E Bike Kit with torque sensor like Yamaha Bosch etc

ElektroCSL

1 µW
Joined
Feb 25, 2020
Messages
3
Hi to you lovers of electric vehicles

Since a while now the electric virus got me and i want to build a bike by my own...

The motor should have at least 1500W with an huge battery, so that i can go at least 200 km by 40 kmh

Im allready using an 2018 Haibike model with the yamaha motor and i really like the sensitive behavior of the torque system

Exists out there a free e bike kit system with an identical torque System like my Hibike?

thank you in advance

Cheers

Daniel
 
ElektroCSL said:
Exists out there a free e bike kit system with an identical torque System like my Hibike?
i can't think of many free ways to build a torque sensing system, unless you have a fair bit of electrical and mechanical engineering skill, as well as skill at finding usable "junk" and repurposing things. if you do, you could take the load sensors out of cheap personal weight-sensing scales when people break them and toss them out, and build a torque sensor similar to the beamts reasonably easily.

if you already have or can get for free an arduino or other programmable mcu system, you could write software to convert this sensor signal into a throttle control, and then find old dead scooters or bikes being tossed out, with generic controllers and motors on them to reassemble into


but the only off-the-shelf ways i know of to get a torque sensing system without using an oem integrated system are not going to be free, or even cheap.

you can either

get a bottombracket torque sensor like tdcm, sempu, thun, etc., or a beamts if you can find one, and then use a cycle analyst v3 to convert it's output to a throttle signal, which you would use to drive a generic controller system and a motor to go with it (both of which need to be able to handle the power you need for your particular ride/terrain/etc), and a battery capable of the range and power you need.

or

get a kt / kunteng / kun teng controller compatible with casainho/stancecoke/etc's opensource firmware project on github (thread on es under casainho's posts), and a compatible torque sensor for those, and then a motor/battery that do what you want.
 
:shock:

If youre intention was to confuse me, than you did a perfect job

seriosly mate, i have literly no idea what you are writing about

Can you shoe me? :)
 
ElektroCSL said:
i really like the sensitive behavior of the torque system

Exists out there a free e bike kit system with an identical torque System like my Hibike?

By free, did you mean costs no money, or something else?
 
john61ct said:
ElektroCSL said:
i really like the sensitive behavior of the torque system

Exists out there a free e bike kit system with an identical torque System like my Hibike?

By free, did you mean costs no money, or something else?

By free i mean, the kit systems on ebay for example, like Bafang etc. The one which you can apply a self made battery.

By the way, please correct me whereever is necessary.... Im from GER and also willing to improve my english

So how you call the systems like bosch, yamaha, shimano, brose etc. ?

Whats the correct expression for such popular systems?

THX Mate
 
by "free" I think you mean a "bolt on kit" , as opposed to an integrated or proprietary system.. yes there are torque sensing bolt on kits such as "TSDZ" and "lightest ebike kit" (search www: "lightest.bike").. as for the responses above I think "free" was interpreted as "open source" .. example TSDZ thread:
https://endless-sphere.com/forums/viewtopic.php?f=30&t=93818
 
Yes, the Bosch etc systems are

"closed" as in proprietary, complete bicycles or at least packaged drive systems sold "off the shelf"

and designed to prevent you from DIYing "mix and matching" with other vendors' components, especially cheap Chinese products.

The more "open hardware" approach allows you (at least in theory, requires some knowledge and skills) to find ways to add electric power to bikes not designed for that

to choose batteries from multiple sources or build your own

to choose a better controller to get more power from a given motor

in short to design your own custom bike the way you want it.
 
ElektroCSL said:
:shock:

If youre intention was to confuse me, than you did a perfect job

seriosly mate, i have literly no idea what you are writing about

Can you shoe me? :)
you asked for a "free" way to do it.

with no further information available, i posted some free, or at least cheaper, ways you could do what you want.

after your reply above, you have now explained that to you, "free" means something completely different than it does in the dictionaries i'm aware of.

but you did not explain this to begin with, so you got the answer you asked for without defining your version of "free". ;)
 
Also what is this?

> Can you shoe me?

Fact is with googling the phrases you don't understand you can learn a lot.

Or you might need a pro to help

or maybe visit a forum where people use your native language?
 
This is a valid topic that deserves some attention. I have been researching torque PAS and have only been able to find two alternatives. Buy the torque sensor that Grin offers along with a cycle analyst to make it work or buy the TSDZ2 torque sensing mid drive motor.
I’ve been trying to put together a bike for a friend for just over a year and have failed to come up with anything that suits her. She has arthritic hands and cannot hold a throttle and wants to use the bike at less than 5 MPH when she walks her dog. Every cadence PAS that I’ve tried is really annoying at that low speed. The other problem is that she has been on a Bosch bike and wants one that duplicates Bosche’s smooth torque sensing.
The cost comparison is a cadence PAS sensor for under $10, several hundred dollars for the Grin parts, or several hundred dollars for the TSDZ2 motor which seems to have some quality drawbacks.
I am a throttle guy and have never felt the need for PAS myself and really don’t want to throw a lot of money into this project.
 
