Before we get into the blunder, you should understand that Shimano has a marching cycle to the production of hardware. Most other
companies do as well, but Shimano’s is readily apparent. If there is innovation or not, the products keep coming. Almost without fail. It’s a waterfall. Release Dura-Ace, then next year Ultegra, then 105, then Tiagra, then Sora, then Claris. There is now an overlap after the 3rd year where it cycles back to Dura-Ace. This harkens back further when there was only a few.
From wikipedia
Something similar happens on MTB with the exception that DI2 seems to have died on mtb with the M9100 and M8100. They also changed the scheme by releasing the M7100 at the same time as M8100. They must have felt the SRAM Eagle Series pressure in the MTB world and want to jump everything to the 12 speed to tackle them. I don’t know if there will be another DI2 on MTB and I won’t get into my “wireless isn’t innovation” speech here (but short is it’s innovation ONLY for OE’s of bikes so they can lower assembly costs).
The FC9100 is, in my opinion, not a pretty crankset. It like you took a crankset and well… see below
Okay so this is an extreme analogy. But hopefully it has that impact. Face merged into an arm blobbing and stretching. Ya, I don’t like those cranks aesthetics or lack thereof, but to each their own. If you like it sweet. If you hate it great! If you’re ambivalent then awesome.
HOWEVER, guess where it went wrong. The company figured “hey, lets build a powermeter, everyone else is!”. So prototypes on the FC9000 and then the FC9100p comes out a few years later. However, have you noticed that there are very few reviews on these new generation of powermeters? Even companies who are in the arm powermeter market? Have you notice that the bigger names are only doing previews and not full in depth reviews? Companies relying on “well it worked on the last one, so if we don’t tell people of the problems then there isn’t any” type of response.
WRONG. Some people are fed up with the terrible and weak responses of companies, and they allude to outliers and cherry pick “this one is slightly off, so your other tests must be wrong!!!” to “it was a firmware problem, it is solved in version x.yy.z”
Companies like Transferwise offer “radical transparency”. I’ve gotten multiple emails saying fee’s are going up and what they are for:
“Margin and growing” = profit. But everything is detailed and they aren’t large and compared to a bank or credit card or similar it always works out to a good deal.
So enough industry ranting! What is wrong. Short, everything!.
That crankset is just hot garbage in terms of a powermeter and anyone with FEA. It’s not symmetric, it’s doesn’t have define spider arms, it doesn’t have consistent internal shape.
Since Ansys (maker of Finite Element Analysis software that simulates the mechanical deformation of stuff) made a new academic version and I have two degrees in mechanical engineering and 12 years of experience using Ansys (and still, not an expert) I threw together some quick simulations. I felt it was important to keep them in the video but here is in short the problems you’ll see me discuss.
So many simulations – ALL FREE TO DO!
- Calibration is NOT consistent. The position of the chainring causes it to change.
- IT MUST BE THE SAME TO MAKE A POWERMETER or else you’re just chasing problems
- You can mitigate it a bit by moving gages further away but…
- Axial is still a problem and can get way worse
- I keep harping on this, but over half the powermeters on arms use JUST a bending arrangement. But if you push axially (as shown in the video) on it it shows as torque because the shape of the arm is not the same on the top and bottom. Therefore Axial force looks like torque and either adds or takes away from your power skewing it. MOST companies fudge their calibration value so that the vast majority of the population reads about right. Some read lower, most read higher to align the numbers and their accuracy claims should read about +/- 5% instead of 2%. Why don’t you notice? Because it’s left only. You told it’s balance. It’s not in some cases. In some it is, but not most.
- It gets worse when you try and fix the calibration problem
- Coupling – When gages are on the face like that of a terrible asymmetric arm you get a weird response because the torque from the left arm to the spider is transferred through a shear mechanism.
- Gages get all messed up when left torque is applied
- Forward and backward response is DIFFERENT
- It depends on position (angle) of the crank from where the chain is coming off
- Chain ring (large or small) can affect it in some cases due to chainring stiffness on sharing the load.
So if you see the R9200 released this year that is more symmetric, more defined spider arm then you’ll know why. Asymmetry shot them in the foot and they didn’t even notice.
So Shane over at GPlama.com (new site!!!) and his youtube channel has been testing and testing and testing this stuff and can’t get ANY to really line up. He’s pointed it out in a few videos already. I believe in “radical transparency” to help answer the question truthfully and honestly which none of these companies seem to be able to.