Great shout. A mate of mine has a pair of the new Vector 3s so will give those a shot and see if the same issue arises. I have a feeling I’ll still see that ~12% difference…
Thanks so much for your insights Phil. Really appreciate that.
I would expect ~10W loss between the pedals and the rear hub. Just the chain is ~6W even if it’s a new chain, no factory grease on it, new cog, new chainring, etc. Throw in some derailleur jockies, bottom bracket, less than ideal chainline…I would never, ever expect a hub power reading to match a pedal power reading. If they match, somebody is telling fib!
Here is some data from friction facts from their experiments with new chains/cogs vs worn chains/cogs:
Here’s some derailleur data. Dura Ace derailleur adds between 3 & 4 watts to the drive train (in)efficiency. 105 derailleur (not shown in this data set) is over 1W worse than the Dura Ace. So between that and the chain/cog you really are looking at 10W in the drive train…and that’s with a Dura Ace derailleur and brand new chaing/cog/chainring. For me, the bottom line is there should be a ~10W difference between a pedal power observation and a rear hub power observation.
Any two difference power meters will give two different readings, I went
- First power meter Kicker , FTP 285
- Swapped to Garmin Vectors -FTP 268
- Finaaly decided to get a tacx neo - FTP 250
This was pretty much going from one to another, all this 1% accuracy is bull locks, just pick one and stay with it. After all the money I have spent on trainers, I which I’d just gone with power meter and fluid trainer
Like has already neem said. there is a difference between reading at pedals and rear wheel as well
Pick one, go with it
All that online data is fine and maybe in general terms certain chains, cogs, and their wear applies to loosing watts. However, until you start testing your own drivetrains and your trainer none of the generalities matter. I would have thought my wheel-on Kickr Snap versus my Quarq XX1 that measures at the spider would be way off. I mean my Kickr Snap has to deal with the chain, cassette, tire, tire pressure and drum. While my Quarq avoids all of that. So in theory I should be losing watts to the drivetrain and wheel, but after 5 workouts at all different power levels and cadences that’s just not the case. Yes there are deviations at times depending on the workout, but the avg power and NP are always within a couple watts of each other.
your trainer vs power meter graphs are better than mine, and yet my walk away is still “don’t use Kickr power”
My take is everyone with a single bike+PM is to have one source of truth - your bike’s power meter. And on the trainer do one of the following:
- use TR with bike’s power meter and trainer in standard mode
- use TR with bike’s power meter and trainer in resistance mode
- use TR’s PowerMatch (bike’s power meter) and trainer in Erg mode
Yeah, I agree. If you’re seeing the same power reading between those two setups, something is wrong. Somebody is telling you a fib! Unless you believe you have a zero-loss drive chain! And if we believe our drive trains are zero-loss, why are we spending hundreds of dollars on ceramic bearings, oversized pulleys, crock pots dedicated to waxing, etc? On the one hand, we spend a lot of time and effort reducing drive train losses…on the other hand, we can’t come to terms with the idea that hub power readings should be different than pedal/crank power readings.
I wonder what the pedal/hub power reading difference would be if somebody put a super tight brush clamp on their chain? Would the hub still measure the same as the pedal? I don’t think it should, but who knows…
No fibs, the power meters are are just reporting inside the margin of error 1-2%. Also, one could theorize that the the Kickr Snap actually reads 2% higher, but then the drivetrain and wheel looses that 2% of power due to efficiency. Which in the end results in the Kickr and Quarq’s powers tracking closely. So does Wahoo know the percentage of watts a wheel on trainer may lose and thus compensates in their algorithms when showing power? In the end it doesn’t matter as long as the powers are consistent.
I had a 20-30W difference between my Neo and Assioma pedals on the first ride where I used power match. Turned out that I had just installed the Assiomas and forgot to calibrate them. After that experience I rode a bunch of rides monitoring the Neo broadcast power and the Assioma broadcast power and analyzed the rides using the DC Rainmaker analyzer. The net result was a very small difference of 1-2% for all intervals a few minutes or longer. For short intervals the Assiomas and Neo agreed quite well as long as power match was turned off. If I turned power match on I would see anomalies in the power output for shorter intervals. I think that the competing servo loops in the Neo and TR do not mesh well for short intervals. TR support told me that they didn’t recommend Erg mode and power match for short intervals. I have found that Erg mode by itself works fine for me in that case. Power match not so much. Finally, I just decided to use only the Neo indoors and ignore the Assiomas and power match. This is because most of my structured training is indoors and what I want there is consistency. I don’t really care about my FTP number except as a way of setting good targets for workouts. Using just the Neo by itself does that- my hard workouts are usually ones that I have to really struggle to complete. So, I figure that the targets are decent. If I had a different trainer that required calibration and changed from day-to-day I would probably do it differently. But my two Neos have been consistent and agreed with each other for over a year without ever requiring calibration or any other attention except cleaning the sweat off them every now and then. Outdoors I use the Asiomas and my experimentation showed that they agree with my Neos to within a couple of percent which is good enough for me.
to my eyes your graphs are showing that the power numbers are consistently inconsistent. In particular test 1, test 2, test 4, and a little of test 5. I had testing similar with a Kickr direct-drive, and then had some outlier tests that look particularly bad. As stated above, if it was me your tests would convince me to use TR with bike’s power meter, using either standard, resistance, or Erg/PowerMatch.
