Drivetrain loss - n=1 study!

As part of transparency for Zwift racing, I’ve started dual recording power data for my rides. I use a Tacx Neo that doesn’t require any calibration and Powertap P1S pedal. Initial recordings below showed that the P1S read a good bit higher than the Neo, which I found unusual as DCR used these two as his gold standard for comparing powermeters:

Powertap ride #1
Neo ride #1

I fully degreased my chain, cleaned up my cassette and lubed with Finish Line Ceramic Wax lube. Here’s the second set of data after cleaning up the drivechain:

Powertap race #1
Neo race #1

Average power on both the pedals and Neo were pretty much identical over the race distance.

I’m pretty meticulous about keeping my bikes clean, but as my trainer bike doesn’t go outside the chain doesn’t get the same love as my road bike. Moral of the story? If you’re using a trainer as your power source, pay attention to your drivetrain. The losses can be significant.


TL;DR: freshened up chain on trainer was good for 15 watts. Good to know…


Interesting, I only use the Hammer (since I have a powertap hub on the road) so maybe I should be more diligent with the drivetrain. Maybe those 120% vo2 intervals are really higher, or at least I can tell myself that when I stink at the 2.5-3min versions lol


While DCR clearly relies on P1 pedals for comparing against other power meters quite a bit, his stance on single-sided power meters of any variety leans more toward “it’s tough to recommend the left-only approach with other options in the same price ballpark that fully capture all power.” I agree, for analysis of this kind I’d want to see real “complete” power at both ends of the drive train compared.

Looking at your first ride, for example, I see a lot of differences that look like they could simply be inaccuracies of a single-sided power meter; I also see a lot of stuff that’s hard to eyeball. I wonder if you could try throwing in the data files for these rides into the DCR Analyzer to make for a more direct, second-by-second visual comparison.

I’m not saying the drivetrain cleanliness/lubrication doesn’t matter, just… this comparison leaves a lot of questions.

1 Like

Absolutely granted there’s probably a lot more to it than just the drivechain. Reason I posted this was lots of people talk about drivechain losses with measuring power at the hub. Only thing I did different was really thoroughly clean the drivechain. I’ll be recording all my races on two sources from now on. I’ll post some other rides in time👍

1 Like

Sorry if this has been answered somewhere on the forum but I couldn’t find anything. I found that over the course of the last couple of workouts the is a slight squeaking coming from the drive chain, which says to me it’s time to relube. However, I was wondering what the science was behind the power loss if your chain is not lubed against a freshly lubed chain whilst on the trainer if there is any at all? Mechanically if the chain is not lubed what difference will this make to the power applied?

Any ideas would be great. This is less of a maintenance “lube your chain it will last longer” question but a performance related one. It will make those workouts feel a lot harder if you’re losing 5-10 watts? For example in Carpathian Peak it can make huge difference. If there is specific data that would be great but to understand the principles for me is more important.



Sorry if this is a dumb question but did I get that right: A dirty group will mean off numbers on a wheel on trainer in erg mode? Asking for a friend who’s bike doesn’t get much love indoors.

@ajh127 I’m not aware of any specific numbers beyond n=1 testing like this, but noise = friction = loss. I’d love to see data, as well

@Triathlete Yes. Any ‘loss’ between your measuring device and your feet isn’t going to show up. Now, for example, if you’re using Vectors or Assioma pedals and having your trainer power match, the numbers you’re seeing are going to be correct, but the trainer will just adjust the resistance to account for the loss. If you’re using a hub-based PM in the same scenario, your numbers will display less since the drivetrain friction will have already been accounted for

1 Like

I use PT hubs, and I have never noticed a big difference in watts between a clean and dirty chain, or between a worn and a fresh chain – maybe, maybe 5w.

I’ve compared to crank-based units and I’ve never seen the “hub will read 10w lower” to be true. If it is, I’m over 5 w/kg at 54, which I would have a difficult time believing.

I think even 5 watts difference as intervals get longer e.g. a 10 min plus threshold interval will make quite a difference in feel, at 105% FTP I can certainty feel the difference between highly uncomfortable and the onset of the burn.

The noise and friction point are definitely salient.

Check out my post there. I recorded with both P1’s and PT hub and saw drivetrain losses increase up to 10% in a very long ride. It did rain at start of ride, which probably reduced the lifespan of the squirt lube on my chain.

After posting, noticed the linked thread is the same thread. Oh well. :laughing:

5w is also within the normal daily variance of most any PM you can buy.

Use these things descriptively, not prescriptively. The truth is in the cells, not the numbers.

1 Like

I can guarantee I’ve seen up to a 20w difference. I need to test VO2+ efforts again because I believe it’s more percentage based and not a fixed wattage. Believe I’ve seen up to 40w difference on 400w+ efforts, but need to retest because that was a while ago. If you are using a pedal, crank, or spider power meter than this really doesn’t matter unless you are racing. I agree I wouldn’t worry about it for training. But if you are using a hub or trainer based power for training, I would concern myself with it and keep a clean and lubed chain.

I got a two hour threshold session scheduled for tomorrow. Curious to see whether my now clean drive train makes it easier. Yesterday’s over and under session wasn’t exactly a walk in the park.

I’ve been noticing that my Assiomas consistently read 10 watts higher than both a Neo and a Kickr.

My chain is kept fairly clean, and gets rotated with a very clean one every couple of weeks.

Is 10w drivetrain loss possible when your chain isn’t really cruddy? I guess I’m wondering if there are 10 free watts on the table if improve my drivetrain!


"Out of the box most chains have friction levels of over 10 watts.

assuming the issue is not just an offset reading problem, then no you can never fully remove drivetrain losses which are the sum of BB, chain, pedal/crankset/frame flex, chainline, cassette etc between the pedal body and power to the ground or whereever the downstream measurement point is. This of course can be compounded if the chain is excessively losing watts but then your variation would not be a consistent 10 watts as chains break in, lube/wax distributes etc