New build: intermittent/random pedal assist, speedometer displays motor RPM

brains · Jun 20, 2021

Howdy-
-I've just cobbled together my first eBike and it's about all there but I'm having some issues with the pedal assist function. The pedal assist delivery seems to be random, or, dependent on my cadence. The assist cuts off when I pedal relatively* fast. If I just lazily spin the cranks without actually pedaling the bike, the motor usually gives me full power but If I actually pedal the thing like I would on a regular bicycle the motor isn't interested in helping half the time. The other half the time I get some sporadic power delivery that cuts in and out.
Also: the display isn't giving me the correct speed. The display looks like it's showing me motor RPM's instead: It goes to 0 while coasting, and when the motor surges during it's sporadic PAS power delivery the speed spikes with it.

Some background...
Build is a super cheap 36v Brainpower controller/lcd display with throttle and pedal assist (assist levels 1-5, with 5 seemingly being "fast") paired with a UPP 36v battery and Suntour HESC 250w motor. The controller and related components were all purchased together as a kit. All traces of the suntour HESC system are gone except for the motor. I am using the external speed sensor that came with the kit- I did not connect the white wire from the motor to anything.

-The throttle works reliably 100% of the time. When the pedal assist cuts out, I can still throttle up so I don't believe it to be any sort of thermal cut off or power drawing issue.
-The speed reading isn't correct... I think. It looks like it's reading the motor RPM's- it reads 0kph when coasting, and If I pedal with the brake cut off engaged it reads 0 as well. My max speed according to the display was 56kph, which seems a bit (a lot) on the high side. The speed sensor doesn't seem to be working, but the controller is still getting a speed reading from somewhere.
-The PAS magnet disc is about 2mm away from the sensor and is even. The power assist seems more dependent on me having a super slow cadence for it to be reliable. I can get some power when pedaling normally, but it is typically punchy: the motor revs up and cuts off. The motor doesn't seem to rev past some sort of max- I can get more RPMs with the throttle, so I don't believe it to be a voltage/amperage cut off.
-The controller has a potentiometer that I fiddled with, but I thought i've read other posts where people figured it doesn't really do much.

So, what's up.
Did I plug my PAS sensor into the speed sensor?
Am I getting a voltage/amperage cut off with the pedal assist that doesn't happen while using the throttle... for some reason?
Is this just how the pedal assist works and I'm not supposed to actually pedal the bike and just let the motor do the work? (it's pretty fun and it works great if I just use the throttle as an assist while pedaling).

E-HP · Jun 20, 2021

More information about your components, motor, and controller would be helpful, but off the bat, I'd say your PAS sensor is either not aligned with the disc magnets, and/or you have too much of a gap between the sensor and disc. You should have about a 1mm gap between the two, and the sensor has to align with the magnets as well.

brains · Jun 21, 2021

What additional info might help?
PAS sensor disc is nice and even now, and about 1mm from the sensor.
The motor is a Suntour HESC rear hub motor. 250 watt, 36-volt, BLDC with a 9-pin plug (three phase wires, then: blue yellow green hall wires, white speed sensor wire, black/red wire). I understand the Suntour motor is probably a bafang of sorts, just with a proprietary connector which I have chopped off and soldered on a standard 9-pin julet.
Controller is a "Brainpower" 36v/48v, and I believe 13 amp controller. I don't know if it's a square or sine wave controller. I wanted a sinewave controller but some of the Amazon-special kits don't seem to advertise what they are. The controller's mounted to the bike right now with the information panel hidden unfortunately... but I could cut some zip ties if I had to.

The motor just seems to want to rev up to it's max and stay there.

The speed sensor (external speed sensor mounted to the fork with a magnet on a spoke) has a little light on it that flashes every time the magnet goes by, so it's doing... something.

*Here's something interesting, if I completely unplug the hall sensors (5-wires in a molex connector) I get the same function, so the controller currently doesn't care if the halls are plugged into it or not. That's a bit goofy right?
This controller has a self-learning function. To get it to work, I colour-matched the wires from the motor to the controller, which spun the wheel backwards. I turned the bike off, plugged in the self learning wires, turned the power on which didn't activate the bike but it did spin up the motor in the right direction when I pressed the display's power button, then I think I unplugged the self learning wires and the motor turned the right way.

E-HP · Jun 21, 2021

brains said:
What additional info might help?
PAS sensor disc is nice and even now, and about 1mm from the sensor.
The motor is a Suntour HESC rear hub motor. 250 watt, 36-volt, BLDC with a 9-pin plug (three phase wires, then: blue yellow green hall wires, white speed sensor wire, black/red wire). I understand the Suntour motor is probably a bafang of sorts, just with a proprietary connector which I have chopped off and soldered on a standard 9-pin julet.
Controller is a "Brainpower" 36v/48v, and I believe 13 amp controller. I don't know if it's a square or sine wave controller. I wanted a sinewave controller but some of the Amazon-special kits don't seem to advertise what they are. The controller's mounted to the bike right now with the information panel hidden unfortunately... but I could cut some zip ties if I had to.

The motor just seems to want to rev up to it's max and stay there.

The speed sensor (external speed sensor mounted to the fork with a magnet on a spoke) has a little light on it that flashes every time the magnet goes by, so it's doing... something.

