This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Distance with autocalibrated speed sensors too low (>= 2%)

I am using my Edge 830 with three different speed sensors on three (road) bikes. They have all been in auto-calibration mode; I never questioned that, because it has been decent on my Edge 1000. However, I just noticed that all sensors report distances short by 2% or even more (so I am loosing 2km per 100km), making me think that this is a systematic issue. I am aware that I can manually set the circumference, but this is inconvenient when swapping sensors in between wheel sets (which I do; I have one sensor per bike, not per wheel set).

I observed auto-calibrated values of one sensor (on a 622x25C tire, which should result in something similar to 2105mm), and I see values as low as 2030mm (100mm off!) and they kinda "jump" all over the place (= frequent changes). Honestly, I don't see a reason for continuous updating of circumference, which makes me think that I just rely on GNSS.

  • Older units used to only auto-calibrate when first pairing or changing activity profiles (IIRC).  IDK how it works now but I think it does some kind of running average.   I generally stick with manual calibration, that way I know what I'm getting.  If you're particular, you do have to go in and change the number if you make a consequential wheel/tire change.   Changing between the usual road bike tire and rim widths (training, aero, climbing) will affect distance less than a few percent.  You might do a rollout with each and if they're not too different, stick with an average value between them.

  • Thanks Do you know how this is done in the Edge 520? A friend of mine uses that devices with auto-calibration (with same speed sensor I am using) and gets pretty much spot on distances (to what route planning, GPX track evaluation etc. say).

    Anyway, I don't get the point of why the autocalibration is done this way. From my perspective, it really doesn't make any sense to calibrate on the fly, all the time. To make affairs worse, using a tool like gpxviewer, Strava, etc. on the recorded GPX track produce a consistent (and probably accurate) distance exceeded the one provided by the speed sensor by the mentioned 2%. I think that Garmin could do better by using compensation or doing the entire calibration smarter. I must admit that I am really disappointed by autocalibration. Looks like it could be a nice thing for those of us running different wheel sets on a bike (and don't want to bother with entering manual numbers), but it's messed up really. ;-)

  • Yes.  I believe the 520 only auto-calibrates initially.  When it does, you get a beep and message around 1/2 mile or 1km into the ride that says auto-calibration complete.  It' keeps this calibration until it's triggered again by changing activity profile or switching between manual and auto calibration (IIRC).   I would choose to make this happen over a relatively uncomplicated track to improve the accuracy. 

  • Thanks, I guess I'll convert to manual settings then. However, it's pretty disappointing and difficult to understand why auto-calibration on the Edge 830 is so poorly implemented (and worse than on older units).