Proof that end grain joints are stronger than side-grain joints

[Reffc]

Interesting video which dispels the myth of end grain joints being the weakest way to join timber.

There are a few things of note though and often end grain joints are made along the long axis where bending moments or shear may be greater than on the short axis, allied with end grain joints always failing on the glue bond. This is explained by the grain fibres themselves being stronger than the glue joint when running lengthways (hence failure at the glue plane). However, the thing of note is that glue is usually stronger than lignin resin in the timber which bonds the cellular wood fibres together, so side grain joints usually fail on the timber, not the glue line, which perhaps explains why the myth of side grain joints being stronger has existed for so long. (ie timber fails before glue joint). The consideration here though is that it's comparing failure forces that should be considered, not whether a joint fails on the glue bond and it's proved from these fairly scientific tests that failure forces of comparable gluing areas results in end grain joints being stronger than side grain joints.

The other take-away (if you watch the 2nd video in the series) is that end grain mitre joints are also inherently strong requiring usually little or no reinforcement. Note in this case that end grain mitres involve bonding of end and side grain so side to side mitres will fail before end grain mitre joints due to greater relative area of lignin bonds. Also, the perception of un-reinforced mitre joints being weak has much to do with relative gluing area and applied moment of long lever arms such as in picture frames where the relative bond area usually comprises a very small cross section of joint.

The final consideration when comparing failure forces is the timber itself. Oak, for example, would be expected to have higher failure forces on end grain joints due to the large grain pore (fibre pore) area compared with closer grained, smaller fibre pore area grained timbers.

I thought it worth posting after considering discussion from another thread where there was mention of end grain mitres.

The decision to reinforce joints by taking advantage of additional strength gained from exploiting inherent strengths of long fibre axis combined with increased gluing area will depend on applied forces in use and for general cabinetry of smaller boxes, plinths or similar constructions not subject to large failure tensile forces mean that no further reinforcement should be needed. The other thing of note is that care to ensure clean and well aligned joints is important.

Video link below:

 

[Simon89]

Admittedly I didn’t watch the whole video, I slipped through as I think I have seen it before…

but when he states that “end grain joints are twice as strong as side grain joints”, Doesn’t he actually mean “end grain joints are twice as strong as side grain”.

As the glue used is clearly stronger than to the wood, he wasn’t able to compare to glue joint as he didn’t appear to observe a side grain glue joint fail.

Perhaps the question of which is stronger is not as relevant as is an end grain joint strong enough?

 

[niall Y]

Oh dear, this old chestnut. As far as I am concerned, no myths have been dispelled. His experiment shows that wood is inherently stronger along its length than across its grain. It also shows that the strength of the glue is constant. Nothing new to see here folks..

It is the various uses to which we put wood that is the variable missed out from the experiment. If we are relying upon timber being structurally sound along it's length, then we are not going to glue it, hence the admonition not to glue end grain. This is where various splice joints are brought into play.( much stronger than a glue joint)

There are many instances where you can get away with gluing end grain. Which is why saying" thou shalt not, in any circumstances glue timber end to end " is pretty pointless.

 

[Jacob]

Couldn't be bothered to watch the vid as the idea is obviously nonsense.
What is a "side grain" joint anyway? Are they saying that a glulam beam would be stronger if constructed as shorter pieces end grain joined?

 

[Adam W.]

Couldn't be bothered to watch the vid as the idea is obviously nonsense.
What is a "side grain" joint anyway? Are they saying that a glulam beam would be stronger if constructed as shorter pieces end grain joined?

They do that with finger joints and have been for over 20 years, so yes, it works well.

 

[Jacob]

They do that with finger joints and have been for over 20 years, so yes, it works well.

As you'd expect. It's a variety of laminating. In fact finger joints are use a lot, but as joined lengths which are then laminated, with the finger joints staggered.

 

[Reffc]

Well, I found it of interest anyway and understood exactly what he was getting at. It's not to say such end-grain joints are suitable for every occasion as clearly that is not the case but there were a few points of interest there for anyone interested who hasn't seen it before.

 

[Hornbeam]

I think you have take all these so called scientific tests with a large pinch of salt. Modern glues can be extremely strong and far more durable than the early adhesives. I will stick with glue to hold together tight well fitting (most of the time) joints. Even for edge joints I always use a reinforcing loose tong or biscuits . On no another thread on how strong are biscuits. I will just go and dunk one in my tea to find out

 

[TheTiddles]

I’ve got a new slogan for a wood glue “stronger than the wood itself”, gonna trademark and be a millionaire, just you wait…

 

[D_W]

The take away from this should be pretty simple. Most people do not wet endgrain well enough if they have trouble gluing it. what happens with a quick thin glue line is not enough glue to adhere to the surface of the wood, and the result is predictable.

Make sure if you try to test this theory, you have fresh end grain - oxidation on the surface of the wood can prevent good surface wetting, also.

There are a lot of myths in gluing, but this is hobby woodworking. I'm sure there are still people who believe edges fail because the bevel above them was hollow ground, that titebond 1 is "not repairable", and that the yellow glues are PVA and have a creep issue (white glue). Oh, and the long made statement that only a tiny thin invisible glue joint of titebond would be strong, and that one needs 150 PSI at a glue joint for it to be full strength. FWW published something along those lines, but titebond didn't say it. They did say, though, if you have wood sitting around oxidizing, you should freshen the surface to remove the oxidation before gluing.

This test is simple enough for anyone to reproduce rather than asserting that the test doesn't show anything, but I've found over the years on here and other forums that if you present something that's true, you'll get a lot of people asserting it's not and nobody proving it's not. If a test could be easier than this to prove for yourself, I don't know what it would be.

