Torque sensor v speed sensor

[Kudoscycles]

Most of the crank drive bikes I have ridden use a torque sensor whereas most of the hub drive bikes use a speed sensor. But there seems no definite reason for this. I understand the theory that a torque sensor effectively replicates the natural riding style but adding some assistance whereas the speed sensor provides power controlled manually via the PAS selector and therefore is a less natural power addition.
To be specific do members think on a high end value hub drive bike they would expect a torque or speed sensor?
Do members think on a high end value crank drive bike they would expect a torque or speed sensor?
What are the implications of a throttle in all applications?
Dave
KudosCYcles

 

[flecc]

I think having a torque sensor is by far the best control method, but it's not too easy to implement on hub motor bikes. However, the TMM type of rear frame pressure sensors do seem to be answering that need now after a number of successful inclusions such as on the newer Giant models. Most agree they now are near enough to matching the best crank drive ones.

Of course the two things are not alternatives, they fulfill different functions and can be used together and are on some crank units. For example, the original Panasonic unit used a magnetic flux torque sensor and separately, a ring of magnets feeding a Hall sensor for speed regulation.

A throttle really only applies to speed sensor only bikes, since the torque sensor is a foot operated throttle so it's not necessary to duplicate it, although a throttle can be added as an override.
.

 

[Kudoscycles]

Flecc, I was confident you would be the first response.
Have you any thoughts on the THUN BB torque sensor on a hub drive bike? Surely the throttle is a useful addition to tired legs which can override the torque sensor and provide a quick access to max power,when say performing a hill start?
The Tonaro seems to be the only crank drive bike which uses a speed sensor and throttle and yet that combination is considered an ideal combination from many Tonaro users?
Dave
KudosCycles

 

[Eaglerider]

It would be ideal to have both.

A torque sensor would feel more 'natural', but I would want the option to switch to speed sensor, or at least the ability to select maximum power regardless of torque applied to suit what I want at the time. If the system would only provide full assistance at high pedal torque, that would be a deal breaker for me.

If my knee decides to have a tantrum when 15 miles from home, I want to be able to take the easy road home!

 

[mike killay]

Pedelec magnet ring and torque sensors do much the same thing.
However, the magnet ring can only sense that the pedals are turning and generally supplies a fixed amount of power by means of the power selector switch on the handlebars (generally with 3-5 power level settings)
So, with a magnet ring, if the power level selector is at maximum, the full power will be delivered as soon as the pedals turn, even if they are 'under pedalled' (ie going slower than the cogs require)
With a torque sensor, the amount of power delivered is proportional to the amount of pedal effort. The amount of power available is also usually controlled by a selector switch.
But as others have observed, if your legs are tired, you will not get full power even if the selector is on maximum.
This is fine for the fit and healthy but utterly useless for the geriatric and arthritic etc.
The Tonaro in reality is a very low powered moped.
It has a magnet ring that can be removed if not wanted. The magnet ring operates through the usual selector switch. It also has a throttle and because it drives through the gears it will go up modest hills without any pedalling. For the steepest hills, light pedalling is all that is required.
So, it is a matter of choice. I suspect that the fit amongst us will prefer a lightweight bike that feels like cycling with unobtrusive electrical assistance when required.
Others will prefer the Tonaro style of controls.
That is why hub motors are at a disadvantage. However if the EU proposal to remove wattage limits is adopted by the UK, the whole scene will change and hub motors may well come into their own.

 

[flecc]

I don't have any personal knowledge of the Thun torque sensor performance Dave, but it's spec shows a good range of torque measurement and the combination of torque and rotation sensing within the bottom bracket is attractive, being water and damage proof.

As for throttles, having only the throttle, with or without rotation sensing, can be tiring after a while on longer trips. This is especially true of Hall throttles which tend to be fast acting so difficult to hold in an intermediate power setting, not much better than an on/off switch when riding on bumpy roads. There's no doubt that a good torque sensing bike is far more restful to ride over long distances, one just pedals and lets the bike sort out the technicalities. Of course having an override throttle together with a good torque sensor is probably the ideal, taking care of all circumstances. However, it adds to the cost and investment in this might not be a good idea until we see the nature of the new legislation due within about a year. That could well ban throttles for the UK, since that has long been the EU position and there doesn't appear to be much call for throttles in Europe where they've lived without them for at least 9 years now.

