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

Power meter connection drop

Former Member over 4 years ago

I am still using a fenix 3hr to record my training sessions. It works perfectly fine with the (2018) Rotor INpower on one of my bikes. However, I started using a Rotor INspider power meter on another bike and am not too happy with the data recording, because the watch-to-pm connection drops very frequently. I cannot tell if Garmin or Rotor are to blame, but the fenix makes the problem worse, since it records the missing power values as "0" digits in the fit file, i.e. the same way it records coasting (freewheeling). Interestingly, some Edge models apparently record "Null", i.e. a string, if the connection to a sensor drops, which makes it easy to curate the data afterwards.

Now I have two questions. 1) How frequently do fenix 5 users experience dropping connections to Ant+ sensors, especially power meters? 2) How does the fenix 5 (or 6) series record missing data in the fit file? ("0" or "Null"?)

Top Replies

Data over 4 years ago +1

Perhaps someone knows this right away, but I'm going to test this and post the result here (it's something I need to know also). I am not using a real powermeter but a PowerCalDuo HR-based "power estimator…

All Replies

0 Data over 4 years ago

Perhaps someone knows this right away, but I'm going to test this and post the result here (it's something I need to know also). I am not using a real powermeter but a PowerCalDuo HR-based "power estimator" but I don't think that matters for this test (watch not receiving any data from the sensor).

0 Former Member over 4 years ago in reply to Data

Any update? I also filed a support request. Maybe they will fix it for the f3.

0 Data over 4 years ago in reply to Former Member

My first test failed. The activity was on auto-pause, so there was no recording at all of the part where the sensor was put away. Second test: first started with PowerCal and Wahoo RPM sensors connected, then put PowerCal away for a while and saw the display going from a valid power reading, to 0 and then to "--". Then put the Wahoo RPM away, same thing (from valid to 0 to "--". In the FIT file, all entries with power sensor not connected are 0, but the cadence field, when sensor disconnected, is absent. So that's not the result you were looking for unfortunately. However I would not really trust this power sensor, it is not a 'real' one. Perhaps someone can redo this with a real power sensor?

0 Data over 4 years ago

If I get a chance I'll test it again with a Forerunner 945, I can borrow one to do that.

By the way, how are you processing your FIT files? I have never seen a string "Null" in a FIT file while processing with Java and the ANT+ FIT SDK, missing data (in the "record"-messages in the activity FIT files) is either a missing field, or has 0 as value.

Although this can depend on device-software versions. My Fenix 3 logged 0 for missing GPS coordinates for a while, later that was fixed by omitting the GPS coordinates fields if there was no data.

0 Former Member over 4 years ago in reply to Data

I use GoldenCheetah and my own Mathematica Notebook offline, and intervals.icu online. The intervals.icu guys told me that missing power data should be reported as "Null"/blank field in the fit file. However, I found only "0". Meanwhile, I switched off the "autocalibrate" function of the power meter. This seems to have improved the connectivity. I have less dropouts. So maybe I don't need to worry about it anymore. (Need a few more rides on this bike to confirm.) BTW, Garmin support replied and told me that the developers would look into it. If it's also an issue with later fenix models, they may actually do something about it. Fingers crossed.

0 Former Member over 4 years ago

Had the chance to test a fenix 5x plus. Works nicely with the INpower, but has no connection to the INspider about 50 % of the time during an activity. (Compared to this the fenix 3HR is flawless.)

0 Luko79 over 4 years ago in reply to Former Member

Can You share a FIT file? I suppose the InSpider is the problem (honestly, only the cranks what the Rotor is made are good, their powermeters are not really...) , probably its radio signal strength is not strong enough (same problem with the GEN1/2 Stages) . I have F3 and it is working flawlessly with pedal based or spider based powermeters without any signal drop.

Dont You have any connection to the InSpider via BT during the ride (with your mobile)? because that decrease the ANT+'s signal strength a bit or overload the BT/ANT+ chip in the powermeter.

0 Former Member over 4 years ago in reply to Luko79

How can I check if Bluetooth is in the way? The INspider also supports it, but I don't know how to chose between ANT+ and Bluetooth when pairing. The only other sensor I use is an old ANT+ only Garmin HRM. But even if this one is disconnected, the problem persists. And there were no othe Bluetooth devices nearby that could have interfered.

0 Data over 4 years ago in reply to Former Member

It's hard/not to distinguish between dual ANT+/BLE sensors while pairing. But after pairing, you can check what type it is by going in the settings, while it is connected there's an "about" that shows the device number and "tech" which lists ANT or BLE. Then delete the BLE sensors that you don't want to use.

You could do a test ride with your Fenix strapped to your ankle, as close as possible to the powermeter. If there are no signal drops, hopefully you can find another way to wear the watch such that it has a better reception.

Needless to say, be real careful riding with a watch on your ankle.

0 Former Member over 4 years ago in reply to Data

Just figured this out a minute ago. Not sure how it was paired before. But when pairing again, the watch and the PM preferred BLE. Took me a couple of times to get the ANT+ pairing done. Will re-check. The f3 is paired via ANT+ for sure. And the connection is stable. Next ride with both watches... If the f3 connection is then dropping too, it's the f5's fault.