A (debatably) better human transportation device for my time at UCLA.

I am now in posession of a Monoprice 3D printer. I splurged for the Maker Select, which is about $100 more expensive than their cheapest printer, since I wanted the larger build volume (20 x 20 x 18 cm). Of course, I had to take advantage of that build volume. Printing Benchy and calibration cubes is fun and all, but I had to make something larger. So I decided to work on a device that would facilitate my transportation at school.

UCLA is a fairly hilly campus, and things on wheels are technically prohibited on Bruinwalk. Plus, my “commute” from the dorms to class never exceeds 15 minutes. Walking can only get you so far, though (pun intended).

I’ve practically lost my bike-riding skills, and by the time I recover said skills, my bike will probably be stolen—UCLA has a bike theft epidemic. Enter: 3D printed inline skates! In addition to learning a new skill and exercising, I’ll get a chance to explore LA from a different perspective.

Why inline skates, and not rollerskates, you ask? Since strap-on rollerskates are already available off the shelf, I decided to try to build something different. Also, inline skates (rollerblades) have a larger wheel diameter, which will aid with cracked sidewalks and UCLA’s hills.

My friend's artistic rendering of me skating (or falling?) down UCLA's stairs.
My friend’s artistic rendering of me skating (or falling?) down UCLA’s stairs.


  • Must be portable: None of the whole “two pairs of shoes” business; my backpack is heavy enough as is.
    • Fit over my current shoes
    • Collapse and be placed into backpack or hooked externally
  • Expected loads: I’m not a particularly large person, and won’t be riding aggressively.
    • 150 lb minimum (120 lb person with up to 30 lb backpack)
    • Impacts: falls, small jumps
  • Durability/maintenance: Time will tell, but I don’t want to be printing components every single day.
    • Resistant to broken pavement, sidewalks, hills: will almost never be ridden on nice skate park concrete
  • 3D print as much as possible: All components besides bearings and axles must be 3D printed. Sure, using aluminum for the chassis would probably increase durability, but I want to fully utilize the tools I have at home.
  • Stretch goals:
    • Motorized: assist with going up/down hills

Literature review


I took a poke around the internet to see what people have come up with. Patrick S came up with this Instructables, which was a great starting point. It mentioned some of the challenges he faced, and included the lessons he learned from multiple iterations.


Wheel assembly

Each wheel on the skate is mounted on the chassis using two bearings, an axle, and a spacer.

Wheels range from 70-120 mm OD, with beginner skates hovering around 80 mm[1]. Like skateboard wheels, they’re typically made from polyurethane, and have a Shore hardness from 70-85 A. It is recommended for beginners to start with slightly harder wheels (82-84 A)[1], with a profile that’s square with rounded corners.

Wheel positioning varies—a cursory Google search shows that the majority of four-wheel skates have evenly spaced wheels, with the exception of aggressive skates, which have a grinding plate in the center.

What is Shore hardness?

Previously, I’d only heard of Brinell and Rockwell, but turns out there’s another standard for measuring the indentation hardness of polymers.

There’s two flavors: Shore A, for elastomers/squishier materials, and Shore D, for polymers. Each uses different durometers that apply different forces, and defines a hardness based on sample penetration, so no direct conversion exists[2]. However, approximations do exist[3]. For example, an 85A material corresponds to 33D.

Due to the 3D printing criteria, I had to investigate the effect of infill on hardness[4]. By varying the infill, different effective hardnesses can be obtained (though material properties will be different).

While poking around for the specs on the Hatchbox PLA I purchased, I learned that Hatchbox makes TPU (thermoplastic polyurethane) filament, with Shore hardness 95A.

Wheel hubs/cores are slightly compliant, act as shock absorber. Since they’re different materials, I was thinking about possibly casting my wheels: printing the core, printing a mold, and pouring in some polyurethane (though single-nozzle FDM has to fail before I consider it).

A typical skate bearing is 608ZZ, ID 8 mm, OD 22 mm, width 7 mm, shielded (higher end bearings have removable shields). The higher the ABEC (Annular Bearing Engineering Committee) rating, the more precise [6]. For example, typical bearings used for hardware applications are ABEC-1, with a maximum deviation of 10 microns from nominal.


Due to portability and weight concerns, I was thinking about using ratcheting loops, a la snowboard bindings, with at least an ankle strap, a toe strap, and a hinged highback: use zipties (maybe printable?) for now, and then try printing the ratchet mechanism using more flexible filaments.

Additionally, since rollerblading involves a bit of forward pressure of the foot, I’ll have to design a toe restraint.

Possible challenges

FEA of 3D printed parts

3D printed parts are anisotropic (different material properties for different directions). It’s totally possible for the layers to separate before the material actually yields. Perhaps it could be approximated by a weak composite (look into ANSYS student)?

Interlocking components: need to print a “Swiss cheese” calibration part to evaluate 3D printed hole precision

sin/cos shelf

Random thoughts/distractions I had while writing this blog post:

  • Different loading directions on an anisotropic part yield issues.
    • We have 5+ axis CNCs, how about 5 axis 3D printers?
    • Can potentially optimize print direction and/or finish
    • Known: NASA has fancy 10+ axis composite printers
    • Øyvind Kallevik Grutle did his Master’s thesis on it!
    • Speaking of pretty prints, how about nice cylinders/spheres? The MakerBot at school printed polygonal “circles”, though that might have been an .STL meshing issue
      • Never mind, reddit says they’re impractical
  • Finishing PLA: can’t take an acetone bath like ABS, so some combination of sanding and filler needs to happen.
  • Use of the first person: feels weird and excessive, and takes too long. In order to expedite blog posts, I’ll try uploading briefly edited notes and adding a TL;DR.

To be honest, I’ve pretty much been spoiled by LaTeX’s automatic everything (numbering, bibtex, etc). If any of these links die, I’ll have to go back, delete, and renumber them. Maybe I should look into whether Markdown supports LaTeX-style bibliographies…