So, I’m having some similar issues with my kickr 16 vs my assioma pedals. I’ve done the advanced spin down, and normal spindown before doing the ride. The Kickr is not only off, but it varies and both over and under reports power at different times. The last interval in particular it was under reporting, and the interval was noticeably harder than the previous ones - my RPE was in line with what the Assiomas were reporting, not what the kickr was (HR was consistent with higher power too.)
Has anyone had these issues with their kickr and resolved them? I’m not really interested in powermatch, as I have a dedicated trainer bike and I don’t want to swap pedals every outdoor ride. The data that I captured on this ride are cannot be typical of kickrs, or they would be unusable without powermatch, so I’m hoping that some adjustment and/or maintenance can help with this. There are tutorials regarding the optical sensors on gen 1 kickrs, but nothing on later models regarding any cleaning. After this data, I have tightened the belt, as it seemed a bit loose (Wahoo gives only vague guidance on tension unfortunately), so we’ll see if that makes a difference, although I would be surprised if a slightly undertensioned belt could cause this kind of power discrepancy.
Lately most workouts (and spindowns, including advanced spindown) has been done in hot garage - typically about 85F/30C. I did several spindowns after the warmup on this ride, and they all reported 22.5-23 seconds on the wahoo app, so the spindown times seem reasonably consistent.
@GPLama Any ideas on what I should look at to try to resolve this?
I genuinely love these challenges! Good data too. I’m under the pump for time at the moment but I’ll throw a few ideas down.
That’s a really interesting drift occurring. The tracking is good, so it’s a slope / offset issue caused by %something%. That last interval is brilliant (not from an accuracy point of view, from a wtf point of view)
First - I assume you’ve done all these, but I’ll include them just in case (and for anyone else finding this post).
Install latest firmware.
Clean the optical sensor (it’s in different locations, iirc Gen 2 moved it to the flywheel/plasic enclosure)
Don’t use the Kickr in direct sunlight or REALLY BRIGHT environments. They’re vampires. (optical sensor issues)
Spindown and advanced spindown (which you’ve already mentioned)
Second - I’d try to eliminate heat. Point a fan at the Kickr and set it on full blast. Try to keep it as cool as possible and run the same tests. The data you’ve posted look like heat… but could be something else entirely.
I had done some of those - I hadn’t found much info on the 2nd gen optical sensors, so hadn’t cleaned them specifically. I had blown some compressed air in various bits.
I have now removed the flywheel, and located/cleaned the sensor and other minor dust from inside that area. It wasn’t too bad, but could have conceivably caused some problems. I’ve included some pictures below of the insides for reference.
I have now:
verified the latest firmware is installed.
Adjusted belt tension (increased it as it seemed somewhat loose compared to guidance on Wahoo video)
Cleaned optical sensor, and marker on flywheel
done workouts in a mostly dark garage
added a black piece of paper above where the optical sensor is to block external light (picture below)
added a blower to blow air directly at the vents in the plastic covering the PCB and optical sensor is located (picture below)
redone factory calibration of Kicker (after 15 minute warmup)
This test was done in a hot garage - high 80s F. I will try to do some tests early in the morning at some point, temps then will likely still be at least 70F.
This test was done in the small chainring, 4th from smallest cog
I unfortunately am seeing the same type of behavior on my latest test (Red lake +8). The main problems seem to be at the ‘higher’ powers (in this case 248 watts), where the power reported by the kickr and the power reported by the assioma pedals varies from matching to being off by about 10%. The lower power sections seem to be off by a much more consistent amount, and is more in line with the type of discrepancy I would expect between two reasonably performing power meters.
Having a 250 watt effort turned into a 270 watt effort is a big difference Given that the Assioma’s are really 2 separate power meters, and the left/right largely agree with each other, I really think this is pretty strong evidence that the kickr is the power meter that is in error (drifting) rather than the pedals. I can deal with a relatively fixed offset (either absolute or some % off), by varying by 10% within a single workout is not reasonably. I’m going to give power match a try, which I expect to work, but requires me to be constantly moving pedals between bicycles. I may contact Wahoo regarding this, but unless this is a known failure signature that they will acknowledge I’m not optimistic about a fix from them. (I am out of warranty anyway…)
(Further down the rabbit hole…)
I did another test today, and tried some different gear ratios. It looks like lower speeds are better for accuracy. When in my highest gear (46/11), the accuracy was worse, and I saw some very noticeable oscillations in power - time 14:00 is a good example of this.
