Smartspin resistance readout + power calculation

[jekkos]

HI, I have been using smartspin for some time but was never able to get it to work reliably. This is mainly due to the following issue

• The bike power meter reading is incorrect. Currently it does not take the resistance into account. I checked the bike electronics and it seems that a potmeter is used to feed the position of the resistance knob to a BLE board. This board also uses cadence as input and calculates a power value which is then sent to the SmartSpin. However currently the reading clearly sees the resistance as constant as I'm seeing very high power values at high cadence (and thus low resistance) which does not make sense

• My idea to solve this is to wire this resistance to the Smartspin PCB on an analog input. Next I'd use my Favero power meter to log data points at different resistance and cadence as 'reference' values. Then I would reconstruct this relation by using the data of these two inputs: resistance and cadence on the SmartSpin. I would try to do this first in python using a curve fit and then replace the heart rate to power function to set the bike Power using Cadence and Resistance as inputs

I wanted to check if this idea makes sense or if you think there is an easier way to get a better power reading from the bike without using a dedicated power meter. I prefer to leave the latter installed on my roadbike and then I don't need swap them every time I want to train indoors.

[doudar]

Check out the Unit_Testing branch. It has power table for power and supports cadence sensors. Instead of a potentiometer, we home the knob instead, so that we know absolute position from "zero". Then we can use an imported power table (created from using a real power meter on the bike) for inferencing power vs resistance position vs cadence.

It's a work in progress, but I'm working on it pretty heavily.

[jekkos]

Sounds great, this solution sounds way better. basically the powertable what I need here. Will check out the code to see how it works.

[jrbarrick]

Yeah, the new solution should be pretty neat to test out once it's up and running. If you're curious about connecting directly to the potentiometer on the bike, I've got a post here about how I did that with my Joroto X4S:
#497 (reply in thread)

I went down the rabbit hole after I built myself a DIY power meter, but it wasn't accurate enough to be useful. The cadence sensor on that meter plus the resistance from the bike directly did let me create some power value calculations that matched up with the bike's power display, was a pretty fun project

[jekkos]

@jrbarrick nice work, that looks very much like what I was planning to do. I'm quite interested int he PCB modifications, what type of wire did you use for the soldering?

I was not aware that this feature was actively being developed. I tried the Unit_Test branch briefly and found indeed that the CSC pairing is working now. So probably all of this won't be needed and I'll just spend my efforts testing what has been developed here.

[jrbarrick]

The wire is .4mm enameled wire, might see it sold as magnet or winding wire. Super handy for this type of work, since the very thin layer of insulation just melts off when soldering, so no need to strip the wire ends. Can get it in all sorts of sizes as well.