 

[Jimod]

I can only speak for myself of course but for me a torque sensor would be a no no on an electric bike. Part of the reason I'm so fond of my present bike is that as long as I'm turning the pedals, no matter how lightly, the power is all there for me. In my area I am either going up hill or downhill with a few short bits in between. It's on those short bits I find it pleasurable to sit back and let the bike do what it was made to do. If I needed to apply torque to make it go then I couldn't enjoy my few bits of level road.

Now most times I'm out I am pedaling hard as I like to go fast, OK 15mph, However even on a nice hard fast run there are times I like to pedal slowly and let the motor carry me along.

As to a throttle, I always thought I needed one but I don't with my bike as it works at full power with the pedals barely moving. With a torque sensor a throttle would be a MUST for me. If I wanted to be applying power from my legs all the time I'd be as well getting a normal bike.

If you take the throttle out of the equation and asked the question in a different manner. Ask it like this.
No matter whether you prefer a torque sensor or a pedal sensor would you buy a bike if it was only available with a torque sensor? Then ask, would you buy a bike if it was only available with a simple pedal sensor?

If you went to a group of people who didn't own an e-bike and ask those questions I think the pedal sensor people would be the majority.

 

[Kirsop stanhope]

I can only speak for myself of course but for me a torque sensor would be a no no on an electric bike. Part of the reason I'm so fond of my present bike is that as long as I'm turning the pedals, no matter how lightly, the power is all there for me. In my area I am either going up hill or downhill with a few short bits in between. It's on those short bits I find it pleasurable to sit back and let the bike do what it was made to do. If I needed to apply torque to make it go then I couldn't enjoy my few bits of level road.

Now most times I'm out I am pedaling hard as I like to go fast, OK 15mph, However even on a nice hard fast run there are times I like to pedal slowly and let the motor carry me along.

As to a throttle, I always thought I needed one but I don't with my bike as it works at full power with the pedals barely moving. With a torque sensor a throttle would be a MUST for me. If I wanted to be applying power from my legs all the time I'd be as well getting a normal bike.

If you take the throttle out of the equation and asked the question in a different manner. Ask it like this.
No matter whether you prefer a torque sensor or a pedal sensor would you buy a bike if it was only available with a torque sensor? Then ask, would you buy a bike if it was only available with a simple pedal sensor?

If you went to a group of people who didn't own an e-bike and ask those questions I think the pedal sensor people would be the majority.

 

[Kirsop stanhope]

I have two bikes. An oxygen which has a throttle and a Kalkhoff proconnect 10 which doesn't. Due to treatment for cancer my legs have become incredibly weak and I desperately need something to override the torque sensor to enable me to get up our 17-20% hills. The oxygen is easier to ride but doesn't have the power to get me up these gradients easily. Does anyone know if there is any way of overidiing the torque sensor on the Kalkhoff?

 

[trex]

I don't think you can.
The only crank drive bike with the power of the Proconnect 10 and with a full throttle is the woosh Krieger.

 

[anotherkiwi]

My understanding is that the torque simulation PAS controllers are able to read both cadence (PAS reads cadence not speed) and speed from the speed sensor and have an algorithm to simulate pedal torque. The more magnets you have the more fine grained is the cadence data giving better dosed power output. From my experience, having moved from 8 to 12 magnets, this is true - smoother start-up and power delivery.

I think that many older square wave controllers are more like on and off switches - the PAS detects the cadence and supplies the quantity of power set on the assistance level without much modulation. Depending on how it is set the PAS has to read a certain number of magnets going by before turning on so you have to turn the pedal further the less magnets your disc holds. Some disks only have 5 magnets meaning you have to turn the pedals more than twice the revolutions to have the same number of pulses!

The above differences seem to explain why people are so adamant that torque sensor is superior to PAS. It clearly was at one point of time but as controllers have evolved with more sophisticated software this may be becoming less and less true.

 

[Deleted member 4366]

My understanding is that the torque simulation PAS controllers are able to read both cadence (PAS reads cadence not speed) and speed from the speed sensor and have an algorithm to simulate pedal torque. The more magnets you have the more fine grained is the cadence data giving better dosed power output. From my experience, having moved from 8 to 12 magnets, this is true - smoother start-up and power delivery.

