Kickr Snap reading low 2021 - data comparison

[edward273 (Edward Hicks)]

My Kickr Snap v3 consistently reads 8-20% low when compared to both my Stages Left only power meter and my Assioma Duos. The Assiomas and the Stages agree with each other very accuratly (within 1.5%).

I’ve received a replacement Snap from Wahoo but the replacemnet is still reading low. I’ve performed spindown after spindown, always after a 10 minute warmup and I’ve even performed a few factory spindowns. These seem to help bring the Snap down from 20% inaccurate to 8% inaccurate but I can’t seem to get it any more accurate than that. Has anyone else been here? I turbo outside on a covered balcony and store the Snap inside so there is an initial change in temperature which could be the problem, but even after a 30 min warmup the Snap still reads low.

I record the same rides with my Garmin watch, headunit and Kickr app on my phone to get all 3 sets of data simultaneously and then compare them on Zwift power, see photos!

[K_DPW]

Make sure you always have the same tire pressure. My blue tacx tire is at 110 psi.
When I performed the factory spindown, my brake factor was 0,95.
I did the factory spindown with 2,5 turns of the blue knob.
Before every ride do a normal spindown, warm at least 10 minutes up before spindown. Blue knob 2,5 turns for normal spindown
My spindown time is between 12 and 13 seconds.
Always the same procedure, tyre pressure, warm-up, 2,5 turns blue knob.

[Justin_Thackeray (Justin)]

All the wheel-on trainers have an intrinsic problem related to exactly how much the tire is pressed onto the drive roller.

First of all, as Kurt says, make sure you’re consistently maintaining the same (good) pressure.

Second, and perhaps more importantly, if you’re pushing the tire even slightly too far into the roller it will deform the tire too much and result in additional force being needed to keep pushing the rubber past the roller.

I don’t think there’s an exact number of blue-knob turns that will work in all circumstances (ie: all brands, sizes of tire and all pressures), you just need to find the right amount for your situation by trial and error. I used a Snap for 3 years and was feeling great about racing as a B, even though the watts I was generating didn’t match (exceeded) my IRL wattage significantly. Then last fall I switched to a Tacx Neo 2T and discovered the Snap had been OVERestimating my output, suddenly the trainer was a bang-on match for my IRL rides, and I was dumped back into cat C.

Bottom line is that the wheel-ons are less accurate than the integrated trainers, and the inaccuracy can be in either direction depending on how carefully you set them up. I bet if you experiment with the blue knob you might be able to find a sweet spot that matches your other data sources.

[Gru (B)]

I have a SNAP 2018 and have found that it reads low if the spindown time is over about 14 seconds. Wahoo says it doesn’t matter if the spindown time is longer, but that is contrary to my experience.

Since you have other power meters available, you might try experimenting with power readings vs. spindown time to see if that helps. Start with 2.0 turns on the tension knob and increase from there and see how that affects spindown time and power accuracy…?

As I’m sure you know, tire pressure and wheel tension are critical. I put stickers on my tension knob so that I can easily tighten it to the same point each time I ride.

[edward273 (Edward Hicks)]

Hi All thanks for your responses.

Just as an FYI last weekend I replaced the whole drivetrain (chain, cassette, jockey wheels, front chainrings, bottom bracket and wheels) they needed replacing and I thought they might be to blame. So don’t think this can be a major factor outside of maybe 1%.

My Assiomas are rated to +/-1% and the Snap is rated to +/-3% accuracy so my target is for the Snap to be within 4% of the Assiomas. Without this the Snap is somewhat useless as a training tool and I might as well have spent less than half the money and bought a regualar old dumb trainer.

So last night I recorded two workouts, and set up my bike and Snap outside in the cold London air (0 degrees C ish) 2 hours before I did the workout so the whole setup could adjust to the outside temperature. Just before starting I then pumped up the cold tires to 110 psi and turned the blue knob 2.5 turns (I normally do 2.5-3). I also calibrated my Assiomas here.

For the first workout “standing starts” I performed a ‘cold’ spindown (without any warmup) before the workout and the Snap was 10% ish lower than my Assiomas. Interestingly this stayed relatively constant, for the first 15 mins the Snap was -8.5% lower and in the last 10 mins it had dropped to -11.8% lower. As no warmup was performed before the spindown the Snap was understandably well off. Pic below.

