Bending Test, Part II --------------------- Mark Drela 14 Feb 00 After the last structural test, the failed glass skin was shaved off the spar sample, the munched section of the core was fixed, and a new 4-layer skin was applied -- twice as thick as in first test. The intent was to induce compressive failure in the carbon fiber spar caps rather than shear failure of the skin. The spar cap still didn't fail in bending-induced compression, but rather the core got crushed. Nevertheless, useful results still came out. The glider spar sizing is redone below using the larger allowable spar cap stress indicated by this new test. A core crushing criterion is also added, since this is what caused the failure with the new test and may be a concern in real spars. New test description -------------------- Same spar cross section as before 4.5 lb endgrain balsa core: 0.35 in wide x 0.48 in high Precured prepreg CF caps : 0.35 in wide x 0.021 in thick (3 plies) 1.5 oz glass skin (+/-45) : 4 layers, 0.010 in thick total Four load points L = 3.7" apart: F/2 F/2 | | v v ==================X=== ^ ^ | | F/2 F/2 The spar failed very early at F = 190 lb at the same location X as the previous failure. Clearly there was some unrepaired damage left over from the previous test. A 3-point load test was then performed on the "good" part of the piece. The failure which subsequently occurred was not dependent on whether a 3-point or 4-point test was performed (i.e. the data is still valid). F | v ===========X========== ^ ^ | | F/2 F/2 The skin started to buckle at F = 280 lb at location X. But before the skin fibers ruptured, the upper compression-side spar cap got mashed into the balsa core. The buckling was clearly precipitated by the diagonal tension glass fibers (oriented like "/") at the failure location literally pulling the cap into the endgrain balsa core. The compression glass fibers (oriented like "\") were buckled and hence useless. The balsa core then had to take up all their compressive load and gave up. Bending moment could not have caused the sparcap failure, since the bending moment at the failure location was only 40% as large as the maximum center value. Stress calculations -------------------- Cap thickness : Tc = 0.021 in Spar width : w = 0.35 in Net spar height: h = 0.50 in 1 spar cap area: Ac = 0.00735 in^2 = w Tc Bending inertia: I = 0.00092 in^4 = h^2 Ac / 2 Shear skin thk : Ts = 0.010 in Shear skin area: As = 0.010 in^2 = 2 h Ts Max bending moment: M = L F/2 = 518 lb-in Spar cap load : P = M/h = +/-1036 lb Cap axial stress : sigma = P/Ac = +/-140000 psi (at center, NOT at failure) Max skin shear load: S = F/2 = 140 lb Skin shear stress : tau = S/As = 14000 psi Core compression stress: sigC = tau As / h w = 800 psi (average) Two important new conclusions: 1) The previous estimate of 100000 psi allowable compressive cap stress was much too conservative. The sizing example is redone below for the larger allowable 140000 psi stress. The spar is now narrower and lighter. 2) If the shear skin's compression fibers are prone to simple buckling, the core must have sufficient compressive strength to pick up their load. 4.5 lb endgrain balsa should take about 1000 psi (from memory, I don't remember the source). The core here is surely stressed more at the edges where the skin is pulling on it, so the 800 psi average failure stress which was observed makes perfect sense.