. There are many good books on Joinery. Taunton press has an excellent joinery book, and the "Wood Joiners Handbook" by Sam Allen is an excellent reference to keep around.

When I used 4/4 I did mean a finished board 3/4" thick. Remember that the proportions I gave were just a starting point. Those proprotions will yield an adequalely strong joint, but may have to be altered based on design, or material selection.

When I plan a joint. I always look at how to evenly distibute stress in that joint. I don't rely on glue strength to hold it together, even though modern adhesives are stronger than the wood. I want to distribute the forces in such a way that no single component has to be relied upon for joint integerity. My belief is that, if I distribute the forces more evenly. I maximize the long term survivability of the joint. All joints will fail in time. Glues will break down, it's just a matter of time.

Are you asking if the mortise should be deeper than the tenon? or wider, and longer than the tenon? The mortise needs to be a little bit deeper than the tenon to accomodate glue squeeze out when the joint is assembled. The mortise is only a few thousandths bigger than the tenon faces. The tenon should be tight in the joint dry, maybe requiring some lite taps with a dead blow mallet, but not so tight that you really have to wail on the mallet to get it together. Modern glues need tight fitting joints to bond properly. Epoxies are a bit more forgiving, but require some additional technique to "fill in the voids" so to speak.

[Mark Singer]

John,
I found an article that really supports the use of biscuit joinery....They state clearly it is not the correct choice for chairs and the joint will fail without warning....see the last paragraph


http://www.woodworkersbookclub.com/images/btn_addtocart_lrg.gif (http://www.woodworkersbookclub.com/checkout/cart.cfm?category=0&sku=705312)
http://www.woodworkersbookclub.com/images/transparent.gifBiscuit joinery is hands down the fastest, easiest and most versatile way to join two pieces of wood together. This book gives you the complete plans to 12 projects, ranging from a Shaker chest of drawers, to an Arts & Crafts bookcase to a modern knockdown entertainment center. But where you would expect to find tenons and dovetails and half-laps, Stack used biscuit joints for all the critical joinery -- saving time, material and frustration.
There are those woodworkers who turn up their noses at biscuit joinery and insist that traditional joints such as the mortise and tenon are the only way to build truly fine furniture. Of course, most of the country's high-end cabinetmakers would laugh at that attitude. A lot of the best custom furniture uses biscuits in one way or another. As many woodworkers have discovered, cutting a perfect biscuit slot requires little layout or expertise to get just right.
Anything with angles just begs to be assembled using biscuits. The tapered dresser is another prime example. Joining the tapered sides of the dresser to the top and bottom would be impossible for beginners and most intermediates without a biscuit joiner. But by changing the fence's angle again, all the joinery problems in this case piece are solved.
While the biscuit joinery will conquer your fear of angles, it also will help you build traditional furniture faster. The Shaker chest of drawers is a prime example. You could spend a week in the shop building all the web frames for a traditional piece of furniture like this if you used mortise-and-tenon joints throughout. But when you use your biscuit joiner for the task, you should be able to glue up a web frame as fast as you can take another one out of the clamps.
Biscuit joints also allow you to build projects that would be almost impossible using traditional or other modern methods. The ash armoire/media center is one of those pieces. Using a mallet, you can knock down this stylish four-door piece in about five minutes for transport in your trunk. Yet, there isn't a single visible joint (like with other knockdown hardware) and the piece is super solid. What's the secret? Special biscuits. Maybe it's time you bought a biscuit joiner. Or maybe you've got one and it's just gathering dust on a shelf because you've never used it to its full potential. In that case, maybe you need this book. The Biscuit Joiner Project Book is bound to amaze and inspire you when you see what's possible with this revolutionary tool. The Strength of Biscuits
Every couple of years one of the woodworking magazines publishes a study examining the strengths of various kinds of joints. Depending on which one you read, biscuits are either the strongest or just one step above joining wood with chewing gum. Here's the truth: Properly glued, biscuits make a strong joint. In fact, because of the properties of modern glues, many times the wood will fail before the biscuit does when the joint is put under enormous stress. But know this: when biscuit joints fail, it is usually sudden and quick. When a traditional joint fails, it is usually a slow process. So biscuits are not the right joint for chairs or other furniture that is subjected to repeated and constant strain. You want a chair to become loose and wobbly before it falls apart with someone in it. But for the vast majority of other joinery applications, biscuits are an excellent choice.

