KickrV5 - drift in ERG-mode?

If the new Kicker V5 auto calibrates when coasting, one option would be to do it after the warm up.

Also, keep in mind that no trainer calibrates during a workout in normal use, regardless of ERG or Resistance mode use. This is no different so it has nothing to do with the mode setting.

I presumed the trainers are tracking temperatures and potentially adjusting along the ride, like our power meters apparently do. Perhaps this is a sign of a bad calculation within the Kickr and it’s not making the best adjustments through longer workouts?

FWIW this is my assumptions only. Wahoo support says Kickr5 is working fine and within tolerance. I wouldnt agree that a 4% drift is within tolerance, but that is basically an other discussion. Wahoo says that Kickr will calibrate after each ride and that I never would need to calibrate the Kickr5 manually. But when I implied that the Kickr5 would also calibrate before ride (when cold) and wouldnt calibrate during workout, the stopped answering. So no more help to get there.

Anyways I dont trust Wahoo support much, if you wanna know why, read this thread:

dlec10:
If I stop pedaling during an ERG-workout, the trainer stops within a couple of seconds, Im assuming that is not enough coasting for Kickr5 to recalibrate. A normal spindown (Kickr18) usually takes 20-30seconds I would guess.

mcneese.chad:
Yes, I can probably do spindown after warmup. Two problems with this though:

  1. Im not sure if the Kickr5 really wants to be calibrated, since its actually “auto-calibrating”. Ive done both successful and unsuccessful calibrations using TR, but Im wondering if this only is a result of me triggering or failing to trigger the auto-calibration. This would also mean that I must calibrate the Kickr5 during each ride, since I have very little control over when it auto-calibrates.
  2. One of the main reasons I bought the Kickr5 was the auto-calibration, when its not working as expected, I really dont see the point of having a Kickr5. When I had the Kickr18 I did one calibration after warmup now and then and that seemed to work ok, Kickr 18 was then normally under-reading a bit during warmup, but reading pretty close to PM after 15min+ as seen in sets under:
    https://analyze.dcrainmaker.com/#/public/76f87f47-7363-43c3-6672-f32d4bdf712d
    https://analyze.dcrainmaker.com/#/public/13079d54-c638-4029-557b-f261477882f3
    https://analyze.dcrainmaker.com/#/public/efa4d820-e96d-4d02-6198-1ec8b091627f

I have no idea if Kickr compensates for temperature better or worse than others. I only have data from Kickr18/5 and Neo2/2T, Neo did not drift, but its also a totally different trainer, no belt etc. I did have a Direto once, but didnt usually dual record at that time, so I only have two datasets of the Direto and I really cant remember anything about temperature or when I calibrated it. I also mostly used powermatch when I had the Direto. The two sets dont show any drift though. FWIW, its pretty cold in Norway atm, so I would guess the temperature in my trainingroom rises quite a bit during a workout these days.

2 Likes

This kind of sleuthing gets me excited. I’ve pinged you on DM @daniebr

3 Likes

Just tossing some data in just to help diagnose:

This was a 3 hour ride from the weekend, not in Erg mode, but in regular Sim BUT it has a total of 7 zeros on the Kickr and they were at the beginning and end of the ride, with a good 2+ hours with no drop in tension on the chain.

Not exactly the same conditions, but there was zero drift in that scenario so maybe it’s specific to ERG?

Most auto-calibration for power meters occurs during coasting from what I’m aware, but honestly I haven’t been impressed with Quarq’s AutoZero either and have always turned off the auto-zero function.

Let me know if I can help, since I have a Kickr v5 and can test some things on my end for you.

2 Likes

Thanks!

I have not seen any drift in sim-mode either, but I have not done any longer than 45min(link in first post).

I dont trust the Quarq AutoZero, so I manually autozero the Quarq everytime, it gives very consistent values though.

I’m not sure if this is a temperature/auto-cal issue or a sim/erg-mode issue.

If it is a sim/erg-mode issue, you should see a drift in longer ERG-mode rides.

If it is a temperature/auto-cal issue I should also get a drift in longer sim-rides when starting “cold” and not coasting. Guess I will have to test that one day. Maybe I will try the Ven-Top tomorrow.

1 Like

So I have done some more testing.

Oct 27: https://analyze.dcrainmaker.com/#/public/666be50e-7b48-41dc-67ab-37579b39a424
Erg-mode, drift from 2,5% to 5%, back to about 3% after TR-spindown at 1h15min.

