I received some KCNC ceramic pulleys from Fairwheel Bikes, a Tucson bike shop known for its eclectic inventory of specialty parts. Aside from being anodized in pretty colors, these pulleys promise better shifting thanks to a bit of lateral float, as well as less drag from its ceramic bearings.
In terms of float, these pulleys fall between Shimano (lots) and Sram (none). My Sram drivetrain did seem a tiny bit quieter with KCNC pulleys installed, but the difference was so tiny that it could have been my imagination.
The real interesting claimed benefit of these pulleys is reduced drag. Fairwheel's site claims that derailleur pulleys are the best place for a bearing upgrade, since they're the fastest spinning bearings on the bike. If you pedal at 100 rpm, your pulleys are spinning at 481 rpm's in the big ring, and 355 rpm's in the little ring (53/39). The faster the bearings spin, the more watts they suck up, so the logic goes. I have my doubts about this, since derailleur pulleys see far less radial load than wheel and BB bearings, and load also increases bearing drag. But since I have no way of testing this, let's just press on and see if these pulleys do decrease drag compared to stock pulleys.
There's two ways to reduce pulley drag: upgrade the bearings, and use a larger, slower spinning lower pulley (you can't put a bigger one in the upper position). A quick spin test immediately raised doubts as to the quality of KCNC's bearings – Sram's pulleys, also with ceramic bearings, had much less drag. Perhaps this is a good time to harken back to what Josh Poertner of Zipp said about ceramic bearings (cheap ceramic bearings are made to withstand heat, and don't necessarily spin faster), as well as the fact that Rich Sawiris of Wheelbuilder.com only recommends Ceramic Speed bearings (Ceramic Speed pulleys go for $140 on Fairwheel's site, KCNC, $35). KCNC's bearings may be ceramic, but they sure don't seem fast.
Next I tested the pulleys on the rollers, with a Quarq Cinqo crank and a PowerTap wheel. Since the Quarq is upstream of the pulleys and the PowerTap downstream, drivetrain drag should show up in the difference between the two power readings. I rode with a constant cadence using a metronome, and I did runs in both the big and little rings. Runs were done on Sram, Shimano, KCNC 11-11, and KCNC 11-12 pullleys.
Sadly, the differences involved were all well within the claimed margins of error of the Powertap and the Quarq. The only conclusion I might draw are that the Sram pulleys seemed 4 or 5 watts more efficient. But to reiterate, this was all within the margins of error for the two power meters.
Lastly, I tested one more theory from Fairwheel's Emiliano Jordan. He thinks that for the same gear inches, you're better off riding in the little ring than the big ring. The thinking is that when you're in the big ring, the rear derailleur is more stretched out, and that extra spring tension puts a radial load on the pulleys which increases drag. I ran the same pulleys in the big ring and small ring, pedaling at 100 rpm in the small ring and 74 rpm in the big ring so that the pulleys spun at the same speed, and got almost identical numbers.
So what's the moral of our story? First of all, things that make sense logically may not pay off in terms of real world benefits. Secondly, all ceramic bearings aren't created equal, but lesser bearings in your derailleur pulleys will probably only cost you a watt or two. Based on my somewhat inconclusive tests, I'd have to say that derailleur pulleys aren't the best bang for the buck for a bearing upgrade. But there's a silver lining. One of my Sram pulleys split, causing my chain to occasionally pop off and lodge between the pulley and the cage. I happily swapped out the composite pulleys for the KCNC aluminum pulleys.
BB30, a new bottom bracket standard invented by Cannondale in 2000, is gaining momentum.
Well, we fell on our collective faces in our attempts at an April Fool's story.
Gene Butcher is a fire fighter who lives in Michigan with his wife and two children. He is not a bike racer.