Hwy89 said:
This is a valid topic that deserves some attention. I have been researching torque PAS and have only been able to find two alternatives. Buy the torque sensor that Grin offers along with a cycle analyst to make it work or buy the TSDZ2 torque sensing mid drive motor.
afaik the kt controllers with open source firmware in the threads here on es offer torque sensing, using various sensors like sempu, etc. you'd have to look at the osfw threads by casainho, stancecoke, etc, to find the links with lists / info /etc. or look on their githubs.
 
Are you saying that latter can be made to work with any version motor, as with CAv3?
 
the answer isn't really a yes or no.


the ca is not a controller. it's just an interpreter/modifier of input signals that then outputs a throttle signal to control a controller. so it has nothing to do with the motor, and is thus "compatible" with all motors, since it doesn't matter what the motor is, the ca doesn't talk to it or connect to it.


the kt stuff is a controller system, that does directly connect to the motor and control it.

so they're not like each other; they don't do the same thing or work the same way (even though with the osfw they may have some overlapping functionality).

whether the kt controller will work with a particular motor or not i couldn't tell you--it depends on which controller you get and what motor you hook up to it. you'd have to check the specific controller and the osfw wiki / info to find out if there are any known incompatibilities, and then check what the motor requires to drive it, etc., and then see if the controller you want to use can do that.
 
OK, I guess the different meanings of "torque sensing" is tripping me up.

The TSDZ2 motor having a torque sensing feature internally is one type, as opposed to speed- or cadence- based control

PAS via BB or pedal torque sensing is something else, the CAv3 or maybe the controller can handle that independently of whether or not the **motor** has torque sensing.

Have I got that right?
 
john61ct said:
OK, I guess the different meanings of "torque sensing" is tripping me up.
there's only one meaning of it, in reference to ebikes--sensing the torque of the human operator input (be that by pedals, handcranks, levers, etc).

The TSDZ2 motor having a torque sensing feature internally is one type, as opposed to speed- or cadence- based control

PAS via BB or pedal torque sensing is something else, the CAv3 or maybe the controller can handle that independently of whether or not the **motor** has torque sensing.

a motor doesn't have torque sensing, the motor doesn't do any sensing, other than some of them use position sensors of one type or another (like the hall sensors common in many ebike motors).

the torque sensing in the tsdz2 is for the pedals.

the tsdz2 controller itself is what processes the torque sensor output.

the ca v3 exists (in this case) to process a torque sensor input for a controller that doesn't know how to (a controller that only has a throttle input).
 
Aha, my wires crossed then.

In my notes trying to puzzle out what defines true FOC I have

accurate access to torque information gives perfect throttle control, very smooth even with huge power levels

but can't find the source.

Another, I guess more accurate:
current sensor on each phase connection in the motor, to provide direct real-time sensing of motor current amps

I know torque is closely related to phase amps, but I guess not that directly?
 
This page may be helpful: https://www.ebikes.ca/product-info/advanced-pas-kits.html#torque-sensors

I love the idea of a torque-sensing PAS that works really well, but i have no patience for one that doesn't work well or isn't well-refined and tuned. I'm on the hunt for a good setup but don't want a bosch or shimano, etc. so i'm looking for a DIY project.

Can anyone share their experience or point out a review or article that really speaks to the fine-tuned control, throttle responsiveness, etc of various torque sensor and controller combinations? I'm looking at mt bike application and ride some rocky, technical trails where very fine control of power at low speeds is crucial.
 
Hope I am not thread-jacking, but I think it is quite related; Does the BB torque-sensors even work with the forces that a typical diy mid-drive system is set up (dual freewheel with an extra chain from front chainring to the motor)? If I understand correctly, the Sempu unit measures the flex in the BB when you pedal, but what happens when the motor starts pulling on the chainring?

Im looking at putting together a DIY MY1016 kit with a CA and some kind of torque-sensing BB for my FS bike if the tsdz2-unit does not fit, but not sure if it will work.
 
john61ct said:
In my notes trying to puzzle out what defines true FOC I have

accurate access to torque information gives perfect throttle control, very smooth even with huge power levels

but can't find the source.

Another, I guess more accurate:
current sensor on each phase connection in the motor, to provide direct real-time sensing of motor current amps

I know torque is closely related to phase amps, but I guess not that directly?
none of that has anything to do with pedal torque sensors or the ca, or this thread.

what you are describing is internal controller sensing of phase currents with current sensors on each phase.

they are not torque sensors.
 
HrKlev said:
Hope I am not thread-jacking, but I think it is quite related; Does the BB torque-sensors even work with the forces that a typical diy mid-drive system is set up (dual freewheel with an extra chain from front chainring to the motor)? If I understand correctly, the Sempu unit measures the flex in the BB when you pedal, but what happens when the motor starts pulling on the chainring?

