The myth probably comes from the fact that with end-to-end, it is the glue that fails and not the wood. Making it LOOK like the joint is stronger with side-to-side, and end-to-side. The glue LOOKS weaker with end-to-end because that is the failure point.
Exactly. The first thing Dad taught me about woodworking when I was a little kid was that the glued joint is stronger than the wood. So it's not that the end grain joint is in itself inherently weaker, but the joint is indeed weaker than the wood, which goes counter to what we've always been told. It goes back to properly engineering the project to avoid building in weakness through poorly located or sized joints.
Go one step further in the myth to see why it exists: The end-grain-glue failure is being compared to the much superior long grain strength (two short pieces made into a long one, then stressed like a beam). Hence, "end to end glue up is weaker than the wood". In a side-to-side glue up, glue is compared to the weaker lignin strength, with stress parallel to the grain. Hence "side to side glue up is stronger than the wood".
Also misinterpretation in memory - people read or heard something, then they remembered it the other way around after many years, so a myth spread... I thought the results were obvious - that's why it's completely fine to miter frame corners and there's no need to look any further.
It's so weird that Titebond or some other big glue manufacturer has never tested this. Such a missed opportunity to say their glue works better on end-grain than edge grain!
Steve: I have been doing research on glues for several years now. When I started, I contacted the Franklin Company, who make Titebond, and asked them for test data that I could compare with mine. Their technical rep told me that they had no test data other than a single ASTM test of shear strength. I found this just as surprising as your reaction above. Perhaps part of the reason is that this company has already garnered the lion's share of the woodworking market, and has focused more attention on the development of multiple advanced adhesives that they offer for industrial use. In light of this, I thought it might be useful if I could provide some factual data that we non-industrial hobbyists could use to guide our projects. I am planning to do more along this line. Thanks for taking the time to tune in to this channel. You have probably gone through more gallons of PVA glue than I have.
One thing that the current political environment has taught me is to pay careful attention to words. Words have meaning. I think you're misinterpreting his results as well as its meaning to wood workers. Watch it again. Wakodahatchee Chris
Great video I'm just a dummy but i don't see how anyone could give this video a thumbs down. It's for sure a thumbs up on just the setup of the test alone. Not to mention the results!! BRAVO Mr. Sullivan
I thought the conclusion of this video was that the glue joint itself doesn't matter, it's whether you are stressing parallel or orthogonal to the fibers, the latter being stronger. Am I missing something?
In a decade of watching RUclips woodworking channels, I have never seen a video create so much discussion instantaneously, all with the purpose and momentum to make the craft better. Lots of videos give to the community as ideas and entertainment, but this video is at a whole other level aimed at improving the craft. The best part is that is is being done in the community and not from a corporate sponsor or an agenda. Really stellar work Patrick, love you videos and your approach and cant wait to see more.
I agree. I guess this is what we call a bonafide teacher. I just started watching woodworking videos (I'm a beginner occasional hobby woodworker stepping up my game) and there are so many pros out there doing some really great work. But I suppose it's not really necessary to understand glue if your joint technology is strong, especially when we consider Japanese woodworking techniques. But now I am curious about advanced glue methodology. What factors enhance the bond? Temperature? Humidity? Pressure? Vacuum? Electro-magnetism? Sound-vibrations? Light?
This was certainly one of the best woodworking videos, and I have watched hundreds. Patrick does an amazing job of clarifying the fallacy of the myth in a manner that leaves no doubt about the voracity of the outcomes. Great job Patrick, and thanks for your devotion to learning and teaching, a life long skill often lost among our youth.
Mechanical engineer and woodworker here. This video is fantastic, well done! You looked at so many factors here that are significant. Joints aren't analyzed just by 'how strong' they are, but also how and why they fail. Wood direction matters, and wood species matters. You hit in all of that brilliantly. Great communication, thanks for the video!
I love how casually he says “… so I built a strain gauge specifically design to measure clamping force”. Absolutely amazing, straightfoward, elegant and, as far as I can see, methodologically sound experiment. Im very grateful for content like this!
I think this myth was also propagated because the failure happens at the glue when end-end joints are made. So the logic was “because the glue fails when end-end, it must mean that the glue joint is weak. But with edge-edge joints, the wood fails…so the glue joint must be stronger right?” Without considering the overall strength! So interesting…
This is exactly what I was thinking too. And of course what he mentioned was that we were comparing long side glued joints with short end-grain joints.
I'd wager that the end to end glued joints get weaker as the force of pressure gets concentrated further from center. I would guess that doesn't necessarily hold true for edge to edge glue.
I'm a retired mechanical engineer and a life long woodworker. I too bought into the myth. Well done Patrick. A very well thought out and executed study.
So as a structural engineer, I loved the last demonstration. The biggest factor that must be considered is application area. Yes end grain joints are stronger, but you'll rarely have an end grain joint that has as much surface area as a side grain joint. The "myth" of end grain joints is really an application of force distribution. This really explains why you cannot "lengthen" a board with an end grain joint since those will typically be subjected to long axis (weak axis) bending where as side grain glue ups will be subjected to short axis (strong axis) bending.
yeah the moment forces on the seam will usually be much larger, coupled with a usually much smaller GSA (glued Surface Area) and then compounded with a usual inability to adequately apply enough clamping force axially prior to curing.
Actually you can lenghten the board. Look at wooden planes, they have wing spars which span the whole wingspan and you won't find structural wood long enough. You just need make a scarf joint with a slope 1:12, preferably 1:15. With that you make end to end grain joint to be almost long to long grain and give the joint much larger surface area.
Although you're right about not applying these findings incorrectly, I think the "myth" probably came about from people observing the joint failing before the material and concluding that the problem must be that the glue is not performing well on end grain. When in actuality the truth is that the glue has roughly the same strength regardless of scenario. It is the strength of the wood itself that is drastically different depending on grain direction.
The last test did not break at the glue line it broke at the wood fibres, thus proving nothing. Any thin end grain piece will easily break at any point. Gluing end grain joints is regularly observed in cheap mass produced wood furniture usually with addition of brads and nails. They always fail eventually, probably related to wood movement over time due to humidity changes. Pointless video, waste of time making it and waste of time watching it
Outstanding work Patrick!! I am also a Mechanical Engineer/Woodworker/Creator and love your approach. Hard to argue data when the process and variabilities are controlled so well. Thank you for your huge efforts and contribution. I'm not sure if application wise I will change my designing of Mortis and Tennon or Mechanical Screws/Nails over an end to end. I will probably still put splices in my 45 degree corners too. I realize you were not saying not to do that, you were simply disproving a myth. You did that fully and I really enjoyed it. Thanks again. I will view all of your videos now.
I just had an end grain joint fail on me and there's no way it was under the kind of stress showed in these videos. Just because it tests strongly a few days after having been made doesn't mean it will hold up over years in the real world. Continue making your joints properly and getting long grain crossing your joints. Preferably, have it mechanically supported by the wood (like mortise and tenon) so that even if the glue failed it would still hold together.
It would be really interesting to see what happens to the strength over time as the side to end grain joint grows and shrinks with changes in humidity. I’ve always thought that this was what caused these joints to fail, not just the differences in the grain direction. The new glue strengthens the joints as you demonstrated but doesn’t the glue fracture and become weak as the wood swells and shrinks over time. This is my concern - that my project is strong when it leaves the shop but a few years later it begins to fall apart.
I'm not a woodworker, but I did work in R&D, a bunch of years, with different materials and tasks. Wood not included. So, in my opinion, this test is a pretty good one and scientific enough, as I see it, to be seriously taken in consideration, as valid. Now, if we are talking about humidity and adequate glue it's used, the rest of the conditions remains the same. The strength of the different wood fibers can be affected by the humidity (or just make them more elastic, which is not necessary a weakness), but the conditions are the same for all the wood species/types used. So I suspect that the 3 methods of gluing will not get those proved weaker in this test, stronger, because of the humidity. As I said, I'm not a woodworker or a specialist of any kind in this field. I'm just thinking loud 😀
That could be a factor, considering that wood dimensions vary perpendicular to the grain more than parallel to the grain, and that affects end grain joints in 2 directions, but side grain in only 1.
@@stefang1087 What you're not taking into account is that wood is a very strange material in that it swells with changes in humidity and temperature like most materials, but UNLIKE most materials it does so anisotropically. In other words, wood experiences much more movement perpendicular to its grain than it does parallel to its grain. The point that Andrew was making was not that humidity might have affected the test - because as you pointed out humidity would have affected all samples similarly. What Andrew is saying is that over time, as wood swells and contracts many times, this badly strains joints that don't respect this unique property of wood, and an end-grain to side-grain joint is a perfect example of this. I do think the results of this test are valid under the limited scope of a 3" test sample that is tested very quickly after gluing. But I guarantee you that if you make a 3 foot wide dinner table and put a breadboard end onto it that doesn't account for wood movement, this glue joint will separate over time even without any external stress at all. The wood movement itself will destroy a long end-grain to side grain joint, regardless of the glue bonding strength. This is why gluing end-grain to side-grain is a bad idea - not because glue doesn't adhere well to end grain.
Been teaching woodworking for 7 years and feel I have some explaining to do for a generation of high school students. I am curious about the dimensional changes of wood with humidity and how that affects the geometric strain on joints. Really appreciate this!
I had too many endgrain joints fail after a few months to not do it that way. Wood expands and contracts from the center out. That's why tables with breadboards split if glued. Chess boards split along the engrain if there's no spine or biscuit. There's a reason why we don't use them
People used to think the Sun went around the Earth... because that's what it looks like as the Sun rises and sets. People used to think end grain to end grain gluing was weaker... because that's what it looked like when the joint failed at the glue line. Things are not always as they look. Patrick Sullivan has busted the end grain myth that so many of us have believed for decades. Thankyou, I take my hat off to you for this outstanding piece of work.
I can honestly say, I never bought into this myth. I "normally" think in what I call "logical mechanical thinking". The side grain of whatever wood, doesn't have any "holes" or "crevices" to sink down into. End grain has tons of little holes to go into as anchor points, almost like a bunch of small dowels. So to me, logically the end to end, would be the strongest across the board. Glad to see someone actually made a video to back up what I already believed to be true. Thanks and good job!
Very cool information. Thanks for the effort. I hope it doesn't have an unwanted effect on people now thinking that wooden joinery is obsolete. It will still be necessary to get joints closer to the strength of the long grain.
right, he still highlighted that the joint is strongest based on the amount of long grain fiber that *crosses* the joint. i.e. basic joinery will always beat straight up glue.
That’s outstanding. And really fairly scientific in the approach. I like it. Kinda throws out the whole notion of needing “sizing” to get extra glue into the end-grain after some has wicked away. I think I might set up some similar tests as you did, if for no other reason than to corroborate your data for myself. At any rate, you earned a new sub, and I’ll be sharing. Thanks! James
While sizing an end grain joint might work OK with hot hide glue, PVA doesn’t stick to cured PVA at all well. It strikes me as a particularly awful concept.
@@lightaces Sizing according to Titbond is giving the end grain surface a quick thin coat of 50% glue and 50% water. That surface is allowed to sit for 60 to 120 seconds and then a coat of normal glue is applied. The end grain is not coated with full strength glue and allowed to stiffen and then recoated. Again, according to Titebond.
