Drivetrain Efficiency Test: Old vs. New

The Effects of Chain Elongation (wear) vs. Efficiency


To determine the efficiency of new drivetrains vs. old (worn) drivetrains and the individual contributing effects of a worn chain and a worn ring/cog.


Twelve 10-speed Shimano Dura Ace and Ultegra chains were tested.  Two of the twelve chains were ‘new’ (they were broken-in slightly on the equipment; see Test Details).  Ten of the twelve chains were considered ‘worn out’.  The chains were provided to Friction Facts for use in this test by local Boulder, CO bike shops.  Prior to being collected by Friction Facts, the worn chains would have been discarded in the bike shops’ trash bins and replaced with new chains on customers’ bikes.


Each of the twelve chains was tested on both a new ring/cog combination and a worn ring/cog combination.  Additionally, the elongation of each chain was measured. The frictional losses of each chain were compared to the elongation levels to evaluate any possible correlation.


- The Frictional losses ranged from a minimum of 6.26 watts, seen in a new chain & new ring/cogs, to a maximum of 10.27 watts, seen in the highest-elongation chain & worn ring/cogs.

- Chain elongation correlates to frictional losses in a generally linear manner.

- Based on the data collected, the frictional losses increase at an average rate 2.02 watts per 1% of chain elongation.  Tests were performed at 250W rider output.

-  The use of a new ring/cog compared to an old ring/cog decreased overall frictional losses by 0.10 to 0.33 watts.  However, the frictional losses of the drivetrain were predominantly dependent upon the amount of elongation seen in the chain, whereas the wear level of the ring/cogs had a much less significant effect on the frictional losses.

Graph 1: Drivetrain friction vs chain elongation.Linear trendlines added.

Sample Chain Elongation over 10 links % elongation Old ring/cog New Ring/Cog
New DA 4.994 0 6.52 6.26
New Ultegra 4997 0.0601 6.81 6.37
#1 DA 5012 0.3604 7.19 7.01
#2 Ultegra 5.072 1.5619 10.27 9.94
#3 Ultegra 5.013 0.3805 8.24 8.23
#4 Ultegra 5.006 0.2403 7.77 7.96
#5 Ultegra 5.012 0.3604 7.94 7.71
#6 Ultegra 5.008 0.2803 7.75 7.7
#7 Ultegra 5.006 0.2403 7.69 7.68
#8 Ultegra 5.039 0.9011 8.54 7.99
#9 Ultegra 5.007 0.2603 8.54 7.12

Table 1: Frictional losses as a function of elongation

Pic 1: The ‘worn’ DA ring used in testing, showing a ‘shark fin’ wear pattern


  • The 10 worn chains consisted of (2) Shimano Dura Ace 10sp chain and a (8) Shimano Ultegra 10sp chains.  In order to minimize variables, the models chosen to represent the new chains were similar to the models of worn chains.  The new chain models chosen were a Dura Ace CN7901, and Ultegra CN6701.
  • Measuring the elongation of the chains:  Each chain was laid flat on a table.  One end of the chain was fixed. A 5 lb weight was hung off the table and attached to the other end of the chain to provide consistent tension across all chain samples.  10 links of each chain were measured at a time.  This was sequentially performed down a single chain a total of 9 times, thus spanning a total of 90 links.  The 9 measurements were then averaged to determine the average elongation of the chain over a 10-link span.
  • The two new chains were broken-in for 4 hours each, at 250 watts load, prior to formal testing.
  • The new ring/cog combination was broken in for 8 hours at 250 watts load, prior to formal testing.
  • In order to minimize variables due to lubrication, a medium-viscosity mineral oil was used on all twelve chains after stripping/cleaning. AGS oil was chosen due to its formulation as a basic oil, with minimal content of solid particle lubricants.
  • The two new chains were stripped of the factory lube in two 30-minute baths of lacquer thinner in an ultrasonic machine.  The new chains were then immersed in AGS bearing oil in the ultrasonic machine and agitated for a total of 30 minutes.
  • The ten worn chains were thoroughly cleaned in (4) 30-minute baths of lacquer thinner.  Each bath consisted of the chains being ultrasonically agitated for 15 minutes, flipped, and agitated for another 15 minutes.  In order to minimize friction contributions from dirt, this cleaning process was performed four times with new lacquer thinner in each bath. After cleaning, the worn chains were immersed in AGS bearing oil in the ultrasonic machine and agitated for a total of 30 minutes.
  • The 10 chains were hung to dry overnight prior to measuring the elongation and friction testing.
  • Both the new ring/cog and old ring/cog were a 50T/12T combination. A worn DA ring/cog was supplied with the worn DA chain (as a complete worn drivetrain).  This worn DA ring/cog was used for worn ring/cog testing.  (see Picture 1)
  • Chains were tested with the Tension Test Method (FTT) at approximately 250 watts simulated rider output.
  • 95RPM cadence was used.
  • The accuracy of the FTT is +/- 0.02W
  • Equipment system losses (the losses due to the ceramic bearings in the equipment) were subtracted from the final results presented in this paper.
  • All chains were tested using the same worn ring/cog combination.  Thus, the chains were not tested with their originally mated worn drivetrain ring and cog.  It would have been favorable to test each chain with the originally mated ring/cog.  However, only one worn ring/cog was available for the test.  Sample #1 DA is the chain that was supplied with the worn ring/cog used for the test.

Pic 2: Pool of worn chains as received from the bike shops.Chains were sorted, cleaned, and re-lubed.Ultegra and DA models of the pool were used for analysis in this test.