Bang on wrong! What is going on? (Assioma vs SB20)

(Edit after figuring it out)
We now know what is the problem:

Cranks are affected from the Platform Center Offset (PCO). Meaning where you push on the pedal matters. Unbelievable, right? Very easy to test and demonstrate. Right crank is reading low because I keep my right leg inward. Left reads high because my left leg likes to go far out. Change your shoes and produced watts change. I repeat. You shoe, cleat position, foot stance chances the power readings. That is why Assioma Duo Shimanos went wild. Too much of a q-factor for the poor Stages Cranks to handle.

Problem in action:

I think this deserves it’s own topic.
I have been feeling cheated and rented a pair of Assioma duos to test my stages sb20’s loyalty.
Results were… interesting.
When powermeters do not agree I usually see some random difference. Especially when there are two separate power meters coming together to form “one dual sided” pm.
In this case I have two of those bad boys. Both Stages and Favero is dual sided.
Result: They are bang on different. Very same power curve, same l-r balance. Only ~12% difference across the board.

Am I doing something stupid? Perhaps. I compared the Stages to old be pro pedals when it arrived and they matched. Sold the pedals and used Stages for a couple of months.

  • both devices are on latest versions of firmware
  • batteries are new or fully charged
  • crank lengths entered correctly in head units and everywhere else
  • weight tested faveros. Pass.
  • hard sprints and a million zero resets
  • installation is correct, used a torque wrench even.
  • used a garmin 810 and 830 ro record. Switched them between pms to eliminate head unit problem.

I did a couple of rides yesterday and today between 20min, 90mins. They all look like this. Exact same l-r balance with same power curve. Just a superbly consistant difference.


I’ve always wondered if the offset pedal mount on the SB20 would result in weird PM results similar to the shimano issues due to differences in the torque application at various points in the pedal revolution. What is your crank length setting?

I have stages single sided crank vs tacx neo and the stages always reads higher, even more than would be reasonable for the difference in location. I’ve suspected maybe Stages reads higher than other meters or higher than “truth”. We have 2 data points, if others with stages and some other meter reply maybe could confirm the theory.

The older Favero Be Pro S has a dynamic calibration as well as a zero offset. I don’t know if its the sane for the Assimoas or the SB20.

I zero reset them in slightly different positions. When pedals are fully horizontal cranks are not.

They should auto zero as the bike is vertical

Likely need a lot of data to confirm. I have the same set up as you. I have compared data files from tracking the stages on my garmin and the tacx neo on the TR app. Stages over many many rides would average 5 watts higher. More or less 2% higher or so.

Interesting observations!!

I think there may possibly be a couple of things happening here.

Firstly, if both systems have the potential to be +/- say 2% or 3% accurate then the overlap between that variance could account for some of the difference and also explain why the tracking is an identical curve.

Secondly, it is perhaps a positive takeaway that they are tracking identical shapes (albeit different value) power curves as this strongly suggests both devices are at least behaving consistently on their measurement, which is really important in terms of you being able to have confidence in the repeatability and reproducibility of the data in terms of how it represents your efforts on any given ride; of course acknowledging that the actual values shown are different (but to which you can adapt / work around if needed).

I have a stages LH crank meter. I’ve noted that it consistently (for outdoors) reports about 5% to 7% higher than the power I record on my (indoors) Wattbike atom. So my n of 1 experience correlates with yours in as much as it suggests different devices record different values and whilst clearly the indoor / outdoor scenario is different to your’s (the whole indoor / outdoor differential has been covered multiple times on the forum) then the takeaway point for me has to be:

“Different makes of power meter record different power values for the same effort exerted.”

Armed with this knowledge, you just need to understand if the difference is consistent or not. You’ve done that piece of work nicely so I think your all good.

Simply note the difference and move on with your training / pacing accordingly with the relevant mental ‘adjustment’ in mind.

By way of example; my indoor FTP is currently 306. So when riding outdoors I mentally adjust this up to circa 320 and that actually works perfectly in terms of when I need to use my power meter readings to pace an element of my ride.

Equally when post ride analysis comes into play, I can downward adjust efforts recorded, to ascertain for my TR Wattbike sessions how my riding compares and again, that seems to work really well as well.