@@1deerndingo well, that makes a lot more sense. Not relevant to my business (see my other post about glue strength not really being important in a well designed joint!), but there you go.
@@lightaces I think the concept of sizing on end grain joints is not so much to cause the two applications of glue to adhere to each other (because we know that doesn’t work), but to fill up or block the porous end fibers, so that when the full-strength glue is applied, it doesn’t all soak into those end grains. That’s my thinking about it, at least 😊
"notion of needing “sizing” to get extra glue into the end-grain after some has wicked away." This makes no sense. If the end grain wicks away the glue, then adding more glue would make sense. Having extra space there won't help.
As a guitar maker, I'm constantly "Breaking the rules," regarding how things are supposed to be glued together. As wood workers, we're constantly told that you don't glue up parts cross grain or end grain to end grain, and yet, in building a guitar, we constantly find ourselves doing those exact things, and that on a very thin, lightly constructed "box" that has to stand up to literally hundreds of pounds of pull or stress. As a result, I've often wondered how we get away with it. I think that, the fact that we keep our materials in humidity controlled rooms, work with those woods in temp and humidity controlled shops and also tell our customers not to expose the instrument to wild changes in heat and humidity might be the key and also possibly a variable that could effect the long term health of an end grain to end grain or cross grain glue up. The big thing in avoiding huge swings of humidity is that you are trying to limit the differing movement of the vairous parts, along their differing dimensions. When one part moves greatly in one direction in opposition to the piece it's glued to, you get breakage. Perhaps then, the rules only become apparent, when a glue joint is exposed to temp and humidity changes over years. I mention this, not as a critique of your methodology, only that there may be other variables that need to also be addressed to determine the long term veracity of "The rule."
Exactly! This controlled experiment does not introduce normal humidity changing conditions over a period of years. Most any joint holds until it does not. End grain joints have a shorter life span than edge grain joints.
Interesting. I work for a hand rail manufacturing company. We end glue and face glue pieces to make blanks for railing rated for indoor and outdoor use. One of our test pieces has been outside with no finish for 3 years at this point, with no sign of de-lamination of any of the joints. With temps ranging from +110 in the summer to -25 in the winter I have few worries about part failure at this point.
where are you gluing end grain to end grain in a guitar in an area that takes load? In fact, where are you doing it at all? The only place I can possibly think of is the neck scarf joint that I suppose classes as end-to-end, and is usually a typical failure point.
I'm late to the party but would like to say that I thoroughly enjoyed this video. I wish there was an exercise like this when doing wood shop. I might have paid more attention to science class. When I was taking advanced woodworking classes in high school we was taught the most important thing in a joint is if it is clean smooth and square to its mating edge. This is a great illustration to show why you consider the direction of load on the joint when designing the project. Thank you so much for making this.
Excellent demonstration of the classic scientific method to prove that "common knowledge" can be wrong! So much for believing opinions expressed by others on social networks like Facebook and Twitter. Unbiased examination is the only thing that can be trusted.
Super interesting results! I would love to see how time impacts the end grain results. One of the reasons we're taught to do things like reinforce miters is not just because of the possible myth that the joint is inherently weak, but also the fact that the joint is subject to a lot of stress as the pieces expand and contract over time. So I wonder what these results would look like in 1-2 yr old samples. I would also love to see these tests with longer boards. I imagine the amount of force required to break two 12" long boards apart at the joint is considerably less because of the additional leverage. Either way, thought-invoking stuff. Thanks for making the video.
For most products there is a way to accelerate testing. Putting test parts in a humidity chamber and cycling from seasonal high to seasonal low relative humidity and doing sat 50 cycles would work. Not sure how long a sample would have to be at a high or low humidity for it to fully penetrate the 3/4" thick piece. I suspect Mark is correct and I have always assumed wood movement is why you are told not to glue end to side grain.
I think you hit it with the last comment. This showed a 3/4 by 3 inch end grain to end grain joint is stronger than the grain of a 3/4 inch by 3 inch long piece of wood. It tells you nothing about the relative strength of the joints. Glue end grain to a board 12 inches long (or 2 inches thick) and you would get very different results.
A friend has one of many glass etched mirrors I did forty years ago. It's not even mitered. Instead, I cut quarter circles to join straight pieces. The mirror weighs in at least 20 pounds and the frame shows zero indication of failure four decades in.
This is an Awesome video, and very well tested. I wondered for a long time why people said this joint was bad, but I did not have a good way to test it. Thank you so much for making this video.
The reason might be because an endgrain joint dries much quicker. A dry joint wont bont nearly as well. The end grain basically wicks up the glue to the point where there isnt enough glue on the surface anymore for a proper glue joint.
As soon as you showed that the glue was stronger than the lignin, I understood why mitre joints (especially with splines) are so strong. Jason at bourbon moth woodworking made a video testing different types of joints, and was surprised that mitres were stronger than boxes and dovetails. Thanks for challenging and expanding our knowledge!
I've never glued anything bigger than Christmas decoration or flatpack cupboards for extra strength and yet I've come across the myth that end to end glue joints are weak. This video clearly demonstrates this to be false. Very well made and presented. Many thumbs up!
This is very interesting. Great to see someone with the proper testing equipment, knowledge, skill and curiosity necessary to bust a myth. Thank you Patrick!
Beautifully explained. In fact, just earlier today i was telling my 13 year old son that the end joint glue is bad idea. Well now i have to show him your video.... lol. Thanks a lot Mr. Sullivan.
I mean, it's all relative to the design of the structure. If you are making a 'T' joint, you'll get a weaker glue-up if you butt the end-grain up against the "front" of the other piece instead of gluing it to the "side", long grain to long grain, assuming the surface area is the same. The long grain breaks in this video are akin to the board breaking feats in karate, where there's no real structural strength there to begin with. The end grain glue joint failed first in a configuration that allowed the full strength of the wood fibers to be brought to bear on the joint.
You weren't wrong, and this video is misleading. Please note, he never actually broke a long grain glue joint. In truth, and design which relies on a 2.25 square inch glue joint to take any kind of force is a bad structural design.
When building traditional balsa framed model aircraft, side to end grain joints are frequent in the 'warren girder' structures employed. I've always been amazed at the resulting strength and rigidity, As a one-time mechanical engineer who has installed many strain gauges I really appreciate this video. Thanks.
Love to hear strain gauges mentioned in a woodworking context. My dad was an engineer and later VP for Micro-Measurements/Vishay. Worked for them 25 years selling strain gauges and helping customers with their applications.
Dear "Mr" Sulliavn, I think your video was perfectly clear as well as your conclusions. I am sorry for the people who understood something else than what you masterfully demonstrated and explained. I thank you for sharing.
Absolutely astonishing. This makes me so happy. The possibilities! So many things I've wanted to do over the years that I was always told would wind up failing, so I never bothered. I'm gonna do all of them now! Thank you for this, Patrick. This is what's up.
I often wondered why everyone said that end grain glue joints were no good, since the wood absorbed the glue into the end of the grain like a straw and they all said that the wood would fail before the glue. It would appear that the size of the cross section being glued is more important than the direction of the wood grain. But ultimately, the direction the load will be applied matters the most. This was fascinating to watch. Thank you,
It was a very interesting test. I think that the only drawback is that due to expansion and contraction of the wood the joint will fail eventually under any kind of stress. I did this mistake with a mitered picture frames. After one year the frame bottom just fell apart, glass pushed it out and broke on the floor. This is where I started to look into woodworking joints and got really interesting in woodworking. Best regards from Croatia
I have many of my picture frames in our summer home that sees high temperatures and high humidity in the summer and sub-zero cold and very low humidity in the winter. None have failed. MItered corners and no splines, dowels, or biscuits.
I would absolutely love to see more testing like this. The more we understand about why joints work (or don't work) the better we can be at designing new projects. Great work!
Wonderful explication!!! I am an engineer and pedantic in mindset,, so when someone does a complicated experimental demonstration with absolutely faultless methodology and reasoning, I am impressed. Thank you for the education and amusement.
Absolutely great testing and reporting! And, no, not too many numbers! As an engineer (albeit electrical, audio, & broadcast) I appreciate the inclusion of numbers. Exposing more people to the numbers helps to get people used to seeing and dealing with them. Paul-Joseph de Werk likely has the answer on why the myth developed.
I was very interested in the video, but when he got to the solid pieces I sat closer to the edge of my seat thinking, "Ok. Now this is going to be really interesting."
Absolutely incredible! Thank you for this new evaluation of a very old issue. Thanks to you, I've learned something new after more than a half-century of cabinetmaking. Of course, I'm not going to give up my mortise and tenon joinery for simple right-angle butt joints, but I have a much better understanding of how the mortise and tenon uses the long-grain strength of the wood in two directions to create a super joint. Thank you for your outstanding scientific analysis.
Thank you, Mr. Sullivan. I learned almost as much about the characteristics of difference wood as I did about gluing them together. Priceless for a rookie woodworker like me. You even provoked me into my first RUclips Comment ever!
Excellent! As a former test engineer, this was exactly the correct approach.... Two additions for your next test.... 1 - Rotate the joints 90 degrees vertically in the press. this would more accurately simulate a mitered corner, or just miter one. 2- do the unglued wood rotated 90 degrees horizontally. This will test the lignin only and should be close to the side to side grain tests. Thanks!
Hi Patrick. Great video with some really interesting results. I think in the future you might want to use a 4 point bend and not a 3 point bend. A 3 point bend places all the stress directly on the joint. A 4 point bend places uniform stress between the two rollers. The effect being that a 3 point bend is used to test a material which is homogeneous whereas a 4 point bend tests 'the structure' between the rollers. Example being: if you want to test bending strength of raw timber a 3 point bend would be fine. If however you wanted to test something like the bending strength of riveted metal truss a 4 point bend would be better. A 4 point bend is also much more forgiving of alignment errors.
Actually this just shows what we’ve always known - that glue joints are stronger than wood in side grain joints, and that wood is stronger than glue in end grain joints. This experiment never broke a side grain joint, because the wood failed first - so we still don’t know how strong that joint is. The end grain joint was the only glue joint to break - otherwise we’re only seeing the comparative low shear strength of wood along the fibers, compared to across the fibers, which we knew anyway.
This is the definitive take-away. Make those boards (end-glued) twice as long, and move the point of leverage to the ends and test it again. I would imagine the results are strikingly different.
@@djlusid And why we would only test the end-grain glued joints with longer leverage? If you did all tests with longer boards, still the results would be the same (not the absolute values of force, but the comparsion between end-grain and long-grain), because why would it change?
Very interesting ! I think that the myth also comes from craftsman seeing the glue fail in endgrain to endgrain led to assume the joint was weaker. On the other hand, if the joint break because the wood fails, the joint itself doesn't look like the issue. Another point, it's hard to get a big surface area in end grain to end grain joint.
That was my thought as well. If I saw those end-to-end breaks in the wild, my reaction would probably be, "wow, that glue joint sucked" not "wow, the wood was extra strong in that orientation."
That sounds like a classic case of confirmation bias. "I Have achieved a result that looks so obvious to any other man or woman, that I need not look deeper."
