|One of the first things to break were the pedals. I printed some new ones, but without bearings. They were PP, so the friction wasn't noticible.|
|I tested a pair out and found that the tread nubs didn't really grip my shoes, so I measured my tread pattern and made the nubs farther apart. Then it gripped just right. I rode these pedals for over a year before they broke off. The spindle only reached about half of their length, so the pedals themselves were taking a lot of bending load.|
I was also going through rear cassettes quite quickly, because it is always wet and sandy here in the winter. I tried printing one out of PP, because it's so tough that it could work, but it just shrinks too much.
This cassette model is fully parametric.
|I tested a small sprocket with a torque wrench and got a decent 25 Nm before the 1/2 drive yielded out. The teeth could have taken more. This was encouraging.|
|Here it is printed on a 3 mm layer of PP. It still wasn't stiff enough.|
|I switched to PETG, which is much easier to print and is also pretty strong. But it's also brittle when cold and especially against shock loading. I did end up testing it on my bike and it just ate the teeth right off. The sprinkling sound was both satisfying and dissapointing, although not too unexpected.|
|More recently I had a deraileur hanger break off. I thought maybe I could make a plastic one. Or eleven.|
|Nope. Not stiff enough to replace the function in the available design volume, even with extremely creative topology (using Autodesk Inventor's Shape Generator).|
|My rear deraileur idler gears also wore out. I replaced them with these.|
|The two wheels were slightly different. I tried using two of the same gears at frist but that didn't work. The upper one needs a kind of skirt to prevent it from derailing.|
|I'm still riding on these (PPGF30). Success.|
|Brake pads are of course another thing that are always wearing out. I have some TPU filament and decided to try printing brake pads to fit in my replaceable pad holders.|
|These last just as long as a "real" set of pads. (You can also print them as thick as you like...)|
I also analyzed the pads using ANSYS' coupled thermomechanical elements. Frictive heating is automatically calculated. Tricky points were staying under the 32k limit, initial rotational velocity and element type selection.
The pads have a pretty low thermal conductivity and so have really high thermal gradients. With 2nd order elements this leads to temperatures lower than the initial temperature, which is nonphysical. I tried switching to the coupled-field first-order bricks (SOLID5) but got a SEGV and gave up. I haven't modelled wear/melting, but apparently that's also possible.
|The rim results were more believable.|