Full Tension Test Method

Friction Facts uses two unique methods of testing chains for frictional losses, the Full Load Test Method (FLT), and the Full Tension Test Method (FTT). Both test methods have strengths and weaknesses respective to each other. As explained above, the Full Load Test Method simulates a true bicycle drivetrain, with a drive sprocket (similar to the crank/big ring on a bicycle) and a load cog (similar to the rear hub/cassette of a bicycle). The chain transfers power from the drive sprocket to the load cog, just as it would occur in a true drivetrain. However, the weakness of the FLT is the accuracy. While the sensor itself is highly accurate, because the energy consumed by the chain (chain friction) is such a small portion of the energy transferredby the chain, the chain friction becomes diluted by the work transfer being done by the chain. Conversely, the Full Tension Tester, offers higher accuracy, but it does not truly mimic the action of a true bicycle drivetrain.

When chains are tested using both test methods, the results provide a valuable combination of very accurate measurements plus true drivetrain conditions.

The Full Tension Test Method has the advantage of obtaining higher accuracy with regard to chain efficiency data. When the chain is tensioned on the FTT, the measured work done to turn the front axle is due to the chain friction only.  No torque load is being applied to the rear axle, as is done with the FLT. With the FTT, a torque sensor can be used with a full-range scale that is is similar to the range of the chain friction. Conversely, with the FLT, a sensor with a much larger full-range must be used to accommodate the entire torque load, but then this large full-range sensor must consequentially be used to measure the very small values of chain friction.

In order to evaluate the very small differences in efficiencies of chains, lubes, etc., the FTT is not only effective, but necessary.

Operational Details:

The FTT apparatus uses two sprockets; a standard crank ring in the front position, and a cog from a cassette in the rear position. The rear shaft pivots on a vertical swingarm in the x-axis. A hanging, sprung weight is connected to the swingarm and used to tension the shaft, which subsequently tensions the chain. The front shaft is turned by a Variable speed AC motor. No torque load is placed on the rear shaft, just tension. The work required to turn the front shaft is due solely to the friction forces seen in the chain, as the tensioned chain is rotated . A single rotary torque transducer is placed on the front shaft. The torque is measured with this transducer and power in watts is subsequently calculated.

Additional details

  • The front ring speed can be varied from 0 to 125 RPM. Friction Facts typically performs tests at 95RPM.
  • The hanging weight can be adjusted to increase or decrease chain tension. The tension applied to the chain simulates the tension found in a chain during pedaling. Lower rider power output equals lower tension, higher power output equals higher tension.  Friction Facts typically tests at a chain tension equivalent to 250W rider output.
  • The hub on the front axle is a 130mm BCD, which can accommodate large and small front rings of various tooth sizes.  Friction facts typically uses a 53T ring.
  • The rear hub is a machined 1.275” diameter drum which can accommodate individual cogs from a common cassette.
  • The front hub can be laterally adjusted outward up to 2” to simulate front ring-rear cog offset, for non coplanar testing.
  • Calibration of the torque transducer is by static beam method. Norbar static beam calibration example
  • The accuracy of the FTT is +/- 0.02W.
  • Equipment system losses due to the four ceramic bearings are 0.24 watts for the FTT.  The Test Results typically have the system losses already removed from the final data, unless noted otherwise.
  • The electronics path is as follows—one strain gauge rotary torque transducer connected to one strain gauge conditioner, outputting an analog signal to an Analog-to-Digital data acquisition device, then connected to a PC running data acquisition (DAQ) software. The DAQ software performs all calculations converting the torque signal into a usable watt measurement. The DAQ software typically averages the data over 1-minute intervals, with a 25 samples/second sampling rate.

Theory behind the Full Tension Tester

As mentioned above, the need existed for a more accurate method of detecting friction forces in chains than the traditional Full Load Test Method. This need led to the development of the Full Tension Test Method.

The energy consumed by a chain (chain friction) in a bicycle drivetrain can be summarized (in simplified form) by the equation-

Total Chain Friction = Sum of the friction at each chain-sproket engagement and disengagement point; given each engagement and disengagment point = [chain tension x angle of link articulation x number of articulations per unit time at that specific engagement/disengagement point].

