You'll get a linear relationship as glue surface increases proportionally to length. Then there will be a hard limit at the breaking strength of the dowel, which depends on its diameter.
Is there such thing as a spiral fluted dowel? Essentially a screw-looking dowel, where insertion is still done via press-fit just like all other dowels, rather than needing to screw them into place. Also the spiral flute wouldn’t have a tight screw pattern, but rather an elongated spiral, otherwise failure would be too easy.
Thank you, Matthias! I use quite a few bamboo dowels because I have them and they are cheap. One thing I have noticed, some of them have a burnished almost waxy surface feel to them. I don't know if it is a substance in the bamboo or a result of the dowel-making process or both, but I feel it may hinder the glue from fully penetrating the surface of the dowel, on those, I give the full-length dowel a light sanding with 220 paper before cutting them to length, I don't know if this helps though.
That's part of the production process: It's when the cutter gets fairly hot that the wood may loose a bit of resin in the process (like it's cooking off). You can sometimes see this if you cut wood as well, then you get a glossy very flat surface instead of a normal cut (and shortly after you see burn marks).
@@RubenKelevra Thank You, that is much how I imagined it happening. Some parts of the bamboo may have less resin which would explain why some have a dry woody-like appearance, whereas some others containing more resin would have the burnished areas with a slight sheen. That being the case perhaps using a solvent would be a better prep for glueing than sanding?
I would like to know, how much would putting a suitable sized washer beneath the screw head prevent the screw head from sinking in and such increasing the strength of the joint. The result of end grain versus cross grain was indeed somewhat surprising.
I used four inch screws driven through the stiles into the rails as my first attempt at fixing some old windows. I used the washer under the screw technique and I was sure that it was better than using just a screw. I abandoned the screw for the dowel technique because I thought the dowel was a better long term fix. Even with out the issue of the screw head working like a wedge there is still the issue of the screw damaging the wood over time because racking from the wind load.
@@davefoc Main weakness of dowels is that they expand and then contract. Not an issue when the force is generated across the joint but becomes an issue then the force pulls the joint.
Great video. It would be interesting to see a "stress-strain", actually force-displacement, diagram for each joint. Then you could get a sense of the failure mode of each, at least from a brittle/plastic mode.
I wonder how 2 or 3 bamboo skewers would compare to one 3/8 dowel pin, if you space the skewers tightly so they do not take up a lot more volume than the dowel pins.
I have been using small bamboo skewers - the ones you can get at a grocery store - as dowels in small projects for a long time. I find them useful especially when joining thin pieces of wood side by side. They work well for me. But I don't have all of the quantitative test equipment that you do, so I have no actual data to share.
Moisture cycling should hurt the strength of crossgrain screws more than that of endgrain screws, since expansion and contraction should stress some threads more than others. Into endgrain screws can be as long as you like. Hoadley specifies half the pullout strength per length of screw into endgrain vs sidegrain. I wonder if the difference with your results is hardwood vs softwood.
This makes me wonder about angles! Toenailing in actual nails at 30°. Or even skew(er)ing the skewers and dowels at angles so they are not in-line with the pulling force.
I'd imagine the fluting adds a good amount of surface area without increasing the radius, which seems especially important bc the dowels seem to pull out rather than fail themselves
My theory is they allow glue to remain during installation, preventing the joint from becoming glue-starved; either way, it shouldn't be too hard to just run a screw up and down a dowel to add some fluting.
My thought was that the fluting allows for a tight surface contact for gluing. The high spots of the fluting are able to be slightly larger diameter than the hole, and deform easily to fit the hole. For a solid dowel, you need to be a lot more precise with the hole size or else the dowel either won't be very snug or won't fit.
@@joelhollingsworth2374 yeah i agree with this assessment. there's just not much glue left in the mating surface area after pushing a standard dowel in.
Perhaps it's time to buy a clip-on displacement gauge allowing you to plot force vs displacement. It would be nice to track the drops in force and jumps in displacement, you can then associate them with what you see and hear during your tests. Some other nice data to be had from such graphs, work done etc. A bit of extra nerdy fun.