In fairness, we don't know how strong the glue joint is if the wood keeps failing first when we try and test it. Maybe end grain is still the weakest, it just doesn't matter in real life because the wood is the real limiting factor
Excellent examination and one that many (most?) traditionally trained woodworkers understood but has been lost to "modern wood machinist" hearsay...As a working Timberwright and someone that only works in "green wood" for most projects, it is refreshing to find someone that is dispelling myths and/or being properly accurate in the description of what is taking place in a given wood joint application...be it a simple glue butting joint of some form or otherwise...I look forward to following your videos from now on...Thanks for the great documentation and explanations...
So really, this is not comparing the relative strengths of the joints, but rather the relative strengths of different wood grain orientation combinations. For end-to-end the difference is that the wood grain is in the strongest possible orientation for both pieces, so the wood glue then becomes the weakest link since there is no lignin boundary to fail first. Great video!
You should definitely give this video a boost. This should you RUclips woodworking viral as much as the Veritasium click bait video. Patrick's channel is great but he falls through the RUclips cracks because of the frequency.
I think it wise to approach these things cautiously. This is either the biggest upset to the woodworking status quo in decades, or there's a delta at play here yet to be uncovered. I'd like to see a third party involved, for peer review if nothing else, but I think maybe a dialog should be opened WITH the makers of TiteBond themselves... I could swear I've seen tests like this in my civil engineering/material sciences class work, with the opposite results. I feel like more study is needed for this to get gospel status.
@@grumblycurmudgeon What I think is key here is that beyond just "Test it and see" - we have a solid explanation as to why this is the case. Every woodworker and guy who's tried to chop firewood knows that it's a hell of a lot easier to split wood along the grain than across it - this knowledge probably goes back to the time of stone tools. The knowledge that all modern woodworkers have that "PVA glue is stronger than (side-glued) wood" is also well known and understood. The conclusion here is that for any glue which is stronger than lignin - the outcome found in this video should be fairly obvious. (Although, I readily admit that I was surprised...but now I think about it...I shouldn't have been). I strongly suspect that this oft-repeated-falsehood about end-grain gluing came about in an era when glue was weaker than lignin - in which case side-grain gluing would indeed have been better (or at least no worse) than end-grain. In all likelyhood, this all goes back to victorian times when the best glues were pretty pathetic by modern standards. So many engineering questions, practices and dogmas were accepted in Victorian times - and never really questioned since then. As an example - I learned that an awful lot of the parts of my (Tesla) car are either glued or stuck together with double-sided tape! My first reaction was that of horror - I expected welding and big-assed bolts...but it turns out that modern automotive adhesives are actually stronger! This eats at our gut-feel engineering skills - and even as I write those words about my car - it's hard to suppress a feeling of deep unease about it. This is why humanity invented science - and why using it's methods teaches us things that our poor evolved monkey-brains are ill-equipped to decide by themselves. So, sure - independent testing and peer-review are important - but no matter what - this single set of experiments has to throw doubt on the the old dogma which does not appear to have any testing at all (at least not with modern glues). There are new questions here though. For example, if the developers of modern glues accepted that end-grain gluing is good - perhaps they'd develop new glues that are optimized for the end-grain situation?
I find these tests great for the general info application. How many woodworkers build projects that will face these types of pressures? An engineer building bridges or towers might use this data. How long will the end grain butt joint last over time without cracking or breaking unaided by a mechanical joint? There is a reason why tradition is a timeless guide
I guess that is why the scarf joint rocks for joining long, thin sections expecting to carry a great load. Combines end-to-end glue joint and natural longitudinal strength.
Great video. This is like a breath of fresh air after recently watching a series of 30 seconds long videos, where a certain youtuber "demonstrated" that glue is stronger than the wood, and the amount of thought and effort he put into making those videos equaled their length.
The glue is still stronger than the wood in most situations. Makes nails and screws seem almost obsolete in most settings unless you want to be able to disassemble!
As a mechanical engineer, I can greatly appreciate your analysis done in the proper scientific method as opposed to the common "jumping on the end of a board" methods. Very well done on all aspects: content, presentation, understandable, videography, ... GREAT JOB!!!! Engineering also has its myths; usually started when some eminent authority first mentioned it in his textbook, then other authors would then simply copy it into their textbooks unchecked giving the myth more weight and "truthfulness".
A proper fantastic job!!!!! As someone with a mechanical engineering background, you made me feel like a fool for falling for this myth all these years!
Blew my mind! Got into proper woodworking rather recently, and after absorbing all the knowledge I could find this was the one message that was taught by everyone. I hope you manage to get this video viral, would love to see the community respond! And hopefully get the message changed from now on.
Very interesting, thanks! When end grain to end grain fails it always fails at the glued section which might have created the myth since people usually do not measure the force needed to break it.
Ya, bit of selection bias kinda like when people say (thing) from long ago was built better because look how long it lasts. No, you are just looking at the survivors.
I wouldn't have seen this video if Stumpy Nubs hadn't "responded" (moreso his scare-quotes than mine) to it. I agree with him that you've buried the lead in your conclusion. I would have a different specific complaint if I hadn't seen his video. It's obvious that you've put a lot of good work into this video and I look forward to the full series.
Wow! I am new to woodworking but cannot count the number of times this myth was suggested to me as truth. Thanks For the tutorial. Very helpful in clarifying the topic. Blessings!
Excellent experimental design. Excellent work eliminating variables and testing only one independent variable at a time. You have done the work of an experimental scientist.
Because there are so many comments I can't spend days or weeks reading them all so I don't know if this has been brought up yet, but I would love to see different glues tested. Because I feel the myth about end grain joints stems largely from the more traditional hide glue used in the 1700s and1800s, or even into the 20th century, Testing various types of glue such as epoxy, hide, PVA, etc might shed more light on the myth. Very well done tests and thorough results! definitely continue to create videos!
What an incredibly well thought out test. The fact you dared question age old "knowledge" is amazing in and of itself. Thank you for thinking outside the box to bring this exceptional exception to misunderstood understanding
More people should do so. It would save a lot of trouble in many many area's. You'd be amazed how much "exceptions" there are. Truths that have been told that sound good enough, but actually presents so much more nuances and half-lies. I'd like to call it science by consensus versus science by facts. The first being science repeated and repeated over and over again, scientists circle pointing to each others findings. Using models, graphs and predictions to calculate the truth. But to actually go out there and actually perform science to verify those truth theories is seldomly done.
Huge thanks for that research, puts everything in perspective. If you are doing any more tests I'd be very interested to see a test on scarf joints using the same width and thickness timber to understand comparative strength. I think in boat building they use scarfs as short as 6:1 or 8:1 and in aircraft they go up to around 15:1.
This is the best testing procedure I have seen. EXCELLENT sir. The test results and reporting is outstanding. I wish most if not all RUclipss could learn from you. You have set the standard........
This was awesome and it completely destroyed what I had always believed about end grain to end grain glue joints. I’ve been woodworking for over 30 years now. Well done science, can’t argue with fact over fiction/word of mouth.
You’re back! Looove your content. My suggestion for a follow up - compare end-to-end glue strength with regular application vs massaging the glue in until completely saturated. I do not recommend letting it dry or applying a watered down first application like some people do. Just massage the glue and press down hard to push the glue into the end grain, and keep applying until it stops absorbing and stays wet. I do this every time (takes much longer) thinking it must perform better, but maybe it’s a myth too.
Late to the party, but awesome video. I think application needs to be considered as well. I can definitely see applications where end to end would yield a joint that will last a lifetime. Table tops, picture/mirror frames, and the like. Where racking force comes into play, I don't think any of us would consider end to end, anyway. You've made me a believer, and I can see where this will make me rethink my approach to gluing up panels, and frames. Thanks!
Be careful though. I just had an end grain joint fail. Lasted less than a year, not a lifetime. I will from now on make joints properly and not rely on glued end grain. There wasn't even that much stress on this joint. Get long grain crossing the joint and preferably design things so that if the glue failed there would still be interlocking wood to hold it together (like tenons).
That was totally fascinating, your methodology was so much more thorough and well thought through than other videos I've seen that claim to cover the same subject. I've subscribed and will look at what else you've shared ! Thanks for providing this test.😊
Someone just passed this video to me. It's exceptionally well done, Patrick. I thank you for challenging the tribal 'knowledge' and doing so with such scientific rigor. And I'm happy to unlearn what I have learned.
I don’t think these results change the conventional wisdom since the natural alternative to an end grain to end grain joint is an unbroken continuous long grain piece which your testing shows to be many times stronger. I would love to see this rigor applied to scarf joints and maybe look at the relationship of glue surface area and strength. Excellent video. Made me think and I definitely learned something new.
@john milton and you were still able to be a dick in your reply... The purpose of the video wasn't to say glued up end joints are the way to go for anything, it was to test if the long held assertion was true or not, and he proved it was a myth. Job well done.
@John Milton, your point is that repeating what you take for granted without trying it yourself is better than showing respect for a very careful science experiment. So what is theory and what is practice?
@@alamaralaa He didn't prove it though. He proved that end to end joints are stronger than wood across the grain, or lignin as he refers to it. He didn't prove that end to end joints are stronger than side to side joints, because the side to side joints never failed. But the fact that they didn't fail in his experiment doesn't mean that they won't fail in the real world. In the real world you don't put significant stress on wood across the grain. Nobody has ever made a table where the grain doesn't run along the length. But in the real world you have years or decades of wood flexing, temperature and moisture. How will that impact the joint? Hard to test, but that's what we have decades or even generations of experience for. And those tell us: Don't glue end grain.
@@cire9984 He was wasn't trying to prove the gluing end grain was stronger than gluing long grain, and I never said he was. His assertion was that gluing end grain was not as weak as people assumed it was, and his tests prove it is a strong joint. Stronger in fact than the wood is strong across the grain. Period. End of discussion. Please don't put words in his or my mouth. You can misunderstand his tests and conclusions and my statement above all you want, but don't tell us we're wrong because you didn't understand them.
The myth could have started back before PVAs were available. Try it using hide or fish glue and see if it holds true. Thanks for an incredible video. I'll be updating some of my designs now :-)
I wonder if the problem is expansion and contraction over time. That would affect end grain quite a bit. Not something that would show up in this lab test though. Unless the wood went through sever cycles of damp and dry and was then tested. There might be truth to the old theory that end grain glue joints are suspect.
@@orion7741 It is, but hide and fish glues are both more susceptible to heat and moisture and deteriorate faster, leading to eventual failure. Besides that, expansion and contraction has taught us to combat it in other ways where needed, and it's not a major factor with a glue joint, as most glues are sealants too, and prevent contraction/expansion of the joint as an added advantage. One shouldn't even use materials or methods for any job where better ones will do a better job, unless you are planning for obsolescence. Oh, and placing a "much" after "significantly" is redundant and either would stand alone, as in this case they mean nearly same thing and one or the other will suffice, and using both is just strange.
The Oak results surprised me. I'd have thought those big pores would make the oak end-grain joint come out on top, but it was beat out by maple, walnut, and cherry. I wonder if the glue viscosity plays a role there: would a thinned-out glue, that can wick deeper into the pores by capillary action, perform better in oak? And I hope Mattias sees this video.
