Newb questions about 1st ebike build ??

[donn]

Everyone's different.

Not that different. Sometimes I boggle at the attitudes people have about pedal assist, like it's some kind of religion. Where it isn't enough that it sort of works for common pedaling situations, it has to be some kind of true path for the righteous. I should check to see if off-road internal combustion motorcycles have adopted pedal assist yet.

 

[E-HP]

3kw is lame for many here, who would not even think about controlling the power of their bike with PAS.

Well 4 HP is decent power for a bicycle...but I noticed as power increases, opportunities to use PAS narrows and throttle control becomes more important. I wouldn't use cadence based PAS on trails, especially on single tracks. Even on the street, I'm starting to opt for the direct control of a throttle, with pedaling, rather than PAS, unless it's a long straight stretch. That said, I could see biting the bullet and trying out a torque based PAS at some point to see how it would perform during trail riding.

 

[MadRhino]

Yep, and every build suits someone's requirements.

 

[amberwolf]

I wouldn't use cadence based PAS on trails, especially on single tracks. Even on the street, I'm starting to opt for the direct control of a throttle, with pedaling, rather than PAS, unless it's a long straight stretch. That said, I could see biting the bullet and trying out a torque based PAS at some point to see how it would perform during trail riding.

If your controller uses throttle to control torque/current, rather than to control pwm/speed(voltage), then it'll be easier to control with any method.

But cadence-based PAS is harder to setup for very low pedal speeds, unless you can use a whole lot of magnets on the sensor, to greatly increase resolution. To really work well you might have to be able to set that number of poles in the PAS settings of your controller, but if it has other settings you can compensate for cadence rate then you can still use this method and have it use "accurate" cadence numbers (if that matters for your application).

 

[amberwolf]

I can't imagine doing throttle only with a 3kw cyclone, a bit too much twist and you're flying lol

Is that with a controller using torque/current type throttle control, or with pwm/speed(voltage) control?

The former should give much finer control than the latter, and some controllers that do the former also allow custom throttle curves for even finer low end control. Some throttles are better than others as well.

 

[benend]

I had a Pedego with a twist throttle, and the previous owner had disconnected the brake sensors.
I never wished they were connected. I had full control over the throttle and never felt like I might forget to back off when braking, probably because dropping the throttle and braking are done simultaneously.

 

[MadRhino]

Other than riding high power fast bikes, I have one more reason to avoid PAS. Once doing a road test with a PAS bike that I had fixed for a friend, the first red light was a bad experience. I have the habit (like many other riders do) to back up my crank with the tip of my foot and then place my foot ready for a start when the light will turn green. I dit it a tad heavy, and the bike started on the red light with full power. I did manage to lay it down and avoid any damage, but made my mind against using PAS.

 

[amberwolf]

That won't happen if you're holding a brake lever engaged while stopped, if it has a switch in it to cutoff the ability to run the motor (in one of several possible ways). If your mechanical brakes are good enough, they'd be able to prevent the motor wheel from turning, even without ebrakes.

 

[donn]

That won't happen if you're holding a brake lever engaged while stopped, if it has a switch in it to cutoff the ability to run the motor (in one of several possible ways).

Though conversely, that can lead to a puzzling failure mode if your reflex is to hold the brake lightly at the beginning of launch. Which sounds weird, but I was doing that on my recumbent. And wondering what was wrong with the dang cheap-o throttle.

I'm surprised, though - wasn't it you who reported that even torque sensors don't give you anything until you get the cranks moving? (in the forward direction, I assume.) For perhaps just this reason, I thought - so there's no accidental launch. And I expect cadence obviously works that way. So I'm surprised that a heavy foot at the stop light could cause this.

I guess "track stands" are out.

 

[amberwolf]

I'm surprised, though - wasn't it you who reported that even torque sensors don't give you anything until you get the cranks moving?

That is only on devices that *also* use the input from a cadence sensor to detect that and require it.

A device that only uses the torque sensor itself would not require (or even know about) crank rotation. For instance, something with a torque sensor like BeamTS, that measures only chain tension. Or at least one of the Stromer models, that measures torque on the rear dropout. Or BionX, that has a strain gauge on the axle (I don't know if it has a cadence sensor).



But a torque sensor (with or without cadence), properly setup, shouldn't cause a launch from rotating the crank with no load, only if you took up all the slack and then pushed on it. It wouldn't generate enough signal before taking up the slack to trigger motor activation. (if you push on it hard enough, then it would engage the motor, of course).

However, a cadence sensor, if setup to be sensitive to a small portion of crank rotation, could easily cause this, just taking up the slack in a chain.


A trackstand isn't necessarily incompatible with a torque sensor, but would probably either require an electrical cutoff (button, brake switch, etc), or a cadence sensor that requires more rotation than you would ever do on a trackstand, or a lower limit on the torque input that is above the force required to do the trackstand. Either of the latter would be a problem for anyone that needed to be able to start the motor going with less pedal force than that needed to get going by themselves...but they probably wouldnt' be doing trackstands to start with.

But you could also create a delay requirement--any input to the torque sensor would be delayed before the mtoor acted on it, and if the torque did not stay at or above that level for the entire delay, it would not activate the motor. I don't know how well taht would work in practice, unless you also tied it to the speed sensor such that it only occured at zero speed, and there was no delay above that. It would produce undesired delay in starting up from a stop, though.


A simple cutoff switch is probably the easiest way to deal with the problem, whichever way you are controlling the motor via pedals.