Would be interesting to see if the relationship between screw strength in end gain vs cross grain stays about the same across different types of wood. Also would be interesting to quantify that comparison in plywood since I'd expect screwing into the side of plywood to be significantly weaker than the other direction.
My experience with repairing wood windows: 1. Two inch long fluted dowels about thirty years old. Mostly the glue from 1976 had failed long before I was working on the windows. 2. Miscellaneous kludge screws to try to fix the failed dowel joints: The dry wall screws were rusted through (the windows were at the ocean) and the zinc plated screws didn't do much better. I also thought that there was a failure mode where the rail/stile joint failed from the screw working against the wood over time from wind loading. 3. 7/16 inch dowels through the stiles into the rails from the outside with titebond glue: As near as I could tell this was a permanent fix in the sense that something else was probably going to fail in the window before the rail/stile joint. I saw no reason for the fanciness of hiding the dowels inside the joint I just drilled a hole through the stile into the rail with the window clamped together. For awhile I was concerned that I wasn't using a fluted dowel. I doubt that was an issue. I put glue on the dowel and into the hole (I used an acid brush to make sure the inside of the dowel hole was thoroughly coated). As an aside I thought about doing the kind of testing that was done in this video. I never did but I thought the testing in this video vindicated my choice of window repair method. One thing though, I could just pull the old windows apart because the original joints had failed. I don't think the guy who tries to repair the windows I fixed will be able to tear the joints apart in 30 years.
An additional parameter of interest could be the rigidity of each joint, which, depending on the application, might be even more important than the tensile strength.
I would’ve liked to have seen you test the pocket hole screws in the same orientation as the other wood screws, as I don’t think the heads would’ve pulled in as far and maybe held just as well or better.
That was my first thought as well. Matthias speculated on the wedge head of the regular screws. The pocket screws have a flat head where it contacts the wood. I wonder if the flat head would avoid the sub-200 joint separation?
Some time ago I switched to using 3/8 inch dowels for shop furniture joinery (benches and such). I don’t find screws any easier to use than dowels. Just clamp up your assembly, drill your holes, pound in your dowels, trim them down when the glue is dry. Simple. And apparently plenty strong. The practice of putting screws deep into counterbores and then covering the heads with dowels for cosmetic purposes mystifies me: why not just use dowels in the first place?
The instant you have visual separation could be considered the yield point. In the case of the screw heads the wood beneath the seat is yielding to the load. Up to that point the wood is probably elastic, after that point the screw hole undergoes plastic deformation. Interestingly, a lot of people think in terms of ultimate strength when the component is often no longer useful after yield. The beauty of ultimate strength is that it provides a little margin after yield and before ultimate is achieved. That provides an opportunity to avoid sudden catastrophic failure. Interesting set of tests.
Interesting results. I like using dowels instead of screws, but they take a little more work to use for joining, and I actually assumed screws were stronger. I wonder what the results would be with a more sheering force on the fastener; turning your test pieces sideways and applying opposing forces on each piece being joined.
This is a discussion that comes up all the time in these tests, and it depends. This is more like the top connection where a seat meets the back of a chair, so it has relevance. As this test showed, stronger connectors (screws) don't always mean stronger joints if they tear through the substrate, which is why the strongest joints are the ones that distribute the forces the most (dowels, box joints, tenons, etc.) That holds for both tension and shear. Given the results, I'm now curious about a "pocket screw" with the wide, flat washer head recessed slightly into the surface but driven up into the endgrain like the other screws. They tend to be skinny so I don't think I'd trust the shear strength, but it would reduce the splitting that caused these screws to fail. Also, he could have run a couple of dowels, skewers, or splines across his samples to reinforce the cheeks, bit that's a note for the next tester.
Good design would minimize sideways forces. For example shelving would be housed in a dado and the screw could keep the shelf in the dado. You would never design a piece where a sideways force on the screw itself could be the failure point. (and by "never" I mean someone is going to do it, but it shouldn't be done).
