Interlocking dowel joint strength test

[zenwood]

In designing a table recently I hit upon the idea of using dowel joints to join the aprons to the leg, and offsetting the vertical positions of the dowels by a few mm so that one set of dowels mechanically locks through the other. The first pic shows the general concept. I use the Ozziejig to drill the holes, and one joint needs to be complete before drilling the holes for the other apron at 90 degrees to the first. The second set of dowels will mechanically lock the first set of dowels. So the first set should be those that would be subject to the most stress: the longer apron of a long table, for example.

There are lots of details I need to perfect. One detail is to make sure that it is possible to clamp the Ozziejig in place after the first joint is complete. The Ozziejig's hole spacing is 22 mm, and so an offset spacer of 11 mm will give a row of dowels with 11 mm spacing. An offset of 5.5 mm between this row, and the row for the other joint at 90 degrees will give the mechanical locking effect for 8 mm dowels.

I tried the approach on a test joint, and decided to load it up to the point of failure. The test joint differs from the concept sketch in that only two rows of dowels were used, and only the inner rows interlocked (the dowels were not quite long enough to intersect at the outer corner. I used Selley's PVA+ glue.

Pic 2 shows the test setup. The leg was clamped to the bench, and a length of scrap hardwood was clamped to the test joint to increase the distance avalable to apply the force. The force was applied at a distance of 50 cm from the joint using a platform suspended on a length of rope. Weight was added to the platform in the form of floor tiles, each of which had a weight of 626 grams. No.1 son and I added tiles until the joint began to show evidence of failure.

Pics 3 - 5 shows the joint beginning to fail. Failure is indicated by the appearance of cracks on the endgrain of the leg, which over a period of about 5 minutes (the joint was under continual maximum load over this time) gradually widened.

[zenwood]

These pics show the process of joint failure continuing.

[zenwood]

For completeness, I decided to destroy the other apron's joint too, by belting it with a large hammer this time, because this one would be weaker than the other. It took two good heavy blows before the joint (or the wood rather) came apart.

The first joint was able to hold 76 tiles before failing, or 48 kg of weight. This corresponds to a force of 470 N applied at a distance of 0.5 m. This is a torque of 233 Nm.

I found this joint amazingly strong, and should be quite adequate for the desk I have in mind. I also found that Bunnings-supplied 8 mm dowel stock (the type that comes in about 2400 mm lengths) varies from about 7.95 mm to 8.3 mm diameter. It may be time to invest in one of those metal dowel plates (number 1 on the shopping list for the WW show).

[wayneo75]

Very interesting.

Dr Karl would be highly impressed that you carried out you own experiment other than expecting somebody else to simply give you the answer!

[Andy Mac]

Hi Zenwood,
Interesting idea and followed through with rigorous testing...better than most of my ideas!
My first thoughts were about the series of drilled holes in-line, surely a source of weakness, a bit like perforations in a stamp for tearing them apart? I think your test piece showed that, the timber was compromised, but would that be any more than an equivelent mortice?
It would appear that the dowels themselves, en masse, have come out of it reasonably well, am I right? In a more usual dowel joint, say two per joint, the dowels themselves would break, but the timber remain intact. So maybe you have to find a compromise between weakening the timber and retaining the dowels?!

Thanks for posting,