Avio PowerSense Powermeter - owner first impression

power-meter

#21

Thanks, I have a Scott foil so I don’t think it would fit for my bike either.


#22

It depends on the clearance, tape a two finger kitkat to the back of your crank and if it doesn’t hit anything you should be fine :joy:


#23

I might stick my gravel bike on my tacx flux and check out the difference, although if one is measuring at the crank and effectively at the rear hub I should expect to see some variation?


#24

Yes, always! To begin with you’ll get drive train losses. Single sided meters exaggerate pedalling imbalances (in either direction) unless you’ve measured your imbalance and applied a correcting factor. If a Tacx Flux is similar to a Kickr (I think it is - there is a flywheel) then it’s likely measuring power the same way by using speed vs. braking rather than strain gauges. IIRC the 1st gen Kickrs used strain gauges to begin with but these were flakey and disabled in a firmware update in favour of speed based measurement.

They’ll be different, definitely, but you should be able to see if there are serious discrepancies which would be a cause for concern. You can properly nerd out here: https://www.dcrainmaker.com/2018/12/how-to-troubleshooting-power-meter-and-trainer-accuracy-issues.html


#25

Here is a longer workout: https://analyze.dcrainmaker.com/#/public/1ec9d0ed-46ad-46a1-7647-9c4203613e3b

If you turn the smoothing up (to, say, 10s) you can see two things really clearly:

  1. The Powersense gets a couple of cadence drops during the intervals.
  2. My left/right balance goes out the window especially as I get tired :cold_sweat:

That said, it looks reasonably consistent but wildly out. It reading much higher overall (40-50w on average!) than the P1s could be a calibration issue as the powersense is self installed, but I did calibrate both meters before the workout on their respective head units. Or it might be some side effect of swapping from using TrainerRoad to record the P1s and an Elemnt to record the Powersense to using an Elemnt to record the P1s and a Garmin to record the Powersense, as yesterday’s shorter test had the power curves basically the same.


#26

It doesn’t look good at all in my opinion. Do you or anyone else have more data comparisons? I need to buy either this or the trusted 4iiii and can’t decide until I see more data. What do you think?


#27

I went for the factory fitted option as it removed any user error or ambiguity with fittting. Until the reviews are in I would be sticking with that if I was purchasing another.

That said, the first test above showed parity whilst the second showed almost identical graphs but with a constant variation which could point to calibration.

If the graphs were inconsistent in the variation in power I would be really worried


#28

Sadly (for me) the answer is yes and it’s not looking great :frowning_face: but I need at least one more test to really nail down what’s going on here.

Obviously please note this is N=1, I’m not Ray Maker and I glued this power meter myself, but I seem to be getting a lot of drift vs my Powertap. I’ve got a few different captures of different lengths now and here is the most recent from a TrainerRoad workout: https://analyze.dcrainmaker.com/#/public/5f3dba8f-14e9-4e8b-7bd4-3700fb61cc61

That was with:

  1. TrainerRoad running on my phone, with a Kickr and running TR’s PowerMatch with a bluetooth connection to my Powertap P1 pedals (the TR fit file is not in the link above because the auto pause really messes with the timing).
  2. The Avio Powersense paired to my Wahoo Elemnt via ANT+
  3. The Powertap P1 paired to my Garmin via ANT+

Both sensors were calibrated at the start of the workout. I stopped and calibrated both sensors after the warmup, in between the first and second main set intervals and towards the end of the final interval. These should be very clear on the graphs :slight_smile:

The discrepancy between the two sensors starts off modest and then balloons over the first 10 minutes. After calibration they’re fairly consistent for a while and the rate of drift slows down. Also there’s an additional complication, which is that my left/right balance decays significantly as I get tired, making it tough to compare the overall power numbers accurately as one is left only and the other is left & right.

Of course the issue might actually be with the Powertap P1s :scream: but I use these for PowerMatch usually and it doesn’t feel like the intervals in my usual workouts get easier or harder artificially.

I do have an analyzer with a similar setup and power data from my Kickr but I forgot to do a spindown calibration on it. I’ll do at least one more test but with my Kickr properly calibrated and with all the power smoothing etc turned off so I’ll have three different power measurements to compare, then maybe I’ll know what is drifting and why.


#29

I’ve been meaning to test mine against a tacx flux but been coughing up lumps all week :frowning:

The plan is to connect the avio to a garmin 810, the tacx to a garmin 1000 ( both via ant +) and then ride a TR session controlled via Bluetooth.

I am expecting to see a variation in recorded power, but would like the graphs to trend the same rather than drifting

I think the three devices route for testing is the only real way to get an idea of which one is ‘wrong’ though


#30

The most recent one is worrying in that respect. Looks like in between the starting (cold) calibration and the second one after the warmup the difference between the meters goes from around 10w (by eye) or 9% above the lower value, to 40w (again by eye) or 16% above the lower value. The way the drift starts off really noticeable then seems to slow down* as I progress through the workout and recalibrate makes me wonder if one or both meters aren’t properly compensating for temperature.

