Motor with integrated torque sensor

[shemozzle999]

Found this while surfing:

Aliexpress.com : Buy 24V/36V/48V E Bike Torque sensor + Motor from Reliable Torque suppliers on Suzhou New Power Co.,Ltd.

http://minshine.cn/updocuments/standard/XMK_EBikeSystem_Manual_EN_V3.pdf

 

[Deleted member 4366]

Interesting find. I was looking at the kit for which they provide a pedal sensor. Seems a bit superfluous if it has a torque sensor. That brings the next question: where do you get a controller that can use the torque sensor?

According to the specs, the motor and controller can manage 20 amps and 48v, so it should be good for about 25mph.

Aliexpress.com : Buy XMK Electric Bike System Kit (Meter+Controller+Motor+Torque sensor+Speed Sensor) from Reliable Torque suppliers on Suzhou New Power Co.,Ltd.

 

[shemozzle999]

They are looking for distributers:

http://www.minshine.cn/index.asp

Controller details here:

http://www.minshine.cn/index.asp?lang=0&nid=2&aid=110&class=standard

 

[Jeremy]

That does look interesting. The torque sensor seems to be very similar to the ones that are integrated into the chainring on some Chinese ebikes. I bought one a few years ago and found it fairly easy to integrate directly to with an ordinary controller. I just hooked it up to the throttle input on one of the very early Xiechang controllers (the ones that used the Infineon XC846), that had the slow ramp filter on the throttle input.

My guess is that the one in this motor would either connect directly to an existing controller, or might at worst need a bit of filtering to smooth out the peaks and troughs. I'm tempted to try and buy a sample to see how this works, as I've long been after a good torque control system. I've nearly finished building a torque activated controller (one where the throttle controls motor torque, not motor rpm) and a torque sensor input might well prove to be the ideal basis for a more intelligent power assist system.

Here's a link to more info on the torque sensor: http://www.minshine.cn/updocuments/standard/TorqueSensor.EN.pdf

 

[Deleted member 4366]

They are looking for distributers:

Suzhou MinShine Electronic Co.,Ltd.

Controller details here:

Suzhou MinShine Electronic Co.,Ltd.

That's even more interesting: Sine wave controllers - the first ones I've seen on the market. I know Speedict are working on one.

 

[Jeremy]

That's even more interesting: Sine wave controllers - the first ones I've seen on the market. I know Speedict are working on one.

Wuxi Xichen Micro-Electronic Co, the people that really make the Greentime/Hua Tong/KU series controllers, have had a sine controller on the market for a fair while now, as have a few other manufacturers, like Golden Motor, Sevcon etc.

TBH, the advantage of sine wave drive for a typical ebike motor is very tiny in terms of efficiency, perhaps less than 1% at the very best over a conventional 6 step trapezoidal controller, but it does tend to make some motors run a bit more quietly. I doubt this will be noticeable on a geared hub, as the gears seem to make far more noise than the motor, but it may well make big direct drive motors run a little more smoothly.

Most ebike motors don't actually have a sine BEMF, it's usually closer to trapezoidal, hence the very small change in efficiency with sine drive. FOC, or similar, is a better match, as that way the waveform created by the controller is likely to better match that which the motor needs. The controller I'm currently testing uses a variation on FOC to match the applied force vectors more closely to the motors angular position at any instant, and does give a very small advantage in terms of efficiency over trapezoidal. The big advantage of this approach, over both trapezoidal and sine control, is the ability to accurately control torque directly, rather than indirectly via a modified speed control loop.

The main problem with FOC and similar control methods is the need to know, or measure, the motor L and R, and to also measure the motor instantaneous phase current. Fast and reasonably accurate instantaneous phase current measurement needs both fairly expensive (in Chinese controller terms) sensors and a very much faster ucontroller than the majority of controllers use. I suspect this is why the cheaper sine controllers don't use FOC, but just substitute sine drive for trapezoidal drive, which doesn't really give any appreciable advantage.

 

[Ken Taylor]

Wuxi Xichen Micro-Electronic Co, the people that really make the Greentime/Hua Tong/KU series controllers, have had a sine controller on the market for a fair while now, as have a few other manufacturers, like Golden Motor, Sevcon etc.

The Sevcon controller is expensive but do you know where the others are available?

I suspect this is why the cheaper sine controllers don't use FOC, but just substitute sine drive for trapezoidal drive, which doesn't really give any appreciable advantage.

Sine wave controllers seem to be noticeably better at no additional manufacturing cost so I wonder why all BLDC controllers are not sine wave?

To present your argument with a different emphasis. A cheap sine wave controller can use the same ucontroller as a trapezoidal controller, is slightly more efficient and quieter. FOC controllers on the other hand require expensive ucontrollers and sensing for a slight advantage. This advantage could also be gained by measuring the shape of the BEMF for a particular motor design and, using a lookup table, modifying the sine wave in a simpler controller to match this shape.

The big advantage of this approach, over both trapezoidal and sine control, is the ability to accurately control torque directly, rather than indirectly via a modified speed control loop.

Is controlling torque indirectly problematical?