Joint strength

[kmealy]

Just ran across an interesting article on joint strength (force applied by racking, one type of stress) with some surprising results.  Issue #203

 

The 10x50 Domino joint was only marginally better than a biscuit (to all the nay-sayers "for alignment only") and much less than an integral or floating tenon.  Also thicker tenons work well (3/8 or 5/16)

 

 

 

[lew]

That is eye opening. I'm really surprised at the miter joint!

[kmealy]

1 hour ago, lew said:

That is eye opening. I'm really surprised at the miter joint!

Yes, me  too.  The picture showing the spline was running across the joint (not in the finished joint, that I typically call a "key").  They did not test, but showed a triangular piece across the back, sort of making it like a half-lap.  They thought that would be even stronger.  That would be very easy to make on a table saw.

 

The other thing they mentioned is some joints are self aligning (e.g., mortise and tenon) and others require clamps in several directions (e.g., half lap joint).

 

In the text, they did warn of  wood movement over time potentially affecting the joint.  IMO, that's also an issue with dowel joints.

 

They also said something I say often in that some joints are strong enough for their application.  There is a big difference in requirements for a chair joint vs. a picture frame joint.

[lew]

1 minute ago, kmealy said:

There is a big difference in requirements for a chair joint vs. a picture frame joint.

For sure.

[John Morris]

How were the stress tests performed Keith? With two pieces of wood, or all four pieces of a frame glued up?

[kmealy]

1 hour ago, John Morris said:

How were the stress tests performed Keith? With two pieces of wood, or all four pieces of a frame glued up?

[John Morris]

That's kind of what I figured. I really dislike that test. It's not practical, nobody builds anything like that. Not trying to be combative Keith  

 

[JWD]

8 hours ago, John Morris said:

That's kind of what I figured. I really dislike that test. It's not practical, nobody builds anything like that. Not trying to be combative Keith  

 

Yeah, I tend to agree with this, but at the same time there are limits on the practicality of making a test realistic.  Testing a frame in the same way multiplies the time involved in making each sample by 4, and if they are doing a reasonable number of samples per joint (say 20 at a minimum) then the time required is going to add up pretty fast.

[Grandpadave52]

Agree JWD. To be truly statistically sound, the sample size per joint type should be at minimum 31 all pulled at random from much larger lot sizes. Not practical for what they were trying to do. Their data is gives some insight, but not credible enough to conclude one joint is superior to others IMO.

[BillyJack]

If you build something correctly, it shouldn’t be under any stress to matter. A chair can differ..

[John Moody]

24 minutes ago, BillyJack said:

If you build something correctly, it shouldn’t be under any stress to matter. A chair can differ..

That is very true to a point, but just wood movement along can cause enough stress to have a failure. IMO

[Masonsailor]

Tests like this are always interesting. It’s difficult to relate the lbs of force that each joint will withstand to practical applications. For instance furniture that is going to withstand repeated heavy loads ( chairs, tables etc) is going to require not only the heavier duty jointery but jointery with tighter tolerances. A very loose fitting mortise and tenon joint will not withstand the same load as a tight fitting one and will succumb to the forces of expansion/contraction. All of these joints are dependent on the amount of area of the glued surfaces and how tight they are fitted. So comparing one #20 biscuit to a single domino is not a good comparison as the domino has a far greater surface area. Especially if you are using the 140mm dominos. The other factor is just practicality. 500 lbs of pressure is a lot. It would probably exceed the strength of the wood which is the practical limit for joinery. Drawers are a great example. A simple rabit joint using tightbond 1 will generally exceed the strength of the wood. The rabit ear will usually break off before the joint gives way. So take all this with a grain of salt !

Paul

Edited March 3 by Masonsailor

[BillyJack]

Butt joints on drawer generally work find for most of the world.. 

 

As far as strength tests…. I have fought with woodworker on numerous forums that say that a Domino is as strong as a M&T. They are pulling a Trump and say it is anyway, when ythey know it isn’t..

[John Morris]

I did a joint stress test with my son in his 10th grade science class, so I know everything. 

