Lab testing confirms: Clean your damn chain

Don’t miss out on the latest CyclingTips updates.

Jump To Comments

I’ve taken some pretty deep dives into the topic of chains over the past few years, and the common theme throughout is that having a clean chain is not only a more cost-effective way of riding, but it’s also more efficient, too.

While the cost savings from using a good lube suited to your riding conditions have been tested and discussed, the impact of contamination on efficiency has been glossed over. So I’m back today with some long-lost data to put efficiency numbers to that gritty chain.

The outcome? Well, it’s a story that’s been told before: research the best chain lube for you, and keep your chain clean.

Opening the Friction Facts archive

Jason Smith, formerly of Friction Facts, is responsible for much of today’s knowledge about drivetrain efficiency. Prior to CeramicSpeed purchasing his independent testing company, Smith would sell his extensive scientific-type findings in online reports.

Winding back the clock to 2012 reveals a handful of efficiency studies on chain contamination. There are many caveats with this data, the main one being that contamination is hugely variable, and the choice of chain lube only exacerbates it further. Still, Smith’s findings give a glimpse into the significance of efficiency losses that can come late in a dirty race or ride.

The first test of its kind looked at the before and after effect of racing a damp-ish cyclocross race in Boulder. This included three chains treated in the original UFO formula (hot wax submersion treatment, quite similar to Molten Speed Wax) from the time and three chains that were given a light oil lubricant. Friction testing, as usual, was done on a calibrated machine simulating 250 watts at 90 rpm.

All figures shown below represent wattage lost through the drivetrain. For example, a loss of 6 W means that only 244 W are received at the rear wheel for an input of 250 W. These figures may be surprising, but it’s a well-proven fact that chain-driven drivetrains are not 100% efficient. Also of note: the rate of loss is not linear, so a higher power input will result in a progressively bigger decrease in efficiency.

UFO V1-treated chains:

Efficiency loss while clean Efficiency loss post-race Decrease in efficiency
Chain #1 6 W 8 W 2 W
Chain #2 7 W 8 W 1 W
Chain #3 7 W 9 W 2 W

Light Oil chains:

Efficiency loss while clean Efficiency loss post-race Decrease in efficiency
Chain #1 5 W 10 W 5 W
Chain #2 6 W 8 W 2 W
Chain #3 7 W 13 W 6 W

The outcome of that (relatively short) test was quite clear: If the lubricant used doesn’t collect contamination and instead provides a barrier to stop grit from entering the chain, there are benefits to be had.

Smith suggests that in truly muddy conditions, such as the recent Mid South gravel race, efficiency losses would be in the 20-30 W range (at only 250 W!). Clearly bike changes in cyclocross races offer more benefit than working brakes, reduced weight (mud gets heavy) and visible sponsor logos.

Another test was conducted in 2013 that focussed on clean, rainy and muddy conditions and the use of a light oil-based lube.

Chain/conditions Efficiency loss Decrease in efficiency vs clean and lubed
Clean and lubed 6 W NA
Rainy 13 W 7 W
Muddy and wet 14 W 8 W
Muddy and dried 18 W 12 W

A more current test

Now wearing a CeramicSpeed T-shirt but using the same machines, Jason Smith recently ran a similar test with more specific chain lubes. This test sought to find the efficiency of CeramicSpeed’s own drip lube when subjected to a controlled, very fine grit, but also provides insight into other popular products.

Lube used Efficiency loss when clean Efficiency loss with grit Decrease in efficiency
UFO Drip 4 W 4 W 0 W
Squirt 5 W 7 W 2 W
Rock n Roll Gold 5 W 8 W 3 W
Morgan Blue Race Oil 6 W 10 W 4 W

A squeaky clean chain is also bad

There’s a common misconception that if a clean chain is best, then keeping it dry and perfectly free of any lubrication (which often attracts dirt) is the best answer. This is simply false, and the articulating metal-on-metal nature of a chain needs lubrication for proper function.

This is something Smith proved back in 2012 by testing the efficiency of six chains. First, the chains were tested after having been methodically cleaned and equally lubed. Then the chains were taken back to bare metals with solvents and tested again.

Efficiency loss when clean and lubed Efficiency loss when stripped Decrease in efficiency
Chain #1 6 W 20 W 14 W
Chain #2 7 W 20 W 13 W
Chain #3 7 W 20 W 13 W
Chain #4 6 W 26 W 20 W
Chain #5 7 W 24 W 17 W
Chain #6 7 W 20 W 13 W

Zero Friction Cycling says wear is watts

Adam Kerin of Australian-based Zero Friction Cycling has taken the torch from Jason Smith in offering independent chain tests. Currently, Kerin doesn’t have the same efficiency-measuring machinery as CeramicSpeed, and so relies on a more time-consuming and chain-hungry test that correlates efficiency with chain wear; something that’s proven to be extremely reliable and quite useful for real-world usage.

Still, Kerin has been sending wear-tested and sometimes gritty chains around the world to various labs for blind friction testing. Some of this data lacks a rock-solid baseline, but it still aligns with what Jason Smith had previously discovered.

Smith’s former testing pointed to an approximate 2 W loss in efficiency for a chain that was worn by 1%. Kerin’s testing suggests this figure is on the low side given the likely wear of low-friction surface treatments and the inevitable increase of stubborn contamination within an old chain.

KMC Digital chain wear gauge
Worn chains are slower, but a dirty chain is far slower, again. A dirty worn chain? Criminal. Somebody stop that (watt) thief!

For example, Kerin had a .78%-worn Squirt-treated chain (ultrasonically cleaned and re-lubed) test at a loss of 9.39 W — 4.59 W higher than the figure published by Friction Facts for testing that lube on a new chain. Meanwhile, a .72%-worn chain that was cleaned and re-lubed with NFS oil tested at 7.65 W — a 2.1 W increase over Friction Facts’ previous clean test.

Sure, 2 W (at 250 W) may not seem like much, but consider that at this point you’re also causing wear to your cogs, and so that loss of energy is costing you money, too.


Is my shortest-ever chain-related article still too long for you? Well, in that case, here are some key takeaways.

In the tests where wax-based (dry) lubricants were compared to liquid (wet) lubricants, and given the same test conditions, the wax-based dry lubes (not to be confused with “dry” oil-based lubes) exhibited less increase in frictional losses.

Oil-based lubes can indeed still be efficient, but the way they attract dirt means you need to be more fastidious with your cleaning regime.

If in doubt, clean your chain.

Seeking a rabbit hole for all things bicycle chains? Follow that link …

Editors' Picks