Strength of Domino... Test Results

[Forrest Anderson]

The results posted by mr_hockey at the beginning of the thread were divided into two categories - Shear and Pull Apart. For ease of reference, the Pull Apart figures were:

M&T 2525
Dowel 1866
Domino 1486
Beadlock 1170
bisquit [sic] 766

My initial reaction was to assume that the the dowelled joint separated (pulled apart) at 1866, and the Dominoed joined separated at 1486, indicating that the dowelled joint could withstand 25% more pulling force than the Dominoed one, before the joint separated.

bassman00 kindly found a video at http://www.woodmagazine.com/wood/story.jhtml?storyid=/templatedata/wood/story/data/1178129509968.xml which accompanies the testing article. It shows how each jointing system works, and although it doesn't show the testing in progress, it does show the failed joints, which are presumably those which were subjected to Pull Apart testing.

I have taken the liberty of posting screenshots from three frames of that video, which show the failed joints.  I have added labels to identify each joint:

Looking at how the joints have failed, it seems that only the biscuited joint has actually pulled apart. The Dominoed, Beadlocked and Dowelmaxed joints have remained intact, and the wood in the horizontal wooden member has split along its length, which stopped the test.

Since it was the wood which seems to have failed in three cases, rather than the joint, I don't believe that the strength figures in the test can be reliably used to compare the strength of the three jointing methods. Instead, they primarily reflect the strength of the horizontal member. I say primarily, because we don't know how deep the various loose tenons were set into the horizontal member. If the end of any of the tenons extended beyond the split in the wood, then the part of the tenon beyond the split has indeed lost its grip in the surounding wood and failed (or the tenon has fractured).

Forrest

 

[greg mann]

Forrest,

Your pictures speak volumes about this. It appears that failure for all but the biscuit joint happened beyond the effective depth of the joint. I can't say for sure but it appears that is the case. To me, this means any of these joints are stronger than the wood, or at least this wood. It also tells me the test parameters were flawed in execution. The pull test should have retained the edge grain board directly adjacent to the joint so that board failure would have had to occur in a more concentrated area, if it was going to fail before the joint itself. The point is: If the joint is stronger than the wood then all else about joint strength is pretty moot.

I do find it annoying that anyone would publish data like this without a thorough analysis of what really happened and what the outcomes really meant. These tests should have been redone correctly or never seen the light of day. At least three of four joints were inaccurately portrayed.

 

[Brice Burrell]

.........Since it was the wood which seems to have failed in three cases, rather than the joint, I don't believe that the strength figures in the test can be reliably used to compare the strength of the three jointing methods. Instead, they primarily reflect the strength of the horizontal member. I say primarily, because we don't know how deep the various loose tenons were set into the horizontal member. If the end of any of the tenons extended beyond the split in the wood, then the part of the tenon beyond the split has indeed lost its grip in the surounding wood and failed (or the tenon has fractured).

Forrest

Here's my take on it, looks to me like the wood split at the end of the loose tenon, bottom of the mortise, I'm not sure but thats how it looks to me, so in my mind this dose reflect the the strength of the joint, just not what we normally think as being part of the joint (the bottom of the mortise). Making a loose tenon,you remove wood to make the mortise, you make the mortise deeper than the tenon is long, so you are left with a weak spot (at the bottom of the mortise). This is why I think the Domino failed before the dowel, you remove more wood with the Domino than the dowel, thus a bigger weak spot. Maybe a better way of thinking about this is not the weaker joint but the weaker structure, remember that is what we are ultimately after, a stronger structure.
Brice

 

[greg mann]

I understand what you are saying, Bruce, but in the example as shown a traditional mortice and tenon would remove material in the board that failed in exactly the same manner that the Domino does, or the other methods too for that matter. IIRC, the traditional mortice and tenon had the best performance. Odd, since all failures were due to separation of grain at what appears to be the bottom of the mortice, or very close to it, then the tradtional M&T should have failed at the same place if all other parameters were equal. In other words, if the failure of the joint always happens in the cross-grain side of the joint then the traditional M&T should be no better than a loose tenon, ever. Your observation about a dowel displacing less wood has a little merit here. I am not sure you are saying this but it sounds like your observation is that the dowel performed as well as it did because it weakened the board less than the other joints did. Regardless of this particular set of test results it makes no sense that a dowel could perform in shear as well as the other joints. I base this on the understanding that an end grain surface is the weak link in this type of joint and a dowel would have less cross section, relatively speaking, and more end grain dependence. Both should be bad for shear strength. Shear strength would be far more dependent on tenon cross section, be it in the form of Domino, beadlock, or traditional M&T. Heck, at this point we don't even know how many dowels were used. 1,2,3? One could bias this test in so many ways. 5mm domino or 10mm?

