Having read the above report and finding it very different from Gates' claims I wrote to them to query the discrepancies.
You may be interested in their answer?
Hi David,
Ah, that old Friction Facts “study”… This thing comes up occasionally, and it is an example of you can’t trust everything you read on the internet. That test was poorly designed, and the results were not entirely correct, and stated in a way that made them more sensational. Both synchronous belts and chains are about 98-99% efficient when properly designed and installed.
Jason was using an older Carbon Drive system, with too high of tension, as he didn’t follow our recommendations correctly. We had already commissioned two other studies before he decided to do his own, and our studies provided different results. We studied his test, and talked with him, but decided not to publish an official response. Later on, we actually hired Jason and his lab equipment to do a new study to prove our other tests were correct, and to test our then already seasoned Center Track system at proper tension levels. Our new test showed the expected results – that belts can be just as efficient as chains, or even more so.
He did get some things right in his report. Belts at very low loads are not as efficient as chains due to the slightly higher installation tension. However, once the loads start ramping up, belts become more efficient than chains. This happens due to the differences between slack side and tight side tension. Chains don’t require significant installation tension, and therefore have two ‘slack’ spans at rest. Belts do require installation tension, and therefore have two ‘tight’ spans at rest. When load is applied, however, the chain gains a tight side (the top span on a bicycle drive). The bottom span on the belt, actually goes slack – transferring the tension to the tight side (top span). How slack the slack side gets depends on installation tension. Our recommended installation tension levels see the slack side approaching zero tension at maximum expected loads from cyclists. There is also frame flex happening, also decreasing the tension on the slack side of the drive. This means that all of that installation tension that was causing inefficiency is now in the top span and transmitting load. This is also the reason that we use an 11mm tooth pitch with a specialized tooth profile. Our pitch and profile were chosen to keep our belt from jumping teeth as the belt approaches zero slack side tension. Smaller and different tooth profiles are more prone to tooth jump at that moment.
Let me try to give you a better idea of what I mean by low loads, and high loads. Loading depends on wattage and rpm. For someone very casually strolling along, imagine say 50 watts at 60rpm. At these loads, you will be losing a very, very small amount of efficiency with our belt drive and recommended installation levels. However, if you imagine an enthusiast or racer who is constantly putting in 200-250 watts at 100rpm, the belt will actually be more efficient than a chain.
Keep in mind, this is with a clean and properly lubed chain. Once chains become dry or dirty, they can lose significant amounts of efficiency, much more than the 1-2% they are rated at when new.
Finally, it is important to understand his reporting of the results. Chain and synchronous belts when designed and installed properly are both about 98-99% efficient. I believe that he stated something like this in his report – I think he called belts maybe 97% efficient? Let’s do an example. For this example, let’s assume that chains are 99% efficient, and belts are 97% efficient. This means that at 200 watts, a chain would lose 2 watts, while a belt would lose 6 watts. If you take 2/6, that is a 33% difference. Jason used this to claim a belt is about 33% less efficient than a chain, even though a belt is really about 97% efficient in this example. That 33% SEEMS like a huge amount, but it’s only a difference of 4 watts, which most riders could not even feel. Keep in mind, that this is all from his original data, which used over tensioned belts, and clean and lubricated chains. In the real world with a bike that is set up properly, and ridden outside of a lab environment, things start to change pretty quickly.
Efficiency is important. It’s also a very difficult topic. When comparing a single speed belt to a single speed chain, the differences on a new, properly installed system (properly lubricated and cleaned chain, and properly tension belt), can be minimal to non-existent, depending on loading and other factors. Or the belt can even be more efficient. But that’s not the whole story, it never is. The benefits of a belt can be huge. Longer life, no lubrication, little to no maintenance, no problem with rain or snow, quieter, smoother, lighter…. For a commuter, these things are extremely important, and worth way more than the tiny amount of inefficiency that you get a very low loads. To the road racer who needs gears, the addition of a heavier (than a derailleur system), less efficient internally geared hub is not worth the fact that the belt will last longer, or doesn’t require lubrication. To the single speed cross or mountain bike racer, the performance of the system in dirt and mud is a definite plus.