*The ‘drift’ towards the end might be more that my left/right balance is noticeably falling to pieces - so maybe it’s barely drifting at all after it’s warmed up?


#31

Could be, if you have a known and pronounced L/R imbalance then that needs to be factored out somehow.

I understood that the avio has an inbuilt temp sensor so ‘should’ compensate for temperature variations… :man_shrugging:


#32

Actually, looking at it again, my left is stronger than my right (assuming that the left and right pedals are correctly identified in the analyzer) which makes it doubly strange that the Avio is reading lower in the final interval. Can’t isolate left power or apply a 50% reduction to the single sided meter in the left/right balance graph, so might try exporting my next attempt to CSV files and work out some metrics myself :nerd_face:


#33

Thanks for the replies!

From DCR:
" Would I buy it: It’s tough. I’ve been using one (or, actually, a pile of units) on and off since spring, and up until late summer it’s definitely felt more like a beta product. I think they’re getting close, and the company in the last few days plans to send over their latest variant. While it is available for sale and purchase today, I don’t have enough time on the latest unit to make a complete determination here. The lack of dual ANT+/BLE is also a tough one as well."


#34

Sitting on my turbo right now, the avio is reading higher than the tacx by a constant 30-35w

Obviously I’m running both garmin as well! Not just taking pics :slight_smile:


#35

1 hr ride on pettit-1

https://analyze.dcrainmaker.com/#/public/ca4eef40-86d3-4f48-7810-627b40542a70

short ride on free ride with some greater power variability

https://analyze.dcrainmaker.com/#/public/c444029b-4528-4f9f-71b3-64087ddc3eba

so, IANAE, but my impressions are.

  1. Both devices suffered some drop outs
  2. power curves are largely consistent.
  3. the tacx is WAY smoother, although I would expect to see that as it is hub mounted and has a flywheel.
  4. variation between the two was less on steday state and grew as power increased, in percentage terms it may be a similar variation though.

What have i learned? er, apart from if I did an FTP test with my Avio I would have a figure that was more in keeping with my hopes and expectations?!

Actually… this does fit with my experience when i purchased the tacx, I did start using it without carrying out an FTP test (I know) and failed a few workouts, my subsequent FTP test did show a reduction in FTP and I have been able to complete worksouts OK since then. That FTP test came on the back of a decent enough summer involving plenty of riding and an autumn of cx rides with some success.

So, is the Tacx under reading? I don’t know but i suspect it is a little. without a third PM it is hard to say much more than that.


#36

Ok, here is an hour workout with data from three things:

  1. Powertap P1 dual sided paired to a Garmin.
  2. Avio Powersense paired to an Elemnt.
  3. Wahoo Kickr paired to iPhone and TrainerRoad*.

Note that the Kickr is in erg mode so, as they have said on the podcast a bunch of times over the years, it is lying through its teeth about the smoothness of the power being output. This doesn’t invalidate it per se, I’ve left it in as a reference value, but it’s artificially smooth.

I did spindown calibrations in TR, Wahoo Fitness and a Factory Spindown in Wahoo Fitness to make sure, this included 10 minutes of warmup then a further three for the factory. So everything was good and warm by the time I started any of the captures, just before which I calibrated the two power meters via their paired computers.

This capture looks much, much better than the ones before. There seems to be less drift, but it’s still there. In the early part of the workout, compare the left Powertap vs the Kickr on the Left/Right Power graph, to the Avio vs the Kickr on the Power graph. They look like a really, really good match! The Avio seems to just dump a few watts compared to the left Powertap at the 31:30 mark and never recovers. Overall though, this is much better! I wonder what the difference is? Maybe my hypothesis about temperature compensation is a good bet, because I don’t think the temperature was very variable during the workout (started in a warm state, didn’t change much).

If I get seriously bored over Christmas I’ll plot the % difference between the left pedal and the Avio and get the temperature graph from everything that reports it. Already got the files exported to CSV :nerd_face:

Anyway here is the link: https://analyze.dcrainmaker.com/#/public/157e6730-7d1d-4795-7f06-92705800cfec


#37

That’s interesting, did you do a full weighted recalibration of the avio?


#38

I couldn’t! One of the annoying things about this meter is that it only has an Android setup and calibration app. I have to borrow my housemate’s phone to calibrate it. Also I was pretty confident I’d got it right the last time. Of my three different scales, one read slightly more than 10kg, one read almost dead on and one was slightly below :slight_smile:


#39

Very tempted to get one of this fitted. Would you recommend it to a friend?
With code can be purchased for 149gbp fitted. Btw how much is it for delivery to Ireland?


#40

I’m not regretting my purchase. I don’t think it’s as mature a product as 4iii or stages. But you are paying less as a result