I still maintain that the failures I see indicate that too much of the test result is influenced by the inherent strength of the board. When the failure is in the board then it is hard to say the joint failed. If the wrong size tenons or dowels are used then, yes, one could weaken the board. That is why there are rules of thumb on tenon size. Without that information this test is too speculative.

 

[fcraven]

I remember another magazine did a test, possibly FWW, where a pinned M&T joint was weaker than a normal M&T joint.  I believe their conclusion that the pin weakened the M&T slightly.

There's a video at the bottom of this page but it doesn't get into the testing.

http://www.woodmagazine.com/wood/story.jhtml?storyid=/templatedata/wood/story/data/1178129509968.xml[/url

Good video. Sure, not that informative about the testing, but the speaker is giddy about using the domino.

 

[Forrest Anderson]

Here's my take on it, looks to me like the wood split at the end of the loose tenon, bottom of the mortise, I'm not sure but thats how it looks to me, so in my mind this dose reflect the the strength of the joint, just not what we normally think as being part of the joint (the bottom of the mortise). Making a loose tenon,you remove wood to make the mortise, you make the mortise deeper than the tenon is long, so you are left with a weak spot (at the bottom of the mortise). This is why I think the Domino failed before the dowel, you remove more wood with the Domino than the dowel, thus a bigger weak spot. Maybe a better way of thinking about this is not the weaker joint but the weaker structure, remember that is what we are ultimately after, a stronger structure.
Brice

Brice

Your points are well taken. Here are two musings for discussion...

If the "Domino failed before the dowel, [because] you remove more wood with the Domino than the dowel", then it could well be possible that smaller Dominos are often a better choice. This tends to go against instinct, where one imagines that larger Dominos would give stronger joints.

Secondly, with all other factors being the same, even if the Domino in the test had been made of a much harder wood, and then been locked into both rails by steel pins running through the sides of each mortice, it would have failed at around the same load.

I'm really just thinking out loud, no more, no less...

Forrest

 

[Brice Burrell]

  Greg, you're right about the traditional M&T, how could it be that much stronger if what I said is true. I bet it failed the same manner. The only thing I can think of for there numbers to make sense is that they made the traditional M&T correctly, 1/3 rule, and used a domino that was too big??? And with the shear test they use a smaller domino and maybe they used one of the larger setting, not the tight fitting mortise on the Domino.

...The following quotes are from Greg:
I still maintain that the failures I see indicate that too much of the test result is influenced by the inherent strength of the board....

I'm not sure, if this is the case why did the traditional M&T do so much better?

....When the failure is in the board then it is hard to say the joint failed....

Unless my theory is right. And I'm not sure it is.

If the wrong size tenons or dowels are used then, yes, one could weaken the board.

Yep, and this could be inadvertently done, see my catch 22 theory below.

One could bias this test in so many ways.

Here you have summed it up prefectly.

From Forrest:

If the "Domino failed before the dowel, [because] you remove more wood with the Domino than the dowel", then it could well be possible that smaller Dominos are often a better choice. This tends to go against instinct, where one imagines that larger Dominos would give stronger joints.

Well it seems that it's a catch 22, I would think smaller domino is better at pull part and a larger one better at shear????? I don't know,

  I'm just trying to make sense out of this test.

 

[Groggy]

I question the value of this test (but may have missed something). Why test a joint in a direction it is not designed for (tension), surely that invalidates the test? You may as well do car reviews by how fast they can go in reverse. A mortise and tenon is best for lateral loads and to stabilise wracking/twisting tendencies. It appears that what has been tested is the area of glue and its relative strength in the mechanical joint.

If I were to suspend my life over a canyon, I can tell you it would not be by a M&T joint unless it was reinforced by a through dowel or two, pegged, wedged or otherwise reinforced. For this purpose a dozen other joints would be superior. For testing M&T joints I would prefer to see wracking forces measured, then you would really see some differences between biscuits, dowels, true M&T and dominos.

 

[fidelfs]

There is a similar test in Fine woodworking March/April 2001.  They don't talk about Domino of course (It didn't exist at that time) but they talk about loose tenon.
I cannot post the article because tauton requires a membership to see the document (I have the membership).

