As previously posted I have a QV (velomobile) with a custom e-assist. (Converted hub motor with Grinfineon /CA3 controls driving a modified Schlumpf (Mountain Drive)bottom bracket and thence to the 11 speed rear cassette) It works superbly but in order to comply "strictly" with the EU rules it shouldn't really have an operator throttle. (The motor output should be controlled by a cadence or torque sensor in the drive train) As the motor ALWAYS drives the pedals via a short chain to the left pedal there is no way of achieving the actual rider input once things start moving. (Assist from stopped can be sensed but of course once things are under way the sensed cadence/torque includes the assist thus creating a "closed loop".) Purely for technical interest I've recently fitted a Garmin Vector 3S kit (the single pedal version) in order to try and ascertain the proportion of total power generated by me vs. the e-assist. That works fine and shows that overall I'm doing abut 60% of the total.(using the CA3 wh/km as comparison with the Vector data) Now it has occurred to me that there might be some way of using data from the Vector's to supply a PAS input to the CA3? Obviously the problem is that the Vector data is only transmitted by BTE and/or ANT+ so some sort of micro-circuitry would be needed to receive the transmission and then output that by wire to the CA3. I'm no electronics whizz or computer programmer but I guess this might be possible using something like a Raspberry Pi device and some clever programming ? Any suggestions??
Garmin Vectors as torque data source for CA3?
- Thread starter Warekiwi
- Start date
as long as you can turn the data you receive from it into a voltage within the range of the torque sensor input of the ca, then you can setup the ca to respond to that.
how to do the data conversion, i don't know (not a programmer), but it should be possible as long as you can decode the datastream to find the actual sensor data, then use that to create a voltage on an analog output of whatever mcu system you use, that then drives the torque sensor input of the ca.
you will also need *cadence* input to the ca, because it will not create any throttle output for a torque input until the cadence rpm is 50 or greater. there are ways around this like feeding a 50-rpm-equivalent pulsetrain into the cadence sensor all the time, etc. but it's not really ideal cuz then your cadence has nothing to do with the control of the system, so it feels different than it should (just as using cadence-only doesnt' "feel" right vs using torque-and-cadence).
the other option is to use the mcu system to create a *throttle* signal instead, and feed that into the throttle input of the ca.
how to do the data conversion, i don't know (not a programmer), but it should be possible as long as you can decode the datastream to find the actual sensor data, then use that to create a voltage on an analog output of whatever mcu system you use, that then drives the torque sensor input of the ca.
you will also need *cadence* input to the ca, because it will not create any throttle output for a torque input until the cadence rpm is 50 or greater. there are ways around this like feeding a 50-rpm-equivalent pulsetrain into the cadence sensor all the time, etc. but it's not really ideal cuz then your cadence has nothing to do with the control of the system, so it feels different than it should (just as using cadence-only doesnt' "feel" right vs using torque-and-cadence).
the other option is to use the mcu system to create a *throttle* signal instead, and feed that into the throttle input of the ca.
Thanks Amberwolf- your reply makes me realise what a challenge this would be! (although theoretically possible) Actually the Vector system does output torque and cadence separately According to what I've read the cadence output automatically overrides any other source of cadence and the actual "power calculation" is done at the receiving unit from the two data streams. Well beyond my capabilities sadly so I can only hope that someone with access to both a CA3 and Garmin Vectors might read this and take up the challenge!
there are some software people on here; you might go to the various open-source project threads to see who is making code for those, then contact them and see if you can interest them in this project.
there are also plenty of other forums specifically for software and projects, that you can also ask at.
i have no idea how easy it is to intercept and decode the datastream this has, but i'd expect it to at least be possible.
there is a chance that they encode the data securely such that even if you can intercept it, you can't decode it (random keys, etc.).
if that happens to be the case, you could create a device that mounts inside the crank-mounted portion instead, and directly reads the sensors, and sends the data separately from the built-in transmitter, to your own external receiver unit that then outputs the signal(s) to the ca.
that would be much more complex, requiring hacking of the garmin hardware and designing sensor interfaces, but it should work if the other simpler version is not.
there are also plenty of other forums specifically for software and projects, that you can also ask at.
i have no idea how easy it is to intercept and decode the datastream this has, but i'd expect it to at least be possible.
there is a chance that they encode the data securely such that even if you can intercept it, you can't decode it (random keys, etc.).
if that happens to be the case, you could create a device that mounts inside the crank-mounted portion instead, and directly reads the sensors, and sends the data separately from the built-in transmitter, to your own external receiver unit that then outputs the signal(s) to the ca.
that would be much more complex, requiring hacking of the garmin hardware and designing sensor interfaces, but it should work if the other simpler version is not.
Similar threads