[Mike Cutler]

Mark.

As someone that was not really an advocate of even loose tenons. I'd like to get back to the original intent of the thread. That being, the technique, and design of the loose tenon that you use in making and assembling your loose tenons joints.

There is one sentence in that final paragraph that caught my attention though.

"In fact, because of the properties of modern glues, many times the wood will fail before the biscuit does when the joint is put under enormous stress."

If I remove the word "biscuit", and simply substitute "joint". It emphasizes a point I was trying to make. Too much reliance on the adhesive properties alone in a joint. There needs to be mechanical integrity in the joint. Having the glue line as the failure point mechanism is not good in my opinion.

I think we have been convinced that an adhesive, of some form or type,is all that is required to join two pieces of material. We still need a good interlocking mechanical joint. I don't care if the glue is stronger than the wood. If I can make a joint that won't break at the glue line. I'm better off. Tenon joinery, loose or integral, accomplishes this.

If I M&T a table stretcher to 2/3 the depth of the leg. That joint cannot fail at the glue line, or the intersection of the two pieces. I have to break a significant amount of material both inside the joint, and outside the joint to completely fail that joint. It's a stronger joint, no matter how we look at it.

Back to your method though. I can't find your original post anymore that demonstrated how you were making the loose tenons. I have a picture, but I personally think that it is something that would be of benefit in this thread. A picture is worth a thousand word and all that stuff. If you could post a pic of the tenon, and the joint assebly, or link to it. I think it would be a good visual for everyone.

[Marlow Wilson]

I'm an novice, but just to chime in on the discussion, I have had excellent results with Lee Valley floating tenons. I am young and impatient and my joinery often suffers from this. I'm a 'why can't I do this with a motor' kind of hack whose projects have improved immeasurably from small floating tenons and my PC biscuts.

http://www.leevalley.com/wood/page.aspx?c=1&p=44779&cat=1,250,43217

These are the Canadian prices but you get the idea.

Cheers and thanks for the great thread.

Marlow

Edit. I just noticed that the photo below breaks some M/T rules so consider it for illustration only.

[Chris Padilla]

For my experience, I like to use biscuits as alignment guides and try to keep them from being the main structural integrity of a joint...it is a personal preference and nothing more. However, it does depend largely on how the project will be used....

I actually used double-biscuits to bond together a panel of 6/4 redwood for a porch. The porch was quite small (12' long and only about 4' wide) and it also contained all the plumbing/wiring for the sprinkler system as well as the main water shut-off valve for the house. Hence, a convenient lift-off panel was needed to access those items and yet I wanted it to match the rest of the porch.

I made that about 6 years ago and even though the redwood panel cupped pretty badly on me (I suspect wet redwood and not paying attention to end-grain patterns when gluing up the panel), it has still held together well using good old yellow glue. Now is that a testament to the double-biscuits or the glue? I dunno...I'm leaning more towards the yellow glue and long-grain joints than the biscuits doing anything extra but who knows....

[Mark Singer]

Mark.

As someone that was not really an advocate of even loose tenons. I'd like to get back to the original intent of the thread. That being, the technique, and design of the loose tenon that you use in making and assembling your loose tenons joints.

There is one sentence in that final paragraph that caught my attention though.

"In fact, because of the properties of modern glues, many times the wood will fail before the biscuit does when the joint is put under enormous stress."

If I remove the word "biscuit", and simply substitute "joint". It emphasizes a point I was trying to make. Too much reliance on the adhesive properties alone in a joint. There needs to be mechanical integrity in the joint. Having the glue line as the failure point mechanism is not good in my opinion.

I think we have been convinced that an adhesive, of some form or type,is all that is required to join two pieces of material. We still need a good interlocking mechanical joint. I don't care if the glue is stronger than the wood. If I can make a joint that won't break at the glue line. I'm better off. Tenon joinery, loose or integral, accomplishes this.