Some people advocate allowing the glue a few minutes to absorb into the end grain, reapply what was absorbed then connect and clamp. I've tried this with pine and noted that the glue obviously did absorb into the end grain, so your suggestion makes a great deal of sense, perhaps along with using Tightbond III which has a longer open time and seems runnier anyway.
"And I hope Mattias sees this video. " I was thinking the same. I remember his glue up tests years ago. I will have to watch again to see his results, but I think they indicated end-grain joints were weak.
It's good to be grounded in reality than say-so. This was particularly good info and will influence future designs for me. Thank you. Please blow up some more myths!
I wondered who paid for all the equipment, but I know I benefited. Amazing photography, organization, and presentation in English a schoolmarm would love! Great!!
Mr. Sullivan, An amazing and accurate design of experiment (DOE). I hope it is appreciated. You not only characterized and quantified the experiment but also continued into the why of the results. Showing both the adhesive failures and cohesive failures led to my concluding that one of the basis of this myth is that any failure of the joint was caused by the glue bond (adhesive failure - the glue alone) and not the interaction of the glue with the wood grain (a cohesive failure). You showed this very clearly in the stress testing and in the micro pictures of the failures that it was the wood that failed. The joint broke in every species case due to the failure of the "grain" adhesive - lignin - and not the PVA. To the casual woodworker (me), when that joint broke, the blame would be placed on the adhesive, not the wood. I can now stop using mechanical fasteners (a "just in case" measure) on my glued end joints. I am a victim of this myth and you have "set me free". Bravo, and thank you.
I would be interested to see this tested with pulling force instead of shear force as wel. I would expect similar results, but I'm not sure. Thanks for your rigorous testing!
It's not shear force. When you apply a deflexion force to the beam, you have one side of the beam where tension (pulling) forces are experienced, and on the other side compression forces. Check out beam theory on Wikipedia or that video ruclips.net/video/f08Y39UiC-o/видео.html But in fact, the situation isn't exactly the same, since there you have some shear stress, where on a pure tension test, you only have tension stress.
Patrick, great job on this video! Definitely super interesting and makes me think twice about gluing butt joints with no reinforcement with mechanical fasteners. EDIT: P.S. you should be able to adjust your camera's shutter speed to largely eliminate the flickering from your shop lights.
Great methodology and great presentation of results. However, your results beget a further question. Namely, what effect does conditioning of the bonded surfaces have, if any, on the strength of the bond? For example, sanding of the surfaces to be bonded with various grades of sandpaper from rough to fine, or cutting of the wood samples using a cross-cut versus a ripping blade.
Very true, I came to this video wanting the answer to 'planned or sawn?' I've seen videos that prove, the more glue the stronger, so a sawn joint with room for more glue and texture to grad to as said in this video would logically be even stronger. I would love to see part 2.
As an untrained inexperienced amateur, I genuinely had no idea that there was a “right” way & a “wrong” way of gluing lumps of dead tree carcass together. Love ya work. Subscribed. Hi from Oz 🇦🇺
Patrick, you misunderstand the end grain rule: 1. If you glue side grain, your glued joint is ALWAYS STRONGER than the wood, so use glue, its OK! 2. If you glue end grain, your glued joint is ALWAYS WEAKER than the wood. You can still glue it, but only if you do not need the full wood strength! I feel you should have mentioned that, because people think it is OK to glue the end grain now in any case! By making a lap joint, finger joint or dove tail the strength is increase dramatically. Maybe you should make a video about that, since you got nice equipment for testing!
This makes a lot of sense. Even though it took much greater force to break the end grain, the point of failure is at the joint itself (the glue), making it SEEM as though it is a weaker joint rather than a failure of material strength. False conclusions that seem so apparent. Great video!
1. I am really impressed by your methodology and attention to detail. 2. I am shocked by the conclusion about end grain glue joints. 3. Other than setting to rest a myth, what would possess you to undertake this tough of a project? 4. If you did this project as your PhD thesis, congratulations, Dr. Sullivan; you passed!
Just found your channel. I realize this video is 2 yrs old. Hope you get it. Very well done. I am impressed by the way it was done. I've been doing this since 1972. I've learned a few tricks a long the way. I always said that the only reason why glue ups fail is because of bad prep and cure time. You just proved my point. The only real joints that break are the ones put together wrong or abused after final builds. Thank you. I look forward to more videos.😊
Absolutely fascinating as a woodworker. Even more so as an engineer in observing the yield point behavior in the various orientations and different species. I found the failure modes of the unglued samples particularly interesting. The cherry and the walnut, especially the walnut, failed catastrophically as the grain failed in almost pure tension, the cherry started splintering, at the surface as the lignin apparently failed in shear, allowing the load on the grain to increase rapidly, then failed catastrophically as the grain failed in tension like the walnut. All the others appeared to fail by splintering, perhaps with the lignin failing in shear with the grain breaks staggered. I wonder what that might mean to people selecting wood for projects that require bending strength.
I think that depends on application of bending. A bow, for example - requires fast recovery response of the fiber to maximize the energy out. A chair however, will be okay with a slower recovery time of the fibers. Softer woods like pine do not make good bows since the recovery time of the fibers is very slow, but the slow recovery time seems to make them more bendable.
No, it gives humans a bad name for making the error to trust error prone humans (AKA believing), instead of following the scientific method that relies on more closely looking at the details, investigation, and following the factual data wherever it leads, and putting their biases aside even if the data points away from their beliefs.
While attending Arrowmont recently, my instructor told me that Titebond 2 was specifically formulated for end grain joints. She learned this during a phonecall with titebond reps.
Great video! You are a very inteligent comunicator. Would be interesting to know which bond is the strongest, side by side or endgrain to endgrain. Becuase the wood grain fails in the side on side example before the glue line does, is difficult to know how strong it is...
@@veri745 if you use a weak material to test a bonding agent that is used on stronger materials and the weak material fails, you have not properly tested the bonding agent
@@mjbailey404 the only "stronger materials" I see available is end-grain, so what exactly has gone un-tested? He covered a pretty reasonable array of wood species
A couple of things I'd like to see tested is smoothness of end grain and old fashioned hide glue. One thought on end grain is without precision cutting or planning the surface, you wind up with a surface that is very rough. So it may not have as much contact as a very smooth joint, esp from the days of hand tools. Same with hide glue. Would it perhaps be pulled through capillary action away from the end joint so there was simply less glue acting? Just some thoughts that some of these beliefs may have started when we used different tools and glues that today.
On a rougher end grain, the two wood surfaces may not have as much contact with each other, but they'd have more surface area for the glue to grab. I think a rougher cut, within reason, would probably result in a stronger joint than a smoothly cut, planed, or sanded end grain.
Hand tools can get a much smoother cut than your table saw or router depending on what, where and how. I'd wager a scraped or planed surface is smoother than a sanded one.
I think this is an improper framing - though granted interesting data. The issue with end grain joints is that it creates a weaker seam than the wood itself, meaning the joint weakens that segment of the piece and becomes the weakest part of the overall piece. Wood being stronger on the grain direction lets you build to the strength of the grain, but end grain joints create a weak point in that direction! Whereas long grain joints simply keep the strength pattern of the wood consistent. The strength of the joint itself is not stronger in end grain, rather an end grain joint is stronger than the long grain wood itself. It's a big difference, and personally I don't want the joint to be the weakest part of anything I make.
WOW this video kinda blew my mind! Totally shattered all I had been told (and believed) about end grain to end grain joints. Thanks for making this video it will transform the way I think about my woodworking from now on.
Awesome testing data. For years I have always used either T & G, dowels or biscuits for either side grain or end grain glue ups. Mostly for alignment but also for strength. Your testing shows I may have been over engineering my joints. LOL. Great work.
Perhaps this myth is so pervasive because of survivors bias. "See kids? That end-grain joint broke right along the glue line! That's why we should never glue like that."
that is exactly right. the misconception being that joints where the wood fails are 'normal', whereas joints where the glue fails must mean the glue is weaker than the wood.
This is looking at the glued joints from the wrong perspective. What should be compared is the strength of a glued piece compared to the strength of a continuous piece. In a side-to-side joined piece the strength is no less than a similarly sized continuous piece where the wood is the limiting factor. However in an end-to-end joined piece the strength is 3 to 8 times weaker than an equivalent continuous piece because the glue fails first. Glued end-to-end joints are bad practice because it is so much weaker than a piece without that joint.
Ok, so hear me out: if you care about the strength of the piece you are making you are going to make sure that the forces that work against the piece will be applied across the grain, not on the lignin. If you have 2 pieces of wood, one 1x1x10 made by 2 0.5x1x10 glued side to side, it will be stronger than a 1x1x10 piece of wood made by gluing 2 1x1x5 pieces of wood end to end, and that is in any direction of the wood (since the glue is stronger than the lignin, assuming the glue is in the middle). Basically: you should try to avoid gluing end to end and instead use larger pieces of wood, but when it comes to gluing side to side it doesn't really matter since the wood strength is not lost.
I tested this myself years ago and was just as surprised. I thought the glue tech must have improved since the myth was established. Your investigation clears this up very nicely!
The myth probably comes from the fact that with end-to-end, it is the glue that fails and not the wood. Making it LOOK like the joint is stronger with side-to-side, and end-to-side. The glue LOOKS weaker with end-to-end because that is the failure point.
I was thinking the exact same thing! Instantly where my mind when when I saw that first end-to-end bond broke alonge the glue line.
Exactly. The first thing Dad taught me about woodworking when I was a little kid was that the glued joint is stronger than the wood.
So it's not that the end grain joint is in itself inherently weaker, but the joint is indeed weaker than the wood, which goes counter to what we've always been told.
It goes back to properly engineering the project to avoid building in weakness through poorly located or sized joints.
Go one step further in the myth to see why it exists:
The end-grain-glue failure is being compared to the much superior long grain strength (two short pieces made into a long one, then stressed like a beam). Hence, "end to end glue up is weaker than the wood". In a side-to-side glue up, glue is compared to the weaker lignin strength, with stress parallel to the grain. Hence "side to side glue up is stronger than the wood".
Also misinterpretation in memory - people read or heard something, then they remembered it the other way around after many years, so a myth spread... I thought the results were obvious - that's why it's completely fine to miter frame corners and there's no need to look any further.
Was about to say the same thing
It's so weird that Titebond or some other big glue manufacturer has never tested this. Such a missed opportunity to say their glue works better on end-grain than edge grain!
Steve: I have been doing research on glues for several years now. When I started, I contacted the Franklin Company, who make Titebond, and asked them for test data that I could compare with mine. Their technical rep told me that they had no test data other than a single ASTM test of shear strength. I found this just as surprising as your reaction above. Perhaps part of the reason is that this company has already garnered the lion's share of the woodworking market, and has focused more attention on the development of multiple advanced adhesives that they offer for industrial use. In light of this, I thought it might be useful if I could provide some factual data that we non-industrial hobbyists could use to guide our projects. I am planning to do more along this line. Thanks for taking the time to tune in to this channel. You have probably gone through more gallons of PVA glue than I have.
