Stages Dual Power Meter Gen 3 LR Accuracy
back to main Stages Dual Power Meter Gen 3 LR Review
There are several aspects of accuracy we can look at. I’ll cover absolute accuracy of power in the conditions/environments that I normally ride in and I’ll touch on power balance and cadence as well. Cadence is the easiest; you perhaps might not be too interested in cadence alone but it is an input into power calculations and sometimes you might find subtle variations of accuracy at different cadence levels…or not.
For those of us with power asymmetries then power balance has some importance but when I’m riding I’m mostly interested in the total power level being ‘actionable’.
Cadence
On this bumpy outdoor group ride in the Surrey Hills, the cadence was +1rpm with the Assioma or -1rpm with the Stages, depending on how you see it, which doesn’t sound like much but it’s over a 1% difference.
That struck me as being strange as in a recent ride with the same Assiomas vs a different PM (Shimano R9100P) it was the Assioma that was UNDER-reporting by 1rpm.
That’s just two devices.
So on another outdoor ride, I recorded a Wahoo cadence sensor on a Wahoo Elemnt. This time the Stages was recording cadence when I stopped for a coffee as the first of these three images clearly shows, that didn’t particularly bother me as the power was still correctly zero. On the second and third of these images, I’ve drilled into time periods when there was pedalling and the Stages clearly over-cooks the cadence at times, as shown here.
Taking the same sensors indoors onto my Kickr, suddenly the cadence picture appears resolved. However, even when drilling down into the smoother sections of the ride, the cadences of all 3 devices varied by 1rpm from the lowest to the highest. This time the Stages was typically recording the LOWEST cadence. This ride is at a higher cadence level (100rpm) so the percentage error is reduced.
I don’t want to dwell on cadence too much as I’m not particularly interested in it and, I suspect, neither are you. I only mention it for completeness and because it might impact on the power numbers.
Power
I will show a few charts to paint a general picture and then look at some specifics with the GPLAMA test, with the files for you to download if you want to.
Here’s part of a ride from February. There’s ‘only’ 3w total wattage difference on this section but that is potentially within manufacturer tolerances. I’ve arrowed a potential dropout from the Assioma but actually this was correct as it was from when I wasn’t pedalling, so it is ghost power from the Stages. Overall though, I’d say this ride had actionable data.
At least superficially, there is a similarly good picture here.
Of course with 20-second smoothing, many of the vagaries are ironed out at such a high level. If we were to drill into the detail from that same ride we could quite easily find specific instants in time where there are much more significant differences – for example, DURING the following 3-minute duration of that same ride there is an instant showing a 30w difference which is part of a whole minute where the Stages is under-reporting…or the Assioma over-reporting.
This ride is also from February but I was faffing around with the sensor pairing before the ride and mistakenly ended up pairing it as left-only, even then there is ‘only’ 1.5% ish difference. I have a bias towards my left-side but, as you can see, the data is not too bad at all when Stages just double the left side. Anecdotally I think when I go above 270w, or so, then both legs seem to even out. It is what it is.
As we roll into late February I started to get some weird results like this one
When I looked at the data, the Stages right side (purple line) just wasn’t playing ball. Like this…
Hmm. It is a strange coincidence that I installed firmware v1.6.1 the previous day. This was the one that allowed the right-side gyroscope to be enabled. Hey, I mean, if a feature is there it has to be enabled, right? Maybe it was the gyroscope that caused the right-side drop, maybe it was the firmware full-stop. IDK. But after that date, I don’t have good data to show you.
I did quite a few failed tests and I can’t devote any more time to this now, so here is my latest GPLAMA test at the start of April (I did 4 in total and had various other issues with the first 3). Below, I’ve given you two sets of data for the one workout where I compare Assioma to Stages LR to Kickr. The first data set shows unadulterated power data and clearly the Stages power is wrong. In the second set, I’ve manipulated the Stages LR power data with Fit File Repair Tool and increased it by exactly 10% throughout. Why? Well, it certainly looks prettier but this also means we can look at what happens to the Left-Right balance on these Shimano cranks. But, 10%…I know, I know. You can download and look at both sets in full.
- Unaltered Data (here)
- Stages Power Data +10% (here) – Note there are no options to change the crank length and it appears correctly in the Stages Power app (iOS AND Android)
Here we go. This looks much nicer with 20sec smoothing. Putting the pesky 10% error to one side these all match surprisingly well throughout, even the under-over efforts at the end are pretty coolly similar.
I’m not going to analyse the various sections of the test because of the need to change the data. However, there IS some meaning to look at the power split. So, let’s look at the LR split for the 200w, the 250w and the under (330w) / over (440w) efforts…and, yes, I know I’m kidding myself they are over-over efforts 😉
Here are those 3 charts
Here is my take out
- I was specifically trying to balance my LR power during the test based on what the Stages showed on the FR945. I thought I did a reasonable job at the time. Bear in mind I think I have a left bias of, maybe, 2%age points (4% spread)
- The 200w section: Shows the Stages RIGHT as lower than left. It also has Assioma LEFT as lower than right (the latter makes some sense as I was probably pedalling more on the right than normal, hence my normal left-bias would have changed)
- The 250w shows the same
- The 330/440 section shows the same for Stages but each Assioma side is roughly equal. (I was just pedalling here rather than trying to balance so this reflects natural pedalling with a bit of sweat)
- Maybe the Stages under-reports the right-side power at ‘recreational athlete’ levels. Maybe the Assioma under-reports the left side? Who knows?
- GPLama’s post on Shimano R8000/R9000-based cranksets suggests that Stages would be under-reporting.
Further Investigation
I have devoted a considerable amount of time between January and April 2020 to accuracy testing and I’m not entirely happy with the data I am presenting here. Stages (Saddleback) validated that my cranks are correct.
I will re-do the GPLama test at some point and change the info here but only when I am certain that the test will work…it takes hours each time.
Summary
- Stages is under-reporting cadence on bumpier surfaces.
- If it wasn’t for that inconvenient 10% discrepancy, the data sets would probably have exceeded my expectations of accuracy. But…
- I can’t be sure that I am seeing an under-reporting of the right-side but what I do see does seem consistent with what others have reported elsewhere.
back to main Stages Dual Power Meter Gen 3 LR Review