Ok, I've received my Nexga S1 and completed a few tests comparing it to my Tacx Neo 2T.
Buckle up, this will be a long post, heavy with data.
I'll start with a question: should a zero offset calibration be completed before starting each and every ride, perhaps even after pauses during the ride? (not talking about accuracy tests)
I usually don't bother doing so and completed at least 150 rides with my Magene P505 equipped bike, at least 80% of them without spinning the crank backwards 5 times (to calibrate the meter).
While this is not scientific, I've never felt like the numbers reported by the meter were off, I feel like I've developed some sense of what power numbers should be (ballpark) for a given effort. Also, there has be no wild variation in my performance over the 150 or so rides, only a very steady (i.e. slow) progression.
I also have 2 bikes equipped with Geoid PM500 power meters and haven't noticed anything weird either. I don't remember if I performed a zero offset calibration for the PM500 test I did (
https://chinertown.com/index.php/topic,6617.msg116827.html#msg116827).
This long introduction leads me to what I believe to be a huge issue with the Nexga S1. It seems to require incessant zero offset calibrations to be performed in order to generate accurate data.I will frequently refer to "realtime offset" numbers in this post to illustrate this point. The companion app allows to export a log (great feature!), which includes "realtime offset" amongst other data.
Zero calibration brings the offset number back to 0.
It is claimed in the manual that the Nexga S1 has an auto calibration feature.
Offset was at 3813 the first time I started the device.
Offset was at 2262 when the first calibration was done, 2h30m after first startup.
Temperature between the first calibration and the last calibration as reported in the log was between 22.1 and 26.8 Celcius. The bike stayed on the trainer in my living room for the entire time.
Firmware was updated to 0.1.5 before installing.
Installation and setup was straigtforward.
Test 0 (initiation)This test was performed right after installing the power meter on my bike. Instructions were to install, then do a few out of the saddle efforts, then calibrate. This test shows this.
Offset was -189 right before this test, which started 3h40m after the last calibration.
Calibration is at around the 2m30s mark. Offset before calibration was -227.
There is a 27 watt gap between the Nexga S1 and Neo 2T before calibration. Keep in mind that the power meter was last calibrated only 3h40m before starting the test. Also, Nexga is much quicker to notice when I stop pedaling.
The gap is down to only 7 watts after calibrating, but it is a longer period.
But it doesn't look that good at the last ERG steps with a 12 watt gap.
Sprints look fine tho with a 5 watt gap and no significant residual offset after them. Again, Nexga is much quicker to notice when I stop pedaling.
Goal here was to test cadence tracking. Both Nexga and Tacx managed to keep up to 156/155 rpm, but Tacx thought I kept going when I actually stopped (notice the power curve correctly sloping down for both devices). For some reason, peak power was 96 watts higher with the Nexga
End of test 0.
I did calibrate the Nexga after the test, offset was at 21 before calibrating.
Test 1This test was attempted the next day, 21h after the last calibration. It was a bust because I configured a custom workout in Mywhoosh that was too intense for my motivation... but
The 79 watt difference is shocking to me. I did not calibrate before the test in order to test how the Nexga would perform if I did not calibrate it like I do not calibrate my Magene / Geoid power meters before most rides.
Offset at power on just before the test was -719, which is quite a drift after 21h.
Test 2This test is with a revised workout and is meant to compare the data a day or so after calibration and the data right after a new calibration. It was performed about 22h after the last calibration.
The difference is pretty clear...
84 watt difference before calibration.
4 watt difference after calibration. Offset was -693 at the moment of calibration.
Zoom on ERG segments and the sprint after calibration. That is the kind of results I'm hoping for.
End of test 2.
Offset at the end of this test was 22.
Test 3The final test's goal was to test if a 1h pause without calibration would effect the data. This test was started 1h after the end of test 2.
It did... Offset was -149 right before the start of this test.
Not calibrating the Nexga for 1h or so resulted in a 17 watt error.
Calibrating the Nexga, again, resulted in data that is aligned with the Neo 2T. Offset was -92 at the moment of calibration, a difference of 57 in only about 5 minutes of pedaling.
Conclusion-
- Good accuracy... if freshly calibrated
- Data isn't affected by sprints
- Data is trash after a ~1h pause
- The problems in tests 1 and 2 were very obvious just looking at my Garmin 1040 connected to the Nexga (going back to the intro and me not noticing anything weird using my Magene and Geoid power meters without calibrating them)
- I suspect that the auto calibration feature does not work as I see nothing relating to it in the log (I only seem calibration events initiated by me), nothing in the app and the test results could suggest it isn't working
I'm obviously disappointed, but I knowingly took a risk...
This seems like an unfinished product that wasn't properly tested... or maybe it's teething issues and I'm paying the early adopter price.
Maybe it could be fixed by firmware? Auto-calibration fix?
Or maybe my unit is faulty? The (shorter) pauses seen in test 3 of Panda Podium's report don't seem to cause problems and the unit holds its calibration over more than 4 hours (
https://www.pandapodium.cc/post/nexga-s1-the-most-accurate-chinese-power-meter/).
I can't use it as it is.
I'll be contacting Panda Podium.
Edits: a bunch of typos, small mistakes and some points in the conclusion
