6-sensor technology

[Gionnirocket]

My ebike motor has sensor sensors. These sensors sense if any of the other sensors are not sensing correctly and then adjust that sensors output to the proper sensed output range.
I have to say... Overall, very sensible.

 

[Jeremy McCreary]

Torque is measured by the controller to deliver the proportional assist power. I can't figure out how the cadence is used? Please share your know how.

In the Turbo PAS Specialized uses in its mid-drives, the primary control on motor power is rider power Pr measured at the crank with a dedicated power sensor.

True cadence (not just yes-no pedal rotation detection) enters this PAS both directly and indirectly. The indirect route comes via Pr according to

Pr = k C Qr,

where Pr is in W, C is cadence in RPM, Qr is rider torque in N m, and k is a unit conversion constant.

But Specialized explicitly states that the PAS also relies on wheel speed and direct cadence sensors. This data likely adds finesse to the primary power-sensing scheme — especially in transient situations like starting out.

You can explore the basic operation of this PAS with the interactive graphical model I presented in much more detail here:

Visualizing and exploring the power-sensing PAS in Specialized mid-drives — an interactive graphical model

For our more math-minded Specialized mid-drive riders, here's an interactive graphical model of your bike's tunable power-sensing pedal assist system -- hereafter, just "PAS". The model doesn't cover complex PAS responses to transient events like starting out or approaching PAS cut-off speed...

forums.electricbikereview.com

As I noted in your earlier thread on this subject, I have many miles on a power-sensing Specialized Vado SL mid-drive and 2 different torque-sensing hub-drives.

The hub-drives' torque-sensing implementations are quite good. Yes, both have slight lags starting out. And both surge very slightly with pedal position in certain uncommon situations. But motor power delivery is proportional and very natural overall.

The SL's power delivery, on the otherhand, is absolutely flawless in all situations — and even more natural. This is an easily observed fact, not theory.

So, time to accept reality: Rider power and true cadence sensors can be used to very good effect in an ebike PAS. A purely torque-sensing PAS can work quite well, but better-performing solutions already exist.

 

[Stefan Mikes]

Torque is measured by the controller to deliver the proportional assist power. I can't figure out how the cadence is used? Please share your know how.
PS, thermal sensor I mean is the overheat cutoff not the ambient temperature.

A torque sensor only knows how hard you are pushing the crank but has no idea how much power you are actually inputting with your legs. To reduce the thing to absurdity, you provide the maximum torque when you are riding out-of-saddle (actually stomping on the pedals) but you rotate the crank at a very slow rate (cadence). Such way of pedalling is called "grinding" and is usually bad to your knees. The same leg power (or even higher) can be provided by the rider by pedalling lightly but at a high rate (high cadence); this technique is called "spinning".

So pedalling lightly at 80 rpm provides the same leg power as pedalling twice as hard at 40 rpm. The point is, electric motors are efficient when they rotate fast, so spinning makes the motor deliver more power at less battery charge consumption. Spinning is easy on the knees and also delivers more blood to your legs, which is beneficial to you.

The idea behind a good mid-drive motor (such as Bosch, Specialized, Shimano, Yamaha, TQ, Fazua, etc) is to amplify the rider's leg power. The torque and cadence (not PAS!) sensors have these two values multiplied and delivered to the e-bike computer as the leg power value. Then, the calculated leg power value is multiplied by Boost (Assist) factor such as 2x (200%). If the rider inputs 100 W with their legs, the motor outputs 200 W of mechanical power. It is the most proportional and natural way to provide e-bike assistance.

Other sensors are vital, too. The e-bike computer wants to know how fast the e-bike is currently riding (speed sensor). The motor speed sensor is essential, too. There are even the battery and motor temperature sensors to protect these components from the thermal runaway! (That's why it is impossible to fry a good mid-drive motor on extreme climbs, or why good batteries don't create fires).

 

[Gionnirocket]

In other words. . .