If you want you can find the document here :http://www.taunton.com/finewoodworking/SkillsAndTechniques/SkillsAndTechniquesPDF.aspx?id=2712

This is from the free part that I can post:
  Testing Joints to the Breaking Point
Surprising results surface in an analysis of commonly used furniture joints
by Bruce Gray
Testing Joints to the Breaking Point
Using the laboratory at the Wood Science and Technology Center of the University of New Brunswick, Bruce Gray tested joints and talked to experts about why joints fail. They tested traditional mortise-and-tenon joints, floating-tenon joints, biscuit joints, and dovetail joints for load, displacement, rate of failure, strength after failure, and rigidity. He explains how the samples were prepared and tested and how to interpret the charts. The article also offers a guide to the anatomy of the joints tested.

It shows the strenght of loose tenon and the test follow a very scientific approach. 

 

[bassman00]

Hopefully that issue of Wood Mag will be out soon so we can find out the methodology used in the tests.  One question I haven't seen asked is about the numbers.  Are all those results far above the norms a piece of furniture will ever see?  I'm also a member of fww online and that article states in it's conclusion that all joints tested, including the double biscuit, were stronger than any load they might be subjected to in normal use or even abuse. 

That being the case, are we talking about the differences of cars capable of going 175 or 200 mph when we would never need to exceed say, 75 mph?

PaulD

 

[Jesse Cloud]

I read the Wood article last night and was disappointed.  They say almost nothing about the methodology of the tests.  My main question would be what size domino was used.  They used a 1" integral tenon, but say nothing about the size of the domi, the dowel or the biscuit.  Also, they used only one domi but two dowels - is that fair?  In any case, all methods except biscuits held longer than the wood did, so I have no problems continuing to use the domino.

 

[Ben West]

I posted this on Sawmill Creek on the Wood Magazine joint strength thread, and thought it would be good here too.

Part of the problem with many of the strength tests published in magazines is that they rely on tests that aren't replicated.  The recent Wood Magazine article, for example, evidently paid a wood products lab at Iowa State to test each of these joints.  To do that research the right way, a researcher would need at least a couple dozen testing replications of each joint on randomly selected wood samples, which would be expensive (I'm in academia -- in a Forestry College with a Wood Products department, no less -- and know for certain that Professors don't do research for small amounts of $$$!).  Magazines obviously don't want to spend the amount of $$$ required to produce reliable research; they want an interesting article for publication. 

So, what you have with the latest Wood magazine is an article that is relying on poor research methodology and, as such, can't be relied on reliable knowledge.  Obviously, we all know and would agree that a M&T joint is stronger than a biscuit, and the article demonstrates that, but to conclude that a dowel joint is 33% stronger than a Domino or that a Domino is 21% stronger than Beadlock conveys a level of accuracy and precision that just isn't allowed by this research. 

I read that Dan tried to do a Google search on this research and was surprised he didn't find an academic reference on it.  You'll never find this research in an academic reference, because it doesn't meet academic standards.  If you look at a scholarly wood technology journals, you'll see research like this with replications, randomization, and statistical comparisons.

 

[James Metcalf]

Ben

All very good points. Now what you need to do is talk the university ,and we both know it is the best university in the state. into running test with the Domino.At night you could take the Domino home,to keep it safe.If that does not work drive on up to Memphis and you can use my Domino for a true test.

P.S. I had nothing to do with Main burning when I was there in 1958.

 

[Dave Rudy]

There's a wealth of good thinking in this thread, and some highly qualified people contributing.  I hope that you will all email or write Wood magazine with some carefully chosen questions/critiques regarding the testing methodology.  As for me, I'm not smart enough to figure out where to get a July issue of Wood.  The May issue is still on all the newstands as far as I can tell.  Sometimes it's hard being at the center of the universe. 

 

[Rocker]

It is very surprising that a University Department would produce such a series of figures on the basis of the results of those tests. As Forrest pointed out, in the pull-apart tests, the only joint that failed was the biscuit. The other joints held but the nearby wood broke. As far as I am concerned, the conclusions drawn from the tests are rubbish.

If I may be permitted a digression, I am not sure about Dave's contention that Colorado is the center of the universe, but I do know that it is the birthplace of al-Qaida. Apparently, the lascivious cavorting of the ladies of Greeley so incensed an Egyptian called Qutb, who was attending a teachers' training college there, that he developed an intense hatred of the US, and went on to found the terrorist movement. When I lived in Greeley, a few years ago, it didn't really seem to be Sin City, but that is just me.

David

 

[ROb McGilp]

I read the Wood article last night and was disappointed.  They say almost nothing about the methodology of the tests.  My main question would be what size domino was used.  They used a 1" integral tenon, but say nothing about the size of the domi, the dowel or the biscuit.  Also, they used only one domi but two dowels - is that fair?  In any case, all methods except biscuits held longer than the wood did, so I have no problems continuing to use the domino.

