What is Startup Torque

[Vihaan]

What is starup torque? I have used the STM32 FOC algorithm using digital hall sensors and the 3 phase current sensing. When ever the 2 wheeler vehicle starts the expectation is it shall accelerate very fast. My clarification is do i need to write a separate algorithm for very high starting torque? I do not understand this.

 

[lnanek]

Wouldn't you need different tech to accelerate very fast? Like a rotary encoder? Hall sensors and measuring back emf are not a precise answer, especially not when starting from a stop.

 

[mxlemming]

You lose about 15% startup torque at most from hall sensors when set up correctly due to the 30 degree angle ambiguity. You shouldn't need separate algorithm for this. There isn't a better algorithm than FOC with mtpa using current control and the best known angle, which with hall sensors at zero rpm is the mid angle of that sensor.

As inanek says, an encoder is superior.

 

[Skyler]

ou lose about 15% startup torque at most from hall sensors when set up correctly due to the 30 degree angle ambiguity. You shouldn't need separate algorithm for this. There isn't a better algorithm than FOC with mtpa using current control and the best known angle, which with hall sensors at zero rpm is the mid angle of that sensor.

Does that mean a sensor less controller is better as no hall sensors ?. I had a 1,000W hub and burned out the sensors and ran it about 2 years until it died with a Greentime sensor less controller. It seemed to have more power but could be the stock controller was 30 amps. The sensor less was 35 amps.

My experience with hub motors is the first hub kits snapped the dropouts especially cast front forks. That is start up torque from my experience. It is too much torque at take off., What caught my eye when I saw this post. The newer hub kits are much better with power assist levels that work like gears.

Even with that I prefer a rear hub motor vs. a front. However if a front hub is used there will be a much greater chance of NOT snapping the front drop out if it is a dual hub motor set up with a rear hub exactly like the front and identical controllers hooked to a single throttle controlling both motors.

With the rear pushing at the same time the stress on the front fork is greatly reduced. Also dual hub motors run cooler as greater heat dissipation. It will greatly increase the life of both motors and controllers vs. a SOLO motor doing all the work. Then there is greater traction and better handling off-road.

My guess is 80 to 90% of the stress on any DC motor happens during start up / take off. With a power meter measuring voltage , current and peak power the start up is the greatest stress on the moving internal parts of a DC motor. Power peaks then and levels off as speed increases. There is half that stress iwith dual motors.

Only draw back is turning at high speed. You need to slow down and let off the throttle when turning. If the front hub is engaged turning will be a lot more difficult. I learned that the hard way. OUUUNCH ? Me and my e bike got a little banged up.. It was crazy.

Instead of a brake sensor someone needs to invent a sensor that shuts down the front hub when the wheel turns at a certain angle. That would reduce the calamities due to human error. Until I invent that just be aware. Dual motors are great. Just slow down when turning. It could save your butt.

I am not supposed to give any advice so only quoting another source from here on ES regarding handling with a front hub.




Thanks.

Skyler.

 

[lnanek]

That's true, I've seen consumer facing controllers ship with a soft start feature enabled for those reasons, forcing users to disable in advanced settings if they want to live dangerous. Old hands here also say equipment like gears lasts longer if you don't accelerate using full throttle. So sometimes the software isn't just trying to achieve maximum theoretical torque

 

[mxlemming]

Man wants high torque from standstill. Telling him don't do that it'll break mechanical parts is unhelpful. Separate problem for a different day. He's on stm32 libraries so he's in a totally different league of development to using AliExpress controllers bolted to a modified bike.

Hall sensors can be interpolated to an arbitrary resolution. Sensorless is not fundamentally better. You trade off wiring, cost, size, reliability performance with various hall vs encoder vs sensorless discussions.

Encoder provides the ultimate in torque and performance but has the most sensitive wiring. Sensorless is most reliable. Hall sensors can provide good starting torque and are cheap but unreliable when embedded in hot zones and have issues with angle offset under load.

Ultimate solution is abi encoder with pwm or SSI for cold start position sensing and backup sensorless for fallback with encoder failure.

Intermediate solution is to interpolate hall sensors and or transition rapidly to sensorless.