Back in the old days, when we were adding a wall to a concrete slab (such as when changing an open carport into an enclosed garage or additional family room), we would use a hand chisel type star drill to put a hole in the concrete and then pound a peg of wood into the hole. We would then drill pilot hole into it and use that the mark the location of the hole for the 2x4 baseplate for the wall. Then we would use a lag screw to attach the baseplate to the peg and thus to the concrete. Later on, we started using lead and/or zinc inserts into the holes, but still lag screws. It would be interesting to see how the strength of those old solutions compare to the modern concrete fasteners.
All these years of watching your videos, my main takeaway is to not use pocket holes. My biggest reason for not get the setup is that it costs too much. Looks like there are better ways regardless. I'm finally starting to actually build stuff, so I'll be getting a dowel jig first. It's cheaper anyways...
I bought both, the pocket and dowel jig at the same time. I used both jigs and can't really say bad things about either. But for quick finishing, the dowel wins, because it is hidden. I made some solid wood pelmets, used dowels and I am super impressed by the outcome.
This is probably the absolute worst scenario to use pocket hole screws in. Their best benefit is that they're quick and flexible. Also, I regret buying a cheap dowel jig
This would be interesting to see with different types of screws. The wedge shape of many seems like a point of potential compromise to me, so, in appropriate applications, like butt joints, i use countersunk cabinet screws, which have an integral washer. While my preference is generally dowel joints, i am convinced that an appropriate screw style can be technically superior in some applications.
Being no engineer, I can't comment on your test methodology, but based on videos of yours I watched a long time ago, I have been using wood-dowel joinery on most of my simple projects, and these have all held up way better than I would have expected, given my clumsy techniques and the crappy pallet or salvage wood I tend to use. This video only reinforces what you have been showing your viewers for ages. One obvious point: Few people make their own screws, but anybody can make their own dowels, using scraps that normally would just be thrown away.
He tested the force to pull out a screw from endgrain at the end, pretty much the same thing that using a washer. So It'll a bit stronger than the other possibilities (around 250kg in his set up).
Just that how big the washer has to be, so it outperforms the threads in the end grain. And also when you bore a bigger hole for the washer+head how big is the negative impact?
I love dowels. They provide great strength in any direction. Any screw with perpendicular force is so weak compared to a dowel. Dowels are great for alignment also. The only problem is jigs for aligning dowel positions is not that great. Someone needs to come up with a great doweling jig!
Could you do a follo-up short testing treaded insert + bolt? I use them a lot for easily dis/reassemble-able joints, but i honestly have no idea how strong of a joint that makes... Get the feeling it'll be stronger than the screws though; seeing as they tended to fail by pulling the head through rather than at the threads. Should be interesting regardless
Very interesting, I wish the pocket holes would have held up better. I'm curious as to whether screws plus glue(or dowels) is a stronger joint than adding the strength of each one separately. Synergy?
For screws in endgrain you'll have brittle failure, whilst you'll have more ductile failure in cross grain (as shown in the experiment). This coupled with potential cyclic loading, variation in moisture and cracks developing over time along the grain. I would be careful with utilising them more than 50% of your numbers for critical connections. Especially for short screws.
Would be interesting to test two screws at different angles, that way pulling out doesn't rely on the screw heads as much. And pilot holes vs no pilot holes.
hey Matt, Workshop Companion on RU-vid showed how to make a "dowel pocket hole joint" where they used a long brad point bit and drove in a dowel, rather than driving in a pocket screw. Any chance you could test those out? I wonder how well they'll hold up vs dowels or pocket holes.
I love these tests... lots or woodworking (and construction for that matter) is based on folklore... I often wonder which type of force is more typical on furniture fails... sheer, torque, pullout, etc...
Did you try screws with glue? What glue and metal combo would work best I wonder? In the end, these tests often ignore the importance of what the application actually calls for, but they are still fun and informative.
