Kickr Core compared to Assioma Duo Power

Hello fellow Sufferlandrians,

I was wondering if anyone else has compared their Kickr Core to Assioma Duo/Uno power pedals?

I have compared them on a few recent workouts and the Assiomas are regularly reporting about 2% lower, not a huge difference but am curious.

I would expect the power in the Assiomas to be higher if anything as they are recording power closer to the input point whereas the some power might be lost by the time it gets to the Kickr Core through the drivetrain.

Thanks,
Kevin

We know for certain that bike drivetrain efficiency is 96-97%. So you lose 3-4% of your power between pedals and rear hub.

So at least one of your power measurements must be inaccurate. If your pedal power was accurate, then your Kickr was actually reading about 5-6% high. But if your Kickr was accurate, then your pedals were reading 5-6% low. Unfortunately it is not possible to tell which is correct and of course they could both be slightly off.

I would suggest re-calibrating both power sources (one at a time) and see if the results change.

I did mean to do a more thorough comparison, but I did compare my Assioma duos to kickr core over the course of a few workouts, and once they were both calibrated I found the pedals were reading a few watts lower. Not sure of the exact percentage but I would have said 1-2%.
My other takeaway was that I should calibrate the kickr core more regularly as it seems to drift quite a bit, even if sat in position between workouts.

1 Like

What @Peteski Pete said …

Then there’s tolerances within each device … then there’s double tolerance issues with single sided pedals/cranks.

In fact if a trainer ends up directly matching another power meter then that would be totally weird as in theory they shouldn’t and all we’d be seeing there is that one device is more out of whack than the other.

And in many ways it doesn’t really matter.

If training indoors … consistency is key - so as long as the method for measuring is the same, under same calibration conditions, then we’re training to whatever number it is. The number itself is kind of immaterial as it’s only to set the trainer to a level that means we’re working appropriately.

And outdoor ‘numbers’ are consistently different fir the majority of the population, so that’s different again.

Back to indoors though - if a trainer was reading low and say a single sided pedal was high in its tolerance band (and that single side was also the individuals dominant(or most powerful) leg, it is possible for everything to be intolerance and see over 10% differences … or even more

2 Likes

I’ll do another calibration, one at a time, and see. I had used my Garmin Fenix watch to calibrate the pedals so perhaps using the assioma app would make more sense. Thanks for your feedback

Thanks. I’d agree, as long as I know when I’m doing an FTP session that it feels like an FTP session etc. and HR is in the relevant zone it’s doing what it needs to. Just curious if anyone else has seen this difference. I’ll try a few calibrations and see if they get closer to each other. In the end though I think they all come back to RPE from the initial Sufferfest videos, your power numbers and heart rate should just back up what effort level you think you are in

1 Like

I have a Kickr Core and Assioma Uno, and frequently compare the two outputs (I calibrate my Assioma every 2-3 days).

What I have found is that depending on the type of workout one can be above, almost the same or below the other. However, the difference is never more than 3-4 Watts (average), which is acceptable for me.

I found that for VO2max sessions the Assioma actually records a bit lower than the Core, whereas for endurance or recovery rides it is the opposite. For usual SUF workouts which have different types of intervals, average power is never more than 1-2 Watts apart.

Hope this helps.

Thanks, yes it does. I don’t calibrate as regularly as that so probably should and hopefully that will reduce the difference. I also see differences at different levels. From 5 second to 1 minute efforts the Assioma is higher, for 5 minute to 20 min efforts the Core reports higher so similar to what you see.

It’s perfectly reasonable that the Assiomas report lower power than the kickr. Think about it this way:

Both devices are calibrated (after manufacturing) in regard to a more accurate device. Because of certain measurement parts assembled into the devices the reported values can alter within an allowed range (specification). As long as they do stay within specification, they are safe for marketing and sale.

A wahoo kickr core is specified with ±2% accuracy; the ASsiomas are listed with ±1%. So in a worst case assumption both devices are within specification and sound but the total error is still greater than 3.0%. And this is not taking power smooting and/or micro dropouts and different sample rate into account.

I’ve the very same setup and for me the Assiomas tend to run a little higher ~0.5-1.5 %, depending on how recent the last kickr spindown was, how stationary the intervals are (power smoothing i.e.).

So imho: Everything looks consistent enough for succesful training. Just be consitent which power meter you use. :slight_smile:

1 Like

Kevin,

A very very complex topic. One I have explored for a long time as someone who loves getting lost in data!

As most have said consistency is the main thing. That includes the calibration protocol.

I’ve previously had Garmin vector 2 and now Assioma Duo pedals. I also have the Tacx Neo. I remember a long time ago this started with setting up a custom workout in zwift with the trainer in erg mode holding 180/240/280w for long periods and overlaying the garmin data. My initial readings were in the 1-1.5% accuracy range. I actually ended up applying a scaling factor to the vectors to get them to align to my trainer…something I now know to be wrong.