Oct28: https://analyze.dcrainmaker.com/#/public/61d5c198-6fef-4495-7366-0e924898dcd1
Sim-mode, Zwift, no drift, Quarq reading about +5%.

Oct29: https://analyze.dcrainmaker.com/#/public/0fdf7836-ed35-4a0b-45bc-4dde416f5a7d
Erg-mode, drift from 2,5% to 6%, trying TR-spindown at 1h28min and 1h35min with no result. That is reminding me of set from oct25 where I tried TR-spindown two times before any result.

This doesnt make me much smarter, the only thing consistent is a drift during ERG-workouts and no drift during Zwift/sim-mode. I have done several TR-spindowns, but it seems like its 50/50 if they have effect or not.

One thing I have noticed though is more differencies in Zwift/sim-mode after I started doing TR-spindowns:

  • oct19 4iiii read +0%
  • oct21 Quarq read +1,5% and +2,5%
  • oct23 Quarq read +1%
  • oct24 - first TR-spindown - Quarq read -3,5%
  • oct28 Quarq read +5%
3 Likes

is there a way to make the links public?

dumb question, but have you lubed your chain during that time?

1 Like

I don’t think spindown/calibration does anything in the new KICKR.

From https://www.dcrainmaker.com/2020/08/wahoo-kickr-v5-2020-smart-trainer-in-depth-review.html (my emphasis):

However, as noted above – this is mostly just for show. One of the new features of the KICKR V5 (2020) is the zero calibration concept. While it’s not quite like the KICKR Bike or Tacx NEO in terms of not being possible at all to calibrate, Wahoo says there is no reason to calibrate it. So much so in fact that Wahoo will quietly override any calibration values it takes within a minute or two using a new continuous calibration system.

My reading of this is that you don’t have to stop pedaling for it to “zero”. So maybe you have something else going on with yours. The Wahoo help desk can’t help?

1 Like

~2 seconds of coasting required.

4 Likes

Done.
Not lubed, cant really see that have any affect, the difference in ERG-mode is pretty similar for the last two weeks.

Wahoo says trainer accuracy is within tolerance, and that no manual spindown is needed.

I think the Kickv5 is meant to be 1% tolerance? Yours seems worse than that.

All your trials and tribulations remind me of my experience with an earlier Kickr. I wasted so much time with Wahoo support, before making two changes. I bought a Tacx Neo and Favero Assioma. Never regretted that decision.

2 Likes

Yes, KickrV5 is +/-1%, Quarq is +/-1,5% and 4iiii is +/-1%.
There are a lot of opinions about what those tolerance really mean.

To quote Wahoo:
"Be aware that no two power meters give exactly the same reading. Multiple identical, simultaneously broadcasting power meters have been empirically proven to report different power numbers during the same workout.

Factors responsible for this difference include:
-Data transmission rate
-Packet loss
-Wireless interference
-Power Meter or strain gauge location (example: the KICKR measures power at the hub, which will result in a lower power reading due to drive train power loss compared to a crank or pedal based PM)
-Temperature
-Humidity
-Differences during calibration
-Battery life
-Accuracy tolerances
-Leg imbalances

Additionally, consider that the KICKR power accuracy tolerances combine with your other power meters tolerance. This means that if your KICKR is rated to report +/- 2% accuracy and your power meter is rated to report +/- 3%, your total combined accuracy tolerance will be +/- 5% at any given time."

So when Wahoo says tolerance is additional to all the factors above, I really dont see why they would give a tolerance number at all (of course they must because of marketing), but it will have no practical meaning. Imo the +/-% given for trainer and PM is supposed to cover all the factors above except drivetrain-loss. In a perfect world, crank/pedal-based PMs should read a bit (2-4%?) higher than a trainer due to drivetrain loss.

So basically Im fine with my PMs reading higher than trainer, but imo 3-5% drift is not within tolerance. I told this to Wahoo after getting the answer above, but I got no new answer. They only set “request status” to solved.

I’m coming from a Neo 2T, my PMs read a bit higher than the Neo aswell, but there was no drift at all. But the Neo had brake-rub and people didnt know what would happen to support after Garmin bought Tacx, so when reading all these nice things about the KickrV5, I switched when I got a good deal. I also liked the feel of the Kickr18 better than Neo2/2T.