Would be interesting to plot the force curve over time for each sample, and then maybe overlay them all together. That would help show the different failure modes. Not sure if you’re logging that data.
I've seen a lot of these ultimate joint strength videos recently on RU-vid but I really question the validity. What I think is more is more relevant to furniture is cyclical loading, or fatigue of the joints. For example, in this test it showed that the screws especially started separating before the joint ultimately failed. Imagine you made a small shaker style side table with no stretchers between the legs- and you only had the top rails directly under the table top. The normal load this will see under daily use is if it's bumped, and the table slides across the floor. This would be especially bad on carpet where the narrow feet sink in. The loads on on the joints would loosen those pocket screws and the table would get really wobbly really quick and drive you insane, and you'd throw it away long before it ultimately failed. If this was made with mortise and tenon joins, and maybe with draw bored tenons, it might not be any stronger in ultimate strength, but it would last a LOT longer, maybe even for centuries before the joints came loose and the table got wobbly.
Now I'm wondering about stats for different kinds of wood. I think you used soft pine on those. I wonder how oak would hold up on, say, end-grain vs cross-grain screws.
I saw a video recently where someone said that the commercial dowel pins are compressed, so that when you add glue and assemble, the moisture in the glue causes them to re-expand a bit, locking them in the hole tighter. I’d never heard that before, and have no idea if it’s true. I don’t have any on hand, but I guess you could test it by lightly wetting one and seeing if it did expand.
So, I assume that using harder woods would result in the screws performing better while the dowels getting little to no gain in performance. Reason for that is the wood is failing, not the screws in your current test so if you use a harder wood it should result in the screws being pulled through the wood less. And on the flip side, with the dowels, the glue / press fit is failing which when going to harder wood, I would assume the glue and press fit would perform the same / similarly.
i'd like to say it's interesting data. one thing that bothers me is that it doesn't take into account side loading (real world) which I believe will split the wood even if the pulling strength is great. end grain seems more likely to split
I have been using little coin magnets in plenty of recent projects. Sometimes magnets that attract to metal and also magnet to magnet. I would be very interested in seeing some testing on those. Especially what little air gaps might do to the final pull strength. I can't always get things perfect and have wondered about eh seemingly exponential falloff with magnets.
air gaps... bad news. iirc, its a cubic inverse ratio. its really hard to set anything ins tyone with magnets as its geometry based, material based... but air gaps are just bad all round. try to use pots if possible. bit costly but well worth it. shields the field externally so no nasty stuff happens, and the field is concentrated between those two iron poles. can make them though? like magnetic door latches were always a simple ferrite between two iron plates or keepers, set back from the outer edge (gives two defined poles, and magnetic "junk" doesnt get in the way so much...), and the latch is a simple iron plate to bridge the gap. no air gap when latched. keep all the field in the iron where it wants to flow. rather than use one pole only, think of "closing the loop" instead. far more efficient. and soft iron lets you manipulate where the field is. magnets LIKE to be shorted out with iron. not just stuck on one pole but shorted out.
Wow, thanks for your suggestions. So pots are like the door latch magnets? I had never heard that. I never thought about shorting out a magnet either.@@paradiselost9946
I don’t consider anything after the screws started pulling out useful. The joint failed when it started separating. The point is to hold the two pieces of wood together, not just close to each other.
excellent content and presentation. I would like to see the bamboo in shear when it is used as a scroll pin. i use this method where i don't want metal fasteners or i want the pin to pull a joint together using a taper.
I always wanted to know how large the difference between screws crossgrain and endgrain is. If I had to guess, I would have said endgrain has half the holding strength (not measured anything, just had a feeling that endgrain fails proportionally more often). Good to see that they are way closer than I thought. So the endgrain failures were more likely due to where they were used (butt joints, which pretty much always are like a lever) than the strength of the connection itself. Also good to see that drywall screws work almost as well as wood screws (and are even more consistent over grain directions). They are cheap and I have a lot of them left from building drywalls during a renovation.