Then I got my Assioma duos. 2 things were important to me:

  1. how close were the left and right pedals to each other for the same loading
  2. What were they reporting Vs the vector Vs the neo…ie were they more accurate?

Question 1 was answered with the static loading test. I went up to an 18kg mass test. the pedals were within 30g (~0.15%) of each other. happy with that.

Then question 2. I initially followed the 180/240/280w test, getting similar results as the vectors, then I discovered this article:
https://tacxfaqx.com/knowledge-base/linearity-showdown/

What this basically means is the flywheel speed (dictated by the rear sprocket selection) can actually affect the accuracy of the trainer to hold a power figure whilst in erg mode. This blew my mind as I just assumed I could do whatever I liked in terms of moving up and down the cassette and the trainer just automatically compensated and the accuracy was constant…very wrong!!!

If I say set erg mode to 250w, largest sprocket on cassette, pedals will read around that figure, but as you step down through the cassette to the smaller cogs, what you find is the neo things its holding 250w put the pedals are saying well in excess of that figure…it can be a long way out. fortunately I run a 1x drivechain, otherwise this would get even more complex.

To cut an even longer story short, I toyed with idea of scaling factors etc (as SUF doesn’t YET have powermatch), but what I now do is do all of my workouts is to do all of my workouts in the 6th sprocket on the back. This keeps the erg target and the pedals around the sweet spot. I also only use the pedals as my power source. Never neo. I test using the pedals, so whilst the neo sets the resistance the assioma pedals provide the quantative data used in my training.

This is a very complex issue. All I suggest is be consistent in how you use your devices. Don’t get hung up with the absolute numbers, but focus on repeatability.

hope that helps!

3 Likes

Sorry to get into “deep talk” with you, but maybe not… :slight_smile:
Do you have a clue wether the accuracy of the trainer is bounded by the given accuracy-statement?
I thought ‘they’ (tacx, wahoo, etc.) where calibrating / verifying there trainers over the whole range stated as working range.
As I do kind of this stuff for a living I’m a little flabbergasted if manufacturers give a best case accuracy and not a worst case approximation.
In my field of work a given specification of ±2 % and a measured accuracy of ± 35% would be a liable offense.

Thanks, before I even started reading your post I looked at the length and thought “I’ve opened a can of worms”. Very informative though.

I wonder does that spill over to the theory that you’re best to stick to the small ring in the front when on the turbo?

I suppose it looks better to give a best case rather than a worse case.

I remember being on holiday many moons ago and being told the local beer is given as ±4% alcohol. So you could have a few and they are 2% alcohol so fine, but then suddenly you get hit with one at 8% and you’re on the floor!

Nope, the data implies the best accuracy is archived at big sprocket small chain ring.

Two aspects with the article: 1) The Assiomas are seen as 100 % accurate measurement normal. This is a good guess, but it’s not necesarrily true.
2) The did a output comparison only at 100 W.
As the Power of a rotating wheel is given with W ~ I * w^2 for a double amount of power the flywheel has to turn four times faster. Thus the measurement error is prone to increase.

1 Like

On a not so deep diving note: This sounds like Segal’s law:

A man with a watch knows what time it is. A man with two watches is never sure."

No offense intended to the ladies around here. It’s just how the saying is :man_shrugging:

Wikipedia on Segals law

3 Likes

But we are looking here at a 5-6% combined error once you take into account drivetrain efficiency. I think that’s about double the worst case tolerance stack up here ie Kickr reading +2% and pedals reading -1%.

If it had been the other way around with pedals measuring 2% higher than the trainer, then that would be well within the expected tolerance range, but this looks to be a fair way off.

This is the main reason I don’t take Zwift racing seriously, lol.

mhhh, the reasons for this are different imho.

From the wahoo app for the core you can get it to use your pedals as the power source. I use my Quarq to control my Kickr core this way.

1 Like

The article you linked to states that a single speed drivetrain is 97% efficient. Where did you get 99.8% from?

In the same test their 2x drivetrain had an average efficiency of 96.2% across its ratios.

you’re right, i understood it as an over all average over ALL gear ratios - sorry for that.
As I remembered these numbers from lectures at Uni I researched the web a little.

Best pairing sprocket / cog: 98.6 % effiency
https://pages.jh.edu/gazette/1999/aug3099/30pedal.html

Here @208W the loss of a chain drive is listet with 2.43 W, which
comes down to less then 2 % as well.

Furthermore, this was not the main point of my argument, being: The accuracy of Powerpedals as Assiomas is not only affected by strain and rotational velocity, but by the fit of the used calibration curve with the single strain gauge + circuit used.
Why else are the giving ± intervals?