After this workout I then performed a ‘warm’ spindown, the Snap having had the previous 40+minutes of the last workout to warm itself up. I then did a second workout 25mins at threshold. Here the readings were a lot closer within 2.6% on average - all good so far!

However, diving into this second workout a bit more, for the first 5 minutes the Snap actually overread by 5%:

Then, for the 25min effort the Snap was -2.7%, perfect:

But then for the last 5 minutes the Snap was reading -6.7% lower:

So although for the second workout the Snap was closer, it was still outstide the combined error by the end, and more annoyingly it actually changed relative to the Assiomas as the workout progressed.

Tonight I did one workout, a Z2 1 hour effort. I set the bike up right before getting on it, 2.5 turns of the blue knob, again on my balcony in 0 degree evening air, and pushed maybe 150W for exactly 10 minutes as a warmup without recording any power etc. I then performed a spindown of the Snap and got straight in.

Full workout, Snap -5.87%:

First portion Snap -3.65%:

Middle portion -6.71%:

End portion -7.55%:

So again the Snap started well but then got more and more out as the workout went on.

To clarify:
Standing starts: ‘cold’ spindown
25’ Threshold: ‘very warm’ spindown
Z2: ‘Recommended 10min warm’ spindown

All of these spindowns gave between 12-15 second times which, on wider reading, seem ideal.

So although making sure tire pressure was 110psi, 2.5 turns of the blue knob, an ‘ideal’ spindown time, and warming up the Snap before hand made the readings more accurate, they still weren’t consistently inside the +/-4% difference that I should expect, and more annoyingly they seem to change as the workout progresses…

Is this the product of doing it outside on a cold balcony do you think or is this basically as good as the readings are gonna get for a wheel-on smart trainer no matter where you do it?

Cheers!

[Justin_Thackeray (Justin)]

I’m impressed by all the data collection and analysis! However, I still think the blue knob needs more experimentation… try 2 turns or even 1.5. As long as the tire isn’t slipping under sudden acceleration or on a steep climb it’s worth trying. The variations during the ride probably aren’t a surprise, because as the tire heats up and cools down it will deform differently, becoming more or less easy to move the drive cylinder.

[edward273 (Edward Hicks)]

So I did another session tonight, 2 1/4 turns of the blue knob, set up my bike outside and then jumped straight on for a 15-20 minute warm up before I started recording…

The full workout 15.5 watts difference @ -5.68%:

The frist 15mins actually look very good, 7 watts difference well within the desired 4%:

But by one of the middle efforts, the Snap was already 23 watts and 6% lower:

And this gadually worked its way out until at the cooldown it was more like 12% out at an 18 watt difference (although at recovery power):

Again I’m seeing the Snap diverge away from my trusted Assiomas as the workout progresses. Given the workout the other day after a 45 minute warmup was more accurate I do think the Snap needs a good warm up - well over Wahoo’s recommended 10 minutes.

I also thought that the watt difference might remain constant but the percentages might change as the workout moves through threshold to recovery efforts etc. However, here the watts difference jumps from 7 up to 23 and then back to 18 so not really.

I also don’t think my unit is faulty and this is just a failing of the Snap design and wheel-on trainers in general as the tyre temperature and pressure along with the Snap’s rollers change too much over the course of a workout for any initial spindown to keep the Snap calibrated for the whole workout.

Maybe if I did a solid 30 minute warm up and then my whole workout consisted of sitting at a single power for x minutes, say 300W for 50 minutes the Snap would stay within 4%. However, this is kind of a pointless test as the trainer should have more utility than a single workout type like that.

[Gru (B)]

Interesting stuff.

0C seems a bit cool to me. I’m happy to ride around 15C. I don’t see an operating temperature range specified for the SNAP, but this could be outside the expected range. (Also, possibly outside the expected range for the tire!) The tire, the wheel, the air, and the metal roller - which is attached to a massive fly wheel - are all at 0C to begin the ride. I’d guess that roller doesn’t warm up very quickly. Do you happen to own an IR thermometer or thermal imaging camera…?

