I performed tests on several multi-speed bikes to determine how much energy is wasted by a dirty chain. The results were significant and consistent. The dirty chains consumed 8 to 12 watts of energy when the power at the crank was tested in a range of energy outputs that would be typical to propel a bike 8 to 14 mph.

What does 8 to 12 watts of lost energy mean when riding a bike? At 15 miles per hour, a typical bicyclist will slow down to 14.5 miles per hour if 12 watts of energy is taken away. This would add about a minute and a half to a 10 mile ride. On a 4 percent grade at 10 mph, wasting 10 watts of energy is like adding 10 pounds of weight to the bike.

I used three different bikes that had been ridden anywhere from several weeks to several months since the last cleaning or oiling of the chain.

In each test I used a DC powered electric drill to drive the crank from the spindle bolt. I measured the motor voltage and amperage at various speeds. I repeated measurements at each speed a number of times. I added two tire driven light generators to increase the resistance on the spinning tire in order to simulate the actual resistance of normal riding. Multiple nicad batteries were used to minimize the voltage drop from the power source.

After measurements were made with a dirty chain, I cleaned the chain, the derailure guide wheels, and the drive gears. I lubricated the chain with a light oil lubricant. No other parts of the bike were changed.

I recharged the nicad batteries and carefully repeated the same sequence of speed measurements, again charting the voltage and amperage.

The energy unit of watts can be calculated by multiplying the volts times the amps. With the tabulated results I could calculate the change in watts. The error fluctuation in each phase of the measurement was in the range of 2 watts. The bike chain that looked like it had the least oil showed the biggest improvement, but even a chain that had been oiled within three weeks showed an improvement 8 to 10 watts saved after cleaning and oiling.

On one bike that had not been oiled in several months, I made the "dirty" measurement and then just oiled the chain thoroughly without cleaning. The freshly oiled "dirty" chain used 10 less watts than the initial measurement. I then cleaned and oiled the chain as in the other experiments and measured the power consumption again. The clean chain did not produce a measurable improvement over the freshly oiled "dirty" chain. This null result is a reflection of the limits of precision in the method and equipment I used.

This experiment was limited in precision by several factors.

- I used a standard bike computer speedometer that calculates speed based on impulses from a magnet sensor. The computer display speed had one mile per hour increments. The unit of the display limits the degree of accuracy with which each repeated measurement replicates an exact speed/power proportion.
- I measured the power draw of the DC motor. The relative efficiency of the electric motor adds to the total energy draw. I am assuming that this energy draw is a constant, but that is not a precise deduction when the motor is actually performing at a range of power levels.
- My survey of literature shows that the power draw from the bicycle transmission increases in a low angle linear fashion. There is an assumption that these measurements at low power levels are reflective of actual riding. The real world riding force on the chain will be higher and the friction may be higher too. The power available from the electric motor I was using limited the range of resistance levels that I could measure and compare.

- A dirty chain consumes significant energy: 8 to 12 watts at low speeds.
- A well oiled chain will minimize this energy loss, even if the chain is not cleaned. But repeated oiling also increases the accumulation of dirt. In many commuter type riding situations the spray of dirt will accumulate rapidly.
- A chain guard system that keeps the chain clean would quickly pay for the added weight and wind resistance.
- On a 4 percent grade, 1 pound of weight adds about 1 watt of energy at 10 mph. A well protected chain is still saving at least 9 watts of energy.
- The wind resistance of a very crude chain guard will consume only 5 watts of energy at 20 miles per hour; and at 12 mph the wind resistance will be 1 watt of energy.

- The energy loss I have measured is the lower limit of probable losses. When force is increased on the chain due to hills or wind, a 50 to 100 percent increase in lost energy is possible.
- A view of an experimental chain guard.