Making a new pedal assist system for ebike

[mourad96]

Hi everyone, I’m developing my new BLDC controller for Ebike and I want to share with you some of my progress, I have been working on the pedal assist system based on the cadence sensor and I tried to improve it and make a new Pedal assist system.
This new system will scale well to your input, it would measure the rate of spin and not just a simple on-off "switch" thus you can Control motor’s speed by the rate of pedaling, additionally you can adjust some parameters to feel that the ebike is responding to your pedaling effort (only in the 5th gear because I set it to be the reference in the ebike ) based only on the pedal cadence.
so basically, the motor controller will sense how fast you pedal and estimates the power needed to be delivered to the motor in order to maintain that speed.
I made a video explaining the details also you can check the results and the test ride from 9:51min, I want to know what do you think of this new system also if you have a question or any good idea to improve the ebike I would like to hear from you.
Video: https://youtu.be/0kd0oFI06xY

 

[pacikla]

i saw your video, a very cool thing you have there.

 

[ZeroEm]

Good luck with this project. I think about PAS a lot and like using it. Would like a mixture of cadence, torque sensing with gear selection in the mix to come up with an output from the motor. I realism how overly complicated this would be but like thinking about it.

 

[donn]

Interesting. It seems like a somewhat awkward fit with common multispeed bicycle gearing. I can't really imagine how a cadence system could really work to provide a natural control, so despite my criticism, it looks like it's somewhat better.

If I understand it correctly, the rider will find the bicycle effectively stuck in the next to highest gear, because the pedal cadence is matched to motor speed at that gear; a different actual gear setting will change the mechanical advantage, but travel speed will be the same in any case, governed by pedal cadence. I guess a higher gear than the reference gear would more or less cut the motor out, because the pedals would always be turning faster than the control speed.

The electric bicycle rentals commonly seen around here apparently don't have multiple gears, so this wouldn't matter at all.

Regarding the demonstration - it appears that real hills may be scarce in your area, but there might be some kind of overpass or bridge that has a significant grade? Also, I don't recall much attention given to starting the bicycle from a stop.

 

[mourad96]

Thank you guys for your feedback,

donn you're correct the pedal cadence is matched to motor speed at the 5th gear if you choose any other gear the controller will give the same speed as before (in our case gear ratio at 5th cog =2.5 that means for 1 crank revolution the motor will try to spin the wheel 2.5 revolution no matter what gear you're in) . also if you want to make it more difficult for you for training purpose you should choose 6th gear and you will feel a drag from the motor.

Also, there some details maybe not so much people has really recognize it which is the method of control demonstrated in the video at 6:02
in which I said that there is 2 method:
-the second method (which I implemented ): the speed will fall off slightly during high current in case you climb a hill (because of the resistance of the motor) but you will feel more natural riding.
-the first method: the controller will get speed feedback from the motor meaning that if you face a hill the motor will keep adding power until the speed matches exactly your pedal cadence
so maybe I should implement those two together and let the rider choose between them with the help of push buttons

as can be seen in the video I'm using gearless direct drive motor despite that the motor runs smoothly (if there is any jerky movement, I will feel it but I have never experienced that)

For the hill i will try to test it in bridge, it's little far from my area
for the topic of starting the bicycle from a stop: in short answer it will start after half crank revolution.
long answer: there is 8 magnet in the sensor and my system before starting needs to check if it is forward pedaling or backward pedaling and that would need 2 magnet to pass the sensor also there is little delay of 2 magnet because of the filter that will add up and in total it will be 4 magnet needs to pass the sensor to start, if i remove the checking of forward or backward it will need only 2magnet to pass the sensor to start

 

[donn]

so maybe I should implement those two together and let the rider choose between them with the help of push buttons

Of course it's up to you, but I suppose you're right that the sag in motor power under load is natural and acceptable.

I mentioned starting because it seems to me to be one of the weaknesses of cadence systems - a when you'd benefit very much from some motor help, especially if you've been riding around in the second from highest gear, but the pedal control is not asking for it. I guess there's nothing to be done for it.