One thing that the current political environment has taught me is to pay careful attention to words. Words have meaning. I think you're misinterpreting his results as well as its meaning to wood workers. Watch it again. Wakodahatchee Chris
Steve, thanks for sharing this video, it was fascinating and really really well done!
Great video I'm just a dummy but i don't see how anyone could give this video a thumbs down. It's for sure a thumbs up on just the setup of the test alone. Not to mention the results!! BRAVO Mr. Sullivan
I thought the conclusion of this video was that the glue joint itself doesn't matter, it's whether you are stressing parallel or orthogonal to the fibers, the latter being stronger. Am I missing something?
In a decade of watching RUclips woodworking channels, I have never seen a video create so much discussion instantaneously, all with the purpose and momentum to make the craft better. Lots of videos give to the community as ideas and entertainment, but this video is at a whole other level aimed at improving the craft. The best part is that is is being done in the community and not from a corporate sponsor or an agenda. Really stellar work Patrick, love you videos and your approach and cant wait to see more.
I agree. I guess this is what we call a bonafide teacher. I just started watching woodworking videos (I'm a beginner occasional hobby woodworker stepping up my game) and there are so many pros out there doing some really great work. But I suppose it's not really necessary to understand glue if your joint technology is strong, especially when we consider Japanese woodworking techniques. But now I am curious about advanced glue methodology. What factors enhance the bond? Temperature? Humidity? Pressure? Vacuum? Electro-magnetism? Sound-vibrations? Light?
This was certainly one of the best woodworking videos, and I have watched hundreds. Patrick does an amazing job of clarifying the fallacy of the myth in a manner that leaves no doubt about the voracity of the outcomes. Great job Patrick, and thanks for your devotion to learning and teaching, a life long skill often lost among our youth.
Mechanical engineer and woodworker here. This video is fantastic, well done! You looked at so many factors here that are significant. Joints aren't analyzed just by 'how strong' they are, but also how and why they fail. Wood direction matters, and wood species matters. You hit in all of that brilliantly. Great communication, thanks for the video!
I love how casually he says “… so I built a strain gauge specifically design to measure clamping force”. Absolutely amazing, straightfoward, elegant and, as far as I can see, methodologically sound experiment. Im very grateful for content like this!
Not as hard as it looks. Spend some time learning to program Arduinos, you can do a lot!
My strain gauge would be adding weights with thinner pieces of wood that break with less pressure. I am so poo I can't afford the r.
I think this myth was also propagated because the failure happens at the glue when end-end joints are made. So the logic was “because the glue fails when end-end, it must mean that the glue joint is weak. But with edge-edge joints, the wood fails…so the glue joint must be stronger right?” Without considering the overall strength! So interesting…
I think this is another case of survivorship bias.
This is exactly what I was thinking too. And of course what he mentioned was that we were comparing long side glued joints with short end-grain joints.
I'd wager that the end to end glued joints get weaker as the force of pressure gets concentrated further from center. I would guess that doesn't necessarily hold true for edge to edge glue.
@@Montrovantis I'd be inclined to disagree. Leverage is quite easy to calculate and will affect both equally.
I'm a retired mechanical engineer and a life long woodworker. I too bought into the myth. Well done Patrick. A very well thought out and executed study.
True
This is a masterpiece of clarity, transparency, and presentation.
So as a structural engineer, I loved the last demonstration. The biggest factor that must be considered is application area. Yes end grain joints are stronger, but you'll rarely have an end grain joint that has as much surface area as a side grain joint. The "myth" of end grain joints is really an application of force distribution. This really explains why you cannot "lengthen" a board with an end grain joint since those will typically be subjected to long axis (weak axis) bending where as side grain glue ups will be subjected to short axis (strong axis) bending.
yeah the moment forces on the seam will usually be much larger, coupled with a usually much smaller GSA (glued Surface Area) and then compounded with a usual inability to adequately apply enough clamping force axially prior to curing.
Actually you can lenghten the board. Look at wooden planes, they have wing spars which span the whole wingspan and you won't find structural wood long enough. You just need make a scarf joint with a slope 1:12, preferably 1:15. With that you make end to end grain joint to be almost long to long grain and give the joint much larger surface area.
@@fifi23o5 no ii8
Although you're right about not applying these findings incorrectly, I think the "myth" probably came about from people observing the joint failing before the material and concluding that the problem must be that the glue is not performing well on end grain.
When in actuality the truth is that the glue has roughly the same strength regardless of scenario. It is the strength of the wood itself that is drastically different depending on grain direction.
The last test did not break at the glue line it broke at the wood fibres, thus proving nothing. Any thin end grain piece will easily break at any point. Gluing end grain joints is regularly observed in cheap mass produced wood furniture usually with addition of brads and nails. They always fail eventually, probably related to wood movement over time due to humidity changes. Pointless video, waste of time making it and waste of time watching it
Outstanding work Patrick!! I am also a Mechanical Engineer/Woodworker/Creator and love your approach. Hard to argue data when the process and variabilities are controlled so well. Thank you for your huge efforts and contribution. I'm not sure if application wise I will change my designing of Mortis and Tennon or Mechanical Screws/Nails over an end to end. I will probably still put splices in my 45 degree corners too. I realize you were not saying not to do that, you were simply disproving a myth. You did that fully and I really enjoyed it. Thanks again. I will view all of your videos now.
I just had an end grain joint fail on me and there's no way it was under the kind of stress showed in these videos. Just because it tests strongly a few days after having been made doesn't mean it will hold up over years in the real world. Continue making your joints properly and getting long grain crossing your joints. Preferably, have it mechanically supported by the wood (like mortise and tenon) so that even if the glue failed it would still hold together.
This blows my mind! Thanks for such a great and enlightening video
@john milton we both live in england so you might be right :P
its logical;
It would be really interesting to see what happens to the strength over time as the side to end grain joint grows and shrinks with changes in humidity. I’ve always thought that this was what caused these joints to fail, not just the differences in the grain direction. The new glue strengthens the joints as you demonstrated but doesn’t the glue fracture and become weak as the wood swells and shrinks over time. This is my concern - that my project is strong when it leaves the shop but a few years later it begins to fall apart.
Worth doing the tests with a drastic change in relative humidity! Could be simulated in a lab environment! Excellent point.
I'm not a woodworker, but I did work in R&D, a bunch of years, with different materials and tasks. Wood not included. So, in my opinion, this test is a pretty good one and scientific enough, as I see it, to be seriously taken in consideration, as valid.
Now, if we are talking about humidity and adequate glue it's used, the rest of the conditions remains the same. The strength of the different wood fibers can be affected by the humidity (or just make them more elastic, which is not necessary a weakness), but the conditions are the same for all the wood species/types used. So I suspect that the 3 methods of gluing will not get those proved weaker in this test, stronger, because of the humidity.
As I said, I'm not a woodworker or a specialist of any kind in this field. I'm just thinking loud 😀
Andrew has a good point.
That could be a factor, considering that wood dimensions vary perpendicular to the grain more than parallel to the grain, and that affects end grain joints in 2 directions, but side grain in only 1.
@@stefang1087 What you're not taking into account is that wood is a very strange material in that it swells with changes in humidity and temperature like most materials, but UNLIKE most materials it does so anisotropically. In other words, wood experiences much more movement perpendicular to its grain than it does parallel to its grain. The point that Andrew was making was not that humidity might have affected the test - because as you pointed out humidity would have affected all samples similarly. What Andrew is saying is that over time, as wood swells and contracts many times, this badly strains joints that don't respect this unique property of wood, and an end-grain to side-grain joint is a perfect example of this.
I do think the results of this test are valid under the limited scope of a 3" test sample that is tested very quickly after gluing. But I guarantee you that if you make a 3 foot wide dinner table and put a breadboard end onto it that doesn't account for wood movement, this glue joint will separate over time even without any external stress at all. The wood movement itself will destroy a long end-grain to side grain joint, regardless of the glue bonding strength. This is why gluing end-grain to side-grain is a bad idea - not because glue doesn't adhere well to end grain.
Been teaching woodworking for 7 years and feel I have some explaining to do for a generation of high school students. I am curious about the dimensional changes of wood with humidity and how that affects the geometric strain on joints. Really appreciate this!
I had too many endgrain joints fail after a few months to not do it that way. Wood expands and contracts from the center out.
That's why tables with breadboards split if glued. Chess boards split along the engrain if there's no spine or biscuit. There's a reason why we don't use them
People used to think the Sun went around the Earth... because that's what it looks like as the Sun rises and sets.
People used to think end grain to end grain gluing was weaker... because that's what it looked like when the joint failed at the glue line.
Things are not always as they look.
Patrick Sullivan has busted the end grain myth that so many of us have believed for decades.
Thankyou, I take my hat off to you for this outstanding piece of work.
I can honestly say, I never bought into this myth. I "normally" think in what I call "logical mechanical thinking". The side grain of whatever wood, doesn't have any "holes" or "crevices" to sink down into. End grain has tons of little holes to go into as anchor points, almost like a bunch of small dowels. So to me, logically the end to end, would be the strongest across the board. Glad to see someone actually made a video to back up what I already believed to be true. Thanks and good job!
Very cool information. Thanks for the effort. I hope it doesn't have an unwanted effect on people now thinking that wooden joinery is obsolete. It will still be necessary to get joints closer to the strength of the long grain.
right, he still highlighted that the joint is strongest based on the amount of long grain fiber that *crosses* the joint. i.e. basic joinery will always beat straight up glue.
That’s outstanding. And really fairly scientific in the approach. I like it.
Kinda throws out the whole notion of needing “sizing” to get extra glue into the end-grain after some has wicked away.
I think I might set up some similar tests as you did, if for no other reason than to corroborate your data for myself.
At any rate, you earned a new sub, and I’ll be sharing.
Thanks!
James
While sizing an end grain joint might work OK with hot hide glue, PVA doesn’t stick to cured PVA at all well. It strikes me as a particularly awful concept.
@@lightaces Sizing according to Titbond is giving the end grain surface a quick thin coat of 50% glue and 50% water. That surface is allowed to sit for 60 to 120 seconds and then a coat of normal glue is applied. The end grain is not coated with full strength glue and allowed to stiffen and then recoated. Again, according to Titebond.
@@1deerndingo well, that makes a lot more sense. Not relevant to my business (see my other post about glue strength not really being important in a well designed joint!), but there you go.
@@lightaces I think the concept of sizing on end grain joints is not so much to cause the two applications of glue to adhere to each other (because we know that doesn’t work), but to fill up or block the porous end fibers, so that when the full-strength glue is applied, it doesn’t all soak into those end grains. That’s my thinking about it, at least 😊
"notion of needing “sizing” to get extra glue into the end-grain after some has wicked away." This makes no sense. If the end grain wicks away the glue, then adding more glue would make sense. Having extra space there won't help.