If I M&T a table stretcher to 2/3 the depth of the leg. That joint cannot fail at the glue line, or the intersection of the two pieces. I have to break a significant amount of material both inside the joint, and outside the joint to completely fail that joint. It's a stronger joint, no matter how we look at it.

Back to your method though. I can't find your original post anymore that demonstrated how you were making the loose tenons. I have a picture, but I personally think that it is something that would be of benefit in this thread. A picture is worth a thousand word and all that stuff. If you could post a pic of the tenon, and the joint assebly, or link to it. I think it would be a good visual for everyone.

Mike,
This is a pic of my door construction. It is not what I am suggesting in this thread which is to oversize the tenon , removing a lot of material for movement allowance. The tenon on the sofa was this larger size, but I don't have a pic...my drawing at the begging of this thread shows it... Interesting though this layout works well for big doors and I have never had a problem:confused:

[John Stevens]

John,
I found an article that really supports the use of biscuit joinery....They state clearly it is not the correct choice for chairs and the joint will fail without warning....see the last paragraph

[skip to quotation from article]


But know this: when biscuit joints fail, it is usually sudden and quick. When a traditional joint fails, it is usually a slow process. So biscuits are not the right joint for chairs or other furniture that is subjected to repeated and constant strain. You want a chair to become loose and wobbly before it falls apart with someone in it. But for the vast majority of other joinery applications, biscuits are an excellent choice.




Mark, that appears to be an excerpt from Jim Stack's book of biscuit joiner projects. As far as I know, Jim Stack has not participated in any quantitative analysis of biscuit joint failure. Of the articles I've read, only one article ( Wagner) reported a sudden failure of a biscuit joint. Gray's article states the opposite, when discussing the graph of load vs displacement of a biscuit joint:



It is surprising to note that although the plastic region (in which irreversible damage occurs) is relatively small, the load curve is gently rounded and the tail is long. This indicates that failure, as defined in this article, is not as sudden as others have suggested. Clearly, the joint is able to sustain significant load over a wide range of joint movement, even after irreparable damage occurs.

So who do we believe, Wagner or Gray? Well, Wagner's article describes some tests performed by Grant Taylor, who runs a custom door business. There was no mention of any expertise Mr. Taylor might have in testing failure in wood products. The fact that Mr. Taylor only tested one sample of each type of joint suggests a problem with sample size, and there may well have been other problems. The Gray article described tests performed in a laboratory facility of the Wood Science and Technology Center of the University of New Brunswick. Six samples of each joint type were tested according to guidelines established by the American Society for Testing and Materials.

Each reader can decide for himself (herself) which article is more likely to have reported reliable info. But even in Wagner's article about Taylor's "home made" experiment, Taylor found that the biscuit joint that "failed with a dramatic explosion" under 2800lb of pressure had exhibited a gap at 1700 lb. So one might wonder whether that sort of gapping would have given an audible warning (creaking) to a user of a biscuit-joined chair long before the chair actually failed.

In addition to that, we should think about how chairs fail. My stepfather is a furniture repairman who works on lots of stuff, from antiques to high-priced Scandinavian furniture. This doesn't make me an expert by any means, but I've seen quite a number of broken chairs. Those that broke at the joints were chairs that were made before the era of modern glues. The joints in those chairs failed because the glue deteriorated. Glue deterioration is a gradual process, unlike loading a joint to failure in a matter of seconds. In the latter process, the wood itself always breaks. In the case of glue deterioration, the joint loosens and begins to creak and show gaps. The creaking and gapping are the warning signs that should alert a reasonable person to the fact that one or more of the joints has failed. But even if the wood breaks or otherwise fails, the odds are slim that it will occur simultaneously in more than one joint.

However, we should keep in mind why we're even talking about using biscuits in chairs. After all, if you look at almost any mass-produced chair today, it's made with pocket screws because they make a very strong joint that is usually aesthetically acceptable because the holes and screws are hidden, and they're very cheap to make in terms of labor costs. The only reason I mentioned chairs was because you mentioned that Sam Maloof doesn't use biscuits because he had problems with them. Maloof makes a lot of chairs, so I replied that Thomas Stender (a studio furniture maker) has used biscuits in chairs for years and never experienced a failure.

