IV. CONCLUSIONS The low number of samples covering a wide array of joint configurations eliminates the possibility for formal statistical analysis. However the results are congruent with "intuitive engineering" The consistent failure pattern and ease of failure when knots were situated at the inside corner of the joint suggested that three elements determine joint strength: the depth of the biscuit, the lateral breaking strength of the wood fibers and the "coefficient of splitting". the last two measures vary with species of wood.
Our research offers clues to the most asked question in woodworking, "How large a gorilla can I let hang from the outside corner of an open 36" ponderosa pine door built in the same way as the test samples?" By extrapolating to the configuration of traditional four-panel door built using construction methods as above, one can make a fair estimation.
The three four-joint sample frames supported a mean maximum mass of approximately 1,600 kg. at the end of a 322 mm lever. The average four-panel entry door has 10 similar joints. starting with the mean maximum load and adjusting for additional joints and longer lever of the 900 mm wide door (772 mm to fulcrum at inside of joint)...
1,600 Kilograms x 10/4 {joints} x 322/772 mm {ratio lever
lengths} = 1,784 kg.If a gorilla of larger than 1,784 kg. hangs on the outer corner of your door, the door will collapse. Further, according to the weight/deflection curves on Figure 6, if a gorilla of only 454 kg. hangs on the outer corner, the door will not operate smoothly.
Other Conclusions. While increased surface area of the biscuit joint modestly increases the strength, additional depth increases strength exponentially. A deeper joint engages a wider strip of adjacent wood. a wider strip of adjacent wood better resists bending and breaking.
Woodworkers can deepen a joint by using a larger biscuit as was successfully tested. If their biscuit joiner tool can cut a deeper slot, they could set the biscuit off-center, deeper in the side grained member. the authors saw few failures on the endgrain/biscuit connection.
A preliminary finding based on the thinnest data is that a stronger adhesive makes a stronger joint. that can be viewed as a mystery considering that the glue makes a joint "stronger than the wood itself". If the failure occurs in the wood, how can a stronger glue make a stronger joint? the clue lies at the bottom of the slot where the area of biscuit/slot contact goes to zero.
Between the biscuit and the slot bottom there usually is a gap. Aliphatic resin glue fills gaps poorly. Urethane blue adheres across gaps much better. Whatever the adhesive, the biscuit should first be set at the bottom of the slot in the sidegrain before the endgrain piece is attached.
The four joint sample group showed that the traditional mortise and tenon joint still holds the title, but the biscuit joints showed significant strength. In addition the two four-sided biscuit samples contributed useful data. A comparison of the two shows that simply pre-moistening the surfaces with water(which catalyzes the urethane adhesive) added approximately 30 percent in load capacity. (If only those automotive carburetor water injectors worked that well.) That'll be an all-time economical performance boost if more data confirms the effect.
The natural variability of wood makes certain demands on the research process. In order to accurately quantify the results, researchers need to run more samples. In addition, the flexibility of wood challenges researchers when trying to convert force to torque. Comparative research with smaller or larger samples have an "approximation factor" introduced. It will be a lesser problem on larger and stiffer material and a greater problem on lighter, more flexible constructions. Accurate data on strength is more important for the hard-to-test specimens.
Avenues for Further Research The authors tested just one species of wood. What about a wood that normally tests much stronger in bending/breaking tests- a wood like white oak. What about a wood not thought of as much stronger than pine but has the interlocking grain like mahogany, making it much less likely to split?
Further research should test some joints by holding the long member vertical and aligning the travel of the ram parallel to it. That would increase the effective length of the "lever" (by as much as 30 percent) and it would decrease the compression occurring when applying diagonal pressure. Stabilizing the "door edge" in a vertical position and testing by rotating the arms closed and open could yield some accurate data on the strength of many types of joints.
Leaving the Door Open on Mortise and Tenon The authors found that the "as tested" biscuit joints had as much as 80 percent of the strength of a mortise and tenon joint. Other tests might show mortise and tenon to be far superior to plate joinery.
Authors did not compare biscuits to mortise and tenon joints by "cycling" them repeatedly in the press to, say, half the maximum load or half the maximum deformation and then compressing them to failure.Testing of new joinery also cannot discern the strength of a door corner joint after is has been slammed by a steamed spouse. Impact cycling and testing could generate data about post-shock strength.
Even after it failed, the mortise and tenon joint held considerable loads. when life, property or tradition fit in the equation, the additional expense of the joint may be justified. A portion of continued research efforts could be directed at improving it.
The following additional information is provided as an adjunct to details provided in the above abstract:
It has now been approximately 3 1/2 years since the testing detailed in the above abstract was completed. I have continued to manufacture doors, gates and other products utilizing the methods set forth in the abstract. Additionally, I have conducted tests in my shop on other biscuit products of various sizes utilizing polyurethane glue from various manufacturers.
All of my testing and the success of the products with not one failure at a joinery location has convinced me of the viability of plate joinery and polyurethane adhesive utilized in combination. I have had an ongoing opportunity to examine many of the doors and gates constructed over a more than three year period. Several of the products have been exposed to severe climatic conditions, and all of the products remain intact without damage or joinery faults.
The reader is welcome to e-mail me, cring@concentric.net, with questions or comments.
C.E. "Chuck" Ring
http://www.abq-sfe.com/enchant
January 17, 1998VI. REFERENCES 1. B. Hoadley, Understanding Wood, Taunton Press, Newtown, CT, 1980, pp.111, 124.
2. J.D.Wagner, Fine Woodworking, 111 (4), 58 (1995)
VII. ACKNOWLEDGMENTS The work performed at the Sandia National Laboratories is supported by the US. Department of Energy under Contract DE-AC04-94AL85000
Additional Support provided by Miles Lasner, Albuquerque Hardwoods and Tim Snyder, Editor, American Woodworker magazine.
Copyright©1997-2004WWA