But no matter the PAS sophistication it can not read your mind, nor see what's in front of you, nor compensate for an unusual situation and that's why the addition of a throttle is invaluable.

 

[fooferdoggie]

Torque is measured by the controller to deliver the proportional assist power. I can't figure out how the cadence is used? Please share your know how.
PS, thermal sensor I mean is the overheat cutoff not the ambient temperature.

a motor is most efficient at its highest rpms. keeping track of it lets the system know how much power the motor can handle.

 

[Jeremy McCreary]

In other words. . .

But no matter the PAS sophistication it can not read your mind, nor see what's in front of you, nor compensate for an unusual situation and that's why the addition of a throttle is invaluable.

Totally agree in the case of my 70 lb torque-sensing hub-drive. Adding a progressive on-demand throttle to that PAS makes for a very versatile power delivery system and significantly improves bike handling, traffic safety, and knee-friendliness in the process. Even for riders like me who always pedal with effort.

But I rarely miss a throttle on my much lighter and much more agile and responsive 38 lb power-sensing mid-drive — with the same set of bum knees. The only exception: When I stop in too high a gear. Took a few months to get good at avoiding that, but I'm fine now with no throttle on this bike.

Bottom line: A good throttle can be a valuable — even essential — riding tool. Its role is way more complicated than the never-throttle and always-throttle partisans understand.

 

[Gionnirocket]

Totally agree in the case of my 70 lb torque-sensing hub-drive. Adding a progressive on-demand throttle to that PAS makes for a very versatile power delivery system and significantly improves bike handling, traffic safety, and knee-friendliness in the process. Even for riders like me who always pedal with effort.

But I rarely miss a throttle on my much lighter and much more agile and responsive 38 lb power-sensing mid-drive — with the same set of bum knees. The only exception: When I stop in too high a gear. Took a few months to get good at avoiding that.

Rarely miss, but you're still without on occasion for what? Ideology?
I rarely use mine but the times I do the ride is more enjoyable especially when pushing the limits of climbing and obstacle navigation.

Bottom line: A good throttle can be a valuable — even essential — riding tool. Its role is way more complicated than the never-throttle and always-throttle partisans understand.

Amen

 

[Jeremy McCreary]

Ideology?

No ideology, just a very good trade-off. Wanted a very light, agile mid-drive climber with gravel bike leanings to complement my hub-drive, and you just can't get an ebike like that with a throttle. Big win despite the few times I miss a throttle on it.

Stll ride both bikes for different reasons.

 

[Chargeride]

Overun is important, but difficult to implement safely, all you need is one second to allow you to stop peddaling till the obstacle is clear.
Of course it could cause mayhem on the street, with old ladies catapulting into pedestrians.
So how could you add it safely?
A button to hold on the bars?

 

[Gionnirocket]

Overun is important, but difficult to implement safely, all you need is one second to allow you to stop peddaling till the obstacle is clear.
Of course it could cause mayhem on the street, with old ladies catapulting into pedestrians.
So how could you add it safely?
A button to hold on the bars?

It's called a brake cut off... Been around since eBikes became a thing.
And even when not braking, I've never found the minimal run_on an issue or safety concern.

 

[Gionnirocket]

No ideology, just a very good trade-off. Wanted a very light, agile mid-drive climber with gravel bike leanings to complement my hub-drive, and you just can't get an ebike like that with a throttle. Big win despite the few times I miss a throttle on it.

Stll ride both bikes for different reasons.

I can definitely appreciate the light bike want..but I'm not a 2 bike guy and though I have a much lighter build in the end I always ride the heavier SUV.
But both do have well tuned throttles.

 

[Chargeride]

It's called a brake cut off... Been around since eBikes became a thing.
And even when not braking, I've never found the minimal run_on an issue or safety concern.