The KickrV5 was perfect on paper for me, auto-calibrating (didnt realize atm that the calibration needed coasting), great frame-compability and no front-wheel block. I must say Im dissapointed though. I genuinely belive there is something wrong with my KickrV5, since it shows two consistent, but different behaviours in ERG/sim-mode. But I really think there is no chance that Wahoo will admit this or help in any matter.

IMO what I’d do is to record a TR workout that is a constant power for 60-90 minutes and send them another graph. It’s a bit hard to see it visually in the DCanalyzer links but a set power (say 200w) over a long enough time should make it very clear that the Kickr is reporting the same wattage but actually drifting to a higher resistance over time.

I think getting the language right to describe it will get the issue escalated up the chain.

Drift is a technically correct way to describe it, but I’d try writing them something like

“Wahoo Kickrv5 reports consistent wattage while steadily increasing resistance over time rather than holding constant resistance in ERG mode, causing workout to get substantially harder even at the same reported wattage”

“Expected behavior: Kickr reports a consistent wattage that results in a consistent offset wattage at power pedals throughout whole workout”

“Actual behavior: Kickr reports consistent wattage but power offset at the pedals went from neutral at minute zero to 15w at minute 45, to 30w at minute 90 which signals that resistance is increasing and that the trainer is not performing to customer expectations”

3 Likes

You are probably right.

Im trying to get a deal at some second hand Assioma Duos. Getting a third datapoint will hopefully confirm my assumptions that it is the trainer that drifts, not the PM.

Getting the language right is probably important, my english is somewhat limited though and Im guessing I can seem a bit harsh. Im not :slight_smile:

1 Like

One more thought - you could try @stevemz’s suggestion of a constant power ride on TR, BUT in standard (level) mode. Do 60 minutes of constant pedaling at a gear/level setting that gives you 200W (or whatever) and see what happens. That might help confirm whether it’s an erg mode problem or a “no coasting” problem.

I have to diagree with Wahoo on this one. The power numbers transmitted over ANT+ are sent using a rolling counter update. So even if you lose a few ANT+ packets, you will still get the correct average power. It will just mean your instantaneous power will be wrong for a second or so.

The list they gave you just shows they know the product is not 1% accurate If a product says 1%, that should mean it works under all these conditions. If it doesn’t, it should be shown in detail in the manual, or on their web site.

I think there needs to be a standard testing/certification setup for all manufacturers to follow. These products are not cheap. Someone is going to get pissed off enough to start broader legal action.

2 Likes

100%. Power meter and trainer companies self-certify. This is a problem. Some who have ‘outsourced’ verification have done it very carefully and cherry-picked data and only published partial information.

The problem is that no company is going to fund a standards/testing body.

Here in Australia we have the Australian Competition and Consumer Commission (ACCC) who are really on the money when it comes to consumer rights. If a company attempts to sell something that doesn’t do as claimed… we can get them involved. Their media release lists are always interesting to browse: https://www.accc.gov.au/media/media-releases

From what I’ve tested first-hand… they’d have a field day with sports tech. It just hasn’t hit their radar yet.

3 Likes

@NickL and @stevemz: The 1h30min@200W in both ERG and standard mode is probably the way to go, will however try to get those assiomas first to get even better datasets. Doesnt sound like a very fun workout, so it better be worth it. I did experiment with resistance mode in TR during a workout earlier, but didnt find anything special.
Oct22, ERG+resistance mode: https://analyze.dcrainmaker.com/#/public/ac624af5-b0df-4eac-4344-cca50a5ae17b

@road_wars: I completely agree with you, I was just quoting what Wahoo-support wrote, and its the same as stated on their website: https://wahoofitness.yonyx.com/y/conversation/?id=7126b950-bf33-11e7-b908-bc764e10c19d

Did a new workout in Sim-mode/Zwift yesterday, Quarq back to reading +1%, no drift.
Oct30, sim-mode: https://analyze.dcrainmaker.com/#/public/d2adf9a6-c03c-4df5-5be6-b4011c386de6

ERG-workout today, back to drifting from 2% to 6%.
Oct31, ERG mode: https://analyze.dcrainmaker.com/#/public/73cb2302-5360-455e-7c09-65043ffa3827

Since I did the excact same workout oct24 and oct31, I tried to compare them, the drift is almost completely similar: https://analyze.dcrainmaker.com/#/public/7df56050-c939-4b25-5591-f143fbfede6b

That’s brilliantly consistent. Which workout was this? I’ll add it to my schedule this week.

1 Like