Torque sensor "throttles" with brush motor controllers

[hsapien]

Hello all. Am a newbie to this forum...though have used and distributed e-bikes for 20 years. Have been wanting to match a torque sensor "throttle" with a brush motor controller, and assumed it would be no big deal since the manual throttles generally work for either brush or brushless motor controllers. Most seem to be Hall Effect type using a 0-5V variance for signaling higher or lower motor RPM. I looked at TommyCat's guide to throttle operation, etc..., which confirmed my understanding. But have also heard torque sensor "throttle" manufacturers have designed in special non-standard interfaces so they can force you to purchase a motor controller with their torque sensor. This happened to me with Thun sensor, sent to me with no clear documentation. I requested the specs but was told I had to buy their brushless motor controller for it to work !!! It wasn't crucial that I adapt the bike at the time so just noted this as yet another screw job by a vender who could care less about their customer.
Now am resuming my quest to find a torque sensor to work with the mid-drive brush motor e-bike I have been using for many years. The original strain gauge on a beam, reading chain tension, is no longer made and I need to replace it. The BB torque sensors should work, but all that I've located cannot give me a definite as to whether it will work with a brush motor controller.
I posed this question to TommyCat, but there may be others who could aim me in the right direction.

[Tommycat]

Hi hsapien,
And welcome to the forum! :-)

I too have been interested in the relationship between the output voltages of a torque sensor (and PAS for that matter...) and how it interfaces with controller and motor output.

PAS is based on electrical digital "pulses" between 5 and 0vdc. Which definitely requires dedicated controller logic... But an analog voltage output from a torque sensor should be usable...?

Here are a couple good articles on these types of interfaces...
https://www.ebikes.ca/learn/pedal-assist.htmland

https://www.electricbike.com/torque-sensors/

They point out the wide differences in outputs between different torque sensors. And the necessity for having a couple outputs for safety. Read them over carefully.

Bottom line seems to be to purchase a Cycle Analyst 3.0. It sounds like it acts as an intermediary between the sensor's output, and controller throttle input. (just what your looking for?)
And being halfway there having purchased a Thun sensor, may be an acceptable solution. As important and reliable that this control input should be, it would be my recommendation.
But please use due diligence in vetting it first. I have not used one to verify that it will do what I've read, or what you desire. At this time I just use a throttle which works well for my riding style.
I've considered a C.A. but the cost, what I would use of it's many features, and it's somewhat DOS looking interface, have held me back.

I'm curious as to how your beam strain gauge work(ed) electrically and how it physically functioned. As in did you like the way it worked to help vary the motor output to your needs?



Regards,
T.C.

[hsapien]

