gear profile? what is that? Here are the profiles:
I already answered why Shimano isn’t using the shifting profile:
Granted I wasn’t expecting them to take away access, but don’t think anyone was expecting that.
And already explained how the data is very different between the power meter data from the shimano crank and cycling dynamics. How do we know they aren’t in touch with the ANT+ Technical Working Group? Do you think a major change to the standard would be fast? The vectors are from 2013
and Garmin announced they wanted to get it into the standard in 2017
But the ant+ standard version of cycling dynamics came out in January 2019 (see the thisisant link above)
So kind of slow to change and Garmin went 6 years without following a standard…
Cycling dynamics is pretty simple for data in that its just 11 pieces of data over a base power meter.
for each leg:
Start Angle
End Angle
Start Peak Angle
End Peak Angle
and
Rider Position
Right PCO
Left PCO
So only needs 3 ant pages to send this data
Shimano’s power data is 49 pieces
for each leg for 12 positions around the cranks:
angle
power
and overall:
rider position
so 8-9 ant pages so easy to see how the data is very different and might even be hard to transmit that much data (the ant+ power meter spec only transmits ~8 pages a second for cycling dynamics data)
Your explanation for my Shimano isn’t using the shift profile is blown out of the water by Shimano’s action. This is exactly why standards should be followed, so idiot corporations don’t hurt consumers.
The scariest thing? It’s consistently high on the small ring, it’s consistent in its behavior of showing positive power for a few seconds when cadence is zero, it’s nearly consistent in underestimating power in sprints: so it’s mostly consistent in its inaccuracies. Over the course of a longish ride, it’s almost certain that the average power you would observe would be close to the average you would have observed from a different power meter – it’s only when you drill down that you might see the problems.
This is why comparing the average power of two power meters over a ride is a very very low bar to meet. What’s important isn’t how close they are on aveage, what’s important is knowing when they’re different, and by how much. Most riders don’t bother to do all this work because it’s boring and they just want to ride – and they shouldn’t have to do all this work, and just riding is the goal we all have for how to use our time.
@dcrainmaker’s review is even more scathing than @gpl’s. I’m really (negatively) surprised that Shimano’s last firmware update was from May 2022, i. e. they did not even attempt to at least mitigate some of the issues.
The second surprise is that the problem isn’t necessarily/only the crank arm design itself, but basically the whole package. They’d be better off just rebadging 4iiii power meters.
Doesn’t 4iiii also claim 1% accuracy even on single sided arms ?
I mean, just the thought of a crank based PM at that accuracy is suspicious/hard to believe at best but on a single sided arm I just find it hilarious as for most people (me included) it simply means that the numbers produced are super accurate.
Back on topic, I also found funny how cadence seems to be an issue even though people are forced to use magnets
Finally, I was interested in getting your thoughts on crank based PMs as to me they are only acceptable really as a budget solution and in the form of left side only.
Stages now seem to work LR, but would you really spend that much for a crank PM over a spider or a set of good PM pedals (speaking only in terms of quality of the data and leaving aside aesthetics etc)?
Didn’t you title your review video and article of the Stages DuraAce 9200 crank power meter with “This one works!”? I watched the video when you released the video, and going off vague memory here, I thought you were generally very happy with it. Or did I get that wrong?
This doesn’t seem to be (only) a calibration issue. Ray and Shane are probably the best independent testers of power meters out there. They know where to poke and what possible causes are. From @dcrainmaker’s video I got the impression that the Shimano unit failed very basic tests that other power meters will generally just pass these days.
The two have lots of industry contact that can help them get to the bottom of things. I don’t see any indication Shimano has been in touch with them: Ray stated that he used the latest firmware from, hmmm, May 2022, i. e. no beta firmware. Shane first loaned a power crank from a bike shop for a day, and then bought one with his own money. It seems Shimano is still upset with him, when @gpl uncovered that all previous-gen Shimano crank-based dual sided power meters were inaccurate. (I find that quite childish.)
If it were a “calibration issue” like you say, this could be something that could be fixed over time. But it seems the issue is deeper. That’s a huge problem.
Two things: companies have to be willing to send pre-production gear to reviewers, and Shimano doesn’t seem to be. And I think it is worse if the problems are “limited to” customer units — that’s what you want to avoid in the first place!
In theory yes, buying off-the-shelf units could be valuable, but in practice, we really haven’t seen that matter in the last decade-ish. As I’ve written before, by and large, early units (not pre-prod, but just early) of almost every tech item have worse performance than later units. But more importantly, the way units are built these days with high levels of automation, it’s almost always going to be black and white pass or fail for electronic devices (specifically focusing on the manuf build quality).
Yes, there can be cases where a company has a high failure rate and they may be hoping that by doing QA it gets a reviewer one of the non-failed units - but even that in this industry has shown it doesn’t work. One only need to see the number of power meters/trainers/etc that we find issues with - it has nothing to do with the physical building of the product, and everything to do with the design/algorithm.
Where you can get into issues is prod-prod units (such as the IQ unit), or whatever, where that may have an element of hand-build to it. I very rarely use pre-prod units in reviews, and note that accordingly. In most cases, I just simply have a unit that’s in the first production batch. And again, referencing above, that almost always skews to a worse experience than a better one.
)
?
