Page 1 of 2 12 LastLast
Results 1 to 15 of 18

Thread: Joint Failure Analysis Thread

  1. #1

    Joint Failure Analysis Thread

    Over on another thread, we were having some friendly discussion on whether dowel joints were strong enough for use on chairs and other furniture whose joints must endure a lot of stress.

    Specifically, Mike Henderson made reference to having seen dowel joints fail in chair joints. I questioned the age of the joint, type of glue, etc.

    Well today, I was at my Mom's having tea with family, and my brother noticed one of the joints was coming apart on Mom's antique coffee table. I thought to myself: "Self, I wonder if that is a dowel joint..." So, as soon as we got the hot tea off the table, I gently pulled it apart. Sure enough, it was/is a dowel joint. (Mom was yelling: "Don't DO that, you're going to BREAK it!!!")

    I don't know how old the coffee table is; just that it is at least 40 years old. Probably more like 80. It appears as if the builder didn't apply glue to the joint itself, but just the two dowels. It is end grain to face grain, so I guess that's not a HUGE crime... I can't tell for sure whether the glue is hide or yellow glue. Half the joint (the opposite side from shown in the attached photos) seems to be either molded or broken away. Probably from multiple moves and dragging it across floors.

    Also, only two dowels were used, and there is easily room for 3 or 4 dowels in that area.

    My judgement says that this wouldn't have lasted even if it were a M&T. I think the joint failed due to the age of the glue and/or rotten wood at the back of the joint.

    The whole point of starting this thread is to provide us Creekers with data. We can all draw our own conclusions based on what is shown here. I'm particularly interested in failures of dowel joints for now, but we can use it for all kinds of joinery.
    Attached Images Attached Images

  2. #2
    The first thing I notice in that photo is that the dowels look to be just plain dowels, not either of the two (fluted or spiraled) that I think allow for better glue distribution.

  3. #3
    Quote Originally Posted by Phil Thien View Post
    The first thing I notice in that photo is that the dowels look to be just plain dowels, not either of the two (fluted or spiraled) that I think allow for better glue distribution.
    Now that you mention it, it looks as if the dowels fit the holes tightly, but smooshed most of the glue out!

  4. #4
    Based on Hoadley's article, I think that any right angle joint exposed to changes in humidity will eventually fail. I think this is true of M&T or dowels.

    Since most commercial (factory made) furniture only uses two dowels in a joint, I think a M&T will last longer because it has more surface area.

    I've seen dowels glued with urea formaldehyde glue fail and UF glue is an excellent long lasting glue. So I don't think we're seeing these joint failures because of failure of hide glue.

    Also, my opinion is that regular dowels are just fine. Remember that the best glue strength is usually obtained from close wood-to-wood contact. By putting flutes or spirals on a dowel all you're doing is reducing the wood-to-wood contact and depending upon the glue to have gap filling properties. Just as we were told that you can't starve an edge joint from too much pressure, I don't think you can starve a properly fit dowel joint.

    Mike
    Go into the world and do well. But more importantly, go into the world and do good.

  5. #5
    I would like to echo Mike Henderson and add,

    Would it not make sense in commercial (factory made) furniture for there be spirals or flues to let exccess glue to go into and eliminate squeeze out (save labour of taking it off) and not becuase it will last longer.

    Adam

  6. #6
    FWIW, it's not the dowels that failed here... it's the glue.

    YM

  7. #7
    Quote Originally Posted by Yoshikuni Masato View Post
    FWIW, it's not the dowels that failed here... it's the glue.

    YM
    According to Hoadley, it's not the glue that fails but the wood in the mortise, which the glue is attached to. I'm not going to say it's never the glue but we've had excellent glue for quite a while and still see joint failure.

    But you're right that the dowels don't fail. A fresh joint is plenty strong enough. It's age that gets the joint.

    Mike
    Go into the world and do well. But more importantly, go into the world and do good.

  8. #8
    Join Date
    Dec 2003
    Location
    Laguna Beach , Ca.
    Posts
    7,201
    I think it probably is a combination of things. The dowels seem short and relative to the size and height of the table are poorly designed to provide an adequate resisting couple to the potential movement and racking force the table might experience. The long grain direction looks solid and the short direction failed. I think the glue failed but there was not much in the joint design to help, not much penetration and not much surface area . There is no redundancy from a lower stretcher that could help resist movement. There is a lot to think about when you design a piece. The size of the resisting members and joints need to be larger if there is only one member resisting.
    "All great work starts with love .... then it is no longer work"

  9. #9
    Join Date
    Feb 2008
    Location
    Northwestern Connecticut
    Posts
    7,149
    Question: Has anyone ever heard of a fox wedged dowel? A pegged dowel joint?

  10. #10
    Join Date
    Dec 2003
    Location
    Laguna Beach , Ca.
    Posts
    7,201
    Quote Originally Posted by Peter Quinn View Post
    Question: Has anyone ever heard of a fox wedged dowel? A pegged dowel joint?
    The pegged dowel joint is just another dowel holding the joint together. Foz wedging is usually done on tenons and is made by creating kerfs installing wedges . This is "blind" and only internal in te mortise. Wen the tenon is inserted the wedges are forced in to the tenon or dowel and it causes it to expand. Usually used for repair and not new work
    "All great work starts with love .... then it is no longer work"

  11. #11
    Join Date
    Feb 2008
    Location
    Northwestern Connecticut
    Posts
    7,149
    While I have never made a fox wedged tenon I have been staring at them in books for some time, both old english joinery books and Japanese joinery books. Looks like tenon meets dovetail, wondered if this was possible with dowels. Seems like a craftsman's way of saying I want this to last longer than glue will allow. Does anyone actually use fox wedged tenons for new work these days?

