Hub Motor Dynamic Efficiency

[HillCruiser]

If someone can help me with this, it would be much appreciated:
Using an arduino output to control throttle; what is the most efficient way to apply power to a hub motor?
I've noticed when I control the throttle when pedalling normally, that my battery seems to drain faster than if I use PAS.
My PAS, as long as I pedal high revs, seems to settle at a certain wattage.
Does a hub motor run more dynamically efficient if a certain throttle level is maintained, versus continuously modulating it around a control wattage?
I can emulate PAS with a step function; keeping the throttle constant until I need more power and then stepping up say 50w and holding there.
I'm thinking that constant modulation involves small acceleration losses (F=ma).
What are your thoughts?

 

[amberwolf]

When using PAS, are you contributing human power? Is it exactly the same all the time, or does it vary according to the load?

When using throttle, are you contributing exactly the same human power in exactly the same way? (people typically do not pedal the same way when using throttle--they compensate with the throttle for the loading instead of pedalling harder, and this is the most common reason for different power usage in one vs the other.)

If you have a torque-sensing BB modulating a Cycle Analyst, you can verify these things by logging the CA TS output to monitor human watts; otherwise you can only guesstimate (unless you have some other human-power measuring device in place).

(if your arduino is doing all the work, you can log it instead)



Another factor is speed vs terrain vs wind. If all of those are not identical using throttle vs PAS, then loading is different and power usage is different.

Another one is stop/start vs continuous riding. The former uses more power (and even more for higher weight) than the latter, for the same distance, cruising speed, and other conditions. It also uses more power for quicker acceleration from that stop (exactly as faster speed up a slope does).


A motor is more efficient at a particular speed vs the voltage and current and load it is run at; this is typically tested with a dynamometer to generate a set of curves for a particular motor and usage scenario.

But if you are running at a fairly constant speed and load, minor variations in this aren't going to make much difference to average power usage vs keeping it precisely constant.


If you are having problems with throttle output from the CA (or arduino) being non-smooth, or it tracks your input too closely, you can change the PI(D) loop parameters in the CA (or arduino if you have those in the equations) to smooth this out, but it can take a fair bit of tuning to get it to respond the way you want. If you do this, be sure to save your entire CA setup before changing anything, so you can go back to where you were if it doesn't work out or gets "out of control".

 

[HillCruiser]

I was focusing on the motor but the enertia of the entire ebike system as a whole should be taken into account. When the system is at a steady-state condition; constant velocity, constant load, there would be no acceleration factor; so force = mass x acceleration = 0.
But wriggling the throttle back and forth, sets up acceleration/decceleration cycles of the whole system, each cycle requiring a finite amount of extra power above the steady-state condition. Is this additional power negligible? Possibly in a normal situation, but let me explain:

My bike is outfitted now with a burley single-wheel trailer carrying a 48v 75ah battery and solar panel. In addition to the axle load, the ebike itself is weighted down with two motors and a 52v 24ah battery of it's own. Plus I'll be carrying camping gear, food, water. So the inertia of this particular system is getting much greater than the normal ebike. My gut feeling (sometimes better than my calculations) is that throttle variations will add up and drain the battery faster. I'm trying to maximize range and there won't always be an outlet to plug into

 

[amberwolf]

I understand, and that's why I posted what I did.

You can diagnose if this difference in input method does make a difference, and why it happens, if you're logging the right data. It will also help tell you which riding conditions (if any) affect the results and in which way.

Or you can experiment with feedback control loops on one input vs the other and tune them to minimize variations in actual response vs the input. The CA has them built in for the purpose of smoothing response to input so a ride isn't jerky, etc.; they can be used for your purpose just as well.


Note that the greater the inertia of the system is, the less your input variations will cause variations in speed, and most likely the less variation in power usage. (because the burst of power in one instant will be more negated by the decrease in power usage the next instant).

 

[sleepy_tired]

What you really want is a watt-meter that records how much energy your motor controller is using. This way you can log the energy usage do simple math to convert it to watts-hours per mile or kilometer and try different techniques to maximize range.

That way you can eliminate the guesswork and figure out the ideal throttle control for your purposes.

 

[HillCruiser]

What you really want is a watt-meter that records how much energy your motor controller is using. This way you can log the energy usage do simple math to convert it to watts-hours per mile or kilometer and try different techniques to maximize range.

That way you can eliminate the guesswork and figure out the ideal throttle control for your purposes.

I listened to you; now every mile, my onboard arduino/pi calculate, displays, logs and plots the Wh/mile for each motor, the total of the two and the Wh/mile power input of the solar panels; based on the data log of two cycle analysts. I put a bar indicator of the total output and total input right next to each other so I could balance the two. Thanks for the comment because it got me out of a mental maelstrom and back into the flow