As a guitar maker, I'm constantly "Breaking the rules," regarding how things are supposed to be glued together. As wood workers, we're constantly told that you don't glue up parts cross grain or end grain to end grain, and yet, in building a guitar, we constantly find ourselves doing those exact things, and that on a very thin, lightly constructed "box" that has to stand up to literally hundreds of pounds of pull or stress. As a result, I've often wondered how we get away with it. I think that, the fact that we keep our materials in humidity controlled rooms, work with those woods in temp and humidity controlled shops and also tell our customers not to expose the instrument to wild changes in heat and humidity might be the key and also possibly a variable that could effect the long term health of an end grain to end grain or cross grain glue up. The big thing in avoiding huge swings of humidity is that you are trying to limit the differing movement of the vairous parts, along their differing dimensions. When one part moves greatly in one direction in opposition to the piece it's glued to, you get breakage. Perhaps then, the rules only become apparent, when a glue joint is exposed to temp and humidity changes over years. I mention this, not as a critique of your methodology, only that there may be other variables that need to also be addressed to determine the long term veracity of "The rule."
and we pay CONSTANT attention to balancing forces!
Exactly! This controlled experiment does not introduce normal humidity changing conditions over a period of years. Most any joint holds until it does not. End grain joints have a shorter life span than edge grain joints.
Interesting. I work for a hand rail manufacturing company. We end glue and face glue pieces to make blanks for railing rated for indoor and outdoor use. One of our test pieces has been outside with no finish for 3 years at this point, with no sign of de-lamination of any of the joints. With temps ranging from +110 in the summer to -25 in the winter I have few worries about part failure at this point.
where are you gluing end grain to end grain in a guitar in an area that takes load? In fact, where are you doing it at all? The only place I can possibly think of is the neck scarf joint that I suppose classes as end-to-end, and is usually a typical failure point.
@@A1BASE right? I was thinking the same thing. I’m a backyard builder and refinisher and cannot think of anywhere end to end would be necessary.
I'm late to the party but would like to say that I thoroughly enjoyed this video. I wish there was an exercise like this when doing wood shop. I might have paid more attention to science class. When I was taking advanced woodworking classes in high school we was taught the most important thing in a joint is if it is clean smooth and square to its mating edge. This is a great illustration to show why you consider the direction of load on the joint when designing the project. Thank you so much for making this.
Excellent demonstration of the classic scientific method to prove that "common knowledge" can be wrong!
So much for believing opinions expressed by others on social networks like Facebook and Twitter. Unbiased examination is the only thing that can be trusted.
Fascinating! Excellent job putting this together!!
Putting this together 🙊
Super interesting results! I would love to see how time impacts the end grain results. One of the reasons we're taught to do things like reinforce miters is not just because of the possible myth that the joint is inherently weak, but also the fact that the joint is subject to a lot of stress as the pieces expand and contract over time. So I wonder what these results would look like in 1-2 yr old samples. I would also love to see these tests with longer boards. I imagine the amount of force required to break two 12" long boards apart at the joint is considerably less because of the additional leverage. Either way, thought-invoking stuff. Thanks for making the video.
Why would the leverage matter, we're measuring the force on the joint, not on the other end.
correct.
as wood grain primarily expands perpendicular to the long grain … end grain glue joint is likely to shift apart as wood moves.
For most products there is a way to accelerate testing. Putting test parts in a humidity chamber and cycling from seasonal high to seasonal low relative humidity and doing sat 50 cycles would work. Not sure how long a sample would have to be at a high or low humidity for it to fully penetrate the 3/4" thick piece.
I suspect Mark is correct and I have always assumed wood movement is why you are told not to glue end to side grain.
I think you hit it with the last comment. This showed a 3/4 by 3 inch end grain to end grain joint is stronger than the grain of a 3/4 inch by 3 inch long piece of wood. It tells you nothing about the relative strength of the joints. Glue end grain to a board 12 inches long (or 2 inches thick) and you would get very different results.
A friend has one of many glass etched mirrors I did forty years ago. It's not even mitered. Instead, I cut quarter circles to join straight pieces. The mirror weighs in at least 20 pounds and the frame shows zero indication of failure four decades in.
This is an Awesome video, and very well tested. I wondered for a long time why people said this joint was bad, but I did not have a good way to test it. Thank you so much for making this video.
It is still a bad joint, but definitely shows that glue is stronger than lignin!
@@daveb8598 Yes ........ and the earth is flat.
The reason might be because an endgrain joint dries much quicker. A dry joint wont bont nearly as well.
The end grain basically wicks up the glue to the point where there isnt enough glue on the surface anymore for a proper glue joint.
As soon as you showed that the glue was stronger than the lignin, I understood why mitre joints (especially with splines) are so strong.
Jason at bourbon moth woodworking made a video testing different types of joints, and was surprised that mitres were stronger than boxes and dovetails.
Thanks for challenging and expanding our knowledge!
I've never glued anything bigger than Christmas decoration or flatpack cupboards for extra strength and yet I've come across the myth that end to end glue joints are weak.
This video clearly demonstrates this to be false. Very well made and presented. Many thumbs up!
This is very interesting. Great to see someone with the proper testing equipment, knowledge, skill and curiosity necessary to bust a myth. Thank you Patrick!
Beautifully explained. In fact, just earlier today i was telling my 13 year old son that the end joint glue is bad idea. Well now i have to show him your video.... lol.
Thanks a lot Mr. Sullivan.
I mean, it's all relative to the design of the structure. If you are making a 'T' joint, you'll get a weaker glue-up if you butt the end-grain up against the "front" of the other piece instead of gluing it to the "side", long grain to long grain, assuming the surface area is the same. The long grain breaks in this video are akin to the board breaking feats in karate, where there's no real structural strength there to begin with. The end grain glue joint failed first in a configuration that allowed the full strength of the wood fibers to be brought to bear on the joint.
You weren't wrong, and this video is misleading. Please note, he never actually broke a long grain glue joint.
In truth, and design which relies on a 2.25 square inch glue joint to take any kind of force is a bad structural design.
When building traditional balsa framed model aircraft, side to end grain joints are frequent in the 'warren girder' structures employed. I've always been amazed at the resulting strength and rigidity, As a one-time mechanical engineer who has installed many strain gauges I really appreciate this video. Thanks.
Love to hear strain gauges mentioned in a woodworking context. My dad was an engineer and later VP for Micro-Measurements/Vishay. Worked for them 25 years selling strain gauges and helping customers with their applications.
Dear "Mr" Sulliavn, I think your video was perfectly clear as well as your conclusions. I am sorry for the people who understood something else than what you masterfully demonstrated and explained. I thank you for sharing.
Absolutely astonishing. This makes me so happy. The possibilities! So many things I've wanted to do over the years that I was always told would wind up failing, so I never bothered. I'm gonna do all of them now!
Thank you for this, Patrick. This is what's up.
I often wondered why everyone said that end grain glue joints were no good, since the wood absorbed the glue into the end of the grain like a straw and they all said that the wood would fail before the glue.
It would appear that the size of the cross section being glued is more important than the direction of the wood grain. But ultimately, the direction the load will be applied matters the most.
This was fascinating to watch. Thank you,
It was a very interesting test.
I think that the only drawback is that due to expansion and contraction of the wood the joint will fail eventually under any kind of stress.
I did this mistake with a mitered picture frames. After one year the frame bottom just fell apart, glass pushed it out and broke on the floor. This is where I started to look into woodworking joints and got really interesting in woodworking.
Best regards from Croatia
I have many of my picture frames in our summer home that sees high temperatures and high humidity in the summer and sub-zero cold and very low humidity in the winter. None have failed. MItered corners and no splines, dowels, or biscuits.
@@burntsider8457 Interesting, mine failed in Mexico city, glue used was Titebond 2 and I did it with pine. Maybe those are the factors.
I would absolutely love to see more testing like this. The more we understand about why joints work (or don't work) the better we can be at designing new projects. Great work!
Wonderful explication!!! I am an engineer and pedantic in mindset,, so when someone does a complicated experimental demonstration with absolutely faultless methodology and reasoning, I am impressed. Thank you for the education and amusement.
Absolutely great testing and reporting! And, no, not too many numbers! As an engineer (albeit electrical, audio, & broadcast) I appreciate the inclusion of numbers. Exposing more people to the numbers helps to get people used to seeing and dealing with them.
Paul-Joseph de Werk likely has the answer on why the myth developed.
The breaking of the solid unglued pieces was so satisfying, especially the walnut.
I was very interested in the video, but when he got to the solid pieces I sat closer to the edge of my seat thinking, "Ok. Now this is going to be really interesting."
Absolutely incredible! Thank you for this new evaluation of a very old issue. Thanks to you, I've learned something new after more than a half-century of cabinetmaking. Of course, I'm not going to give up my mortise and tenon joinery for simple right-angle butt joints, but I have a much better understanding of how the mortise and tenon uses the long-grain strength of the wood in two directions to create a super joint. Thank you for your outstanding scientific analysis.
Thank you, Mr. Sullivan. I learned almost as much about the characteristics of difference wood as I did about gluing them together. Priceless for a rookie woodworker like me. You even provoked me into my first RUclips Comment ever!
First!
Excellent! As a former test engineer, this was exactly the correct approach.... Two additions for your next test.... 1 - Rotate the joints 90 degrees vertically in the press. this would more accurately simulate a mitered corner, or just miter one. 2- do the unglued wood rotated 90 degrees horizontally. This will test the lignin only and should be close to the side to side grain tests.
Thanks!
People like you help humanity move forward and advance. A very scientific approach and very well demonstrated. Thank you Mr. Sullivan.
Hi Patrick. Great video with some really interesting results. I think in the future you might want to use a 4 point bend and not a 3 point bend. A 3 point bend places all the stress directly on the joint. A 4 point bend places uniform stress between the two rollers. The effect being that a 3 point bend is used to test a material which is homogeneous whereas a 4 point bend tests 'the structure' between the rollers. Example being: if you want to test bending strength of raw timber a 3 point bend would be fine. If however you wanted to test something like the bending strength of riveted metal truss a 4 point bend would be better. A 4 point bend is also much more forgiving of alignment errors.
Actually this just shows what we’ve always known - that glue joints are stronger than wood in side grain joints, and that wood is stronger than glue in end grain joints. This experiment never broke a side grain joint, because the wood failed first - so we still don’t know how strong that joint is. The end grain joint was the only glue joint to break - otherwise we’re only seeing the comparative low shear strength of wood along the fibers, compared to across the fibers, which we knew anyway.
Good point. Thank you for sharing. Have a blessed day.
Yes, this.
This is the definitive take-away. Make those boards (end-glued) twice as long, and move the point of leverage to the ends and test it again. I would imagine the results are strikingly different.
Exactly, this whole, lengthy presentation is an example of cognitive bias.
@@djlusid And why we would only test the end-grain glued joints with longer leverage? If you did all tests with longer boards, still the results would be the same (not the absolute values of force, but the comparsion between end-grain and long-grain), because why would it change?
Very interesting ! I think that the myth also comes from craftsman seeing the glue fail in endgrain to endgrain led to assume the joint was weaker. On the other hand, if the joint break because the wood fails, the joint itself doesn't look like the issue.
Another point, it's hard to get a big surface area in end grain to end grain joint.
Thats what I thougth also
That was my thought as well. If I saw those end-to-end breaks in the wild, my reaction would probably be, "wow, that glue joint sucked" not "wow, the wood was extra strong in that orientation."
That sounds like a classic case of confirmation bias.
"I Have achieved a result that looks so obvious to any other man or woman, that I need not look deeper."