Great testing. In regard to the reported speed, do you have ERG Mode Speed Simulation enabled? I’m trying to work out what’s causing the deflections 1/2 way into those intervals (1/2 way into the 4th 250W effort something changes)… but it doesn’t look like cadence or speed changes much?
Lower flywheel speeds are usually better for accuracy. Most trainers have an optimal working zone for flywheel speed and power reporting. This isn’t something that much is known about… we’ve stumbled across it more recently. In particular the Neo 1/2 will under read 15-20W in ERG if the flywheel is >40km/h. I found that one trying to answer why so many people report that underreading… usually they’re new to ERG and slam down the gears (which ERG allows).
Keep me posted on how you get on resolving this one.
I’ve had problems with my Kickr (purchased April 2016) since updating the firmware in April 2018. After installing the update, my Kickr totally locked up. I had to perform a couple of factory spin downs before I could turn the the pedals. Since then the power readings seem to be way too low and I notice a change in the resistence during intervals even though there shouldn’t be one. I had my FTP tested 2 months ago in a lab for 278W. For my ramp test on the Kickr not long after the lab test I managed 233W. Wahoo support were no great help. The worst thing is not being able to trust that I’m putting out the right watts for my training zones.
The Assioma pedals can measure power differently than a trainer or other power meters by up to (4-5%) And that’s per Favero. Assioma measures instantaneous angular velocity while other power meters/trainers calculate power using average angular velocity per rotation.
I’ve done some experimenting on this and found the NEO2/KICKR and other power meters measure power consistently regardless of what gear combination you select. The Assioma however, was a bit different.
I do have erg mode speed simulation enabled, so I guess there won’t be a direct relationship between cadence/speed in that case.
I’ll update this thread if I make any further progress. I think the next thing I will try is to do some intervals in a low gear and see if that helps. I’ll also see about a cooler morning workout.
The 4th interval was the most puzzling. I wasn’t intentionally doing anything different while riding it. My cadence did increase a little as the power was increasing, but I don’t know if that is a cause or an effect. It became a very hard interval towards the end
Hi Tariq - I just watched your video, and found it quite interesting. I have done a workout experimenting with gearing, and it looks like it has a significant effect on accuracy/consistency for me. The inconsistency is a much bigger issue for me than a stable offset would be.
For now I’m just logging data while doing my TR assigned workouts. Below are the more interesting parts of Baird +6 - I did each set of 5 VO2max intervals in different gearing, at about the same cadence.
Based on this 1 ride, it seems that for my setup a higher gearing is more accurate and consistent. I will be doing more testing along these lines, as the higher gearing seemed better at the higher powers, it looked like it resulted in more varied power at the lower power intervals.
Here is the first set, in 36/21 gearing. The dip/rise in the 3rd interval was noticeable by feel, so the Assiomas are tracking ‘reality’ better there than the kickr. On all these intervals not only is there a big difference in power, but it varies significantly over the duration of the interval.
This is pretty bad, and well beyond the claimed accuracy. I doubt that Wahoo would say that this is how things are supposed to work
Final set in 46/13 gearing, quite close on all of these. (The first 3 of the 5 were at reduced power, as I was getting tired and +10 watts on these was brutal. I upped the last two back to normal power since the power was tracking better.) Not perfect, but pretty much acceptable I think.
I’ll be doing some more experiments to try to come up with a configuration that gives me stable enough power numbers from the kickr to be able to use the built in power measurement rather than having to swap bikes or pedals all the time.
Ahh now I see it. I kept looking at the DCanalyzer but couldn’t find it the first time. That’s around the speed where my Assioma started drifting upward. However, my crank powermeter was consistent with the neo and kickr.
It will be interesting to see if you get the same results using the 46/13 gearing again. If so, then you found your answer.
