Seems like Wahoo doesnt log anything 
I just looked over the datasets I’ve kept, and noticed something interesting. In nearly all of the Zwift rides I’ve done, the power difference corresponds almost exactly to a drop in virtual speed to around 15 km/h or less.
For example, when hitting the bottom of Alpe du Zwift, although my power at that point doesn’t change that rapidly, the power difference develops at precisely the moment that I slow down to climbing speed. Same thing can be seen beginning an ascent of Ventoux, and in a recovery session in NYC (at around 20 minutes; again, no significant change in power output here).
This makes me think it has something to do with the speed of the KICKR’s flywheel.
As for ERG mode, this is all done in one gear, so there’s not much change in the flywheel speed there. However, you’re supposed to do it in a low gear, and therefore with a low flywheel speed, which could explain why it’s commonly seen during ERG mode workouts.
Has anyone else observed this pattern?
Yep… exactly this with the initial release firmware. But it wasn’t virtual speed, it was the flywheel speed being <20km/h… (which usually correlates to low virtual speed) . This was addressed (mostly) with recent firmware updates.
What version of the firmware are you on?
Huh, interesting, did not know that. Guess I should’ve done my research before buying 
Still, I’m on the latest firmware (4.2.1).
Having read DCR’s report on this, it looks like the KICKR was originally over-reporting power, while here I’m seeing the opposite.
When you say “mostly” fixed, are there still known issues?
I think the drifting is a different issue to the cadence issue that was present in the initial firmware.
Drifting seems more dependent on the temperature of the trainer.
Right, but at this point I’m not sure which is which… At least in the data I’ve collected, you can view it both ways. Most of the hard efforts I do that might cause heat-drift are either in ERG mode in a low gear, or going up a steep hill, both of which also have a slow flywheel. So these conditions are themselves correlated.
I’ll try a flat TT some time this week (high power, fast flywheel) and see what happens. Another thing I want to try is pointing a fan at the KICKR’s flywheel to help it cool.
Add me to the list of the afflicted, lol. From what I’m seeing, higher power = more drift.
OK, today’s test was a 60 minute flat TT in Zwift, again in sim mode. The goal was to get the flywheel spinning pretty fast, but at a sustained high power output.
https://analyze.dcrainmaker.com/#/public/dda225a0-0e3b-4d3d-656d-df05fadd69ea
Pacing issues aside, the power matches damn near perfectly – the difference is less than 1 W over the duration of the effort. With 60 second smoothing, it looks like this:
P.S. Please ignore the device names that DCR Analyzer reports – they are the wrong way around. zwift.fit contains the power meter data and elemnt.fit contains the KICKR data (unlike my previous datasets).
I had some weird feelings with my kickr v5 Tuesday, the resistance kept feeling too easy for a fraction of a second then went back to normal.
Checked the wahoo all and there was an update which looks like it specifically addressed this issue (& might show as power drift maybe??)
I’ve been using the Kickr 5 as a baseline recently for other meters and observed some differences that were resolved by following the below:
Tips:
- If the numbers don’t line up, do a factory spindown after riding the unit easy-ish for 10 minutes. Don’t do it while it’s white-hot after a hard workout or after an extremely hard warm up.
- If you’re on a quick-release bike, don’t lock the bike in hulk-tight. Not sure if this was causing issues I was seeing, but I had the bike locked in crazy tight. Firm-AF-tight seems to be better.
- Check the freehub isn’t packed with shitty grease or is dry. I have had trainers arrive with no freehub grease. I strip and reapply freehub grease every few months.
- Check your hanger. Make sure the bolts are firm into the frame holding it on. Make sure it’s straight.
- Other obvious ones: Make sure your chain is the right length for the gears/chainrings, chain is clean, b-tension is set correctly, both jockey wheels spin nicely, and use a straight chainline when doing steady-state comparative tests.
- Switching to a better-than-stock BB might also help (well, it can’t hurt…). I switched my test bike to a Wheels Manufacturing BB a little while ago.
Here’s today’s data from the Kickr 5, a set of Assiomas, and a spider meter:
*Note: The Kickr will still smooth/fake ERG watts even with ERG Smoothing disabled.
6 Likes
I got my Kickr and had to swap from Shimano freehub to SRAM XDR. No excess grease in my Kickr. Reassambled and rode 2 hours last night, doing only a couple of short pauses to let Kickr auto calibrate to do its thing. And of course zero reset Quarq before starting.
- Quarq in blue - recorded with Garmin 530
- Kickr V5 in black - recorded with Zwift on AppleTV
Pretty good.
Another chart from ZwiftPower:
This was a 2 hour endurance ride in Zwift, with Zwift controlling Kickr in sim mode to simulate outside. And my Garmin 530 was showing workout and recording Quarq:
Here is when I gave a little gas to Fuego Flats Forward Sprint to keep things interesting (just like outside):
Close enough, just want Zwift to simulate riding outside. And use my Garmin 530 for the workout, just like when I train outside on flat roads.
Pretty darn happy and so much better than my grocery store bike and Kickr V3. I’ll continue to freewheel every now and then, to let the Kickr’s auto calibrate do its thing instead of having to stop and do a spin down.
2 Likes