Hi T.C.,

Before reading your recommended links I'll say the beam torque sensor on the bike works great if you like the feel of a regular bike. If your cadence is even, the motor power is even...which is a great way to teach cyclists how to be efficient. Uneven pedaling causes motor to be uneven, most noticeable at the highest power selection.
The bike's controller is complex and over-engineered in my view (maybe to prevent copying, which has not occurred), incorporating lead acid battery charger, low voltage cut-off based on 10.5V for each 12V stack, thermistor temperature control, 5 power levels (including Stand-By), error codes to isolate problems, ability for rider to have personal PIN code, etc... The software also provides for interesting features like boosting power to the maximum only when the bike slows to a crawl. The designers did this to prevent power drain on hills unless you absolutely need that extra boost. You would be going a little faster, but draining the battery quicker without this feature. The system allowed for the best range of any e-bike produced a the time. Battery life lasts quite well since high current draw is rare as the motor always gets muscle help.
As far as I know it was the first torque sensor based e-bike to be produced (about 20 years ago). In my view it still has advantages over other "state of the art" designs. I should mention also I distributed over a thousand of these bikes at deep discounted prices after the original manufacturer sold them off. The discount was based on how much the bike would cost at a shop without any of the electrical components in the event I could not support them properly with electrical parts down the line. The top quality cro-moly frame, double-wall alloy wheels with internal gear hub, etc, all adds up! An excellent non-powered bicycle with all the electrical removed.
The "Charger" bike was produced by 2 partners, one of whom designed the electrical portion of the bike and did not want to cooperate refusing to share schematics and code details. General Motors had bought into Aerovironment, who designed their EV-1 car, which was leased to customers for a few years followed by a forced return so GM could scrap the fleet. A real crime in my view, as also reported in the book, "Who Killed the Electric Car". GM attorneys would not let Aerovironment release the "Charger" bike details for liability reasons I was told. In effect, that fewer bikes on the road meant less chance of law suits. If there hadn't been another partner involved, GM would have buried the fleet of "Charger" bikes alongside the EV-1 cars.
Fortunately the bikes have held up quite well with one customer having reporting over 20K miles a few years ago...still without having to replace motor brushes, one-way Sprague bearing, or speed reduction belt. Like any electronics there have been some circuit board failures, though not as many as anticipated. I think most failures have been from careless use in wet weather, as the power pack could have been better designed to prevent moisture penetration. We provided a "rain sock" for this reason, which helped considerably. The torque sensor is more vulnerable than it might have been and careless use of chain cleaning solvents can eat away at the moisture barrier causing a short and permanent damage to the strain gauge and small amplifier board. Jumping on pedals can actually snap the beam. I was able to rebuild these with replacement beam/gauges, but after this source went out of business assumed I could find a different type of torque sensor to also work with a more standard motor controller.
So to answer your question, the torque based system is far more desirable and efficient than a manual throttle in my view. You either love it or want a mo-ped. Customers who have modified the bike to use manual throttle have the best of both worlds... I have details on how this can be done if anyone is interested, along with the free-wheeling crank sets needed so your pedals don't move unless you're pedaling.
My strengths are mainly mechanical and the limited electrical knowledge is enough to get me in trouble. If purchasing a Cycle Analyst could help to interface a BB torque sensor with a common brush motor controller, this would be fine if the information could lead to modifying the BB sensor and/or controller so the C.A. brain isn't required for each bike. Your average rider doesn't need or want a C.A., nor will justify the cost.
After the last recession forced us out of a warehouse in the San Francisco Bay area, my business has mainly been selling a few mid-drive motor kits I designed for recumbents and other long frame bikes. There have been a few rare "Charger" bike sale now and then while supporting existing customers.
The past 5 years my business has been almost entirely inactive while over 200 bikes and a mountain of parts rack up storage expenses. It is my desire to make some use of these bikes and parts at this stage, along with possibly manufacturing a new cargo bike beginning with the over 200 wheel sets collecting dust. Ideally, the circuit boards and torque sensors from the remaining original bikes would then be used as spare parts for the existing bikes floating around the world. A replacement BB sensor with motor controller for these original "Chargers" would pave the way to selling these off. The "Charger Mark II" would incorporate readily available replacement parts.
Since the original circuitry has a 10.5V per 12V battery power cut-off, this has made it a hassle to use lithium battery packs that are bundled as 24V and must be tapped into to separate the stacks for the controller to work. I did a few for customers but no longer because of hazardous shipping issues, lithium cell capacity inconsistencies, plus BMS and cell cooling reasons. A standard controller would allow for standard lithium packs to be used.
OK, there you have it folks. Any suggestions on how to get these last 200 bikes on the road with BB sensor and standard 24V brush motor controller would be appreciated. I'd like to make more bikes and kits with my good friend's patented design using a motor with integrated freewheel as a mid-drive. His design is more efficient than the current Chinese (incl Bosch) mid drives and it's possible to adapt to many existing non-powered bikes with standard BB. Funny how the Chinese and Europeans regarded the "Charger" bike as a freak when exhibited at their bike shows 20 years ago. Hub motors were the rage and few could understand the value of mid-drives or even heard of torque sensing "throttles". Now that they've seen the light, it's the rage.

[Tommycat]

Very informative write-up!

Sounds like an interesting project, all be it a bit beyond my scope. ;-) If no one steps up to the task... have you tried E.S.? As these threads allude to, your answer may be found there.


Throttle re-map...

https://endless-sphere.com/forums/viewtopic.php?f=2&t=107047

Torque Sense...

https://endless-sphere.com/forums/viewtopic.php?f=2&t=107023

Best of luck,
T.C.