Chalo said:
The most power efficient motors both now and back when I started messing around with these things are still the big slow direct drive hubs, once you factor in efficiency losses due to reduction gearing.
In a 150kg gross weight vehicle, the weight penalty of those is something like 2% of the total compared to most of the lightest alternatives. And even that difference only comes into play when accelerating or climbing.
15kg bike + 85kg me + 2kg battery + ( 4kg vs 11kg ) motor. 6% weight difference and 3% efficiency due to gears.
Now let's climb a 10% hill at 10kph. P = mv sin( hill ).
9.81*106 * 10*1000/3600 * sin(10/100) = 16495 j/s
9.81*113 * 10*1000/3600 * sin(10/100) = 17584 j/s 6.6% more power required.
According the motor sim:
a Q100 500w geared hub at 39% throttle to balance torque against the load, requires ~350w and delivers it at ~28% efficiency.
A Golden 500w DD hub at 59% throttle to balance torque against the load, requires ~410w and delivers it at ~21% efficiency.
So, the 3% efficiency lost through gearing is more than made up for by the combination of: torque multiplication by those gears; rpm being closer to optimum (higher due to gearing); and lighter weight.
Unless your ride(s) consist of predominantly long, flat, high speed runs with no need to stop-start or climb, a lighter geared hub is substantially more efficient even with the gearing loss.
In 1885 Karl Benz built a 500w gas powered tricycle with a single speed transmission. 10 years later, Panhard & Levassor introduced a 3 speed crash box, which remained state of the art until the 1928 Cadillac synchromesh came out.
As someone once said:
Furthermore, the engine provides its highest torque and power outputs unevenly across the rev range resulting in a torque band and a power band. Often the greatest torque is required when the vehicle is moving from rest or traveling slowly, while maximum power is needed at high speed. Therefore, a system is required that transforms the engine's output so that it can supply high torque at low speeds, but also operate at highway speeds with the motor still operating within its limits. Transmissions perform this transformation.
That applies (almost*) equally to small electric motors as is does to the ICEs it was written about. You either go big, or get gearing, and ebikes are all about staying small and efficient aren't they?
Chalo said:
The fact that ferrofluid cooling and supplementary heat sinks seem to appeal to the same folks who use heavy draggy wheels/tires and heavy energy-wasting suspension instead of lighter and more efficient alternatives tells us what this stuff is really about: fetishism. It's hot rodding for the sake of hot rodding, like putting a big turbo on a little economy car engine, or overclocking and water cooling an outdated CPU.
I was agreeing with you that extra cooling (via extra weight and drag) is the wrong solution to the problem.
(*The whole max. torque at 0 rpm is widely misunderstood. Power = torque * rpm. If rpm == 0, power== 0! You need rpms to develop power, and at low speeds, that requires gears. The torque multiplication is a bonus.)