Understanding torque sensors

[SandwichTech]

I am just starting to give myself a crash-course in ebike technology, so bear with me if these questions seem obvious. If I want any electrical motor power to be cumulative with my pedalling efforts and never "overtake" me (so that my pedalling becomes freewheeling), am I correct in assuming that the only way to achieve this is with a torque sensor setup? And that this effect is equally possible with a mid-drive, front hub, or rear hub; geared or gearless?

Also, I have seen that some torque sensors are within the BB, some are in the rear dropout. It would seems that while the BB sensor would be representative of the bikers effort, the rear dropout sensor's reading would differ by the changing gear ratio. Are the dropout sensors only used in combination with a cadence sensor to provide power-matching instead of torque-matching assist? If so, is power-matching generally preferred? It seems like a more "fair" way to distribute assistance.

Update: I am now realizing that I should have posted this in ebike general discussion. My apologies - the post was originally more motor related and then changed as I talked/edited myself through my questions.

 

[Voltron]

Usually the point of the gears is to keep the chain load about the same as your leg power... Whether you're on flat ground in a high gear, or climbing in a low gear, the load on the chain is the same in either gear, so it doesn't make much difference whether you measure the bottom bracket doing the pulling, or the dropout getting pulled, it's still feeling the same load. Some aftermarket systems use a spring loaded set of derailleur jockey wheels setup that looks like a chain tensioner to measure chain deflection.

Alternately, you could just use a throttle, and powermatch seamlessly with your legs on the fly, with no lag time. It just takes the self discipline to not crank it all the time!

 

[stancecoke]

The use of the pure torquesensor signal is not satisfying.
It gives to much power from standstill and to less at higher speeds.
Better you use motor power = factor × human power. The factor can be depending on the speed.
Human Power = 2×Pi×Torque@pedal×cadence = 2×Pi×torque@wheel×wheelspeed. So it doesn't matter where the torque is measured.
But you need a controller the can handle the torque and speed and cadence signals. I recomment the open source projects for Kunteng or Lishui controllers.
Here you can find a collection of technical information about torquesensors:
https://translate.google.com/translate?hl=&sl=de&tl=en&u=https%3A%2F%2Fwww.pedelecforum.de%2Fwiki%2Fdoku.php%3Fid%3Delektrotechnik%3Adrehmomentsensoren

regards
stancecoke

 

[SandwichTech]

Voltron, thanks for the reply. It seems that chain load WOULDN'T stay the same (I am ignoring the possibility of added chain load from a mid-drive motor). I would think the goal would be a consistent leg power and cadence from the biker, resulting in steady torque at the BB (Power = Torque*RPM). To achieve consistent cadence and torque you have to change gears (length of torque arm) meaning the chain tension is changing proportionally. Small gear in front (slow uphills) has high chain tension while large gear in front (fast flats) has low chain tension.

 

[Voltron]

Not if you're pushing as hard in the flats as you can go against the aero drag. Separate from the motor, if you're the type of rider that always pushes near your top leg power output on the flats, then as you go more uphill the increasing load overpowers your legs, then you gear down to bring it back into your leg power range. So basically hard leg output equals higher chain tension, regardless of what gear or which end is measured from... You could ride a different way, but load wise, the system can't tell high gear pedaling into air drag vs low gear pedaling up a hill.

 

[Voltron]

People tend to spin higher cadence in the flats and take a break, and save lower cadence grinding for uphill, but still load is load. I tend to ride the exact opposite of that, with pushing large gears at low cadence and high chain tension into the wind on the flats, then spin my way uphill at lower chain tension, until it's so steep I'm out of gears and forced back into grinding.

 

[SandwichTech]

The use of the pure torquesensor signal is not satisfying.
It gives to much power from standstill and to less at higher speeds.
Better you use motor power = factor × human power. The factor can be depending on the speed.

Stancecoke, thank you! That all makes perfect sense and making the power factor a function of speed is a great idea. Open source firmware was a topic in the back of my mind that I hadn’t started researching. Your link looks like a great starting point.


Sent from my iPhone using Tapatalk

 

[Voltron]

But what if you want high power from a near standstill, like mountain biking, where you need a shot of higher power to clear a steep but low speed obstacle?

 

[docw009]

I am just starting to give myself a crash-course in ebike technology, so bear with me if these questions seem obvious. If I want any electrical motor power to be cumulative with my pedalling efforts and never "overtake" me (so that my pedalling becomes freewheeling), am I correct in assuming that the only way to achieve this is with a torque sensor setup? And that this effect is equally possible with a mid-drive, front hub, or rear hub; geared or gearless?

Not the only way. Some cadence sensors are set for a max speed in a particular level. You can pedal a cadence sensor bike right around that limit, and have it so the motor and your pedaling are contributing. You can also turn off the pedal sensor off and if you have a throttle, you can feather as much power as needed.

Sure, it's easy to let your speed drop on a cadence sensor bike and then any motion of the pedals has the motor doing all the work. Doesn't have to be that way unless you got a crap system. I've owned one controller that went all the way to 20 mph in all PAS levels with any motion of the pedals. That controller got taken off and scrapped.

