Finding a good controller

[badboy1999]

Because the biking computer in my bike store pedelec failed, I'm now running my bike with a $15 ebay 36v controller. It works allright, but the throttle is not throttle like in a car (where you adjust the amount of air going to the engine) but sets a target speed, so for a nice tailwind-like experience you really have to keep the throttle in an exact position relative to your speed, where it's on the border of not doing anything at all. Usually I just shove it to 100% and then 'pulse' the throttle.

This isn't the greatest experience and I fear it's terrible for my battery. What controller should I buy that works differently? Or maybe the easiest solution would be to modify my controller somehow? It's hard to find what exactly the throttle does on ebike controllers. Ideally I'd want pedal assist back, but I'm not sure how that's implemented and can't seem to find it either.

If someone could point me in the right direction, that'd be greatly appreciated!

 

[badboy1999]

Thanks for the extremely helpful reply!
The Grin Phaserunner seems to be $300, which is a bit expensive for what I want, which is just 1 additional feature.
The cycle analyst looks very interesting, but it's $130 with $40 shipping so that's also a bit much.
The lebowski look really cool, but it's huge and a lot of work.

I'm currently attaching a arduino (that will receive the throttle signal, and shunt signal) to my ebike with the following program flow:
if the measured current is lower than the target: output a logical 1 to the controller.
else: output a logical 0 to the controller.

I think this will lead to a nice control loop with the current flow as a proxy for power usage of the motor. I hope I'm not overlooking something that will lead to oscillation . Anyhow it should be finished today so I'll know for sure then.

 

[crossbreak]

Adappto and Kelly,too

 

[crossbreak]

This on/off logic wont work. Throttle response is much too slow...maybe 10Hz or so..this will oscillate badly. You need real PID control. Or a lookup table that is defined carefully. The lookup table will have much faster throttle reply but is a lot more work to set up I think

The PID thing you can simply implement by downloading the corresponding library

What PID does is the same as the other controllers do. Only difference is that these have this thing implemented internally so the throttle response is much faster. Which leads to much better results, i.e. faster response. Additionally, with bad PID values the throttle may a) oscillate or b) response will be very slow. Have a look into the cycle analyst manual to find good PID values description and how to set them right for an ebike

 

[crossbreak]

however, if you want a real torque throttle, you need to measure RMS phase current. So this is what it should look like:

Hall current sensor ->RMS DC converter ->Arduino ->motor controller

the hall current sensor could be an allegro 50A model for example
the RMS converter could be a AD737, not sure if this chip would work, as the allegro has an offset output. So 2.5V correspond to 0 amps, while more or less than 2.5V would be positive or negative current, which has to be rectified somehow to get an RMS value from it

but maybe you are fine with the PID battery current mode throttle. just try it...did that too and it was fun But at the end of the day I bought a kelly controller :lol:

 

[badboy1999]

Hah, I'd be perfectly fine with a PID battery current mode throttle, I'm only a biologist not a rocket scientist
I'm new to this PID thing. Took me a while to get it (especially the differentation term of the equation), I couldn't get the PID library working correctly, I'll try again next weekend.
Thanks also for the Kelly Controller tip. I thought I'd researched them enough and that they were expensive but I found a $60 controller that should work swimmingly with my setup:
http://kellycontroller.com/kbs36051x25a24-36v-mini-brushless-dc-controller-p-501.html
I'll probably buy this one next weekend if I can't get the PID working.