Im looking at putting together a DIY MY1016 kit with a CA and some kind of torque-sensing BB for my FS bike if the tsdz2-unit does not fit, but not sure if it will work.


Yes. Proportional to the power and the speed desired by user. ( how hard you pedal)

I love my TCDM. Makes a slow throttle only ride ( boring) into a nice strong long range ride. workout too.

Properly set up they work great in a myriad of gears, and teh power applied is tuned through the CA for my diffent riding modes. I like to keep tension on the chain and input power ( human) to the wheel, and I have my PAS setup to just let a ratio of watts assist to the watts human made.

It is a 0-5v sensor, and 2.5v corresponds to a known calibrated tq ( in my proper setup).. and the tq consumed for a motive tractive force is universally distributed through the gearing no matter what gear you are in. Its the tq... and the crank force.... Only the top speed changes dependant on the gearing and input force...

Go slow, still need to pedal.... ( like 5mph, 34-34 (1:1) ratio, ).. I make 200w and for every watt I make, the pas throttles the motor for me.. incrementally.... so I input 5w.. motor makes 25w... I input 10w... motor makes 50 w.. and so on. So the harder I pedal the more the motor pulses. The softer, I pedal, and the motor follows suit.

Funnest part was tactically tuninig to find a good place where the motor wouldn't overwhelm my human cranking... by careful manipulation of the input values to the CA3. Say, if the PAS asked the controller to input say, 10 watts for every watt I made human, ( on a 100w human input, motor pulses 1000w) the motor would just hunt and oscillate in the range where it was overwhelming the human tq.. then slow down, and do it again.

So I changed that to only 5 watts per every human watt.. ( on a 100w human input, the motor pulse 500w)... and the oscillation and overrunning of the motor stopped and the pedaling was the dominant motive force... and the pedaling "assisted" (the ) system... didn't work against t me., my range went up, and the tq sensor would pulse the motor gently in cadence ( step) with the (main) human input.
 
HrKlev said:
If I understand correctly, the Sempu unit measures the flex in the BB when you pedal, but what happens when the motor starts pulling on the chainring?

NO.
When the motor pulls on the chainline, the Tq sensor output goes down ( cause you are no longer inputting power above what the motor is) ... and subsequently the sensor output drops...

Doesnt measure flex, to speak, but measures the tq produced ( through the flexing)... Measures the Tq created from input and makes this into a 0-5v signal. Yes, it utilizes flex, but not in the BB... the BB is only the case for the sensor... It measures the chainlint force (( lever arm length) x (force ON end of LEVER)) ( hint: when you are changing gears, does the lever arm length ( length of crank arm to BB center) grow or shorten? No. The output ratio changes ( through the gears to make certain desired speed) ( aka measured chain force) ... not the input ( force x rpm)...

( will not "flex" without "pedal cranking"...). It triggers the motor throttle input, and it is up to you ( the tuner) to make sure the power from the motor doesnt overwhelm the pedaling......

So if the motor is triggered, you are no longer inputting ( as much ) tq, and the power ( assist and human) goes down...or up, depending on desired speed... Motor can easily overwhelm human input.. and without careful tuning, this lurchy ON/OFF of the motor output ( on a graph) is a "Hunting Oscillation" of power input vs desired speed.

Fact of the matter, it is designed to "assist"..

and not to say a 1000w motor wont overwhelm a rider in an instant...

It will.... so you set the motor to not go over say 250-500w ( pas assist power) .. and the primary tension will always be on the chainline ( at human power levels) and contributing to the PAS output, and the CA input, then to the CA output, and finally the throttle input to the controller... where it gently pulses in ratio to the human power input, of 2-200w give or take.

I mean, hit the throttle on a 8Kw bike and there is no way you can pedal at all with any significant input, so it is designed to use the ( range of) HUMAN input force to "throttle " the power out to the wheel.

Im looking at putting together a DIY MY1016 kit with a CA and some kind of torque-sensing BB for my FS bike if the tsdz2-unit does not fit, but not sure if it will work.
 
Thanks a lot for a very informative answer! I will have to read it a couple more times when I have the time :D

Kind of makes me want to buy a CA and a torque sensor anyway, to try it out. What kind of setup do you have? And do you think this system is responsive for technical single trail riding?
 
amberwolf said:
john61ct said:
In my notes trying to puzzle out what defines true FOC I have

I guess more accurate:
current sensor on each phase connection in the motor, to provide direct real-time sensing of motor current amps
what you are describing is internal controller sensing of phase currents with current sensors on each phase.

they are not torque sensors.
Thanks for clarifying that.

Are there controllers that feature the above, but that are not FOC?

IOW do not calculate the rotor angle / position in real time from that data stream?

Or is it the case that only FOC controllers have current sensors on each phase?
 
For example, some say PowerVelocity are not truly FOC, just Sinusoidal Commutation,

but PV states "FOC implementation in VESC is likely the best in its class. It is in fact FOC done the right way."

I realize it's "just a label" and the real-world benefits are more important, but accurately defined labels are a useful shorthand for such discussions.
 
Back
Top