I wish you could test my favorite: 2.5mm "bamboo nails" (bbq skewers). They are a really easy method of reinforcing rabbet joined boxes for example. But I'm not sure how much they provide strength, especially if the rabbet is already glued.
Do we know if the wood glue would deteriorate with time? If I made a stool with screws and another with dowels, would the stool made with dowels still be stronger after 100 years?
just cross pin the dowels and the glue will be good till termites or rot gets it. i have a bookcase i built 50 years ago with no nails, glue, or screws. just dowels cross pinned with dowels. i made it that way so i could tap it apart to move. still stiff like new except where patches chewed on a corner.
Thank you for doing all this testing and video-making. This information is invaluable to the carpenter and woodworker. I wonder if you might be able to do some testing with hardwoods? I feel that the hardwoods (walnut, hard maple, cherry, hickory, oak) might be able to withstand the head pulling through better than softwoods. The other consideration is that furniture typically isn't made with softwoods. To be honest, most fine furniture is made with loose tennons or dowels instead of screws, but that's a whole other thing. I really am curious to see how hardwoods can resist pulling the head of the screw through them. I can send you some sample hickory and walnut. I might be able to send some cherry as well. I don't really have any oak or hard maple though.
Wonderful video. Only improvement would be if you had a way to measure the strain(% extension) the same way you are measuring force(mass). That would allow you to show force vs strain. Stess vs strain also with known area.
Very cool. I think you'd get better results with the drywall screws if you used coarse threaded ones, the ones you used are more for use with thin metal studs.
@@matthiaswandel Wow, that's surprising! Does anyone know why that would be? The whole reason/theory behind coarse-thread drywall screws is that they're supposed to hold better in rough construction lumber, softwoods like pine/fir/&c, right? I guess another issue is likelyhood-of-splitting as distinct from holding power, but still?
9:02 When you think about the way the wood is 'built', not that surprising. Even though you're severing the long fibers every mm say with a screw thread, there's still a lot of parallel fibers holding things together, sortof glued together. And the same holds through for the side grain screw. I think at least.
Used to be you could get spiral-fluted dowels at most builder's supply places. These days they're all straight fluted. I'm sure the spiral type is still available...and probably makes for a stronger glue joint.
With the weakest joint holding for 100kg (about 200lb), my conclusion is that choosing a joint is not at all about the strength, but rather about what is the most useful and convenient method for the application. 100kg strength is more than good enough for anything made out of wood. Pocket hole joinery is sometimes the most useful and convenient and sometimes it's the screws. You didn't do dovetails and such. why? I think I may know the answer...
there is straight load, and then there is impact load, which is momentary load, easily 10x higher. Also, for example, joining apron rail to table legs, there is about 10x leverage. So accidentally bump into a table (impact load) with full force, and that's enough to bust a lot of joints in the wood.
Now you just need to make a pocket hole bamboo skewer jig to get 100% test coverage ;-) And in a stroke, Matthias wipes out the pocket hole industry in a single video.
Speaking as a professional, amateur engineerer I bet the bamboo adds extra strength as its grain will run diagonally though the wood! Can I get my name on the patent with you please? ;-)
Thank you for doing this. I would love to see how lag bolts or construction screws would hold up (thinking GRK fasteners). I’ve considered using them and then covering the hole with a dowel for an easy joint.
I'd be curious how much pinning a 3/8" dowel with a bamboo skewer, i.e., at each end drill a hole equivalent to the skewer size through the dowel perpendicular to the direction of force and glue it in, effects the results. I've seen this suggested for joints where you want some mechanical fastening even if the glue joint on a dowel or biscuit/domino/loose-tenon fails.
I'd be curious to see the impact of screwing/doweling into quartersawn vs. plain sawn vs. rift sawn. Does the ring orientation of the wood make a difference (especially in the screw head pull-thru)?
Is it possible that the wood glue is reinforcing the skewers within the test subjects? Bamboo has a very porous grain so it may absorb the glue all the way through. I would think that could contribute to the tensile strength significantly.