In fairness, we don't know how strong the glue joint is if the wood keeps failing first when we try and test it. Maybe end grain is still the weakest, it just doesn't matter in real life because the wood is the real limiting factor
Excellent examination and one that many (most?) traditionally trained woodworkers understood but has been lost to "modern wood machinist" hearsay...As a working Timberwright and someone that only works in "green wood" for most projects, it is refreshing to find someone that is dispelling myths and/or being properly accurate in the description of what is taking place in a given wood joint application...be it a simple glue butting joint of some form or otherwise...I look forward to following your videos from now on...Thanks for the great documentation and explanations...
So really, this is not comparing the relative strengths of the joints, but rather the relative strengths of different wood grain orientation combinations. For end-to-end the difference is that the wood grain is in the strongest possible orientation for both pieces, so the wood glue then becomes the weakest link since there is no lignin boundary to fail first. Great video!
Well done. Very well thought out and very well worded description. If you ever want to do collaboration let me know.
Yes please!
You should definitely give this video a boost. This should you RUclips woodworking viral as much as the Veritasium click bait video.
Patrick's channel is great but he falls through the RUclips cracks because of the frequency.
I think it wise to approach these things cautiously. This is either the biggest upset to the woodworking status quo in decades, or there's a delta at play here yet to be uncovered.
I'd like to see a third party involved, for peer review if nothing else, but I think maybe a dialog should be opened WITH the makers of TiteBond themselves... I could swear I've seen tests like this in my civil engineering/material sciences class work, with the opposite results. I feel like more study is needed for this to get gospel status.
@@grumblycurmudgeon What I think is key here is that beyond just "Test it and see" - we have a solid explanation as to why this is the case.
Every woodworker and guy who's tried to chop firewood knows that it's a hell of a lot easier to split wood along the grain than across it - this knowledge probably goes back to the time of stone tools. The knowledge that all modern woodworkers have that "PVA glue is stronger than (side-glued) wood" is also well known and understood. The conclusion here is that for any glue which is stronger than lignin - the outcome found in this video should be fairly obvious. (Although, I readily admit that I was surprised...but now I think about it...I shouldn't have been).
I strongly suspect that this oft-repeated-falsehood about end-grain gluing came about in an era when glue was weaker than lignin - in which case side-grain gluing would indeed have been better (or at least no worse) than end-grain. In all likelyhood, this all goes back to victorian times when the best glues were pretty pathetic by modern standards.
So many engineering questions, practices and dogmas were accepted in Victorian times - and never really questioned since then.
As an example - I learned that an awful lot of the parts of my (Tesla) car are either glued or stuck together with double-sided tape! My first reaction was that of horror - I expected welding and big-assed bolts...but it turns out that modern automotive adhesives are actually stronger!
This eats at our gut-feel engineering skills - and even as I write those words about my car - it's hard to suppress a feeling of deep unease about it.
This is why humanity invented science - and why using it's methods teaches us things that our poor evolved monkey-brains are ill-equipped to decide by themselves.
So, sure - independent testing and peer-review are important - but no matter what - this single set of experiments has to throw doubt on the the old dogma which does not appear to have any testing at all (at least not with modern glues).
There are new questions here though. For example, if the developers of modern glues accepted that end-grain gluing is good - perhaps they'd develop new glues that are optimized for the end-grain situation?
I find these tests great for the general info application. How many woodworkers build projects that will face these types of pressures? An engineer building bridges or towers might use this data. How long will the end grain butt joint last over time without cracking or breaking unaided by a mechanical joint? There is a reason why tradition is a timeless guide
I guess that is why the scarf joint rocks for joining long, thin sections expecting to carry a great load. Combines end-to-end glue joint and natural longitudinal strength.
A scarf joint, especially a long scarf joint, of perhaps 1 in 20 slope works well because it has a very large surface area.
Great video. This is like a breath of fresh air after recently watching a series of 30 seconds long videos, where a certain youtuber "demonstrated" that glue is stronger than the wood, and the amount of thought and effort he put into making those videos equaled their length.
The glue is still stronger than the wood in most situations. Makes nails and screws seem almost obsolete in most settings unless you want to be able to disassemble!
Well, it was just 2 videos. Not really a series.
@@SteveRamsey lol
As a mechanical engineer, I can greatly appreciate your analysis done in the proper scientific method as opposed to the common "jumping on the end of a board" methods.
Very well done on all aspects: content, presentation, understandable, videography, ...
GREAT JOB!!!!
Engineering also has its myths; usually started when some eminent authority first mentioned it in his textbook, then other authors would then simply copy it into their textbooks unchecked giving the myth more weight and "truthfulness".
A proper fantastic job!!!!! As someone with a mechanical engineering background, you made me feel like a fool for falling for this myth all these years!
Blew my mind! Got into proper woodworking rather recently, and after absorbing all the knowledge I could find this was the one message that was taught by everyone. I hope you manage to get this video viral, would love to see the community respond! And hopefully get the message changed from now on.
Very interesting, thanks!
When end grain to end grain fails it always fails at the glued section which might have created the myth since people usually do not measure the force needed to break it.
Ya, bit of selection bias kinda like when people say (thing) from long ago was built better because look how long it lasts.
No, you are just looking at the survivors.
I wouldn't have seen this video if Stumpy Nubs hadn't "responded" (moreso his scare-quotes than mine) to it. I agree with him that you've buried the lead in your conclusion. I would have a different specific complaint if I hadn't seen his video. It's obvious that you've put a lot of good work into this video and I look forward to the full series.
Wow! I am new to woodworking but cannot count the number of times this myth was suggested to me as truth. Thanks
For the tutorial. Very helpful in clarifying the topic. Blessings!
Excellent experimental design. Excellent work eliminating variables and testing only one independent variable at a time. You have done the work of an experimental scientist.
Because there are so many comments I can't spend days or weeks reading them all so I don't know if this has been brought up yet, but I would love to see different glues tested. Because I feel the myth about end grain joints stems largely from the more traditional hide glue used in the 1700s and1800s, or even into the 20th century, Testing various types of glue such as epoxy, hide, PVA, etc might shed more light on the myth.
Very well done tests and thorough results! definitely continue to create videos!
What an incredibly well thought out test. The fact you dared question age old "knowledge" is amazing in and of itself. Thank you for thinking outside the box to bring this exceptional exception to misunderstood understanding
More people should do so. It would save a lot of trouble in many many area's.
You'd be amazed how much "exceptions" there are. Truths that have been told that sound good enough, but actually presents so much more nuances and half-lies. I'd like to call it science by consensus versus science by facts. The first being science repeated and repeated over and over again, scientists circle pointing to each others findings. Using models, graphs and predictions to calculate the truth. But to actually go out there and actually perform science to verify those truth theories is seldomly done.
@@patrickd9551 is "seldomly done"... and was not done here.
Huge thanks for that research, puts everything in perspective. If you are doing any more tests I'd be very interested to see a test on scarf joints using the same width and thickness timber to understand comparative strength. I think in boat building they use scarfs as short as 6:1 or 8:1 and in aircraft they go up to around 15:1.
This is the best testing procedure I have seen. EXCELLENT sir. The test results and reporting is outstanding. I wish most if not all RUclipss could learn from you. You have set the standard........
This was awesome and it completely destroyed what I had always believed about end grain to end grain glue joints. I’ve been woodworking for over 30 years now. Well done science, can’t argue with fact over fiction/word of mouth.
You’re back! Looove your content. My suggestion for a follow up - compare end-to-end glue strength with regular application vs massaging the glue in until completely saturated. I do not recommend letting it dry or applying a watered down first application like some people do. Just massage the glue and press down hard to push the glue into the end grain, and keep applying until it stops absorbing and stays wet. I do this every time (takes much longer) thinking it must perform better, but maybe it’s a myth too.
Late to the party, but awesome video. I think application needs to be considered as well. I can definitely see applications where end to end would yield a joint that will last a lifetime. Table tops, picture/mirror frames, and the like. Where racking force comes into play, I don't think any of us would consider end to end, anyway. You've made me a believer, and I can see where this will make me rethink my approach to gluing up panels, and frames. Thanks!
Be careful though. I just had an end grain joint fail. Lasted less than a year, not a lifetime. I will from now on make joints properly and not rely on glued end grain. There wasn't even that much stress on this joint. Get long grain crossing the joint and preferably design things so that if the glue failed there would still be interlocking wood to hold it together (like tenons).
That was totally fascinating, your methodology was so much more thorough and well thought through than other videos I've seen that claim to cover the same subject. I've subscribed and will look at what else you've shared ! Thanks for providing this test.😊
Well done Patrick! I am a retired structural engineers, builder, and wood worker. I just learned something. Cheers from Canada.
Someone just passed this video to me. It's exceptionally well done, Patrick. I thank you for challenging the tribal 'knowledge' and doing so with such scientific rigor. And I'm happy to unlearn what I have learned.
I don’t think these results change the conventional wisdom since the natural alternative to an end grain to end grain joint is an unbroken continuous long grain piece which your testing shows to be many times stronger. I would love to see this rigor applied to scarf joints and maybe look at the relationship of glue surface area and strength.
Excellent video. Made me think and I definitely learned something new.
@john milton and you were still able to be a dick in your reply... The purpose of the video wasn't to say glued up end joints are the way to go for anything, it was to test if the long held assertion was true or not, and he proved it was a myth. Job well done.
@John Milton, your point is that repeating what you take for granted without trying it yourself is better than showing respect for a very careful science experiment. So what is theory and what is practice?
@@alamaralaa He didn't prove it though. He proved that end to end joints are stronger than wood across the grain, or lignin as he refers to it. He didn't prove that end to end joints are stronger than side to side joints, because the side to side joints never failed. But the fact that they didn't fail in his experiment doesn't mean that they won't fail in the real world. In the real world you don't put significant stress on wood across the grain. Nobody has ever made a table where the grain doesn't run along the length. But in the real world you have years or decades of wood flexing, temperature and moisture. How will that impact the joint? Hard to test, but that's what we have decades or even generations of experience for. And those tell us: Don't glue end grain.
@@cire9984 He was wasn't trying to prove the gluing end grain was stronger than gluing long grain, and I never said he was. His assertion was that gluing end grain was not as weak as people assumed it was, and his tests prove it is a strong joint. Stronger in fact than the wood is strong across the grain. Period. End of discussion. Please don't put words in his or my mouth. You can misunderstand his tests and conclusions and my statement above all you want, but don't tell us we're wrong because you didn't understand them.
The myth could have started back before PVAs were available. Try it using hide or fish glue and see if it holds true. Thanks for an incredible video. I'll be updating some of my designs now :-)
Hide glue is significantly much stronger than any PVA glue though.....
I wonder if the problem is expansion and contraction over time. That would affect end grain quite a bit. Not something that would show up in this lab test though. Unless the wood went through sever cycles of damp and dry and was then tested. There might be truth to the old theory that end grain glue joints are suspect.
@@orion7741 It is, but hide and fish glues are both more susceptible to heat and moisture and deteriorate faster, leading to eventual failure. Besides that, expansion and contraction has taught us to combat it in other ways where needed, and it's not a major factor with a glue joint, as most glues are sealants too, and prevent contraction/expansion of the joint as an added advantage. One shouldn't even use materials or methods for any job where better ones will do a better job, unless you are planning for obsolescence.
