Shimano power meter inaccuracy, and assymetry

Discuss light weight issues concerning road bikes & parts.
RocketRacing
Posts: 888
Joined: Thu May 10, 2018 2:43 am

by RocketRacing

Gplama has reported on this, as has this guy. https://titanlab.co/shimano-crank-design-blunder/

https://m.youtube.com/watch?v=M0tWTUwpt1k

The quarq was felt to be a very good design. How about the new axis version? Like shimano, they went to an assymetric 4 bolt setup. Are they going down the same rabbit hole, or am i missing something? (Edit, the quarq is a symmetrical 4 bold setup, where the shimano 4 bolt cranks are far from symmetrical).

Just looking for answers as i want one... but would go with pedal options if there would be an inherant issue...
67F325B2-EED0-452E-B4AC-12FE39778F4D.jpeg
0C0D4AD8-2142-4430-882A-A640F1468779.jpeg
Last edited by RocketRacing on Sun Oct 06, 2019 5:26 pm, edited 3 times in total.

Cemicar
Posts: 147
Joined: Mon Aug 21, 2017 7:40 am

by Cemicar

Meanwhile, Pioneer seems to overcame this 9100 issue by silently upgrading their sensors last year. And nobody has complained about the accuracy or the L/R balance indeed.

by Weenie


RocketRacing
Posts: 888
Joined: Thu May 10, 2018 2:43 am

by RocketRacing

Cemicar wrote:
Sun Oct 06, 2019 5:02 pm
Meanwhile, Pioneer seems to overcame this 9100 issue by silently upgrading their sensors last year. And nobody has complained about the accuracy or the L/R balance indeed.
I think it would need to be properly tested to know for sure. That is part of the issue.

TobinHatesYou
Posts: 4196
Joined: Mon Jul 24, 2017 12:02 pm

by TobinHatesYou

Cemicar wrote:
Sun Oct 06, 2019 5:02 pm
Meanwhile, Pioneer seems to overcame this 9100 issue by silently upgrading their sensors last year. And nobody has complained about the accuracy or the L/R balance indeed.

Nope. AFAIK no Shimano R9100/8000/7000 based dual-sided PM is accurate. The very nature of strain measurement and the asymmetry of the crank arms and bolts prevents that. Users of Pioneers can’t know if their L/R measurements are accurate without having at least two other power sources to test against. How many people have gone that far? The Pioneer update was mostly just to keep up with the Joneses by adding true BLE support.

velolive
Posts: 17
Joined: Sun Sep 09, 2018 6:49 pm

by velolive

It's impossible to fully overcome this issue, right side will be still less accurate.

gurk700
Posts: 314
Joined: Fri Jan 20, 2017 7:40 pm

by gurk700

Correct me if I'm wrong but Quarq and the new AXS aren't crank based. They are spider based. I don't think they have the inherent problems crank based PM's (specifically right side of newer shimano cranks) suffer from. They measure power completely differently, at the spider.
So I don't see any reason why AXS should have issues. They probably carried a lot of technology from the older Quarq DZero and besides some edge cases, those are as accurate as PM's get besides some smart trainers.

TobinHatesYou
Posts: 4196
Joined: Mon Jul 24, 2017 12:02 pm

by TobinHatesYou

^ Correct.

philipeleven
Posts: 56
Joined: Mon Sep 23, 2019 2:01 pm

by philipeleven

Shimano use a far more complicated solution that normal 4 strain gauges. technically it could detact bend and torque of every directions. The problem might be the way they calibrated the amplification factor. They might use a simple static calibration instead of dynamic calibration.

and……spider PM is a complete different scenario. The main concern is to prevent value difference between big ring and small ring
Attachments
微信图片_20191002183609.jpg

RocketRacing
Posts: 888
Joined: Thu May 10, 2018 2:43 am

by RocketRacing

philipeleven wrote:
Sun Oct 06, 2019 6:45 pm
Shimano use a far more complicated solution that normal 4 strain gauges. technically it could detact bend and torque of every directions. The problem might be the way they calibrated the amplification factor. They might use a simple static calibration instead of dynamic calibration.

and……spider PM is a complete different scenario. The main concern is to prevent value difference between big ring and small ring
One of the issues raised in the video above is that the assymetric crank design of the shimano crank means that the callibration is constantly changing based on where one is in the stroke. And factoring in the left side data makes it even worse. It is doing a lot of power estimation, and thus becoming less of a power meter, and more of a power estimator.

It is a long watch, and mostly over my head, but check out the video.

Cemicar
Posts: 147
Joined: Mon Aug 21, 2017 7:40 am

by Cemicar

TobinHatesYou wrote:
Sun Oct 06, 2019 5:53 pm
The Pioneer update was mostly just to keep up with the Joneses by adding true BLE support.
I read that Pinoneer has been aware of the 9100 issue and has silently replaced the strain gauge from the end of 2017. The 2018 autumn update included another gauge change alongside of the Bluetooth smart support and the improved waterproof (IP68).

Don't know it's still somehow inaccurate even after that. Very few people compare multiple power sources (for instance, I'm curious about the comparison of the L/R balance between Quarq and truely dual-sided powermeters, but nobody bothers).

TobinHatesYou
Posts: 4196
Joined: Mon Jul 24, 2017 12:02 pm

by TobinHatesYou

Cemicar wrote:
Sun Oct 06, 2019 10:17 pm

Don't know it's still somehow inaccurate even after that. Very few people compare multiple power sources (for instance, I'm curious about the comparison of the L/R balance between Quarq and truely dual-sided powermeters, but nobody bothers).

It's those of us who have bothered that noticed. Not only is the L/R balance wrong, the total power is incorrectly reported.

Hexsense
Posts: 936
Joined: Wed Dec 30, 2015 12:41 am

by Hexsense

It can be quite accurate if their dynamic compensation match your pedaling characteristic. That's the assumption which is hard to evaluate. How do they know how much to dynamic compensate at each angle? Based on estimation from historic data of normal pedal stroke? How much deviation your pedaling is from 'normal'?

Anyone remember some early power meters which is inaccurate for some reviewers because they have 'unusual pedalling circle'?

TobinHatesYou
Posts: 4196
Joined: Mon Jul 24, 2017 12:02 pm

by TobinHatesYou

^ This is somewhat tangential. Cadence isn't the issue here. Shimano-crank based PMs could have the same IAV measuring Assiomas have, and they'd still be inaccurate because it's more of a dynamic calibration issue rather than resolution issue.

RocketRacing
Posts: 888
Joined: Thu May 10, 2018 2:43 am

by RocketRacing

TobinHatesYou wrote:
Sun Oct 06, 2019 11:00 pm
^ This is somewhat tangential. Cadence isn't the issue here. Shimano-crank based PMs could have the same IAV measuring Assiomas have, and they'd still be inaccurate because it's more of a dynamic calibration issue rather than resolution issue.
Exactly.

by Weenie


kwakekeham
Posts: 92
Joined: Fri Jul 22, 2016 8:03 pm
Contact:

by kwakekeham

philipeleven wrote:
Sun Oct 06, 2019 6:45 pm
and……spider PM is a complete different scenario. The main concern is to prevent value difference between big ring and small ring
And somehow avoiding SRAM patent US8505393B2 infringement. Claim 1 is a a dozzy!

Post Reply
  • Similar Topics
    Replies
    Views
    Last post