[kmealy]

14 hours ago, John Morris said:

That's kind of what I figured. I really dislike that test. It's not practical, nobody builds anything like that. Not trying to be combative Keith  

 

This is a common stress on chair and table legs.  This was pointed out in the article, "What is racking stress?"  (Gee, I just repaired 4 tables this morning with broken leg joints)

 

In all test cases, the damage was a split in one of the components, not in the mechanics of the joint parts.

I did a presentation on this a couple of years ago.  Two main issues:
* Most every test I saw (including this one) the stress is racking.  That's only one type of stress, but OK, maybe it's the most common.

* Is "strong enough for this application" important?   I remember seeing a This Old House episode years ago when Tommy was putting Dominos all over an edge glued piece that would be a bench or shelf in an offset alcove.  My comment to myself was, "Well, there's a $1500 tool (no doubt donated and sponsored) doing a job that does not need to be done."  A good edge glued joint will be stronger than the surrounding wood.

[John Morris]

7 minutes ago, kmealy said:

This is a common stress on chair and table legs. 

 

I'm sorry Keith, didn't see where it said the test was related to chairs and tables, I guess I didn't read far enough, but I'm not very observant at time either.

I thought the test subject was a frame sort of structure, since we are using 3/4 lumber in the test I think? A mortise and tenon joint in a chair done well using say 8/4 lumber, and pinned or drawbored is different than a frame type of joint. But I get the jist of the article. I enjoyed it, thanks for sharing Keith.

[kmealy]

2 minutes ago, John Morris said:

 

I'm sorry Keith, didn't see where it said the test was related to chairs and tables, I guess I didn't read far enough, but I'm not very observant at time either.

I thought the test subject was a frame sort of structure, since we are using 3/4 lumber in the test I think? A mortise and tenon joint in a chair done well using say 8/4 lumber, and pinned or drawbored is different than a frame type of joint. But I get the jist of the article. I enjoyed it, thanks for sharing Keith.

If you look at the left edge of the first photo, you can see a small part of the information about why they tested this (you will see stress applied to a table leg).

[JWD]

I'm not a huge fan of biscuits, but I have found uses for them.  I had an old crappy biscuit joiner when I lived in Tucson, got it used from Kent's Tools, and sold it there again when I moved in 2005.  Never replaced it, there's always another way to do anything in woodworking (if not 12 other ways!) so I just never made it a priority.  I did find half a dozen biscuits in a bag the other day, looking for the bag of 1/4" dowels.  Couldn't find those so I used 3/8" and broke the jig I was making as I enlarged the hole.  The dowel was a plug filling a misplaced screw hole - kinda overkill but I wanted to re-use the location for a screw just a little to the left.... I thought of the right way to do it after trying one of the thousand or sow wrong ways first!

 

Is it any wonder I never throw stuff away?  

[Zack]

Never know what you’ll need on any given day.

[kmealy]

I may have said this before, but when biscuit joiners were all the rage I bought one.

 

My first use was to butt join vertical and horizontal parts of a 7' tall bookcase.  To my amazement, it was very sturdy right out of the clamps.  Any other method I'd used had wobble until I attached the back panel.

 

Since then, I made 35 bookcases for a church library and staff offices, 3 for myself,  made about 10 tall cabinets for my children, and even used it to tie together headboard/footboards on a bunk bed (doubled up the joints.)   There has never been a failure.  I've also used them on my touch up kit boxes (that weigh about 40-50 pounds loaded) and a number of other small boxes.

 

I have never used them for edge-to-edge joints.

 

So I moan when I hear the common response, "They are for alignment only."

 

I replaced my initial Porter-Cable one with their version 2.0 and gave my original to a son-in-law.  I had problems getting the blade off to switch to the FF blade.  I took it into the service center, and apparently, they did too.  So at the time, P-C sent me a new one, then fixed the original one.  Somewhere along the line, I got a benchtop Delta model used for little money and it works really well for small pieces and I used it for repairs.    I've gotten plates from P-C, Lamello (that I really like), and to get a minimum order together from a local dowel supply (the one that for a while owned or co-opted with McFeeley's).  Theirs are made from what looks like baltic birch plywood, so I use them only on less stressed joints.