Yesterday, I narrowly avoided what I now believe would have been an engineering disaster. I had planned to repair the frame ends by replacing them with parts machined from phenolic plastic sheet. Yesterday morning, I awoke with a vague worry about its strength in applications subject to shock and vibration while under tension; I undertook a brief experiment to allay my fears.
Introduction:
Oak has been used in marine construction since before America was founded. The scantlings of modern ships and the intuition of boatwrights are based on their experience with its properties; the plans for RUNE specified that its ribs be built of oak steamed and bent into shape.
Since the second World War, plastics have been applied to diverse engineering situations. Phenolic resin is reinforced with cloth, formed into sheets under pressure and heat, and available for easy machining with woodworking tools. Its strength is measured with standardized tests, and its properties are predictable. The tensile strength cited in the literature is comparable to that of oak, and because it does not have a favored grain orientation, heavy phenolic sheet should be ideal for making the curved shapes of RUNE's frames.
On the other hand, its thermoset resin seems stiff, and I worried that it might be too brittle for the varying loads applied to a boat hull. Oak can respond to cyclic and impact stresses without catastrophic failure; is the same true of phenolic sheet?
Therefore the null hypothesis is that parts made of phenolic resin sheets are comparable in strength and fracture toughness to oak.
Procedure:
I cut the same shapes in oak and phenolic. The shapes resembled the cross-section of an I-beam, with 2 inch by 4 inch rectangles at the top and bottom, connected by a 6-inch stick, with a cross-section 3/4 inch square. The three zones blended into each other with fillets of 1.5 inch radius.
I drilled two 5/16-inch holes in the top and bottom rectangles to connect them to chains. The chains were shackled to nylon straps and the linked arrangement was used with a fork lift to raise a large oak timber an inch or two above the ground.
After the timber was raised and lowered, I drilled a one-eighth inch hole through the square section, and raised it again. While the timber was being held by the test shape, I tapped sharply on the side of the piece, aiming for the side of the central stick near the eighth-inch hole. I then lowered the timber and drilled another hole in the same plane.
I repeated the cycle of drilling, raising, and rapping until the piece broke.
Results:
I first tested the phenolic piece. I drilled a first hole front to back at the mid-line of the stick. I then drilled side to side above the first hole. Third, I drilled a second hole, front to back about an eighth inch away from the first, and parallel to it. After the third hole, the phenolic ruptured abruptly, in the plane of the two parallel holes.
I then tested the oak piece. I raised and lowered the timber, drilling first front to back at the middle plane, then side to side above it, then I returned to the middle plane. I drilled four parallel holes at the mid-plane, and then two side to side holes through the same plane, before the oak splintered and parted.
Discussion:
The mass of the timber, sized 13 inches by 15.5 inches by 147 inches, was calculated to be about 802 lb using a density of 46.8 lbf per cubic foot. If the tensile strength of the phenolic were 6000 psi, it should be able to lift a load of 3375 pounds at its full cross section. A wide range of strengths are reported for oak, but it should also be able to lift the entire timber at its full cross section.
Calculating the effective cross sectional area is difficult for holes that are not co-planar. In this discussion I calculate the cross sectional area by assuming that the fibers do not adhere to each other at all, perpendicular to the tensile stress. The area is calculated by projecting all the holes to a single plane perpendicular to the pull.
The cross section of the phenolic test piece after three holes were subtracted was 0.312 square inches. Since the sample broke when the area decreased to that point, the maximum tensile strength for the phenolic sample would be 2570 pounds-force per square inch.
The equivalent calculation for the oak test piece yields a cross-section of 0.094 square inches, from which a maximum strength of 8560 lbf/sq. in may be inferred.
Conclusion
Considering the uncertainties in the testing method, it can be concluded that the tensile strength of phenolic laminate, when subjected to impacts, is about a third that of oak. Since the design of this boat is predicated upon the characteristics of its wood frames, it would be unnecessarily risky to substitute the phenolic sheet.
Without having repeated the tests, I cannot infer the accuracy of these measurements.
Acknowledgements:
All the sample preparation and forklift operation was done by Roger Hambidge; I received significant suggestions for safe operating procedures from Wade and Joel, and the debate with Dave Snediker focused my thoughts on the need for this test.