Iike a 10% diff. I was ave 144 on kickr and 157 on vector. I have just relied on the kicker for indoor I’m starting to experiment with the vector outdoor so I don’t know if it really matters for my purposes but it would be nice to be able to correlate.
I experience same behavior with my KICKR with multiple powermeters (Stages, Powertap, and Vector 3S). What I found after using @dcrainmaker analyze tool is that :
KICKR is reading a bit low on lower intensities (under 200w)
More accurate around 250/300w
Reads a bit high for higher intensities
HOWEVER, part of the difference could also come from some imbalance in your pedalling. Are you using L/R pedals, or just Left ? In my case, I tend to have an higher imbalance on lower intensities (55/45) which disappear almost entirely above 320w (51/49). So that explain part of the difference I get (not all of it however!)…
PS : I always used two powermeters at the meantime (pedal + crank arm) to make sure that the problem is not coming from “here”
So I’ve gotten around to collecting some more data on this. I filed a support ticket with Wahoo, and they had me do a workout at different intensities and send them the file. The kickr always agrees with itself, so naturally they found nothing wrong. They were not interested in mismatches with other meters because being off compared to other meters is ‘expected.’ I’m not terribly surprised they don’t want to get into issues like this, but I am disappointed.
Here is my latest data, from the ‘Beech’ workout. The garage was 72 degrees F this time, so much cooler than the previous tests. This has 3 15ish minute endurance segments. I calibrated the Assioma (L/R) pedals before the start of the ride, and the kickr after the warmup, and again after the first endurance section, as the power had drifted way off. The 2nd calibration did not help - the power was off for a few minutes more, then it drifted back. I have 2 different DCR analyzer sets - one that aligns for the beginning of the ride, the other for the portion after the 2nd calibration.
The big problem that I am having is the difference between the Assioma (L/R) changes multiple times during workout. In this case the kickr is reading high both times, but I have had workouts where it reads both low and high in different sections of the workout. The error is 10-12% in the bad sections, so well beyond what can be explained by the combined in-spec error of the pedals + trainer. When the kickr and assioma’s drift apart, the assioma L/R data stays together. Also, while the RPE on these efforts was low enough I couldn’t really notice the changes, the HR data tracks the power reported by the Assiomas. On other harder workouts I have noticed RPE changes, and those again have tracked with the numbers that the Assioma pedals report.
Given that the Assioma pedals are independently measuring power at each pedal, I would be surprised if they would both drift this much off in unison. This behavior is the same in all error cases I have seen - I never see a divergence in only 1 side on the Assioma.
I hope to borrow some powertap P1 pedals to test with to see if I can reproduce the same kind of errors. I am also curious to see if they show the same short duration L/R differences that show up on the Assiomas.
It really looks like erg mode with its +/- 10% accuracy on my kickr is useless. My graphs look nothing like what @GPLama and @dcrainmaker are able to get (As an aside, what smoothing interval do you guys use when showing the data you have on youtube?)
I’ve borrowed a pair (dual sided) of powertap P1 pedals to compare my kickr too, and they show similar results as I have had with the Assiomas. The ride this morning was one of the most obviously off that I have had, as the drop in resistance was very noticeable when it happened at about 15 minutes in. For this ride the kickr varied from +17% to -5%. It does track fairly closely for some sections as well.
Trainer was warmed up and spun down before this workout, and pedals were calibrated as well. Garage was about 62f, so heat should not be an issue.
Again, like with the Assiomas the L/R track relatively well, so since each pedal is measuring power independently, this is really 4 power meters agreeing that the kickr is wrong. I’m going to file another ticket with Wahoo to see if they will acknowledge a problem, or if I’ll get more of the 'there are lots of reasons for these differences that have nothing to do with us" response.
@GPLama@dcrainmaker Have you guys seen behavior like this on a kickr? (mine is a '16 model) I’m not optmistic that power match will work well with this, as the error varies so much that it will complicate the tracking.
Nope. Looks shagged. Given you’ve performed more testing indicating the on bike power meters are ok, all eyes are on that Kickr being a Shitr. If it hasn’t been mentioned already - Aim a fan directly at the unit to keep it cool and test again. I have a suspicion it’s heat related… then again it could be something even more obscure.
Wahoo Support should be all over this as an interesting challenge to solve.