Login required to started new threads

Login required to post replies

Prev Next
Which powermeter to trust?
Quote | Reply
In the past year i have been trying to look at power to structure my training and determining training intensity.

Becuase the first two weeks if april has been all about hail and snow where i live, i decided to put my TT bike on my Tacx Flux S for a bit. The bike has a single (left) pedal powermeter, Assioma Favero. Read out of the Tacx using the Tacx app on my phone. Readpout of the Assioma on my old Garmin Edge 800 that i use as my TT bike dashboard. I am seeing a big difference between the two and wonder which one is the more accurate.

On an easy one hour recovery ride last night with a couple of intervals in the end because i got bored i got:
Tacx Flux S: avg143/max515 Watt.
Assioma Favero: avg115/max580 Watt.

Any ideas on why i get this big difference and how to fix it??
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
Hi,

I will try and answer this. You have some differences inherent to where power is measured. Generally speaking, the closer to you (read the crank) the closer the power measurements are to your 'real' power. You have to also keep in mind that a left side power meter takes your the measurement from you left leg and doubles it. Most people have an imbalance (my right leg is much stronger than my left) so a left side power meter would mean that I would lose 25% in power measurements (yes the imbalance is that big and I am working on correcting it). If you do single leg squats, for example and see a significant difference you may find that partially explains the imbalance. You could look into making sure that your power meters are calibrated and have temperature compensation which can also create some differences. In the end, you should only compare apples to apples (left side power meter to another left side power meter and tax with a Kickr for example) with what I mentioned in mind. I have a Kickr Core and a 4iiii left side only (which I will upgrade soon to dual) and I am sure that if I paid attention to the numbers they would be quite different but I don't. Most of the reputable power meters are usually accurate to +/- 2% and what you really want in a power meter is consistency/reliability. When riding indoors I suggest only using the trainer for power and when riding outdoors use the left side only, keeping in mind you might see 10-15% difference and just ignore it as long as the wattage is consistent for the left side (you can compare on your most common routes over time).

I hope this helps and stay safe.

"see the world as it is not as you want it to be"
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
I use Rotor INspider.
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
Ok, firstly the max power is where it's really expected to be fair differences, and so actually those are really close. It's the average that's the one which is unusual.

But really to understand you need to record a ride or two using two head units (use an old device, borrow one from a mate). And then get the *.fit files and upload into DCRainmaker Analyzer to compare.

At low powers then depending on the gearing you use then I've found the tacx trainers (neos) pretty meh in terms of power. It's near their floor. Equally one legged powermeters when just pootling could also be at their least accurate depending on your pedallign style/skill.
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
You cannot expect any degree of precision with a single sided PM

Within a single ride, my R/L balance can swing from 45/55 to 55/45

I’d use single side as a directional estimate and the flux as your source of truth
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
My first reaction is 143W vs. 115W is a massive difference. Something must be wrong.

I lean toward the Flux S as the one OK, since a 115W workout is barely touching the pedals. So, look in your Garmin for signs that the Assioma was dropping out or irregular. Maybe it will be obvious there. Then look for the same in your Tacx file. The nut is, do the charts appear to generally represent what you did, or do they have any spots where they jump around. But, other than that, there is no sure way to know which one is misbehaving.
Quote Reply
Re: Which powermeter to trust? [mvenneta] [ In reply to ]
Quote | Reply
But...As long as it's consistent it's OK!
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
In no particular order

- Double check you have the crank length properly set or the pedals will be off.
- Double check that you have properly put the washer on the Assioma pedals when needed. If the black "cylinder" is touching the crank arm your numbers will be off
- Try doing intervals at 150, 200, 250, 300.....and compare the intervals. You may gain insight
- The L/R imbalance may change at different power levels. At threshold I am more 50/50. At easy efforts, I tend to be right dominant.
- Contact Favero support. They can do remote diagnostics. They were amazing with me and it turned out to be the washer issue above
- You can place a weight on the Favero and check calibration
Quote Reply
Re: Which powermeter to trust? [mvenneta] [ In reply to ]
Quote | Reply
mvenneta wrote:
You cannot expect any degree of precision with a single sided PM

Within a single ride, my R/L balance can swing from 45/55 to 55/45

I’d use single side as a directional estimate and the flux as your source of truth

Nope.

