gvcormac wrote: ↑August 5th, 2022, 9:02 am
My understanding is that the PMx monitor underestimates the power of your first several strokes, because it uses some default parameters for drag. So in trying to optimize your performance, you are actually trying to do two things, which may be incompatible: maximize your actual energy output, and maximize the extent to which it is captured by the PMx.
Your completely right on that one, for more reasons than one. Two things happen in the PM5 in the first stroke: the first indeed is that the PM5 doesn't have a dragfactor yet, so it assumes a default for the time being. I haven't detected if it corrects this retroactively once the first valid dragfactor comes in (which could be done technically). As such correction tends to be in the centimeter range, one wouldn't be able to detect it with layman's equipment, but you would notice it in short sprints in a 100 meters. However, everybody has the same handicap, although my estimate is that the PM5 assumes a default DF of 120, so when you row your sprint on a DF of 225, you might get a disadvantage.
Another thing that happens, especially at the start, is that C2 uses incomplete formula's for its power calculation (and thus for speed and distance), to keep things simple. University of Ulm detected this when they hooked up a robot to the C2. The PM5 assumes that the power can be calculated using the cube law based on the flywheel speed and uses this to calculate linear speed and distance (i.e P = k*ω^3 = c * u^3). This works well when the flywheel speed is more or less the same at the same point in the stroke. However, the complete formula is P = (Inertia * ( dω / dt ) * dθ + k * ω^2 * dθ) / t, which can be transformed in P = k * ω^3 + Inertia * ( dω / dt ) * ω. The second part of this formula has an additional factor "Inertia * ( dω / dt ) * ω" is especially important at the start, as dω /dt (i.e. angular acceleration) is considerable but isn't taken into account in the PM5's calculations. As the team in Ulm concluded, the effect is that the PM5 "rewards" stroke-to-stroke consistency and really "punishes" stroke-to-stroke inconsistency, but also is pretty bad at estimating the speed and distance when the flywheel still accelerates from stroke to stroke (i.e. the first three to five stroke at starts).
As I see it, to be as efficient as you can on the PM5, you need to get to a steady flywheel speed as soon as possible. So exploding at the start to get that flywheel up to speed would be most effective from theory. Whether people can do that, I guess it would depend on the person holding the handle....