I updated my above graphs with more stats.
I’m not concerned with a few watts difference here or there, they are close enough to get the full intent of the workout. The later tests I went to higher gearing(from 32x21 to 32x16) to see if inertia had an effect. I will probably end up using Powermatch (or at least give it another try), but I had my issues with using that on the ramp test so I was looking for a way to avoid it. After testing it appears I can without concern either indoors or outdoors. Powermatch doesn’t like cadence drops in my experience and that’s a problem for my riding style.
During my testing I was seeing pretty consistent 3-5W difference, and then later in testing had some outliers. That was enough to abandon testing and switch to using the bike’s power meter only.
While I’ve had good experiences with PowerMatch, if not then I would try using the trainer in resistance mode (if doing Erg, its what TR switches to during 8-min or 20-min FTP tests). My Kickr also supports standard mode.
If you have concerns about Erg and cadence variations, or you happen to like shifting during indoor sessions, give resistance mode a try.
I think I’ll just keep as is with the Kickr’s Erg as it’s been good and consistent to me for 1.5 years. Again, I don’t care about 5-10 watt swing difference. My FTP is only an estimate, the power meters at best are accurate within 1.5%, so it’s all a wash in the end. I welcome the subtle variations, it will only make it more representative of outdoor riding. Now maybe if the Kickr Snap and the Quarq had a 20+ watt difference I would reconsider.
Variations are the reason I prefer using TR’s PowerMatch and training in big chainring. With the direct-drive Kickr, the little ring and Kickr as power produce an unnaturally smooth ride. I can’t find any outside rides in flats or climbs that have power that smooth. To my legs and brain, the little ring and Kickr as power source feels nothing like riding outside. Wheel-on trainers will feel different.
All I ride is MTB and all I have is a little ring. Also, I have power smoothing turned off in the Wahoo app which TR support said does affect how Wahoo controls the trainer. I can feel the subtle variations constantly and rarely does it feel smooth, but maybe that’s just the result of a wheel-on trainer. I have no experience with direct drive so I can’t speak to that.
I also had 10%+ difference between the kicker and quarq. But because I trusted the consistency of my quarq, I’ve used power match for erg or standard mode and just moved on.
Here’s the data when I was trouble shooting the difference. Cliff’s notes, kicker was consistently higher than my PM. Use the quarq to drive the trainer.
@philsjones What do you torque your Vector 3s to? And do you have a recommendation on torque wrench or is one just like the rest?
@mrlavalamp The vector 3 torque recommendation is 25lbf/f 34Nm See https://www8.garmin.com/manuals/webhelp/vector3/EN-US/GUID-DA17FC98-ED3B-4B9D-8028-0E8F94AC0767.html
I have a lifeline torque wrench I bought from Wiggle, that has a max torque of 25 lb/ft and came with a collection of useful 1/4" hex and star sockets. It is small and neat enough to pack with my other tools for races etc. Handy little set and seems well made. I could not see the same brand, but this set looks pretty identical https://www.chainreactioncycles.com/x-tools-essential-torque-wrench-set/rp-prod155414.
(Just don’t buy the deflection/beam type torque wrenches - they are not as accurate and much harder to read. https://www.amazon.co.uk/Performance-Tool-W3001C-Torque-Wrench/dp/B0002KO1I0/ref=asc_df_B0002KO1I0/). There are some recommendations (sponsored no doubt, here https://road.cc/content/buyers-guide/254944-6-best-torque-wrenches-tools-help-you-get-bolts-just-right-tightness).
You need a claw spanner to convert the 1/4" drive to the pedal nut size https://www.sigmasports.com/item/Park-Tool/TWB-15-Crowfoot-Pedal-Wrench/5PFM?co=GBR&cu=GBP&glCountry=GB&id=266242&gclid=EAIaIQobChMIko_ztrTe4QIVhZztCh2hrAE1EAQYASABEgIqqvD_BwE&gclsrc=aw.ds
However the claw spanner is usually for a 3/8" socket so I also bought a set of various converter drives though you only need one. https://www.ebay.co.uk/p/Hilka-06200700-Universal-Adaptor-Joint-Set/2299240156?iid=171215807458&chn=ps (I have a 1/2" torque wrench and loads of 1/2" and 3/8 sockets as well as 1/4" so buying the set made sense.)
I have a basic rule. Get the right tool for the job, and buy decent tools… that way they will last a lifetime… and always be useful.
Final thought - if you are pressing hard on or holding firmly, the nut end of the torque wrench when you are turning it you might, on a low torque setting, alter the resistance. So I make sure I just hold the handle. Also the lifeline one adjusts with the handle, so be careful that in tightening a number of bolts you have not altered the setting. Just stuff to watch out for…
Is that enough info?
It seems like many of the powermatch issues seem to be people using pedals? Crankbased options seem to play better for whatever reason. Maybe there are less external factors like torque?