My single-sided Faveros are ANNOYINGLY precise (reproducible). Like literally down to the watt. I wish it would vary upwards, but nope!

It's so precise that I bought two singles for two bikes.

I also suspect they are highly accurate, as I cannot tell the difference when I swap between the two Assiomas I have. I was expecting some wattage differential but nope, also annoyingly similar down to the watt.

I've done several FTP tests on Zwift where I'm not tracking the power target as I ride and not in erg mode, and my Assiomas again - like literally down to the watt.

A 10% watt variation is totally unrealistic in my usage on Faveros with SS. Even 5% is totally unrealistic. If there's any watt variation for me, it's definitely <5%.

If I had a 5% wattage upward variation in my FTP tests, which would be around 10 watts at my FTP, I'd be cheering hallelujah!! If I were seeing +10 and -10 for identical efforts on FTP tests and all my other rides, I'd be calling my Favero SS unusable. Clearly that's not remotely the case.
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
@OP - You might be having some smart trainer power drift as the unit warm up or cools down, but I'm just extrapolating from my experience with the highly regarded 1st gen Kickr, which while every review says is 95% accurate, for sure has power drift for me as it warms up over the course of an hour. I can literally feel it, as at the start of my ride, the Kickr might say 200watts and it feels definitely too easy for me, then I look at my Faveros which I always run at the same time, and it's reading 170-180watts. For mine, it takes about an hour of riding (unfortunately) for that gap to close or get really small. So I end up always recording my (lower) Assioma power rather than the Kickr power, as it feels more correct to me.

I suspect quite a few other (?most) smart trainers have similar problems to lesser degree, even the top-rated ones; people are amazingly silent about discussing it, probably because they're seeing it too and afraid to acknowledge that there are difference between their smart trainer and their separate powermeters after dropping so much money on them.
Quote Reply
Re: Which powermeter to trust? [lightheir] [ In reply to ]
Quote | Reply
Have you tested against a dual sided power meter / crank based / accurate trainer?

I used precision, when I should’ve said accuracy.

For example ... I’ve seen when folks sprint they typically have higher degrees of power imbalance than steady state efforts, and then that’s slightly more imbalanced than a noodle pace

So if on sprints you’re 54/46 and then steady state you’re 50/50... any extrapolation or power curve will be off

Yah?
Quote Reply
Re: Which powermeter to trust? [mvenneta] [ In reply to ]
Quote | Reply
mvenneta wrote:
I used precision, when I should’ve said accuracy.
You are OK in your use of terms.
  • Precision = ability to recreate consistent results (even if the results are incorrect). If you have a consistent power balance, a single sided PM could be precise, but still inaccurate.
  • Trueness = ability to report close to the actual value. If you have a variable power balance, a single-sided PM could be true but inacurcate.
  • Accuracy = combination of Precision and Trueness. Unless you have both a 50-50 power balance that never changes during the course of a workout, only a dual-sided PM can be accurate (both True and Precise).
A single sided PM may be Accurate (or perceived to be accurate) for some. But they are not for the general population.
Last edited by: exxxviii: Apr 14, 21 8:29
Quote Reply
Re: Which powermeter to trust? [mvenneta] [ In reply to ]
Quote | Reply
mvenneta wrote:
Have you tested against a dual sided power meter / crank based / accurate trainer?

I used precision, when I should’ve said accuracy.

For example ... I’ve seen when folks sprint they typically have higher degrees of power imbalance than steady state efforts, and then that’s slightly more imbalanced than a noodle pace

So if on sprints you’re 54/46 and then steady state you’re 50/50... any extrapolation or power curve will be off

Yah?


My Kickr 1 by most reviewers including DCRainmaker is 95+% accurate. It lines up extremely well with both my duos when it is warmed up as I mentioned,

As for accuracy, it's hard for me to believe my accuracy is off when all 3 separate devices line up nearly identically (and the Faveros to the freaking watt!). Having multiple devices that read the same is a pretty good sign of accuracy.

