Trainer and PM disagree

Discuss light weight issues concerning road bikes & parts.

Moderator: robbosmans

Post Reply
lemos
Posts: 13
Joined: Fri Jul 06, 2018 11:55 am

by lemos

So, I have a Tacx Flux S Smart and Favero BePro pedals. Getting instant readings from both, on a stable pace, I get something like a 20% deviation, which is freakin huge.

How do I know which one is right? I've calibrated both at the same time.

by Weenie


TobinHatesYou
Posts: 6849
Joined: Mon Jul 24, 2017 12:02 pm

by TobinHatesYou

lemos wrote:
Sun Nov 22, 2020 10:57 am
So, I have a Tacx Flux S Smart and Favero BePro pedals. Getting instant readings from both, on a stable pace, I get something like a 20% deviation, which is freakin huge.

How do I know which one is right? I've calibrated both at the same time.

Record the same ride with both, upload to ZwiftPower Analysis or analyze.dcrainmaker.com.

Observing instant power differences visually on different devices is not that useful. You’re better off using 3s or even 10s power to get a good gauge on how far off one PM is. As for knowing which is more accurate...you’re going to need a third power source.

brearley
Posts: 400
Joined: Mon Aug 11, 2008 10:08 pm
Location: Hull

by brearley

I'd just take it for granted the flux pm is wrong.
Mine was almost always 10% low of crank based.

Take into account the flux doesnt actually have a pm inside it, power is worked from an algorithm based on cadence and flywheel speed

demoCRIT
Posts: 107
Joined: Sat Jun 01, 2019 3:04 pm

by demoCRIT

brearley wrote:
Sun Nov 22, 2020 11:46 am
I'd just take it for granted the flux pm is wrong.
Mine was almost always 10% low of crank based.

Take into account the flux doesnt actually have a pm inside it, power is worked from an algorithm based on cadence and flywheel speed
Had fluex for over a year - was pretty close to my garmin vector pedals.
Neo (currently owned) reads 2% lower vs pedals.

User avatar
Lucendi
Posts: 165
Joined: Fri Jul 19, 2019 10:37 pm
Location: Sweden

by Lucendi

There will always be some difference when comparing any two power meters. 20 % is not an insignificant difference though!

But generally a few percent difference does not matter one bit as even if the reading is slightly high or low, the power meter should still be reliable (showing consistent results).

As long as a power meter is consisten it is a useful tool for training.

TribesMan
Posts: 22
Joined: Sat Oct 19, 2019 5:51 pm

by TribesMan

Smart trainer is not a "power meter", it just estimates the power depending on the flywheel speed.

And because the estimation depends a lot on temperature and how (or when) the spin down is performed, it can be very inaccurate.

Pedals on the other hand should be very accurate since they measure torque and cadence directly.
Only thing that could influence that is wrong crank length setting.

Also you should be albe to check the accuracy of the pedals using some weights attached to the pedals (at least you can do that with the Assioma)

TobinHatesYou
Posts: 6849
Joined: Mon Jul 24, 2017 12:02 pm

by TobinHatesYou

My smart trainer (Hammer) is pretty damned accurate. All of my trainer sessions are recorded on two additional power meters.

GothicCastle
Posts: 276
Joined: Mon Jul 04, 2016 1:52 am

by GothicCastle

Why do you think they are off? Just looking at a head unit, or have you examined ride files? I'd suggest riding for 10 minutes before calibrating the Flux, then zero offset the pedals according to the instructions. If workout files differ significantly after that, you have something to analyze/ discuss with Tacx.

Stueys
Posts: 486
Joined: Sat Nov 22, 2014 1:12 pm

by Stueys

TribesMan wrote:
Sun Nov 22, 2020 4:11 pm
Smart trainer is not a "power meter", it just estimates the power depending on the flywheel speed.

And because the estimation depends a lot on temperature and how (or when) the spin down is performed, it can be very inaccurate.
Not entirely convinced that's true across the board, for instance the Neo is well regarded as having a highly accurate power meter.

With anything PM related I tend to use the DCrainmaker analyser toolset, gives you a nice visualisation capability. I would say that comparing two devices is fairly pointless though (unless you have a proven baseline on one of them), you need 3 devices to start to determine what is actually correct.

ParisCarbon
Posts: 1602
Joined: Mon Feb 05, 2007 5:39 am
Location: Winnipeg Canada

by ParisCarbon

Stueys wrote:
Mon Nov 23, 2020 4:38 pm
TribesMan wrote:
Sun Nov 22, 2020 4:11 pm
Smart trainer is not a "power meter", it just estimates the power depending on the flywheel speed.

And because the estimation depends a lot on temperature and how (or when) the spin down is performed, it can be very inaccurate.
Not entirely convinced that's true across the board, for instance the Neo is well regarded as having a highly accurate power meter.

With anything PM related I tend to use the DCrainmaker analyser toolset, gives you a nice visualisation capability. I would say that comparing two devices is fairly pointless though (unless you have a proven baseline on one of them), you need 3 devices to start to determine what is actually correct.
I agree with Stueys 100% on this.. my Neo is dead accurate with 3 SRMs and a P2MaxNG... the original Neo is the most reliable of the units...
20% quoted above is actually a HUGE difference.. especially in the online racing world.. 20% in the wrong direction @ 280 watts means you are now doing 330+ watts .. probably explains though why there are so many cat 4 A+ category riders on Zwift..

brearley
Posts: 400
Joined: Mon Aug 11, 2008 10:08 pm
Location: Hull

by brearley

As I said earlier, the flux doesn't have a power meter. It is one of the reasons it is cheaper than the neo which does.
It is possible to use a bit of trial and error on the flux with spindown to get power reading in line with your pm

Basically either stop the flywheel early so the spindown finishes earlier so the power is higher, or keep it spinning for lower numbers

Another thing to check out on the flux is for belt slipping which can be cured with a new belt or a bit of belt spray and there are also some alignment issues also which affect the power numbers.
All this requires taking the unit apart and make sure you put the screws back in the same holes theyre different lengths

Ps I do not advocate doing this to cheat on zwift :lol:

by Weenie


TobinHatesYou
Posts: 6849
Joined: Mon Jul 24, 2017 12:02 pm

by TobinHatesYou

brearley wrote:
Mon Nov 23, 2020 9:37 pm
As I said earlier, the flux doesn't have a power meter. It is one of the reasons it is cheaper than the neo which does.
It is possible to use a bit of trial and error on the flux with spindown to get power reading in line with your pm

Basically either stop the flywheel early so the spindown finishes earlier so the power is higher, or keep it spinning for lower numbers

Another thing to check out on the flux is for belt slipping which can be cured with a new belt or a bit of belt spray and there are also some alignment issues also which affect the power numbers.
All this requires taking the unit apart and make sure you put the screws back in the same holes theyre different lengths

Ps I do not advocate doing this to cheat on zwift :lol:

All modern trainers use RPM vs resistance vs acceleration to estimate your power, including the Neo. The last notable trainer to have an actual strain gauge power meter was the original 2014 KICKR, which grew less accurate over time as the module loosened up on the frame.

Post Reply