 

[mourad96]

especially if you've been riding around in the second from highest gear, but the pedal control is not asking for it. I guess there's nothing to be done for it.

I'm sorry. I don't understand that part could you please explain it

 

[donn]

I'm sorry. I don't understand that part could you please explain it

Maybe it matters more to me than to most others, because my bicycle is a recumbent - I can't stand on one pedal and "push off" while I mount the bicycle, I have start in a seated position using only regular pedal power. At this point, the pedals are stopped, so like any cadence system, your system means the controller doesn't call for motor power until I have started the bicycle all on my own. My battery and motor are quite heavy, and this force at slow pedal cadence can be hard on my old knees, so I'm glad I have a thumb throttle instead. This problem is aggravated when you've come to a stop while in a high gear.

In principle, a torque sensor would do much better here, but I've seen reports that some configurations with a torque sensor are also active only when the cranks are moving, maybe because the designer is concerned that the stopped bike may run away due to some false sensor signal.

 

[AussieRider]

I've always thought the logical place for a cadence sensor is on the gear cluster / cassette. Stops when you stop pedaling & matches motor to wheel speed, not pedal speed so all gears can be used. My thinking must be wrong somewhere...

AussieRider

 

[ZeroEm]

by donn » Nov 09 2020 7:14pm

mourad96 wrote: ↑Nov 09 2020 4:04pm
I'm sorry. I don't understand that part could you please explain it

Maybe it matters more to me than to most others, because my bicycle is a recumbent - I can't stand on one pedal and "push off" while I mount the bicycle, I have start in a seated position using only regular pedal power. At this point, the pedals are stopped, so like any cadence system, your system means the controller doesn't call for motor power until I have started the bicycle all on my own. My battery and motor are quite heavy, and this force at slow pedal cadence can be hard on my old knees, so I'm glad I have a thumb throttle instead. This problem is aggravated when you've come to a stop while in a high gear.

In principle, a torque sensor would do much better here, but I've seen reports that some configurations with a torque sensor are also active only when the cranks are moving, maybe because the designer is concerned that the stopped bike may run away due to some false sensor signal.

Your not alone, have an Etrike and like having help to get rolling when stopped and did not down shift. I switch between wanting my throttle to work and worrying someone will grabbing my controls to move the trike or just checking it out. Running 35A @ 72V and the trike being empty they would get quite a surprise or get hurt. At the moment my Etrike needs to be moving .5 kph for the throttle to engage and a bit faster for the PAS to work.

 

[mourad96]

I'm sorry. I don't understand that part could you please explain it

Maybe it matters more to me than to most others, because my bicycle is a recumbent - I can't stand on one pedal and "push off" while I mount the bicycle, I have start in a seated position using only regular pedal power. At this point, the pedals are stopped, so like any cadence system, your system means the controller doesn't call for motor power until I have started the bicycle all on my own. My battery and motor are quite heavy, and this force at slow pedal cadence can be hard on my old knees, so I'm glad I have a thumb throttle instead. This problem is aggravated when you've come to a stop while in a high gear.

yes you're right, that problem cannot be solved only with cadence sensor.
some ebikes add a push button instead of throttle (i think because of regulation) that helps the rider to get to 5Km/h

 

[amwildman]

Can you utilize the speed sensor input as a cross reference? It would add more complexity, but would also give the controller more info to work with.

 

[mourad96]

Can you utilize the speed sensor input as a cross reference?

you mean the speed of the motor,right ?
if that's the case, I made somthing similar to what you suggest, I made some updates on the software and now the controller will calculate the voltage drop that means when climbing a hill, the controller will read the current and estimates the fallen speed and it will give you more power to compensate for that so no need to pedal harder ridding a hill or ridding in rough road to give more power, the controller will try to maintain your speed of pedaling no matter what the situation is

 

[amwildman]

I actually meant the wheel speed.

 

[donn]

In RPM, revolutions per minute?

 

[mourad96]

Yep, reading the motor speed in rpm than converting it to Km/h to get wheel speed