 

[amberwolf]

If I want any electrical motor power to be cumulative with my pedalling efforts and never "overtake" me (so that my pedalling becomes freewheeling), am I correct in assuming that the only way to achieve this is with a torque sensor setup?

It should be possible to do it with a cadence sensor, if the controller or intermediate electronics (like Cycle Analyst v3) is used to interpret the cadence in a way that controls motor power or speed or whatever factor you need it to, to keep it like this, vs your gearing.

A torque sensor will also work, but might not be necessary. If you like you could try it out with cadence first, then add a torque sensor if you find it doesn't quite do what you want with cadence alone.

There is (was?) a chain-tension torque sensor called BeamTS that mihgt be closer to what you want to do, if you wish to use pure TS.

Most of the pure cadence systems don't do anything lke what you're after, because they just activate power on or off at a level you set via PAS level on a display. But tehre are some that can scale the cadence itself to create a continously-variable control of motor power (or speed, etc etc). This is what I use the Cycle Analyst v3 for.

And that this effect is equally possible with a mid-drive, front hub, or rear hub; geared or gearless?

Any hub, yes. But it depends on your middrive design if it could work as desired (or at all), and also on where your torque sensor is located.

Also, I have seen that some torque sensors are within the BB, some are in the rear dropout. It would seems that while the BB sensor would be representative of the bikers effort, the rear dropout sensor's reading would differ by the changing gear ratio. Are the dropout sensors only used in combination with a cadence sensor to provide power-matching instead of torque-matching assist? If so, is power-matching generally preferred? It seems like a more "fair" way to distribute assistance.

All of the torque sensing systems I know of (except maybe Bionx, but those are irrelevant to you since they are out of business) use a cadence sensor as well-just that some of them (all the BB types I know of, for instance) have the cadence sensor integrated into them, rather than being separate.

How the assistance needs to work depends on exactly what you need it to do.

In my case (physical inability to do the work myself), I need it to do *all* of the work of starting me up from a stop, so I need it to *not* look for cadence at all, at first, only torque. But then I need it to obey cadence once started, so I can control speed with the cadence RPM. Presently I don't know a system that does this (because my use case is not common, though it's not unique), so I'm pondering various ways of doing this using the Cycle Analyst v3 and some external electronics (to fake a cadence signal when at a stop, but use the real cadence when moving, since the CAv3 "assumes" at least 50RPM cadence signal).
https://endless-sphere.com/forums/viewtopic.php?f=4&t=37964&p=1384020

 

[ZeroEm]

SandwichTech » Dec 24 2019 1:56pm
I am just starting to give myself a crash-course in ebike technology, so bear with me if these questions seem obvious. If I want any electrical motor power to be cumulative with my pedalling efforts and never "overtake" me (so that my pedalling becomes freewheeling), am I correct in assuming that the only way to achieve this is with a torque sensor setup? And that this effect is equally possible with a mid-drive, front hub, or rear hub; geared or gearless?

Was where you are 4-5 months ago. I started off just using the throttle to assist but that gets tiresome so I wanted a torque advanced pas system that cost a lot more than a simple Cadence sensor. I was convinced to try the cadence first and have been happy with it after learning to adjust it.
I have 24 magnet sensor into a CA3 with a digital control for pas levels. You can adjust the timing when it kicks in not to early or late. You can have it add power with cadence or have it subtract power.
I started with it adding power but you need less power when peddling faster not more so now have it subtract.

I like high peddle speed when starting up a hill as the hill starts slowing me down my pas adds power.
While peddling on flats if I want less help I peddle faster and get less power and if I feel lazy slow and let the power com up.
my average cadence is 47 and my dig switch first setting is off or 0 watts and my high is around 700 watts and can peddle at 25 mph/40kph. But still tuning and that changes with my abilities.
Still have my throttle and use it as needed but if I leave it alone only use 10 watts a mile. I average 9w/km per charge but that adds my quick store runs 1.5 miles at 30mph/50kph or don't use it if I ride with other non E-bikes.

Pas is a great way to extent battery distance.

 

[stancecoke]

making the power factor a function of speed is a great idea.

With the Bosch Nyon you can create your own profiles. Here some suggestions for different ride situations:
https://translate.google.com/translate?sl=auto&tl=en&u=https%3A%2F%2Fwww.pedelec-biker.de%2F2015%2F09%2Fbosch-nyon-individuelle-fahrmodi.html

e.g. a mountain profile:

regards
stancecoke

 

[Toorbough ULL-Zeveigh]

If I want any electrical motor power to be cumulative with my pedalling efforts and never "overtake" me (so that my pedalling becomes freewheeling), am I correct in assuming that the only way to achieve this is with a torque sensor setup?

am i correct in assuming you're an old school engineer?
that you think there's ever only the one way of doing anything; i.e. the rightway chaloway.
well, you did ask.

not the best way & not at all what you wanted but no chain to over run.

[youtube]4JKOKpg21dQ[/youtube]