Jesse, I think you and Ben have it right. The test is really a one off "a v b v c v d" system comparison. Anyone used to dealing with Scientific data and reading this would, probably cry or roll on the floor laughing. The number of dominoes and dowels needed to create a joint as strong a an M&T has to be taken into consideration, otherwise the test is biased. In addition, the standard of the beadlock mortise should also be quoted. I certainly wouldn't like my name in the header.  

Regards,

Albert

 

[Ben West]

Ben

All very good points. Now what you need to do is talk the university ,and we both know it is the best university in the state. into running test with the Domino.At night you could take the Domino home,to keep it safe.If that does not work drive on up to Memphis and you can use my Domino for a true test.

P.S. I had nothing to do with Main burning when I was there in 1958.

James, just sent you a PM.

 

[Matthew Schenker]

There's a wealth of good thinking in this thread, and some highly qualified people contributing.  I hope that you will all email or write Wood magazine with some carefully chosen questions/critiques regarding the testing methodology.  As for me, I'm not smart enough to figure out where to get a July issue of Wood.  The May issue is still on all the newstands as far as I can tell.  Sometimes it's hard being at the center of the universe. 

Dave,
I'm waiting until the review is published and all of us have had time to read it over.  But if it turns out to be faulty, I definitely think it is a good idea for members of this forum to write to WOOD and share our opinion.  It is in our best interest to respond to the assessment procedures, using science and logic, and not sound like people who are just upset because our favorite tool received a bad review.

When we have all read the WOOD review, and if it turns out to be an appropriate response, I would draft a letter.  Members of the forum can read the letter, comment, and through this process we can produce a letter that reflects our ideas.  This would be really effective because we can have people who represent professional woodworkers, people who can comment on the science, and others.  Then we can send the letter to WOOD.  That would have a much stronger effect.

For now, let's wait for the review.

Matthew

 

[Dan Clark]

I agree with Matthew, especially about:

and not sound like people who are just upset because our favorite tool received a bad review.

I have that copy of Wood magazine and have read the review.  Twice.   To be fair, they were very positive about the Domino.   The Domino was rated #1!

From their article:

[some comments about how fast you can make an Arts and Crafts rocking chair with the Domino and then...] "Couple that with the quick indexing, ease of use, choice of five tenon sizes, and complete dust and chip extraction (with an attached vacuum), and the Domino becomes our first choice among the four tools."

Also, they did NOT state that the testing done by the faculty or staff of the Iowa State University.   They said:

We then broke them by applying force using a hydraulic testing machine at the Iowa State University's Structural Materials Facility, testing in the same manner as the "Wood-Joint torture test" from issue 173 (November 2006).

I.e., they did testing for an article, not as a rigorous scientific test.  This raises obvious concerns about the validity of the strength testing.

Let's focus on the key issue that has been raised - the strength of joints and the testing methodology used to support their conclusions.   After examining the article text and the two pictures showing the joint failures, three things struck me. 

First, it appears that they used different pieces of wood for different joint types.  I.e., they did NOT cut chunks from the same piece of wood to ensure consistency.   

Second, it appears that they tested only one instance of each joint.  Any inconsistency in the wood quality could result in failures due to the wood and not the tenon type.   

Third (and most glaring), the shear test in the Domino shows the rail wood stuck to the stile's wood with the Domino still in the stile, but at a slight angle to the right.  How the Domino have pulled out if it was glued properly?!?   How could the Domino have twisted right it was installed in a mortise the same width as the Domino?!?   

Combine all three of these issues and you have to question the tests.  This is NOT a Domino-specific issue.   It is about the relative strength of all of the joints.   Did the tests provide valid, statistically significant results?  If someone else did the tests would the results be the same?

A summary of the my key questions/issues/concerns:

• What were the specific Dominos, tenons, dowels, and dowel maxes?  Length, width and height?
• Were the rails and stiles cut from the same board?
• How many times was each test performed for each joint type? Once?  Multiple times?
• If tests were performed multiple times, how were the results computed?   Average?  Median? Min/Max?
• When the Domino mortise was cut, what width setting did they use?  Narrow, medium, or wide?
• How were the joints glued?  Just the tenon?  Just the mortise?  Both?  Were the joints glued to match the qualities of each joint type?  (I.e., to give best results for that joint type.)

These are my questions.

Regards,

Dan.

 

[Matthew Schenker]

Dan,
All good points!  I have not seen the article yet, but I'll seek it out today at my local book store.

As more people read the review, I would be interested in hearing responses to the basic methodology and validity of this kind of testing.

Matthew