 

[Reffc]

I don't really claim anything definitive except that for my purposes, clean fresh cuts on ply surfaces for mitres results in very acceptable strength joints. I did a trial last year on a large panel mitre (36mm thick and over a metre long) and tested it after several days curing and it proved immensely strong,certainly strong enough for purpose but it was quite an appreciable gluing area. I was confident enough when gluing up the customer commission that combined with more than adequate bracing (tongue jointed) it was going no-where. Here's a photo of the part assembled construction with the main mitre being to the front panel (which is face down). This one was done using Everbuild 502:

My view on this is that there are different approaches people use and the right one is the one that works dependably every time. I've done a fair few of these and haven't yet had one fail on me.

This was the assembled and veneered construction:

 

[Jones]

Here's what I learned:
1. Glue works
2. Wood can be split quite easily
3. A plank is stronger along the length than across the width
To be fair I knew all that already.

 

[Reffc]

We all did I hope!

 

[DrDarren]

On a more logical note, you cannot prove with a piece of evidence. You can only contribute evidence that counts as a fact towards a claim given sufficient warrants and absence of counter-examples. You can however, falsify with one piece of evidence (see Popper's work on science and falsification).

 

[Reffc]

As we're being logical, it is also widely accepted that a "proof of theorem" is a logical explanation (not description) of why a certain theory is true even if it can be explained more fully. There are false theories where it has been found that the evidence is challenged and disproved but whilst the evidence proves the theorem, arguing differently may be said to be pedantics. You can prove a theory works as predicted by testing it. In this case, there is no common ground to the debate since different timbers, different loads and different glues all add to the melting pot. What I can say with some confidence is that for my own purposes, bonding some specific joints this way just works and is easily able to withstand greater loads than those experienced in use and handling. There are indeed better and stronger ways of achieving the same end result and whether those methods are selected is a matter for personal choice, assessing appearance, required strength and longevity. I'm sure none of us would embark on any project and deliberately chose an unwise jointing method. Horses for courses.

 

[Spectric]

If you want a stronger end grain joint then just put a dowel through it and then screw into this or use barrel fasteners https://www.accu.co.uk/2270-cross-dowel-barrel-nuts?page=1

 

[TRITON]

They do that with finger joints and have been for over 20 years, so yes, it works well.

I think you are well aware Adam that the vid is talking about flat square ends, not something machined into interlocking fingers.

 

[Sam Kilby]

I don't really claim anything definitive except that for my purposes, clean fresh cuts on ply surfaces for mitres results in very acceptable strength joints. I did a trial last year on a large panel mitre (36mm thick and over a metre long) and tested it after several days curing and it proved immensely strong,certainly strong enough for purpose but it was quite an appreciable gluing area. I was confident enough when gluing up the customer commission that combined with more than adequate bracing (tongue jointed) it was going no-where. Here's a photo of the part assembled construction with the main mitre being to the front panel (which is face down). This one was done using Everbuild 502:

View attachment 145926

My view on this is that there are different approaches people use and the right one is the one that works dependably every time. I've done a fair few of these and haven't yet had one fail on me.

This was the assembled and veneered construction:

View attachment 145927

Disregarding any additional jointing methods that may have been used as that rather muddies the waters of the original point, you can’t really bring ply into the conversation..
ply is a cross lamination where each lam is laid at 90 degrees. So every cut edge of ply presents both end grain and edge grain. There’s fire each cut edge of ply would be a halfway house gluing wise, and they should all behave the same.
Solid timber is a different proposition

 

[Sam Kilby]

In fact, come to think of it.
A long mitre on ply, all things being equal (properly mated, appropriate clamping pressure etc) should provide a better gluing opportunity than a butt joint, as by its nature the mitre will have more gluing surface area.
A mitre joint on the edge grain of timber should be the same. A mitre on the end grain, no.

 

[RobNichols]

I found the video originally posted really useful in helping me appreciate the physical properties of wood and how it relates to glue. Basically wood is a series of rods (fibers) joined together in a matrix of organic material (lignin). The torsional strength comes from the fibers.

Modern wood glue is typically stronger than lignin, but weaker than the fibers. The glue is only a little stronger than the lignin, but a lot weaker than the fibres.

If you glue long-grain to long-grain the glue joins lignin to lignin, and as the lignin is weaker than the glue, the lignin fails before the glue. That is, the wood fails first.

If you glue end-grain to end-grain the glue joins fibers to fibers and as the fibers are stronger than the glue, the glue fails first.

The key takeaway is that the strongest joints will be those which take most advantage of the strength of the fibers. A joint where the fibers bridge the joint will be the strongest. That is exactly what occurs in the classic mortise and tenon. Items such as dowels and loose tenons add strength because they effectively bridge the joint with wood fibers.

It says to me that arguments along the lines of it being a waste of time applying glue to the parts of a joint when grain meets grain in a particular orientation are just rubbish. Yes, some joints may only be as strong as the lignin, but that is still pretty strong - almost as strong as the glue. There will be situations where that is strong enough.

This way of thinking has changed the way I think about glue. If I need the strongest joint, I know I need one that best utilises the strength of the fibers. In these joints the role of the glue is really to keep the fibers in place - to replace the lignin in the artificial abutment. The role of glue in a mortise and tenon joint is to stop it pulling apart. It's role in adding torsional strength is minimal - it's the fibres doing the work there.

It has led me to looking at wooden projects and thinking "how can I preserve the flow of the fibres". How can I ensure that long grain bridges the areas of a structure where the stress is greatest. It's led me to worry less about the glue, and more about making the most of the inherent strength of the wood.