Great article and discussion; thanks. Two comments and a question:
1) All the discussion about whether the strain energy is returned as propulsive force or lost is, it seems to me, sidesteps the main issue–mentioned more than once early on–that any small differences in the crank mechanics is swamped by the messy issue of human kinematics. If one could measure it–difficult, but not impossible; with what level of uncertainty in that measurement is the question–I think it’s quite possible that much of the stored energy IS returned as propulsive force. Maybe some clever researcher could set up a controlled experiment using human subjects and cranks with variable stiffness as well as elastic and inelastic properties and see how the rider’s power output changes at a given HR? One advantage of such a scenario is that you could set it up so that the variations are large enough to detect–i.e., more than 1.6 W, the difference between the most- and least-stiff cranks in this test. OK, have at it.
2) It’s been mentioned that carbon may have damping where metal (e.g., Al) has almost none within its elastic limit. To add to that, carbon is highly anisotropic, and even batch-to-batch (or arm-to-arm, talking about crank production) variations in carbon layup and resin content and distribution can cause significant variations in the stress/strain relationship. A perfect example of this issue is the new Stages crank arm-mounted power meters. Note that you can get them with a SRAM Rival arm (Al), but not a Red arm (CF). Why? Attaching a strain gauge to a CF arm doesn’t produce a consistent stress-strain output. Measuring the damping characteristics is similarly complicated by the material and part-to-part manufacturing variability.
3) Where’s the EE crank?