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

The first PM I purchased was a LH Stages DuraAce power meter. I always sensed it read low comparing my power numbers to friends who I would do the same rides with, but could never be sure given all the factors that could affect discrepancies. However, after I purchased an indoor trainer (a Wahoo Kicker) I could confirm that the stages consistently gave me 12-15% lower power. Because it was consistent and not erratic, I could still use it for training, but knowing it read low was still a bummer to know. I kept it for a bout 2 years then purchased Garmin Vectors, which have been right on top of my KICKr in readings and very consistent in general. Out of unnecessary caution I place the pedals at the same position before every single ride and calibrate the Vectors. Haven’t had an issue with them and it’s coming up on 3 years. I use them to give me both my indoor/outdoor numbs to keep things consistent.

This is the expected outcome and what we are paying for.

I hear you but despite comparing all of these power meters, the only real conclusion is that they are all different. There is just no way to know that the Quarq is the most accurate. The best you can do is to find three that are close.

Maybe you should reconsider the Assiomas. They have a feature in software where you can adjust the power to match another power meter.

1 Like

Good topic. I’ve recently become aware that my Quark PM reads about 10% high. My suspicion began a few months ago when I started switching between my KICKR’s power reading and the Quark during workouts
 the effort felt a tad easier with the Quark. So, I started experimenting:

I borrowed a pair of Garmin Rally XC pedals from my LBS and put them on the bike. With the bike on the trainer, I ran a TR workout showing me the KICKR’s power, and my headunit displaying the Garmin Rally’s power. They matched perfectly.

Then, in the TR workout, I switched from the KICKR to the Quark to compare it to the Garmin pedals. Sure enough, the Quark was reading 10% high.

So, I went in the SRAM app and started adjusting the Quark’s slope. I settled on -5% slope, but am still dialing it it. (I need to borrow those pedals again!)

2 Likes

This thread has been a really good read, for someone yet to buy a powermeter (using fluid trainer and virtual power). This has solidified my decision of getting 1 powermeter in form of pedals, and use it both indoor & outdoor. Clearly despite all manufacturers claiming ±1% accuracy, that’s not the case! I guess I’ll settle for 1 PM with repeatable results, where it doesn’t matter if the number is ±10% of the “real” number.

6 Likes

To be useful, a power meter only really needs to measure consistently. As long as it does that ride to ride you can test your ftp and set your training zones.

The Assiomas have the feature precisely so you can achieve what you want - multiple power meters that all read the same.

These are $500-1000 consumer products good enough for training on a bicycle not $20,000 scientific instruments.

3 Likes

My 2 cents.

Had two Stages PMs.

Swapped to Favero’s.

Stages read way higher, make believe numbers. Vanity power.

Based on weight, VAM and numerous comparisons with trainers and other riders, I’d vote the Favero’s accurate and the Stages way too optimistic.

Your mileage may vary


Dual sided power meters reading that far apart is absolutely not normal. Crank, spider, and pedal based PMs should agree when working properly. This week I mounted Assioma Duo pedals on my SB20. On my ride this morning they were 2 watts apart for the final average (199 for the Stages, 201 Assioma) . Maybe I am just lucky, but never had anything like a 12% discrepancy between my SB20 and a single sided pedal PM. Even there the two were within 5%.

1 Like

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

.

They are trying to prevent cheating. They want to power meters that are close but they don’t really care about the actual number. Esports cycling is trying to provide the illusion of fairness for a fake sport.

You have not made your case. As long as a power meter is consistent, how does it prevent an athlete from testing themselves, setting zones, or executing training?

Other than ego, it really doesn’t matter what the actual number is as long as the athlete is oriented to the right training zones.

4 Likes

Same here with mine. For the longest time I thought outside rides were just easier because I got to be outside and not so static. Perhaps to a degree but stages usually 10-15watts lower for me

I think the reality is that a lot of us have accepted that they are not going to match. Many have functionality that allows you to pick one as the “standard” and adjust the others to be close enough to the “standard” so they all “match” close enough for the purpose of training. You’re asking for perfection. We’d all love it if it existed, but it doesn’t, so at some point you have to accept it and move on so you can get back to training.

I hope the new pedals are closer to your expectation and I’m glad you have two that are working well!

3 Likes

In my experience - yes, this is too much to ask. For all the claims of 1-ish% accuracy, the number of people that I know (myself included) that have had 10+% errors between different power meters and smart trainers, and the fact that power match is a key part of TrainerRoad tells me that these claims are largely false.
If you want this, you should only get power meters (such as Favero) that allow you to adjust the reporting of the power. Then you at least have a chance of getting a set of multiple power meters that report very similar numbers. Without that, you’re just stuck with what the PM reports.

