Phew, thanks for all that.
The 10 mph figures seem pretty consistent between riders, but around 20% higher than most power/speed calculation assume (36W)
The higher figures are very hard to pin down, due to this saw tooth profile you mention, which I also get. Power rises and rises even though speed, and presumably drag, etc. are steady, for over a minute. It would be interesting to understand the algorithm Bkool use and therefore the assumptions they factor in. Since my Powertap reliably shows the same affect (with lower figures) I guess the resistance must be changing, and since there is no drag, and speed is constant, it can only be the resistance that the Bkool trainer applies, perhaps something to do with the electromagnets?
I should have picked 25 mph, 30 mph is a bit of an ask! I checked the power/speed figures out with default weights, drag, etc. and no way could I sustain 30mph on a flat road, never mind reach 40 mph, as that needs well over 1,000W, so at the high end it seems clear the Bkool is struggling to simulate the real resistance that would be experienced. I guess this is also true of very steep gradients, so Col de la Madeleine is probably much harder in real life (gulp). Think I'll wait for a howling tail-wind if I ever give any of these alpine climbs a go in reality!
I would guess a velodrome has much lower rolling resistance, as the flooring is chosen for much different reasons to a tarmac road. I suppose they may also assume a different bike is used, with lower aero drag, as is usually the case on a velodrome (no brakes, TT bars)? I'll try my test out in the velodrome, having first disabled the bots!
This is the Saw Tooth effect I get at higher speeds (not at 10mph) - (Y-Axis) yellow is power, green is speed, (X-Axis) time in minutes...
10 mph - no saw tooth
View attachment 76073
20 mph - power keeps rising for around 1 minute before flattening out
View attachment 76074
30 mph - power keeps rising for over a minute, before I ran out of energy!
View attachment 76075
On the trainer Salcombe is a lovely interval for power training. In reality it is deeply scary at c.20% for most of the last half mile and I did exactly as you said, crawled up it at 4 mph with my heart rate right at my max (185) and the front wheel lifting off the ground due to the torque required to keep the 34-28 gear turn over. Would be much better to use a 30 or 32 and get a little more cadence. One for the lightweights, I think. I intend to try it again as soon as the weather improves, and get a video up on the site. Also the climb out the other side of Sidmouth to Peak Hill, which is the same 20% but even scarier due to the amount of traffic!
Having become slightly obsessed with this, I was thinking about it while trying to get to sleep last night, and have a few things to try on setting up the Powertap, which might explain the discrepancies.
Geoff