How a built-in torque sensor works

[shemozzle999]

Courtesy of an Endless Sphere thread:


http://endless-sphere.com/forums/viewtopic.php?f=3&t=46608#p681423

Link to motor manufacturer:

http://www.szxfmotor.com/en/display.asp?id=295

 

[trex]

It looks like a modified spring loaded PAS sensor.
Compared to their normal 250W motor, XOFO charge $100 extra for this - a bit over the top?
As I was reading the other thread (http://www.pedelecs.co.uk/forum/electric-bicycles/14818-me-bike.html), I am wondering if torque sensor compares well against throttle.
You can't adjust the springs to suit individuals, if it reaches its maximum level too easily, it's useless.

 

[shemozzle999]

It appears to me that they have 2 extra hall sensors monitoring the 2 rings of magnets.
I assume they take the output signals from both and feed into a circuit which then produces a linear difference signal, in the range 1v to 3v depending on the crank pressure, which in turn is fed into the controller to control the torque level of the motor for which ever power level has been selected.

http://www.szxfmotor.com/en/display.asp?id=295

 

[trex]

How does it compare against a Thun torque sensor or the autotorq fitted on the Momentum?

 

[shemozzle999]

The Thun and Autotorq measures direct off the BB and gives an constant output independent of which gear you are in, where as this system deflection rate varies with gear changing.

Atmosphere are dealers for the Momentum range, maybe they will have one on demo at the Bristol event so members can try it out.

 

[trex]

So in theory the XOFO output should be better for hill climbing? For example, if I input 10NM at the crank, the Thun would translate it to say 1V, whereas the XOFO outputs say same 1V at 12T at the rear cog, then if I change to 24T at the rear (climbing hill), the XOFO would send out 2V?

 

[shemozzle999]

I theory yes but we are restricted to a 250W system for the foreseeable future and you can't get a quart out of a pint pot.
It is really all down to the bike owner to learn how get the best from his/her bike and try to run it at it's most efficient in all varying conditions.

 

[Deleted member 4366]

So in theory the XOFO output should be better for hill climbing? For example, if I input 10NM at the crank, the Thun would translate it to say 1V, whereas the XOFO outputs say same 1V at 12T at the rear cog, then if I change to 24T at the rear (climbing hill), the XOFO would send out 2V?

Whatever power you get has nothing to do do with what type of sensor you have or where it's located. Power comes from the controller. There's software in there that interprets the sensor and it gives whatever power profile that the programmer programmed into it. As far as I can tell, the torque sensors on Chinese bikes are used as little more than a switch that goes on when you pedal. You have a contril panel to set the speed, and the power algorithm is based on the difference of your actual speed from the speed set on the panel.

The Cycle Analyst v3 can interpret the signal from the Thun torque sensor, and it can convert it to a throttle signal for a normal controller, but I don't know yet what algorithm it uses.

 

[trex]

The xofo torque sensor output is connected to the throttle control of their sine wave controller (they sell the motor with a 15A controller and LCD as a set), presumably programmed to link the output signal to the motor current. Other than that, their kit still requires brake sensors because the torque sensor will not cut the power to the motor.
Being placed at the rear wheel, the signal is multiplied by the gear ratio.
The response curve shows a straightforward linear V response versus input torque.
Perhaps someone will want to compare the benefits/disadvantages of placing the sensor at the wheel. I think the xofo may be better at interpreting the cyclists' wishes.

 

[Arbol]

Up with this thread, cross-linking with:

http://www.pedelecs.co.uk/forum/threads/torque-sensor-bike-with-panieers.17612/page-2#post-217554

 

[Deleted member 4366]

There's so many different types of torque sensor. Some measureccrank torque, some chain tension, some frame stress, some axle torque, plus others. Like I said before, you shouldn't get hung up on them. It won't convert your motor into a Bosch. They use a crank torque sensor plus a crank speed sensor plus a wheel speed sensor plus some very sophisticated software. Whayever torque sensor you have, you need a special controller matched to it. You can't mix and match.

 

[mike killay]

Exactly what is the point of these sensors?
Both my Tonaro and Woosh have magnet rings and throttles. I can give exactly the amount of assistance that I want using the throttle.
After all, it is not as though the motor is 50 bhp

 

[trex]

Exactly what is the point of these sensors?

The output of the sensor (in the case of the XOFO XFL) is a voltage 0-5V, therefore theoretically can be used to replace the throttle signal. Imagine reducing your wiring to just one cable from the LCD to controller, without the need for brake sensors and PAS sensor.
The controller may need to be programmed not to use PAS.

 

[Arbol]

trex, I have asked xofo and they have told me their XFL needs PAS and e-brakes.

When you say "The controller may need to be programmed not to use PAS", how is a controller programmed? Using something called Arduino, or is something different?

 

[trex]

The controller XOFO uses for their kit needs PAS, but common e-bike controllers don't require PAS to be fitted. Some members run throttle only on their kit.