*Here's something interesting, if I completely unplug the hall sensors (5-wires in a molex connector) I get the same function, so the controller currently doesn't care if the halls are plugged into it or not. That's a bit goofy right?
This controller has a self-learning function. To get it to work, I colour-matched the wires from the motor to the controller, which spun the wheel backwards. I turned the bike off, plugged in the self learning wires, turned the power on which didn't activate the bike but it did spin up the motor in the right direction when I pressed the display's power button, then I think I unplugged the self learning wires and the motor turned the right way.

I'm going to assume it's the 866 display (please confirm). What are your P11 and P12 settings?
The speed dropping to zero seems to indicate the external speed sensor isn't being used to calculate the speed, but using the motor magnets, and that your motor is a geared hub, therefore no speed reading when the motor (as opposed to wheel) isn't turning. What are P6 and P7 set to?
The motor working without the halls indicates you have a dual mode controller, sensored and sensorless operation.

brains · Jun 21, 2021

The controller model wasn't advertised in the listing I got it from, but it looks and acts like a Jp-lcd1 model. 4 buttons: power(also selects, toggles) up arrow, down arrow, and a "C". I've got the speed sensor working by setting the P3 (speed detection signal in the manual) to 0, but I've still got the choppy pedal assist. It just really seems dependent on my cadence- pedal too fast, assist cuts out. Slowly turn the pedals so as not to actually provide any input, motor provides power up to a certain point that seems to be dictated by what assist level I'm at.
-Is that just normal operation for a magnet ring PAS bike? I assume the assistance cut off should be dictated by speed, not by my cadence.

I found out the potentiometer has the potential to make the throttle not work if you wind the screw out a lot, so there's that exciting data point for anyone interested.

E-HP · Jun 21, 2021

brains said:
The controller model wasn't advertised in the listing I got it from, but it looks and acts like a Jp-lcd1 model. 4 buttons: power(also selects, toggles) up arrow, down arrow, and a "C". I've got the speed sensor working by setting the P3 (speed detection signal in the manual) to 0, but I've still got the choppy pedal assist. It just really seems dependent on my cadence- pedal too fast, assist cuts out. Slowly turn the pedals so as not to actually provide any input, motor provides power up to a certain point that seems to be dictated by what assist level I'm at.
-Is that just normal operation for a magnet ring PAS bike? I assume the assistance cut off should be dictated by speed, not by my cadence.

I found out the potentiometer has the potential to make the throttle not work if you wind the screw out a lot, so there's that exciting data point for anyone interested.

For the PAS, check the alignment of the PAS sensor relative to the magnets on the disk. They have to be aligned and the distance between the sensor and the disk magnets should be around 1mm (as close as possible without touching).

The speed sensor is a separate issue. What is P1 set to?

docw009 · Jun 24, 2021

I initially bought a Brainpower 13A controller in 2017. The PAS was horrible. Then I found that I had bought their model 1 controller where it was supposed to be horrible. Any pedal movement would give max speed. I bought the model 2. PAS worked, but a 13A ebike is not much fun. So I scrapped that one too. I installed a 20A KY controller. Part of my learning curve. Wasted 80 bucks on Brainpower dreck.I didn't learn though. Bought a "never used" Brainpower controller with an S866 disolay. The latter worked, but the controller didn't. Got a refund. By the way, the S866 has the cheapest buttons I've ever pressed on a display.

ALl I can say is maybe you bought the wrong model of the Brainpower.

brains · Jun 26, 2021

an update just in case anyone has a similar problem-
-we tried changing the magnet disc from a 8-magnet to a 5-magnet and that made the pedal assist function much better. The motor doesn't turn off at my normal cadence anymore- it reliably provides assistance up to a speed determined by the assist level. The pedal-assist assistance isn't proportional to my cadence, it's more of an on/off switch up to a certain speed.

I would guess that at my normal cadence maybe the magnet sensor is seeing the magnets go by too quickly with the higher-magnet-count disc and it just assumes, "well there's a magnet here all the time... they must not be pedaling..." and it turns off.

There is no cut off at any reasonable cadence any more, so I'm happy with it. So the build was:
Amazon special ("B19teenteam" branded this time) controller and component kit. Jp-lcd1 display, full twist throttle, 5-magnet PAS sensor, 36/48v sensor or sensorless controller (13a) The controller is set to the factory defaults, except the speed input signal is set to "0" so the external speed sensor works instead of the motor's internal sensor.
Motor is a Suntour HESC 250w 36v (this bike is an izip e3 SIMI, but the motor shows up on the Carrera vulcan and some Raleigh ebikes where the sountour controller, battery, and torque sensor are rubbish). We've cut off the higo z919 motor plug and soldered on a more common julet 9-pin connector.
Battery us a UPP 36v super-shark style thing.
Tonight was my first ride on it with the now-working set up. Assist level 1 gets me up to 12mph, assist level 5 (the highest) tops out at around 21mph.

longlegs · Sep 24, 2021

So I guess the number of PAS magnets should coincide with the number of magnets you can find in the settings?

New build: intermittent/random pedal assist, speedometer displays motor RPM

brains

1 µW

E-HP

10 GW

brains

1 µW

E-HP

10 GW

brains

1 µW

E-HP

10 GW

docw009

1 MW

brains

1 µW

longlegs

100 µW

Similar threads