I think that many older square wave controllers are more like on and off switches - the PAS detects the cadence and supplies the quantity of power set on the assistance level without much modulation. Depending on how it is set the PAS has to read a certain number of magnets going by before turning on so you have to turn the pedal further the less magnets your disc holds. Some disks only have 5 magnets meaning you have to turn the pedals more than twice the revolutions to have the same number of pulses!

The above differences seem to explain why people are so adamant that torque sensor is superior to PAS. It clearly was at one point of time but as controllers have evolved with more sophisticated software this may be becoming less and less true.

That's not correct. Torque simulation is Chinese for current control.

With the old type of controllers, the levels on the LCD/LED display relate to speed, i.e. they represent fixed throttle positions. As an example, level 1 might be 8 mph, above which the controller won't give any power. Below it, the amount of power given is based on an algorithm that increases it the further you are below the target speed up to the maximum power. That means that at take-off (zero speed), you get maximum power. In this context, power is current.

With modern controllers (torque simulation), the power fed to the motor is constant for each level. As an example, level 1 might be 60watts (=1.5 amps), which doesn't change with speed, except that when you reach a certain speed, depending on which motor you have, the back EMF from the motor will start reducing the current until zero at maximum rpm.

The two systems have a different feel. The speed control ones, give a high initial torque, while as the current control ones give a constant torque, which would normally be much lower at take-off than the speed control system. Some people find that a lot more controllable and less scary.

 

[Alan Quay]

That's not correct. Torque simulation is Chinese for current control.

With the old type of controllers, the levels on the LCD/LED display relate to speed, i.e. they represent fixed throttle positions. As an example, level 1 might be 8 mph, above which the controller won't give any power. Below it, the amount of power given is based on an algorithm that increases it the further you are below the target speed up to the maximum power. That means that at take-off (zero speed), you get maximum power. In this context, power is current.

With modern controllers (torque simulation), the power fed to the motor is constant for each level. As an example, level 1 might be 60watts (=1.5 amps), which doesn't change with speed, except that when you reach a certain speed, depending on which motor you have, the back EMF from the motor will start reducing the current until zero at maximum rpm.

The two systems have a different feel. The speed control ones, give a high initial torque, while as the current control ones give a constant torque, which would normally be much lower at take-off than the speed control system. Some people find that a lot more controllable and less scary.

Torque simulation is what I'm after for my light assist bike.

Does it use a speed sensor, or is it just constant current?

Thanks for the explanation.

 

[anotherkiwi]

That's not correct. Torque simulation is Chinese for current control.

Thanks for the lesson d8veh - don't ever believe what one reads on Chinese web sites I should have guessed the software wasn't as sophisticated as they said. It could be which is a shame. Does the Speedict work as advertised?

 

[Deleted member 4366]

The Speedict is a pretty good product. The basic things like speed, distance, watts, battery info. mapping, logging all work very well and are very useful for understanding how your ebike works. With a Speedict, I could see how the torque simulation controller worked. The power levels are absolutely flat as you can see in the image below. You can see it cutting the power at 15 mph too. I'm not sure about the other functions, like speed limit. I think they probably work as long as you use the special throttle cable that's an optional extra. I tried it without the cable, but it wouldn't hold a constant speed. It's because the throttle output from the speedict is PWM, which controllers don't like. I think the special cable has a low pass filter in it to smooth the signal into a simple DC level.

 

[Deleted member 4366]

Does it use a speed sensor, or is it just constant current?

Thanks for the explanation.

It uses a normal pedal speed sensor, so that it only gives power when you're pedalling. The sensor only gives the controller the information that the pedals are turning.

 

[Tomtomato]

Does anyone know if there is any way of overidiing the torque sensor on the Kalkhoff?

I don't think you can connect other controls to the Impulse 2 motor. However, you can ask 50 Cycles to change the settings from "standard" to "Dynamic", which will make the bike require less inputs from you, while delivering more assistance.

 

[Cyclezee]

Having tried lots of different systems my personal preference.....at the moment..... is for a torque sensored hub motor.

If we are talking about the BB type my brand preference.....at the moment..... is for TDCM rather than Thun, that preference is based on cost and performance.

If we are talking integrated within the motor hub then it would be the Ansmann RM 7.0 DD motor.

 

[anotherkiwi]

What controller would you marry that to?