There is no way at all that my Unos are more than 10 watts off in precision (5%). I would notice that immediately. To even get a +5 watt increase in my ride averages requires a noticeable increase in average effort, every time. And for FTP tests, it's literally down to the watt. 1-2 at most. As said, I wish it would from time to time give me +5 on that FTP just to make me feel better, but nope. Literally +/- 1, MAYBE +/-2, but that's it. +5 FTP is like hallelujah land for me, really!
Last edited by: lightheir: Apr 14, 21 8:36
Quote Reply
Re: Which powermeter to trust? [lightheir] [ In reply to ]
Quote | Reply
I’m envious of you!! :)
Quote Reply
Re: Which powermeter to trust? [mvenneta] [ In reply to ]
Quote | Reply
mvenneta wrote:
I’m envious of you!! :)


Don't be envious when gaining a paltry 5 watts is so freaking hard for me nowdays.

It's sad when I read some of those aero testing sites and see that I might be able to equal the wattage gain from a whole block of hard training by just moving my water bottle from the lower frame to BTA =

That said, while I'm far from a speedster on the bike, I've gotten to the point of diminishing training returns such that I really notice aerodynamic changes despite insisting to myself over the past 7 years that as a front of MOPer, that training >> aero for me. I've learned the hard way that at least where I've gotten to myself now on the bike, aero speed gains can exceed my training gains. Road vs TT bike is like 4 minutes/hr. Add race wheels, another 2 min/hr. Add TT helmet, another 1 min/hr. Change from regular bike jersey to tight trisuit, another 1 min. (Put on not-that-flappy cold weather jacket, slow by 1min/hr).

Put all those together and I'll need over 40 watts of training to account for that gap between aero-TT vs road-bike nonoptimized.

I've got a new bike coming soon which should get me better aero position and I have yet to try latex tubes on race day, but that's next.
Last edited by: lightheir: Apr 14, 21 8:44
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
I only trust power meters I can perform a static torque or force check on (and preferably adjust)...or can compare against one that has been checked...and that measure the contributions of both legs.

http://bikeblather.blogspot.com/
Last edited by: Tom A.: Apr 14, 21 9:13
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
As others have said, you might expect a small amount of variation due to using one-sided Assioma. You might also expect a small difference when comparing measurements at the crank/pedals vs the trainer due to drivetrain losses, but your results are the wrong way around for that.

As each source of power is being reported into a different "head unit" could there be anything going on there? Not counting zeros, "smart" reporting intervals vs every second, averaging etc?

Cheers,
Rich.
Quote Reply
Re: Which powermeter to trust? [Tom A.] [ In reply to ]
Quote | Reply
Tom A. wrote:
I only trust power meters I can perform a static torque or force check on (and preferably adjust)...or can compare against one that has been checked...and that measure the contributions of both legs.


"Konnen sie die uberprufen, Herr Anhalt?"

"9."
Last edited by: RChung: Apr 14, 21 9:37
Quote Reply
Re: Which powermeter to trust? [RChung] [ In reply to ]
Quote | Reply
RChung wrote:
Tom A. wrote:
I only trust power meters I can perform a static torque or force check on (and preferably adjust)...or can compare against one that has been checked...and that measure the contributions of both legs.


"Konnen sie die uberprufen, Herr Anhalt?"

"9."


NatĂĽrlich

http://bikeblather.blogspot.com/
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
I find my trainer reads ~10% lower than my quarq power meters on 2 of my bikes. I gave up on having them match. They are both consistent which is more important. An increase in FTP with one power meter is also an increase in FTP with the other. That's more important to me than the absolute number.

Are you running some kind of training software to control your trainer? Many programs allow for you to control the trainer based on your power meter numbers, that way it wont matter how much they differ. I prefer to train to the quarq numbers since I also will race with the quarq.
Quote Reply
Re: Which powermeter to trust? [j.shanney] [ In reply to ]
Quote | Reply
j.shanney wrote:

Are you running some kind of training software to control your trainer? Many programs allow for you to control the trainer based on your power meter numbers, that way it wont matter how much they differ. I prefer to train to the quarq numbers since I also will race with the quarq.

+1 here. I rarely use the Neo for any power feed - relying on the quarq instead. I have compared the two, and the use of the quarq is the one that made a lot more sense to me for several reasons when considering the bigger picture of what I was attempting to achieve from training.

