Throttle too sensitive after installing new controller

[eminko]

Hi everyone,
I recently installed a new controller on my electric motorcycle, but now I'm having an issue with the throttle response. It's extremely sensitive — even a small twist causes sudden acceleration, and it's very hard to maintain a steady speed. When I slightly reduce the throttle, the motor almost completely cuts out instead of decreasing speed smoothly.


I measured the throttle output:

• With the throttle disconnected from the controller and powered externally, the signal ranges from 0.8V (idle) to 4.2V (full throttle).
• But when it's connected to the controller, it only reaches 3.5V at full throttle.

This makes the throttle feel more like an on/off switch rather than proportional control.
Would adding a buffer op-amp (like LM358) between the throttle and controller help smooth the signal?
Has anyone else experienced this problem after swapping controllers?


Also, I noticed another issue:
When I'm in 3rd gear and go over 65 km/h, if I suddenly release the throttle, the motor creates a strong braking effect — it feels like engine braking or even a sudden regen brake, and the bike lurches forward sharply.
This only happens at high speed — at lower speeds, throttle release is smooth and doesn't cause this behavior.


Thanks in advance!

 

[fechter]

What kind of controller are you using? Many of them have programmable settings.

 

[eminko]

I don’t know the exact model of the controller. In Turkey, this type of controller is sold as a “performance controller with custom software.” It’s a Chinese-made product. The main microcontroller has “XCM-X32” written on it. The controller has 18 MOSFETs and is rated for 72V, 42A.
Unfortunately, the manufacturer or seller does not provide any information about programming or configuration.

 

[Tipu20]

What kind of Screen you are using there might be manual for that, the manual might allow you to enter into settings menu “sometimes also called P menu” in which you might be able to control the sensitivity.

 

[Chalo]

You could use a potentiometer throttle that sweeps from 0 to 5V, which would give a different feedback than the Hall effect throttle. But not necessarily better.

You could also try a different controller, since that's often the cheapest component in the system.

 

[Philosoraptor]

Second on config. Figure out who makes the controller and reach out for info.

 

[nalu808]

Sudden throttle response, probably throttle setting to close to the throttle resting voltage.

 

[amberwolf]

I don’t know the exact model of the controller. In Turkey, this type of controller is sold as a “performance controller with custom software.” It’s a Chinese-made product. The main microcontroller has “XCM-X32” written on it. The controller has 18 MOSFETs and is rated for 72V, 42A.
Unfortunately, the manufacturer or seller does not provide any information about programming or configuration.

If you could show the *outside* of the controller, including all it's labels and markings, it'd be more likely to help us help you find out something about it, or see if we already know something.

A link to the sales page for it may also have more info.

From what's been posted so far, It is likely that the controller itself has no user-alterable settings, and that it's throttle control only controls the speed of the motor, not the torque. With a high torque system this can cause the problems you're seeing.

One possible way to help tame this is to use a device between your throttle and controller that also measures battery current between battery and controller, and is set up to limit throttle into the controller if battery current goes up too fast or goes above some limit you would setup.

The Cycle Analyst from ebikes.ca can do this but it is rather complex to set up, because it does a bunch of other things too, and you have to be sure they are all set up for the way you want to use it. There was also a device call the Throttle Tamer from ZombieSS here on ES, though it doesn't monitor battery current, just lets you adjust the throttle response curve. I think another device was made by a user here, but can't remember the specifics.

There's a thread about the Tamer's development, but I don't think it has enough detail to replicate it. It used a small MCU and some potentiometers to do the job, but if you're good with component-level electronics design you could do something similar in analog electronics. Probably easier with an MCU and coding if you can do that. (I can't)


Alternately, replace the controller with an FOC type controller that uses throttle-torque-control, and set it up to respond the way you want. Simply because it uses throttle-torque control it would behave better than a throttle-speed-control type system, but most can be tuned for best response for your setup and usage. But, like the Cycle Analyst, you'll have to completely go thru all its' settings and make sure it's correctly setup for your system, motor, battery, etc, as they're not really just plug-and-play type devices.

 

[TommyCat]

Just wanted to re-visit this for a moment to clarify testing...

I measured the throttle output:

• With the throttle disconnected from the controller and powered externally, the signal ranges from 0.8V (idle) to 4.2V (full throttle).
• But when it's connected to the controller, it only reaches 3.5V at full throttle.

Great place to start!

Often when bench testing, a voltage source of 5vdc is used do to its easy availability, and typically what's mentioned for testing.
At WOT a hall sensor throttle will drop ~ .8 of a volt at its output voltage.
So, if you're reading a WOT output of ~ 3.5, its voltage supply is probably 4.3vdc.
The reason for the lower throttle supply that is typically used. Is to allow for headroom in deciding if the throttle is reading higher than the typical output in normal range. Or high enough to trip the controller safety lock-out for a throttle fault. (See bottom chart.)

I also wanted to verify how the signal is ranging from Home to WOT... does it look like this?

A fairly linear rise that starts close to the biggening and ends close to the end of the throttle twist?

If that looks good, I recommend you do a motor actuation test in a stand and see exactly where and when motor is powered. As clarified by this chart... I.E. When the motor starts, and when it reaches full output.

Looking at the results of each test. You should be able to easily determine if the problem is with the throttle, or controller/programming.


For further context and information, see these threads from where these two charts came from...

Guide to Hall Sensor Throttle operation, testing, and modification. - Electricbike.com Ebike Forum
and
Using an On/Off-Momentary Push Button Switch that’s normally open to replace a hall throttle.


Regards,
T.C.

 

[amberwolf]

Often when bench testing, a voltage source of 5vdc is used do to its easy availability, and typically what's mentioned for testing.
At WOT a hall sensor throttle will drop ~ .8 of a volt at its output voltage.
So, if you're reading a WOT output of ~ 3.5, its voltage supply is probably 4.3vdc.
The reason for the lower throttle supply that is typically used. Is to allow for headroom in deciding if the throttle is reading higher than the typical output in normal range. Or high enough to trip the controller safety lock-out for a throttle fault.

FWIW, a "linear" (analog output) hall sensor normally outputs around half it's supply voltage with no magnet field or at "half throttle", and at most could theoretically output it's full supply voltage (depends on the specific sensor).

So the lower the supply voltage, the lower the overall output, and the less change in output for a given mechanical movement. That makes the typical cheap throttle *more* controllable (or rather, it makes it more smoothly able to control a high-torque system) than if it got it's full supply voltage.

There is often a diode inside the controller in series with the 5v supply to the motor halls to protect the 5v supply against induced voltage spikes from the phase currents in the motor phase wires (and from shorts in the motor wiring from cable damage), and sometimes the throttle is powered from this line or from a separate diode-protected line to protect the 5v supply from shorts in the throttle cabling for those systems that have battery votlage (to a meter, or switch, display, etc), in case of cable damage that allows them to short together, so you only lose the throttle itself instead of the whole 5v system and controller.

So the 5v supply drops down to 4.3v or thereabouts, and the throttle votlage range drops by that much, too, which happens to also make the system more controllable....but it does also drop the maximum amount of throttle that can be input into the contorller, so if the ocntrolller has a usable input range that goes up above the actual throttle output, you can't get full output from the controller.

In this particular case, in this thread, that isn't an issue.