Yrs, Im sure you find it easy, but some 80yr old on the pavement, takes off in high power mode, stops pedalling and the bike carries on surging, panics and finds themselves on top of a pram.
No characters in this post are supposed to resemble Jeremy

 

[Jeremy McCreary]

Yrs, Im sure you find it easy, but some 80yr old on the pavement, takes off in high power mode, stops pedalling and the bike carries on surging, panics and finds themselves on top of a pram.
No characters in this post are supposed to resemble Jeremy

Yeah, really hate landing on prams. Some of those babies bite!
;^}

 

[Gionnirocket]

Yrs, Im sure you find it easy, but some 80yr old on the pavement, takes off in high power mode, stops pedalling and the bike carries on surging, panics and finds themselves on top of a pram.
No characters in this post are supposed to resemble Jeremy

Here we aim to land on the mommy

 

[Neto]

A torque sensor only knows how hard you are pushing the crank but has no idea how much power you are actually inputting with your legs. To reduce the thing to absurdity, you provide the maximum torque when you are riding out-of-saddle (actually stomping on the pedals) but you rotate the crank at a very slow rate (cadence). Such way of pedalling is called "grinding" and is usually bad to your knees. The same leg power (or even higher) can be provided by the rider by pedalling lightly but at a high rate (high cadence); this technique is called "spinning".

So pedalling lightly at 80 rpm provides the same leg power as pedalling twice as hard at 40 rpm. The point is, electric motors are efficient when they rotate fast, so spinning makes the motor deliver more power at less battery charge consumption. Spinning is easy on the knees and also delivers more blood to your legs, which is beneficial to you.

The idea behind a good mid-drive motor (such as Bosch, Specialized, Shimano, Yamaha, TQ, Fazua, etc) is to amplify the rider's leg power. The torque and cadence (not PAS!) sensors have these two values multiplied and delivered to the e-bike computer as the leg power value. Then, the calculated leg power value is multiplied by Boost (Assist) factor such as 2x (200%). If the rider inputs 100 W with their legs, the motor outputs 200 W of mechanical power. It is the most proportional and natural way to provide e-bike assistance.

Thanks for sharing this info.
This could be the reason why I don't feel instantaneous assist during takeoff... Torque x (0) cadence = (0) leg power... That means no assist. Assist will come as you turn the crank few degrees.
Is there a way to disable this feature?
I understand pedal spinning is preferred than grinding.

 

[Stefan Mikes]

Thanks for sharing this info.
This could be the reason why I don't feel instantaneous assist during takeoff... Torque x (0) cadence = (0) leg power... That means no assist. Assist will come as you turn the crank few degrees.

It is a way more sophisticated than you might think. Several sensors together detect what happens to the e-bike. At the beginning of pedalling, torque on the pedal is detected. There is an extremely short delay from the pedal push to the motor activation (as short as the rider cannot even notice that). When the sensors detect the e-bike movement (which happens very fast), the full power amplification logic is applied.

In a regular mid-drive motor the delay must exist to protect the e-bike and the rider against inadvertent "launching" the e-bike as it were a rocket.

Is there a way to disable this feature?

Giant e-MTB with Yamaha motors advertise the "Zero Cadence" feature. That is, the motor is activated on a stationary e-bike as soon as the rider sets the foot on the pedal (you can actually feel the motor "cooking" under you when you do it). A similar feature is available in Specialized mountain e-bikes; you can control the motor activation delay between zero and some very short value with an app. The reason is the following: The rider might remain stationary uphill, and needs to initiate a climb onto, say, a rock. There mustn't be any delay between the pedal push and the motor action in this situation. It is why only e-MTBs are programmed for "zero cadence". Any other e-bike type with a mid-drive motor must have a delay of (say) 100 ms to protect the motor against frying and the rider against an unexpected e-bike forward movement.

I understand pedal spinning is preferred than grinding.

Yes, precisely. Mid-drive motors feel lively above the cadence value of 70 rpm. Grinding makes the motor providing less power than expected and leads to the excessive battery use. Proper use of the derailleur is necessary (usually, I downshift three gears before a planned stop, for instance at an intersection).