Hope this helps :+1:t2:

1 Like

This looks like almost every power meter comparison I’ve ever seen. One reads higher and one lower. DCRainmaker has shared tons of these. Nothing unusual.


I used to do the same as I suffer about a 15% differential between my indoor Garmin Vector PM and my outdoor Stages PM, but outdoor workouts with Adaptive Training lead to incorrect data being fed to the algorithm for PL adaptations and the new FTP Detection will also be affected. Not as simple as doing some mental arithmetic any more unfortunately.


I’m suffering a 15% discrepancy between my Garmin Vector 3 pedals (left side only) and a Stages Left PM. Looking at a lot of data points on climbs etc, and normalising for wind over several efforts and comparing them to online climb calculators, I think Stages is more accurate. Currently working through the issue with Garmin tech support.

Left side power meters take one reading from one side, then multiply it by 2. I have a stages left crank power meter and it reads about 10 watts higher than my garmin rally XCs. When looking at my cycling dynamics I use my left leg more at lower intensities and when I go into higher intensity my right leg works more. That’s something a one sided power meter can’t record. So it’s expected the ftp numbers to be different.

EDT: you have two left sided power meters but one thing that throws my garmin off is the pedals aren’t tight enough on the cranks.

My discrepancy is across all intensities and pretty consistent at around 15%. It’s a sensing issue - I just need to find out if it’s cause by a hardware fault, which is what im working on with Garmin, or something more systemmic with the PMs I have.


That is a big discrepancy for sure. And I agree that as the power meters are both L+R=total meters, and I assume your L/R balance is around 50/50, it’s unlikely that either PM is mechanically faulty.

I’m assuming its the crank lengh setting somehow - but I dont know how.

Could you post the cycling dynamics data from the Faveros? torque effectivness, smoothness, power phase etc.

1 Like

No, he did not. 12-13% difference across the board renders one of them useless.

no, they must be on top of each other. They claim 1% accuracy.
Or they can start charging 50$ instead of 550$

In general, people sound like they gave up on accurate power meters. C’mon guys, we need to be a lot more demanding considering the prices we are paying.

this. My balance is more like 55-45. Does not change the result, they both report this figure.
There are 4 isolated power meters and 2vs2 they track each other perfectly. Just off by a constant percentage. What are the odds for that? It has to be a wrong multiplier somewhere but no crank length difference is that big. I set crank lengths everywhere but even setting 165mm instead of 175mm can be 6% if I am not mistaken.


Your original post said your concern was that

I was pointing out that in most analysis of multiple meters, I do not see that. They are extremely similar and move as one, they just have different bases.

That’s what your picture shows too.

My point was not that I disagree that this is a huge difference. I agree with you there. However, I don’t agree they’re useless. You either choose to use one indoor and one outdoors and have different FTPs, or you adjust the bias and make them equal.


This absolutely jives with my SB20 experience. DM me if you’d like.

@kpedal The workaround, which blows goats, is to change the crank length in the StagesPower app and 100% always-always-always zero reset the cranks in the app before starting anything with TrainerRoad.

(In my case, I run 172.5 crank length and sadly have jacked the app setting to 175. I have a crap-ton of experience with PM’s and know my RPE intimately. Bottom line: the power balance and RH crank on my SB20 is whackadoo. Yes, I have licked the batteries clean, etc. etc… No, Stages has not responded one iota to a service ticket logged on their site). This is not my first rodeo.


I’ve tested my Stages gen2 against a Tacx Vortex, Wahoo Kickr, and a Powertap hub. The Tacx is wildly inaccurate. My Kickr constantly reads 5-10 watts higher than my Stages. My Powertap hub reads 5 watts lower than my Stages.

Yes! Personally, I just think it’s a fools errand to try and train with multiple power meters. Use a single power meter indoors and out. Same bike, same power meter.

If you really really have to use two different ones then you need to test and validate them against each other and then adjust them to read the same (assuming you even have this feature).

I don’t know. It’s been a race to the bottom on price with power meters. This might be a valid point if people were still buying SRMs for $2,000. Even living with an SRM was easy for people - needing to send it in every couple of years for battery service and calibration.


contact support and try to get a new set, IMO.

A fair observation and duly noted. My own point is that despite this, what people seemingly report experiencing is that different makes of PM do in fact vary, which I guess is the disappointment that underpins what you’ve noted :+1:t2:

1 Like