However, that gap between the averages is on the upper side of what I think would be expected, even taking into account one sided imbalance at 'tick over' and low power precision of tacx trainer power estimates. (OP: If you were in the big ring for this test at power that low I'd say this is a likely issue.)

As sad above, if you want to understand this then you'll need to do some work yourself, and potentially share some details/files. Twin recorded structured workout with some defined power intervals done in varying gear combinations is going to give you the answer in most cases, although at some point you may also need to run a dual sided powermeter against the trainer too.
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
Some bad advice so far. Running a bunch of tests to compare the two is a waste of time. The only way you are going to know what is going on is to compare your devices to known accurate power meters. That would require you to have a friend who is very accommodating about loaning equipment and has the necessary knowledge to understand if their devices are actually accurate. Then you could, say, compare the pedals to a powertap or crank based power meter, or borrow known accurate pedals and compare them to the Flux. The likelihood you can arrange all of that is probably not very high.

In my household we are two for two for one-sided power providing junk data, as found from testing a Stages that came with a second hand bike. We've had good consistency between three Quarqs, a set of Assioma Duos, and a Saris H3. Also providing junk numbers is a Wahoo Snap trainer. Your Flux might be accurate or it might not be, depending on when it was manufactured and some luck of the draw.

Personally, if I was you, I'd just bite the bullet and upgrade to Duos. Then use them as your trusted power meter. If the trainer doesn't match, use the pedals as your power source on the trainer.
Quote Reply
Re: Which powermeter to trust? [tttiltheend] [ In reply to ]
Quote | Reply
"The only way you are going to know what is going on is to compare your devices to known accurate power meters." my question to you is how would one actually determine accuracy? In some fields there is such a thing as a certified standard against which one can compare a device reading to determine accuracy, but in bicycle power meters everything I have ever seen is by inference. What I mean by that is for instance you hang a known (and presumably accurate or standardized reference) weight/ mass from the pedal and then calibrate the force gauge. That still does not guarantee accuracy, because the power you read is a calculation of a bunch of input numbers and also depends upon the temperature coefficient of things such as the strain gauge etc.... accuracy in power meters seems to be a tough nut to crack.

For the situation above you are suggesting a comparison and assume that at least if 2 agree then that is the best you can say. Neither can be said to be accurate, just that at least they are the same or in agreement.

The problem that the OP is facing is which of the two results he wants to believe. From a training point of view precision is most important and therefore if he uses the bike power meter outside then uses it also inside to calculate the power he is producing at least it is apples and apples. Whether the trainer is the same makes very little difference once you decide to believe the bike power meter and as pointed out you can in many programs use the bike power meter to power match likely for this very reason. As for single or dual sided that horse has been beaten to death.
Quote Reply
Re: Which powermeter to trust? [Remco] [ In reply to ]
Quote | Reply
@OP

Easy, correct answer here:

Just trust and use the Favero Uno powermeter numbers. They are 100% legit. Sure, you may get some L/R variance, but there is literally nobody on this form with hard numbers showing wide uno variation in practice, nor are there any of the (thousands) of users like myself that see such variance (like 10+ watts) on rides that we do literally every week, for months straight.

The Favero is an extremely respected and well regarded unit, AND it will be what you will be using on race day to target and measure power.

Just use the Favero. -endthread
Last edited by: lightheir: Apr 14, 21 11:14
Quote Reply
Re: Which powermeter to trust? [s5100e] [ In reply to ]
Quote | Reply
s5100e wrote:

For the situation above you are suggesting a comparison and assume that at least if 2 agree then that is the best you can say. Neither can be said to be accurate, just that at least they are the same or in agreement.

Not true. As I described above, IF you can actually check the accuracy of the static torque or force reading (and, ideally, "tweak" it), then you've just taken care of the largest contributor to power meter inaccuracy. Accurately reading rotational velocity is pretty trivial, and most PMs do a good job of that.

So, doing a torque slope calibration (i.e. NOT just an offset zeroing, which is what head units call "calibration"...sigh) on, let's say, a Quarq, an SRM, or a PT hub, is going to give you an accurately reading PM. I've been told some PM pedals also allow for a force check, but I haven't done that sort of check myself yet, so I can't really say how well (or not) that works.

http://bikeblather.blogspot.com/
Quote Reply

Prev Next