Have you discussed your observations with Wahoo tech support?

[chris_benten1 (chris)]

The Snap heats up quite a bit if watts are high…(say consistent over 200). Even with a fan blowing or in cold temps. After a hard ride it is warm enough to fry an egg. I would imagine the changing temps would play havoc with the calibration. I finally just let my P1s be the source to control the Snap. Much more consistency using that configuration.

My H3 is much more stable. I rarely calibrate. Even in 35F temps the other evening I was dead on in the 150-250 watt range. At the peak watts they can be separated by 12-15 watts but are usually within 10. Good enough for me. I do calibrate the P1s prior to each ride. The temp changes do affect them.

[edward273 (Edward Hicks)]

Hi all, thanks for the info. No I don’t own any thermal image cameras, but you’re probably right about the roller. It does seem that when the Snap doesn’t overheat and has had a significant warmup so it’s all at equilibrium, then it stays within the 4% error margin for at least some period of time although this diverges outside the margin before the end of a session.

Yesterday I tried an indoor cadence build session, 2.5 turns of the blue knob, 10 minute warmup and spindown of the Snap + zero reset of the assiomas. As I keep the bike and the Snap in doors then the whole system will have been at the ambient temperature to start with. Interestingly, this was one of the Snap’s worst performance and actually got closer to the Assiomas at higher power (see average vs max):

I have talked to Wahoo tech support but they have already sent me a replacemnet Snap (this current one) and seem to think that the external factors + drive train loss are to blame. I’m going to ask to exchange my Snap for a Kickr Core and pay the difference I think

[BenJones (Ben)]

you can post as much data/charts as you want, but unfortunately wheel on trainers are notoriously inaccurate, if you care about it as much as you seem to, I would highly recommend returning or selling the Kickr Snap and investing in a decent direct drive smart trainer if possible, seems like you would benefit from it a lot if you care about your data that much ;D

[Gerrie_Delport_ODZ (Gerrie)]

But if you have power pedals why worry about the trainer power.

[K_DPW]

Maybe he wants to do a lot of workouts in ERG mode?

[Gerrie_Delport_ODZ (Gerrie)]

You can use the pedals as power and the trainer as controllable for workouts. It work fairly well.

[K_DPW]

You may take a look here:

Is my KICKR Snap miscalibrated?

This user on trainerroad did a bunch of tests, and his Snap was very accurate compared to his Quarq power meter.

[Craig_D_Mart (Craig)]

My Snap read low compared to my PowerTap wheel, and the Core that has since replaced it. I was happy with that as I wasn’t cheating while racing.

[edward273 (Edward Hicks)]

So Wahoo have very kindly offered to replace my Snap with a Core at no extra cost - Kudos to them and their customer support team (who must be sick of me haha)!

I think in conclusion my Snap unit wasn’t faulty as such - if left to fully reach equilibrium with the surrounding temperature, given a 40 minute warm up, and then given a factory spindown then if kept at a constant wattage for my 25 min pieces above, the Snap was within the 4% error for the whole effort, just not when I trainstioned to recovery power…

However, I think that a changing environmental temperature (i.e. outside in an evening), changing roller, flywheel, and tyre temperature as a workout progresses and changing tyre pressure resulting from this means that it’s very difficult to keep a wheel-on trainer reading accurately for a whole workout which is what I discovered. If just set up outside having been stored inside, given a quick 10 min warmup and a spindown then it was rarely within 4% for any of the workout.

r.e. ERG mode, using the pedals as power and the trainer as controllable is only available in Zwift, Sufferfest which I also use doesn’t provide this and so in ERG mode you end up just pushing 5-10% harder - pretty useless when following a workout. I’ve also found that when using this in Zwift the trainer over compensates one way and then over compensates the other way so you end up oscillating too high and too low the entire time

Thanks all for your input!

Ed

[Dean]

@edward273

So Wahoo have very kindly offered to replace my Snap with a Core at no extra cost

Geez, the support from Wahoo just keeps impressing.

[edward273 (Edward Hicks)]

First ride with the Core and the data is matching really well, very happy!