The [tension x angle x number/sec]  portion of the equation represents the friction seen at a single point of engagement (or disengagement) between a link in a chain and a toothed sprocket. For the total bicycle drivetrain, the chain engages/disengages a toothed sprocket at eight unique points  These eight points consist of two points on the front ring, two points on the rear cog, and 4 points on the two derailleur pulleys. To determine the total forces, the above formula would have to be applied to each of the 8 points, and then the individual friction values of the 8 points summed. A significantly more detailed explanation of forces acting upon a bicycle chain can be found in “Effect of Frictional Losses...”, by Spicer et al. Spicer's paper

When looking at the components of the force equation, it is apparent that the friction forces seen in a chain are not solely dependent on a rotating load and the chain actually performing work, but rather the resultant tension that is seen in the chain, regardless of the manner in which that tension is created. It so happens in a bicycle drivetrain and the Full Load Test Method, this tension appears on the chain as the result of the chain doing work on a rotating load. However, there are many other ways of introducing tension in a rotating chain to evaluate the friction forces.

Finding another method of tensioning to increase the accuracy of chain efficiency measurements is the driving factor of the design and development of the FTT- creating tension in the chain without the need for a rotating torque load. The test apparatus places equal tension on both upper and lower chain spans. When the chain is rotated, chain friction still arises due to the tension and articulation of the links under tension, but the tension and thus friction appears without the need of a torque applied at the rear shaft. Theoretically, the chain is not biased as to why the tension is present.  The same frictional equation applies to a chain whether it is performing the work or the chain is simply under tension, not performing external work.

The FLT (and a true bicycle drivetrain) has a an asymmetrical distribution of chain tensions in its four unique chain sections. Higher tension is seen on the top chain line (drive line from the front to rear sprocket), and the bottom chain line and chain lines through the derailleur are subjected to a much lower tension. Meanwhile, the FTT has only two chain sections, which experience higher total tension, and are symmetrical in nature. Regardless, this distribution of tension, whether symmetrical or asymmetrical, does not affect the measured total friction losses in the chain. Recalling the formula, the total losses are the sum of engagement and disengagement points. Ultimately, the total chain friction is not dependent on the relationship between the chain sections, nor the individual engagement and disengagement points. Each engagement/disengagement point acts independently of one another with regard to its contribution to the total chain friction. Thus, the different force diagrams of each tester should not affect the results of the forces being investigated, i.e., the complete chain friction forces.

Proving the Theory

The above Theory was simply a theory. To the best of Friction Fact’s knowledge, an FTT style chain friction tester has not been produced let alone data published using this method. Substantiating a linear correlation between the measured chain frictional losses on the FLT versus the FTT was necessary prior to the design, development, and usage of the FTT method.   (On a side note, Wipperman has used a similarly-designed apparatus to accelerate chain wear rates, but not to test frictional losses. http://www.youtube.com/watch?v=RGjcD8xEu8o)

Upon informal preliminary testing of chains between the FLT and FTT, it became apparent that the Theory did indeed hold true. Similar friction losses are seen in a chain regardless of how the tension is created. A linear correlation between the two testers does exist.

This correlation is quantitatively documented in the formal Chain Efficiency Test, Part 1 and Part 2.  A quick glance at the FLT results graphs and the FTT results graphs show similar trends between test methods

Design History

The first generation prototype FLT and FTT apparatuses are similar to the final production apparatus. A third tester was built to test chain losses at low tension with high articulation. This third tester was essentially an apparatus using a light hanging load with the chain going back and forth over multiple cogs/pulleys. This tester is not used. It was realized during initial testing that a chain requires higher tension to determine frictional losses accurately; third tester could not provide the tension level required.

The design for the second generation apparatus functionally combines both FLT and FTT methods into a single piece of equipment. As seen in the CAD drawing below, for FLT functionality, the bar is locked between the front and rear shafts, keeping the distance between both shafts fixed, simulating the bicycle drivetrain. To change to FTT method, the shaft bar is unlocked, the rear torque load disengaged, and the rear swingarm assembly is slid rearward. This second generation design is more cost effective, and possibly provides a higher level of accuracy, since the same piece of equipment is used for both measurements.