To make the rider aware of the proper cadence, the Mastermind TCD displays of Specialized e-bikes are colour coded. There is a strip at the bottom of the display with the cadence value inside a diamond. When your cadence is >0 but too low, the strip turns brown; it turns green as soon as you have entered the optimum cadence range. (The overall colour theme reflects the assistance mode you are in: pink for Smart Control).

A 2021 review of Specialized Tero when both the e-bike model and Mastermind display were a new thing.

Necessary to mention: a mid-drive motor e-bike does not require the brake cut-off as the motor power is automatically cut off when you stop pedalling. The throttle is not the mid-drive motor thing (whenever a throttle is applied on a mid-drive motor, it is usually very cheap Chinese junk). It is because you control a good mid-drive motor with your legs. (The throttle makes some sense on a hub-drive motor e-bike).

 

[Neto]

It is a way more sophisticated than you might think. Several sensors together detect what happens to the e-bike. At the beginning of pedalling, torque on the pedal is detected. There is an extremely short delay from the pedal push to the motor activation (as short as the rider cannot even notice that). When the sensors detect the e-bike movement (which happens very fast), the full power amplification logic is applied.

In a regular mid-drive motor the delay must exist to protect the e-bike and the rider against inadvertent "launching" the e-bike as it were a rocket.

Giant e-MTB with Yamaha motors advertise the "Zero Cadence" feature. That is, the motor is activated on a stationary e-bike as soon as the rider sets the foot on the pedal (you can actually feel the motor "cooking" under you when you do it). A similar feature is available in Specialized mountain e-bikes; you can control the motor activation delay between zero and some very short value with an app. The reason is the following: The rider might remain stationary uphill, and needs to initiate a climb onto, say, a rock. There mustn't be any delay between the pedal push and the motor action in this situation. It is why only e-MTBs are programmed for "zero cadence". Any other e-bike type with a mid-drive motor must have a delay of (say) 100 ms to protect the motor against frying and the rider against an unexpected e-bike forward move

Thanks again for these additional info. The more I know, the more I don't know (Aristotle).

I'm more inclined to Zero cadence feature with a brake cutoff to prevent accidental takeoff and cooking the motor.

Including cadence, what do you think about this algorithm? (torque-a) x (b+cadence) = leg power, a and b are constant.
This could be applied to both e-MTB and regular mid-drive. Instant response and safe. With brake cutoff still.

 

[Jeremy McCreary]

Including cadence, what do you think about this algorithm? (torque-a) x (b+cadence) = leg power, a and b are constant.
This could be applied to both e-MTB and regular mid-drive. Instant response and safe. With brake cutoff still.

Interesting approach! Constants a and b are positive, correct?

Let X be the adjusted rider power index you called "leg power" above. Would like to hear more detail on how your proposed PAS would use X to dole out motor power in starting out and normal riding. Thanks!

 

[Stefan Mikes]

I'm more inclined to Zero cadence feature with a brake cutoff to prevent accidental takeoff and cooking the motor.

No one needs that. Unless you are on a technical trail, you just set your foot on the pedal to ride, not to support your foot

I have no idea what to think of your algorithm. Millions of mid-drive e-bikes ride in the world without any issue.

 

[Neto]

Interesting approach! Constants a and b are positive, correct?

Let X be the adjusted rider power index you called "leg power" above. Would like to hear more detail on how your proposed PAS would use X to dole out motor power in starting out and normal riding. Thanks!

Yes, they are positive.
(torque-a) is a typical minimum value to start any action. (a+cadence) is also a minimum value to allow instant start. There should be none or minimal effect on normal riding.

Please note that I'm only trying to learn the technology out there cause I'm shopping for a new bike lighter than my old 64lbs cruiser.
I still like to learn how the speed and accelerometer sensors are used.