Oversized derailleur pulleys are like Hansel - So hot right now. Sure the benifits of an oversized pulley are probably pretty marginal (maybe a couple of watts... according to the folks trying to sell you them), but putting an oversized pulley jockey on your bike just gives it that certain je ne quoi that says "I take my bike parts seriously"
Unfortunately, the entry price on oversized pulleys from companies like Kogel and CeramicSpeed start around $150 and goes up astronomically form there (See this $700 system from AbsoluteBLACK which easily costs more than some of the bikes in my garage are worth). So, off to my $15 DIY version.
A few notes on safety
I've 3D printed some bike parts before, but before I use a 3D printed part on my bike I usually ask myself these questions:
* What kind of loading will the part be under when I ride?
* If this fails, what will be the consequences?
For the pulley jockey wheels, they're operating on the side of the chain with minimal tension, and only undergo lateral loading during shifting. I've had pulley wheels fail during rides before. While suboptimal you can limp out of a ride with a failed pulley wheel. All the testing here was done on my bike trainer, so a failure would mean a 30 second pause while I swap out the pulley wheel.
V1 - 14t pulley
I started with a quick sketch of a 14t pulley that had a narrow-wide tooth pattern. similar to the 12t stock part. The hardest part was sourcing a bearing that would fit using the existing hardware on my SRAM Rival cage. After some searching, I chose to design around this 6x22x7 stainless steel bearing w/ rubber seals and was ready to print. I started on the Markforged 3D printer at the ITLL because I've been pretty stoked on the strength of parts off the Markforged in the past... but for this part I was not pleased w/ how the surface finish turned out. I reprinted the pulley wheel on the Form Labs SLA printer in the Idea Forge. The surface finish on this machine blew me away, and without any post processing I press-fit the bearing and thru the part on my bike. (see photo below)
After around three hours on the trainer riding at about 185 watts, the only thing I noticed was a little more noise, which made me think the tooth engagement of the new part wasn't on point. Additionally, I noticed I still had plenty of clearance in the cage, so I decided to move on to a larger design.
V2 - 15t pulley
Given the noise, I decided to re-shape the teeth on the pulley wheel. The narrow-wide tooth layout seems to make a lot of sense on the large chainring of a 1x bike, but seemed like unneeded complexity on the pulley wheel. The narrow-wide profile is pretty ubiquitous on pulley wheels from SRAM, CeramicSpeed, Kogel, and absoluteBLACK (but interestingly: not Shimano), so I'd love to know the benefits of this. Alas, I didn't have clearance to go for a 16t pulley so I scrapped that design and redrew the part with a slightly modified tooth design.
I also removed some "spokes" from the design, which in retrospect I might regret as it will decrease the stiffness of the part and I am concerned about how the shifting will compare with this larger and less-stiff version. We'll see... I re-printed on the Form Labs SLA printer and immediately noticed the bike was quieter but haven't had a chance to ride it for any meaningful amount of time. Stay Tuned
Conclusions
From the Table below, I think the DIY version is probably just as good as the stock, and maybe better. I'll be curious to get some more hours on the bike and see how the part wears. When the weather is warmer i'm also curious to see how the part holds up to riding in an outdoor environment.
| Stock | DIY Version | Aftermarket | |
| Cost | $45 (for two) | ~$15 (for one) | $150++ (for two) |
| Bearing Type | Sealed SS | Sealed SS | Sealed Ceramic |
| Teeth | 12 | 15 | 14 |
| Power Savings | 0 | ?? | ~1% (?) |
As stiffness is important, I may further iterate the part to make it a bit stiffer if shifting seems to be suffering, and I'm on the lookout for a better bearing to try building around. I also might go back to the 14t version if I become convinced there's any advantage to the narrow-wide tooth design.