That said, here is one more thing to maybe check: Try changing the crank length for the pedals to something ‘wrong’. This should lead to a proportional change in power displayed. If my ‘gut feeling’ physics is working today, if you set the Favero crank lengths longer than they actually are, this should result in higher power. I don’t know if power will scale linearly with crank length, but if you set the Favero crank length 10-15% longer than correct, you should see a significant increase in power. This at least will confirm that the crank length settings are being correctly used by the Faveros. I would imagine a similar effect on the stages crank length.
If you change crank length 10+%, and you don’t see a noticeable change in the power reported, there is something going wrong with the power meter using the crank length settings.

I don’t think this is likely, but incorrect crank lengths being used by the power meters (even though the head units are correct) would explain the behavior you are seeing.

1 Like

Ha - your’re one step ahead of me
 That’s why I suggested this. More than a few times debugging something the problem ended up being something that was ‘known good’ and was only checked after many dead ends.

What I do find odd is that everything, including L/R, are scaled fairly nicely. This makes it really seem like some kind of systemic miscalibration/misconfiguration in on ‘pair’ of power meters. If it was a hardware failure/error I would expect it to only affect one of the power meters, not both in one set.

2 Likes

So when a brand claims ±1% (or whatever), is there something like an ISO, TUV standard or similar? Or is it just whatever they feel like testing it against? Is that test a static test of the strain gauges, to measure the strain and comparing it against some known value?

Like others have pointed out, do we really need it to be within 1% of the real world actual power at the crank? If it’s close enough, but repeatable, isn’t that enough for training, pacing and so on?

What’s the benefit of knowing with confidence that your FTP is 250 versus 225 (10%) difference? And is that benefit worth paying significantly more than what we are paying today?

That’s my take. I have never seen any official description of what this means, or now it is tested.
Is it 1% at all power ranges? All cadences? What time length? They don’t even specify the most basic parameters for what they mean, so my guess is that each manufacturer has some test that if they squint at the results in just the right light, they get 1%-ish accuracy. I’m sure they have something that at least sort of justifies their claims, but that doesn’t mean that their claims mean anything in practice.
Are they really claiming that every 1 second sample is within 1% of some ‘standard’ ? I doubt it, as with the current head units only recording at 1 Hz, I doubt that this is even possible unless you somehow synchronized the sampling of the two meters.
1% accuracy at 50 watts? Unlikely - are they really within 0.5 watts?
My guess is that they look at 3-5 seconds as the absolute shortest interval of interest, likely longer to get 1%. Some of this is very reasonable due to sampling issues, but they should publish what they mean by 1%.
Keep in mind there is no way for a consumer to check their products, so they really have no accountability. For a scale I could buy a reference weight, or take it to a calibration center and get it checked, and have strong evidence the the product meets (or doesn’t) meet its specification. All we can do is compare it to a competitor’s device, and complain about a difference, which is when the finger pointing starts.

Right on! If there is no published standard, then how do we know that their claim is wrong?
The idea of using a weight at the consumer level to calibrate the strain gauges sounds particularly wild. I imagine the strain seen on crank arms must be in the order of micrometers for the loads we are talking about. I don’t think any typical commercial scale for at home use comes close to the required accuracy


Well, in the case of the data OP is presenting, I don’t think that for any reasonable claim of “1% accuracy”, that it can be true that both the SB20 and Assiomas are both meeting those claims in the testing he is doing.
“We one day had one ride on one power meter where the ride average was within 1% of our reference, and based on that, we claim all our power meters are within 1%”. That is a method for testing power meters and making claims, but it is garbage. Even with the lack of a accuracy test standard we can make reasonable assessments of accuracy claims. We don’t know in the case of the OP whether the Assiomas or SB20 is ‘wrong’, but one or both of them are. I think we can confidently say that in this case they are not both meeting their accuracy claims in any reasonable way.

For most training we need to know two things:

  1. are we above our tested threshold

  2. are we below our tested threshold

Possibilities:

Get three pairs of Assiomas and set them to match your Quark. That is a sure fire solution.

Or buy more Quarks and test them against each other. Not guaranteed.

SRM has been know for accuracy. Call them and ask if they can sell you four power meters that read identically. Not cheap at all.

Skim read this thread but for what it’s worth DCRainmaker tested the SB20 against Assioma and the results were very close (in fact he was very impressed at how close, especially at high power where things often drift apart), so maybe there’s something else at play here.