    I have some old maple chairs who's legs are attached through the seat with round tenons that are wedged. The glue has long since given up but except for a bit of wobble under a heavy load the joints remain solid. Thought maybe a dowel system that uses a similar approach might be interesting.

    I have studied some timber framing joints with draw bored pegged tenons where glue was not used and failure was not acceptable. My friend did a barn restoration last summer and gave me a tour. Some of the horizontal timbers were attached with what looks like a pegged dovetailed half lap. Really interesting to see joints built to hold massive loads remain intact after more than 125 years. Obviously these are not elegant furniture joints but they sure are impressive to look at.

    I guess as I learn more about woodworking I find myself spending less time trying to diagnose failures and more time trying to understand successes. I see a dowel as a quick joint prone to failure. I don't want to practice getting really good at making things that wont last so I stay away from dowels mostly. I see these other types of joints that rarely fail and am intrigued by their strength and beauty. Seems the time they take to prepare is paid back in longevity of the work.

    Anyone have a favorite special joint that doesn't normally fail?

  12. #12
    Join Date
    Apr 2007
    Location
    Phoenix, AZ
    Posts
    1,417
    I'd love to hear some more data from antiques specialists; or better to see the dowel/tenon/other joints testing done full up with good controls and number of tests, with several parties involved to ensure test design is fair and no hidden assumptions/errors are built in.

    While I have no irons in this fire (no Dowelmax, no mortising maching, never used dowels as a joint), the evidence given against the dowel joints seems anecdotal, and the extrapolation of a few observed failed joints which happen to have dowels involved, to the belief that the dowel itself was somehow causal, to the statements that dowels are inferior, seems weak and poorly supported to me.

    There are many much more likely cause-effect reasons for these small data-sets of example. Type of furniture--chairs especially; available size of joint members which might pre-dispose a weak joint and also necessitate dowels for installation and sizing reasons. Historical accuracy of boring tools vs. chisels. Old glues vs. new glues.

    Until I see data that discusses statistically valid sample groups, ratios of chairs or other "failed pieces" produced with one joint vs. another, checking of failures caused by the joint vs. other reasons, correlation of real joint failures is similar pieces with the amount/percentages of pieces produced with these different joints....
    .... it's all just conjecture and bad science.

    Sorry, I'd LOVE to be able to easily read and believe "dowels are bad, tenons good", but thus far I've never read sufficiently supported discussion to really trust that statement. Not saying that some of YOU out there don't know or have that data, just that you haven't presented it. Furniture repair specialists who also know an antique total market (not just the bad joints they see) might have good info here.
    Last edited by Dave MacArthur; 03-29-2008 at 2:54 AM.
    Thread on "How do I pickup/move XXX Saw?" http://www.sawmillcreek.org/showthread.php?p=597898

    Compilation of "Which Band Saw to buy?" threads http://www.sawmillcreek.org/showthre...028#post692028

  13. #13
    I'm afraid that just like the drug companies who only published studies that supported their drugs, the only people who are likely to publish studies of dowels or loose tenons are the manufacturers who make products that install dowels or loose tenons. And those studies need to be viewed with skepticism because of the conflict of interest.

    While someone with great experience in repairing antique furniture can provide valuable insight, that evidence is still anecdotal, and as the old maxim says, "The plural of anecdote is not data."

    I think each of us will have to make up our own mind about which joinery to use in our projects, and take our chances.

    Mike
    Go into the world and do well. But more importantly, go into the world and do good.

  14. #14
    If I use dowels I always use the spiral dowels. One advantage that they have is that when the glue dries it dries with that spiral patten. And even if the joint becomes loose the spiral pattern prevents the dowel from coming out. Just like I put glue into a screw hole to form glue threads. A great trick to keep screws in MDF BTW.

  15. #15
    Join Date
    Jan 2006
    Location
    Frederick, MD
    Posts
    322
    I have to respectfully disagree with some of the other posts - I believe I can see some spiral pattern on the dowel. The spiral may not be as deep as modern dowels - but there's something spiral there (middle picture).

    What I jumps out at me is that the surfaces of the joint appear to have little or no glue. And end grain to long grain joint isn't going to hold very well anyway - but that certainly looks like a glue-starved joint. At a minimum, there should be more squeeze out around the dowels.

    I don't think more dowels would have helped. All it would do is leave less "meat" in the joint.

    If that table is 40 years old - it would have been made in the late 50's. Looks older than that to me. I'd guess pre WWII / post depression era.

    It's not visible in the picture - what does the INSIDE of the joint (Underside of the table) look like? If you wanted a joint like that to stay together - the right way would be to use some corner blocking underneath/inside. If theres no block (or the block has fallen out) - that's a major contribution to the joint failing.

    Do you intend to try and fix the piece? It should be possible. I'd suggest taping off the finish side (blue tape - to control the squeeze out), re-glue the joint and install some backing blocks with screws to hold everything together.

Similar Threads

  1. Failure Analysis Help
    By Vaughn McMillan in forum General Woodworking and Power Tools
    Replies: 17
    Last Post: 01-08-2006, 8:37 AM
  2. Frued Box Joint Blade - Works As Advertised
    By Hank Knight in forum General Woodworking and Power Tools
    Replies: 5
    Last Post: 10-10-2005, 12:20 AM
  3. Blind finger joint?
    By Jamie Buxton in forum General Woodworking and Power Tools
    Replies: 53
    Last Post: 06-30-2005, 8:05 PM
  4. How do you edit a threads title?
    By Frank Pellow in forum Forum Tech Support
    Replies: 8
    Last Post: 12-08-2004, 5:21 PM
  5. Question: Using a thread die in a drill press
    By Douglas Robinson in forum General Woodworking and Power Tools
    Replies: 10
    Last Post: 04-27-2004, 3:29 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •