If the problem is related to a hall sensor failure then it might not be related to overload. Those little hall sensors can go anytime. Even the manufacturers have a disclaimer written in the data sheets.
http://www.honeywell-sensor.com.cn/prodinfo/magnetic_position/installation/p88781_2.pdfThese products are NOT suitable for protection
against personal injury.
To AVOID personal injury or death, DO NOT
USE these products as emergency stop or safety
devices, or in any other application where human
safety is at risk.
Hall sensors in 3 phase brushless motors are not for protection against personal injury, but the warning does shed some light on why they sometimes fail.
I am contemplating buying 2 of these mini motors one for the daughters bike and a spare.
Not being able to run sensorless is a bummer. Isn’t the free wheel lock and gears outside the realm of motor former and controller.
The only thing that I can relate this problem BikeMad has suggested might exist with mini-motors is the lower current requirements and the magic controller.
My controller I am using ATM is rated at 72v@40amps and mod for 48v. This 40amp thing is more than my HBS36R will draw, I first my sensorless controller to my hub it ran like a noisy unreliable pig. Refusing to take off from 0 start almost 50% of the time. I opened the controller and cut one of the two 20 amp shunt pipes and viola, The bike never fails to take off. Simply matching the current requirement allowed through the controller with motor improved things massively. It’s a little sluggish on take offs, but I relate this to it being a native 72v controller running at 48v. FET gate charge biasing have not been modded and isn’t too hair triggered I think.
And/Or
I was told my controller would not work sensorless on "geared hubs" as it is a problem with digital controllers and MCU clock frequencies unable to provide bandwidth for the sensorless 3 phase phase drive signals from faster RPM geared motors.
Simply put, analogue controller and analogue sensorless integration would be worth a look with geared motors. Or upgrade to a bigger MCU with clock frequencies exceeding 50MhZ.
The latter has been discussed also at endless sphere forums..
Large cooling units or active cooling methods may provide opportunities for one to overclock MCU strobe freq, however this may be may be limited by memory performance. I cant say without looking into the CPU and memory bus config of the magic controller.
(edit:)
"Adding a cooler to an MCU and proving more voltage (>10mV) and faster clock gen will give better MCU performance."
..
I do plan on making my own controller in the next few years but I want it to be analogue. It said that analogue controllers need to populate more components than digital but not if you break down an MCU into its basics components.
Advantage of a digital controller is that a single MCU can replace many components and software replaces adjustability endlessly and this from a manufactures stand point is too valuable. However EV tech parameters are IMO usually refined compared to a lot of applications. A DIY controller is not really a viable and cheap project, but to an avid enthusiast like myself an awesome thing to learn how to engineer.
Their isn’t that much that cant be changed with a few trim pots etc., and if anything one could just omit sensor logic and do this analogue and get fet switching speed to do just this.