Try pan head screws. Also, anyone who joins wood with a butt joint to carry a load greater than 20 kg, regardless of method, seems you would be asking for trouble.
Not sure if this was covered, but did you try course thread drywall screws? It appeared you were using fine thread, often used for fastening to metal studs.
IMO as soon as the joint has opened a visible gap it's failed. So I'd put the screws as failing at around 100kg. Which is quite a bit less than dowels.....My next project will be building a subwoofer box, after seen this vid I may use dowels instead of screws.
I think TOIDS DIY AUDIO did a subwoofer box feature where he added in a bunch of internal baffles to increase strength and decrease panel resonance. Also look into including at least one outside panel that is non-parallel to the opposite baffle...eg: front panel is slightly angle in relation to the back panel. This actively helps to break up any internal standing waves that could cancel each other out and cause null point nodes in your frequency response. Good Luck!!!
@@mikewatson4644 I agree any of these methods would be strong enough for a sub box...But screws cost money, dowels can be as cheap as free and are equally as difficult to install as screws providing you intend to drill pilot holes for the screws. The next benefit over screws aside from being clearly stronger is I can run a round over bit on all the edges of the box after assembly without any fear of hitting a screw. Same when I'm stapling carpet to box later or installing other hardware I won't be running into screws that aren't there.
Leaned something more about imperial system. Or something I should have known. Stupid me though #8 screw would be 8mm screw , but that would be too simple for imperial. It's arbitary number 0.164" or 5/32" (they are not the same of course).
Makes one wonder how screws and dowels would fair using hardwood. I also suspect pan head (or round head) screws, especially with a washer would do. Not tested, Domino style or floating tenon type joint. Then you could go completely around the bend and test other glues (Titebond type glue has a fair amount of creep) like Plastic resin glue, epoxy, and resorcinol glue. Your average woodworker likely would only consider epoxy or urethane. After seeing some tests in Fine Woodworking I'm typically OK with Titebond. Good stuff! Thanks for doing this!
I tried drywall screws, yeah, cheaper alternative, slightly better head, vs flat head They suck, they shear off. I much prefer to use regular construction screws and add a small washer, because even the extra expense of 1/8 (3.2mm) ID washers is cheaper than washer head /pocket hole/structural screws, and MUCH cheaper than the frustration of extracting sheared off drywall screws. I'm on the verge of converting to ye olde Lag screws.
Thanks for looking at this often asked question. Very appreciated! Can you address pocket hole screws with the various common wood glues? (Titebond, Elmer's, maybe a 5 minute epoxy, etc?) Can you show us how your test platform is set up?
Well... I guess I'm going to have to eat more Japanese food now. Bummer. 😬 Anyway, thanks, Matthias! Stay safe there with your family! 🖖😊 And happy holidays!
interesting stuff, although maybe a bit unfair to screws, they were quite a bit smaller than the dowels in surface area the dowels would not fit where the screws would. maybe larger diameter screws would be more fair?
Hello Matthias, thanks for video, Super video with stats, it blows my mind the things you able to do, either machine building or testing. Thank you, very informative. Regards
Did he mention what pilot hole size he used? That could be important. I think he had done a previous test on that but can't remember if that included end grain.
@@F0XD1E l don't think he mentioned it this time, but usually you wanna go for about the minor thread diameter, or the thickness of the core of the screw between the ridges
This is totally unrelated to the subject at hand, but maybe someone here has an answer. Recently, all of the RU-vid posts have a thumbs up of 304. Yes, all of them. Hmm, suspicious number since it has a slang meaning. ???????
its very important to provide pressure to the gluejoints. the more pressure it cures under,the stronger the joint! you also MUST wet the endgrains!!! not much but enough so that the glue gets sucked into the fibers. capillary action,baby!
You need a displacement (z-axis travel) versus force graph to have these tests mean anything, your conclusion of “glue doesn’t make pocket holes stronger” years ago was incorrect for that same reason. The force to required to create displacement and the opening of joints is more important that ultimate failure strength.