the plain Miter then add splines and it goes nuts

 

 

Thanks!

Interesting video.  Interesting test.  Relative to each joint the results don't surprise me.  The force at each corner though would vary though if the boards were a different length.  The energy actually applied at the corner is Force times Distance.   With the distance being the horizontal width of the joints standing on end as tested. Shorter boards and it would have taken more weight to break each.  Longer boards and less weight would have done the job. 

 

In addition in all my years teaching furniture design I've never come across a student design that would try to pry a corner apart in use. Even the daring cantilever tables that show up every decade or so have more counterbalance and and structure holding the extended top.  We have the students make 1/4 scale models for their first critique, and designs that are likely to fail will demonstrate that even at 1/4 scale. Changes get made to eliminate failing tendencies before making a full size version.   

4D

I'd guess that chairs might be the one piece of furniture that would require really strong joints. I've only built Morris chairs. Those were constructed of white oak with mortise and tenon joints. To my knowledge, none of those have failed. I've built a ton of boxes of various wood or, wood like material, during my cabinet building days. Some with pocket screws but, most with various iterations of dado joints. Which, when you think about it,  is a sort of a mortise and tenon. 

Recently, my son and I built a vanity. The face frame was pocket screwed butt joints...with glue, of course. It was, in turn, pocket screwed to the carcass. The drawers were built using the quarter, quarter, quarter, method. Lots of glue surfaces, there. 

For small gift boxes and, such where tremendous joint strength isn't a necessity,  a miter  with splines is my go to joint. I had no idea it was such a strong joint. Those splines add visual interest, which is the aim. And, much easier and faster than dove tails. And, apparently, just as strong. Though, I sometimes use a dovetail bit for the spline. 

All this talk about joints brings back some fuzzy memories of my youth.

 

What I found more interesting than the joint failures was how often the top (load bearing plywood) board broke.  Speaks to just how much force is being employed.  Of course, what isn't tested:  time.  Little kids (some big ones) rocking/jumping/moving the items.  I think you can also declare a winner:  pretty much all of them.  Even the lowest-tested-weight failure held well enough to be successfully employed in most wood joining needs.

Thanks Cliff.  Definitely an eye opener.  Wood glue has always amazed me with its strength.

My issue with tests like this is that

• they only measure one of four types of stress on a joint
• they only do one type of joint, wood can meet in various ways and various angles
• strength of wood varies by axis and species
• what to do if wood failure causes joint failure
• tests only use one type of glue
• they are conducted on fresh joints, not those with several seasons of wood movement
• there is little consideration for sudden vs. gradual failure (important if you are talking about something like a chair)
• there is little consideration for sudden vs. gradual stress applied
• is it strong enough? you don't need the same amount of strength for something like a dining chair vs. a picture frame
• variations for too tight, too loose, or optimal fit in the joint
• a few samples is not statistically valid

 

Other than that...

 

• 

 

 

Edited May 3, 20224 yr by kmealy

I found that fascinating! Thank you.

Thanks for sharing. Very interesting. 

16 hours ago, kmealy said:

• a few samples is not statistically valid

Aside from the additional good points you make, this point is most significant. In a controlled and validated test environment a minimum of 31 identical samples for each joint would be required statistically.

 

The first thing that grabbed my attention was the disclaimer "this channel received money or free things for making this video. While both interesting and entertaining it truly doesn't determine which joint is "a winner." He does some interesting stuff on his channel, but first and foremost, he is a "Content Creator" using woodworking as his medium. Just my $.02. YMMV. Just call me "Davey Downer."

 

Such tests should not be taken in absolute terms, but in relative performance.  While the applicability of any joint configuration test is the basis for argument when applied to a specific project need, broadly speaking the tested joint with the highest performance will usually perform better in other configurations.  As I commented at the family reunion:  it's all relative.

Well. Don't you all think that his white smock ascribes a good bit of scientific credulity?

30 minutes ago, Gene Howe said:

Well. Don't you all think that his white smock ascribes a good bit of scientific credulity?

I have one of those as well.  but use for let us say, examinations.

I am just saying, enjoy the entertainment, but take the results with a grain of salt.

 

I am reminded of another YouTuber that wanted to do similar tests and contacted one of the dowel jig manufacturers to get a sample to run the testing.  This particular one claims that their joints are stronger than <whatever other joint>.    In order to get sent a sample, the tester need to agree to a bunch of criteria.  That included certain woods, certain joints, certain sizes, testing within a time range after joinery, etc. etc. etc.  I guess you would call that specifying the test to determine the outcome.

1 minute ago, kmealy said:

specifying the test to determine the outcome.

I do (or did!) a lot of engineering forensics.  There are a lot of "tests" that are designed to sell products.  Remember paper towels vs hand dryers using air?  Completely bogus methodology, conclusion invalid.  I see a lot of that.  So I keep right on taking the baby aspirin.

Interesting video and comments. However I think the real test will be time and usage. For most joints we make they will be also supported by other joints and we are not likely to ever put the amount of stress on them as this test did.

3 hours ago, Gerald said:

... we are not likely to ever put the amount of stress on them as this test did.

True, unless someone is leaning back in a chair or dragging a dining table across a carpeted floor.   But then, the miter is not the right joint for those jobs.    I am safe with my jewelry boxes with mitered and keyed joints, I guess.

On 5/4/2022 at 10:00 AM, Grandpadave52 said:

or each joint would be required statistically.

I ain't goanna be doin' no statistical analysis no how, no time soon.

On 5/4/2022 at 1:02 PM, kmealy said:

but take the results with a grain of salt.

Take everything with said sodium chloride.

On 5/7/2022 at 1:53 PM, Cliff said:

I ain't goanna be doin' no statistical analysis no how, no time soon.

Aw, come on Cliff. Where is your sense of adventure? X-bar/r charts, c- charts, p-charts, u-charts, bell curves, +/- 3 sigma, Upper/Lower control limits, sub-groups, sample sizes, all those common house-hold terms...old QC/QA statisticians never die, they just become insignificant