But Maloof, Stender and chair-making are only tangentially relevant to this thread, because they deal with the question of how strong a biscuit joint is. The central topic of this thread is about using loose tenons to join wide boards. Specifically, loose tenons mounted in mortises whose cheeks are about 1/4" thick. Based on my limited experience and the articles I've read, I believe that a biscuit or two, set 1/4" below the surface of the joint, would be strong enough to serve as the floating tenon(s) in this type of joint. I understand that you and many people here disagree. Thank you for taking the time to engage me in a fact-based conversation about this subject. Even if I haven't changed my mind as a result of our exchange, I still think you've helped improve my understanding of woodworking.

[Mark Singer]

John,
The bisquit ar a 1/4" setback would not achive the same result of reducing the potential for differential expansion of elements whose grain is perpendicular.... It was actually 3/8" setback....a 3/4" tenon in 1 1/2" element....
It seems strange to me to even consider bisquits for chairs...I have always used mortise and tenon and there is a part of me that says this is a higher level of woodworking and it should be done correctly in a proven manner. I am not saying the results of the testing is wrong.....it really doesn't feel right....and that is a big part of woodworking as you do lots of projects , you tend to stick with what has worked in the past...and thats just the feeling inside , if you know what I mean....like Mike Cutler, I like the mechanical aspects of a joint to add integrety and not depend as much on the glue..

[Mike Cutler]

Mark.

Thanks for posting the pic. I think it will help with a visual, even though it was not specifically what the this thread was about. Thank you. The "divided" loose tenons in the door pic would seem to work in the same manner as traditional divided tenons. Relieve the stress of one single large tenon and divide it into two. The same amount of area is in the joint, but the stresses are individually diminished. I essentially used the same physical joint for my Arts & Craft door project, but the tenons were integral divided tenons that were mortised ino a floating tenon and then joined to the door.
In your original post the oversized loose tenon is able to be a single width, instead of divided, because the loose tenon by nature removes the stress component of the large integral M&T joint in my opinion. I like it. I've seen some books on oriental joinery methods that are very similar to what you have presented. The floating tenons went very deep into the mortised member. They too were wider than western convention.

John.

It's not a matter of changing minds. I have always advocated that someone should stick to the methods and materials that they are comfortable with. If you are comfortable with the results you are obtaining that's all that really matters. I have read your posts in this thread more than a few times. You have presented the biscuit joiner in a different lite for me. I admit that I used to view the biscuit joiner as a pretty limited tool, that was being overused due to marketing hype. Heck.. You even had me in a Borders Bookstore until 9:30 last nite reading books on biscuit joinery. I don't doubt the initial strength of a biscuited joint. I'm just uncertain of the long term reliability when referencing chairs as the model. I think that a chair and a door present some of the more difficult obstacles to overcome in woodworking. One, the chair, sees tremendous stress and abuse( I'm referring to a dining room chair) in it's lifetime. The other (the door) has to remain perfect in a single plane, with no mechanical forces to aid in a perpindicular direction.

[Mark Singer]

Mike and John,
I really appreciate the tone of our discussion...it has remained objective and thoughtful....unlike some of the "my brand is better than yours" threads. In woodworking each of us finds techniques that work and we tend to use as a default....if they work , that is probably a good thing.
Finishes are much the same way....if you have one you like stick with it..
I enjoy looking at all the ideas and keping an open mind...then going with what feels right..for me that would either be true mortise and tenon or loose tenon consruction for all frame joinery but, panel or sheet good construction ,or large surface of solid wood construction where biscuit joinery works really well and I am comfortable with it

[Mark Singer]

Mike the dimensions on the expasion relieved joint I used on the sofa were 1 1/2" x 8" member....3/4" thick tenon ,cut square not rounded by 6 3/4"....It is important to leave enough wall thicness to resist shearing through. Although my premise is that the mortise walls and tenon become a composite...here the adhesive was West Systems epoxy....which is the best for Teak wood and insures the development of a true composite...when I say composite, the analogy would be like plywood...each layer by itself is weak and flexes....together the surface resistance and developed section modulus produces a stable and strong group that acts together