do you have your pedals set to the correct crank length? those discrepancies seem...large.

I did the same today and was crushed to find my new Tacx 2T reads significantly lower than my Quarq Elsa RS.

I can deal with the psychological blow, but am wresting with how to convert training power (Tacx) to real world power (Quarq).

Don’t want to hijack your thread, so will just follow along for now and see what advice pops up. If nothing else, it does highlight the Lions’s point.

My Quarq DZero reads a bit higher than my Kickr, but not by that much. It was on the order of 3-4% at lower wattages and 1-2% at higher ones. One solution is to use the Quarq as the power source in Zwift. I’m interested to hear of others.

Just use the quarq as your power source. Trainer power seems to be generally questionable, and that way you're always using the same device to measure

Yes, I changed the crank length in the computer from 162.5 to 155mm a few rides before Sunday’s ride. I did a few easy spins and noted the numbers seemed to be off more after the change but no official analysis.

I recently went to 155 cranks and initially forgot to change the crank length in the garmin.

2 weeks ago I did a ramp test when the garmin computer was still set at 162.5mm before I remembered to change it. The analysis from the kickr and pedals are almost the same throughout that test.

I have PowerTap P1 pedals. I own an Elite Direto direct drive trainer and the trainer is pretty consistently about 2-3% higher. However, before I bought the Direto I had the chance to use a friends Kickr (generation 2 I believe), and it was also slightly higher than my pedals. I don’t remember the exact amount, but probably in the 2% range.

I was a little surprised by that, as I would have thought the pedals would give me the higher read, given that they don’t experience drivetrain loss. But they’ve read lower than both trainers I’ve tested against.

Most modern power meters are able to do on the fly re-calibrations for temperature changes but this wasn't always the case. Before mid-ride temperature corrections it was common to see substantial drift over the course of a ride just due to temperature changes. If you look up any of the long term reviews of the original Garmin Vectors you will see how badly the lack of temperature corrections can affect things.

Smart trainers can't due this and I suspect it is part of your problem especially over a 5 hour ride. Even if the ambient temperature didn't change much the trainer will have built up a lot of heat. The recommendation is always to calibrate the trainer after 15-20 min so its warmed up but 5 hours is well outside the expected usage conditions of a turbo so it doesn't surprise me the trainer started to drift. On a Tacx trainer you can do a re-calibration without having to re-start Zwift, Trainer road etc and I think the Wahoo app allows you to do this too. Over the length of a Zwift race I don't think this is necessary and I would just calibrate the trainer during the warmup period. For much longer rides I would either do a mid-ride calibration or just use your PM.

The power differences are also coming from where the power is measured (Cranks vs hub/wheel of the trainer) So Crank based or pedal based power meters will normally read higher than the trainers because of the drive train losses. But the 7% differences seem too high for that so it could be calibration.

Were you in the big ring by chance? It's a known issue with the neo that at high flywheel speeds it becomes more inaccurate. There is also a new firmware update to address this issue for the 2 & 2T though. But for erg mode training, putting it into the small chainring and running the trainer at low flywheel speeds helps. This does not help for zwift racing though, so just use your quarq for that.

To me, the obvious answer is to use the bike equipped meter to control the trainer. But that may not work for some Z races demanding a certain setup. But, for training consistency that is what I would do. Then there's zero discrepancy.

Unfortunately, pretty sure most Z races that "care that much" require use of the trainer's power output.

Maybe just not the difference, train using the bike meter to control the unit, then race in Z within the rules and just know what you're seeing different.

I have always struggled to get my head around this because I don't understand what is and is not accounted for in the spin down calibration tests.

Trainers like the Kickr require spin down tests because the trainer doesn't measure power in the same way a standard hub based power meter does such that what the trainer measures in not really what it is going into the system. In a spin down test you put get the fly wheel up to a very specific speed and because you know the weight of the fly wheel you know its kinetic energy. You then measure how this energy dissipates with time after inputs into the system stop and this helps calibrate the system. The thing is you also need to make some assumptions about how much power was required to get that fly wheel spinning up to the desired target in the first place i.e. what is the transfer efficiency is into the fly wheel. It may be that trainers only try and work out the losses between the freehub and flywheel but I think most factory measuring/calibration rigs use some sort of modified cassette and chain system. That means they are probably taking into account some fraction of drive train losses but I am not sure where they measure the power input.

My suspicion is that there are some secret sauce algorithms and methods involved that trainer companies don't want to make public involved. But without knowing what actually goes on its hard to say for sure where the errors are in the system.

I do not know if this sheds any light on this discussion, but at the 4 hour mark of this ride I went up the Alpe d Zwift. If I just isolate that section here is what it looks like.

Kickr still higher power
3.99% difference.

I am going to calibrate the kickr and see if that changes anything. Appreciate all the input, I’m not too concerned about it, I have not done a zwift race yet but would like race something this year!

From the information i could gather and my own experiences : tacx neo are known to track precisely precision power meter so the quark mentioned is probably a bit high.
About the direto i own one and it was quite precise when new but now drift about 20 watt optimistic, proper calibration and belt tension.
I've only read and experienced good things about Favero assiomas precision, so for those who have those choose them i guess.

For the record, wanted to share that a firmware update for the Tacx narrowed the gap for me. Still far from perfect, but more manageable.

Maybe I’ll just train with the Tacx and race via Zwift with the Quarq 😉.

I updated firmware for my Assioma pedals yesterday and did my first zwift race. The numbers are 1.3% difference at 5,10 and 20minutes, so much better for me also.