Oh, and placing a "much" after "significantly" is redundant and either would stand alone, as in this case they mean nearly same thing and one or the other will suffice, and using both is just strange.
@@Bob-of-Zoid It's not necessarily redundant. They might have meant that the fact is significant, but the glue is much stronger
Since you have the test apparatus, I’d love to see a comparison of different joining techniques (dovetail, biscuit, etc.) on end to end joints.
There is an endless amount of mythology in the usage of so many materials in all the trades. Your information was perfectly presented, thank you!
Dear Sir, please let me send you a big thank you for the work you show us here. Your video should be in carpentry schools! Best regards from France
Asking questions where everybody sees only well-known answers is really something!
The Oak results surprised me. I'd have thought those big pores would make the oak end-grain joint come out on top, but it was beat out by maple, walnut, and cherry. I wonder if the glue viscosity plays a role there: would a thinned-out glue, that can wick deeper into the pores by capillary action, perform better in oak?
And I hope Mattias sees this video.
Some people advocate allowing the glue a few minutes to absorb into the end grain, reapply what was absorbed then connect and clamp. I've tried this with pine and noted that the glue obviously did absorb into the end grain, so your suggestion makes a great deal of sense, perhaps along with using Tightbond III which has a longer open time and seems runnier anyway.
This is called sizing the joint, and it is recommended practice if you have to glue end-grain together to prevent glue starvation.
@@tomhargreaves8820 i know the video u talking abott
@@nickgreen5982 You definitely do this when building balsa aircraft open framed structures; seems to work OK !
"And I hope Mattias sees this video.
"
I was thinking the same. I remember his glue up tests years ago. I will have to watch again to see his results, but I think they indicated end-grain joints were weak.
It's good to be grounded in reality than say-so. This was particularly good info and will influence future designs for me. Thank you.
Please blow up some more myths!
I wondered who paid for all the equipment, but I know I benefited. Amazing photography, organization, and presentation in English a schoolmarm would love! Great!!
Mr. Sullivan,
An amazing and accurate design of experiment (DOE). I hope it is appreciated. You not only characterized and quantified the experiment but also continued into the why of the results. Showing both the adhesive failures and cohesive failures led to my concluding that one of the basis of this myth is that any failure of the joint was caused by the glue bond (adhesive failure - the glue alone) and not the interaction of the glue with the wood grain (a cohesive failure). You showed this very clearly in the stress testing and in the micro pictures of the failures that it was the wood that failed. The joint broke in every species case due to the failure of the "grain" adhesive - lignin - and not the PVA.
To the casual woodworker (me), when that joint broke, the blame would be placed on the adhesive, not the wood. I can now stop using mechanical fasteners (a "just in case" measure) on my glued end joints. I am a victim of this myth and you have "set me free".
Bravo, and thank you.
I would be interested to see this tested with pulling force instead of shear force as wel. I would expect similar results, but I'm not sure.
Thanks for your rigorous testing!
It's not shear force. When you apply a deflexion force to the beam, you have one side of the beam where tension (pulling) forces are experienced, and on the other side compression forces. Check out beam theory on Wikipedia or that video ruclips.net/video/f08Y39UiC-o/видео.html
But in fact, the situation isn't exactly the same, since there you have some shear stress, where on a pure tension test, you only have tension stress.
Patrick, great job on this video! Definitely super interesting and makes me think twice about gluing butt joints with no reinforcement with mechanical fasteners. EDIT: P.S. you should be able to adjust your camera's shutter speed to largely eliminate the flickering from your shop lights.
Great methodology and great presentation of results. However, your results beget a further question. Namely, what effect does conditioning of the bonded surfaces have, if any, on the strength of the bond? For example, sanding of the surfaces to be bonded with various grades of sandpaper from rough to fine, or cutting of the wood samples using a cross-cut versus a ripping blade.
Very true, I came to this video wanting the answer to 'planned or sawn?' I've seen videos that prove, the more glue the stronger, so a sawn joint with room for more glue and texture to grad to as said in this video would logically be even stronger. I would love to see part 2.
Great wood glue study, perhaps the best I have ever seen. Very helpful for woodworkers who strive for perfection with their work.
Outstanding production, enlightening, grateful for the watch. Go Patrick!!!!
As an untrained inexperienced amateur, I genuinely had no idea that there was a “right” way & a “wrong” way of gluing lumps of dead tree carcass together.
Love ya work. Subscribed. Hi from Oz 🇦🇺
Patrick, you misunderstand the end grain rule:
1. If you glue side grain, your glued joint is ALWAYS STRONGER than the wood, so use glue, its OK!
2. If you glue end grain, your glued joint is ALWAYS WEAKER than the wood. You can still glue it, but only if you do not need the full wood strength!
I feel you should have mentioned that, because people think it is OK to glue the end grain now in any case! By making a lap joint, finger joint or dove tail
the strength is increase dramatically. Maybe you should make a video about that, since you got nice equipment for testing!
This makes a lot of sense. Even though it took much greater force to break the end grain, the point of failure is at the joint itself (the glue), making it SEEM as though it is a weaker joint rather than a failure of material strength. False conclusions that seem so apparent. Great video!
that was my thought as well.
1. I am really impressed by your methodology and attention to detail.
2. I am shocked by the conclusion about end grain glue joints.
3. Other than setting to rest a myth, what would possess you to undertake this tough of a project?
4. If you did this project as your PhD thesis, congratulations, Dr. Sullivan; you passed!
Just found your channel. I realize this video is 2 yrs old. Hope you get it. Very well done. I am impressed by the way it was done. I've been doing this since 1972. I've learned a few tricks a long the way. I always said that the only reason why glue ups fail is because of bad prep and cure time. You just proved my point. The only real joints that break are the ones put together wrong or abused after final builds. Thank you. I look forward to more videos.😊
Absolutely fascinating as a woodworker. Even more so as an engineer in observing the yield point behavior in the various orientations and different species. I found the failure modes of the unglued samples particularly interesting. The cherry and the walnut, especially the walnut, failed catastrophically as the grain failed in almost pure tension, the cherry started splintering, at the surface as the lignin apparently failed in shear, allowing the load on the grain to increase rapidly, then failed catastrophically as the grain failed in tension like the walnut. All the others appeared to fail by splintering, perhaps with the lignin failing in shear with the grain breaks staggered. I wonder what that might mean to people selecting wood for projects that require bending strength.
I think that depends on application of bending. A bow, for example - requires fast recovery response of the fiber to maximize the energy out. A chair however, will be okay with a slower recovery time of the fibers. Softer woods like pine do not make good bows since the recovery time of the fibers is very slow, but the slow recovery time seems to make them more bendable.
Very interesting video. It turns out, human error was giving end grain a bad rep
No, it gives humans a bad name for making the error to trust error prone humans (AKA believing), instead of following the scientific method that relies on more closely looking at the details, investigation, and following the factual data wherever it leads, and putting their biases aside even if the data points away from their beliefs.
While attending Arrowmont recently, my instructor told me that Titebond 2 was specifically formulated for end grain joints. She learned this during a phonecall with titebond reps.
That is so cool I had no idea!
As someone who taught for years in an unrelated technical field I just want to say this is an excellent video lecture.
It's 03:30 (3am), I'm in Sweden, I'm watching a lecture about wood, grain, glue... and I'm quite fascinated. Cheers!
Great video! You are a very inteligent comunicator. Would be interesting to know which bond is the strongest, side by side or endgrain to endgrain. Becuase the wood grain fails in the side on side example before the glue line does, is difficult to know how strong it is...
Not sure the strength of the bond really matters if the wood fails first
@@veri745 if you use a weak material to test a bonding agent that is used on stronger materials and the weak material fails, you have not properly tested the bonding agent
@@mjbailey404 the only "stronger materials" I see available is end-grain, so what exactly has gone un-tested? He covered a pretty reasonable array of wood species
@@mjbailey404 The bonding agent is static in all tests. Regardless, the glue isn't being analyzed, it's the application.
A couple of things I'd like to see tested is smoothness of end grain and old fashioned hide glue. One thought on end grain is without precision cutting or planning the surface, you wind up with a surface that is very rough. So it may not have as much contact as a very smooth joint, esp from the days of hand tools. Same with hide glue. Would it perhaps be pulled through capillary action away from the end joint so there was simply less glue acting? Just some thoughts that some of these beliefs may have started when we used different tools and glues that today.
On a rougher end grain, the two wood surfaces may not have as much contact with each other, but they'd have more surface area for the glue to grab. I think a rougher cut, within reason, would probably result in a stronger joint than a smoothly cut, planed, or sanded end grain.
Hand tools can get a much smoother cut than your table saw or router depending on what, where and how. I'd wager a scraped or planed surface is smoother than a sanded one.
I think this is an improper framing - though granted interesting data. The issue with end grain joints is that it creates a weaker seam than the wood itself, meaning the joint weakens that segment of the piece and becomes the weakest part of the overall piece. Wood being stronger on the grain direction lets you build to the strength of the grain, but end grain joints create a weak point in that direction! Whereas long grain joints simply keep the strength pattern of the wood consistent.
The strength of the joint itself is not stronger in end grain, rather an end grain joint is stronger than the long grain wood itself. It's a big difference, and personally I don't want the joint to be the weakest part of anything I make.
WOW this video kinda blew my mind! Totally shattered all I had been told (and believed) about end grain to end grain joints. Thanks for making this video it will transform the way I think about my woodworking from now on.
Awesome testing data. For years I have always used either T & G, dowels or biscuits for either side grain or end grain glue ups. Mostly for alignment but also for strength. Your testing shows I may have been over engineering my joints. LOL. Great work.
Perhaps this myth is so pervasive because of survivors bias. "See kids? That end-grain joint broke right along the glue line! That's why we should never glue like that."
that is exactly right. the misconception being that joints where the wood fails are 'normal', whereas joints where the glue fails must mean the glue is weaker than the wood.
This is looking at the glued joints from the wrong perspective. What should be compared is the strength of a glued piece compared to the strength of a continuous piece. In a side-to-side joined piece the strength is no less than a similarly sized continuous piece where the wood is the limiting factor. However in an end-to-end joined piece the strength is 3 to 8 times weaker than an equivalent continuous piece because the glue fails first. Glued end-to-end joints are bad practice because it is so much weaker than a piece without that joint.
Ok, so hear me out: if you care about the strength of the piece you are making you are going to make sure that the forces that work against the piece will be applied across the grain, not on the lignin. If you have 2 pieces of wood, one 1x1x10 made by 2 0.5x1x10 glued side to side, it will be stronger than a 1x1x10 piece of wood made by gluing 2 1x1x5 pieces of wood end to end, and that is in any direction of the wood (since the glue is stronger than the lignin, assuming the glue is in the middle). Basically: you should try to avoid gluing end to end and instead use larger pieces of wood, but when it comes to gluing side to side it doesn't really matter since the wood strength is not lost.
What a wonderful analysis of glue strength! Should be a chapter in a woodworking textbook.
I tested this myself years ago and was just as surprised. I thought the glue tech must have improved since the myth was established. Your investigation clears this up very nicely!