The resin gear that was beautiful and useless

Choosing a process for surface finish and losing the property that mattered.

A resin gear can be genuinely beautiful, its tooth profile crisp and its surface glassy smooth straight off the printer, and genuinely useless at the same time, because a gear tooth's real working life depends on surviving thousands of small repeated impacts as each tooth engages the next, a demand that has almost nothing to do with how fine or how smooth its surface happens to look.

Introduction and overview

A gear was chosen to be printed in resin specifically because resin holds a sharper, more accurate tooth profile than filament ever could, and the finished part looked exactly as good as that reputation promised, every tooth crisp, every flank smooth, nothing about it suggesting a problem. What that appearance never revealed, because appearance has no way of revealing it, was how the material would actually behave once that same tooth profile was put to work meshing against another gear thousands of times a minute, a use case that tests toughness under repeated impact rather than dimensional accuracy under a single careful measurement.

The crystal-glass comparison

A crystal wine glass, thin-walled and precisely finished, looks and feels like the superior object next to a thick plastic picnic cup, and in almost every respect it is, clearer, lighter, more pleasing to hold. Knock the crystal glass against another glass repeatedly, the way it might genuinely be knocked during an evening of real use, and it can chip or crack at the rim long before the humble plastic cup shows any damage at all from the same treatment, because the plastic cup's real advantage was never visible in how either object looked on a shelf, it was how each one tolerated a small sharp impact repeated many times over. A resin gear tooth meets exactly that same test on every single revolution, engaging and disengaging the tooth beside it with a small impact each time, and a material chosen purely for how fine an edge it can hold has said nothing yet about how it survives that impact repeated thousands of times.

Why toughness and surface finish pull in different directions

Resin's fine detail and smooth finish come from curing a liquid into a tightly cross-linked chemical network, and that same tight network, which is exactly what gives resin its dimensional sharpness, is also what tends to make it more brittle, less able to flex slightly and absorb a sudden load the way a tougher plastic can. A gear tooth's root, the narrow curved section where the tooth meets the body of the gear, concentrates stress on every single engagement regardless of material, and a brittle material meeting a concentrated, repeated stress at exactly that point is a close match for how fatigue cracks actually start and grow, accumulating damage a single static measurement of the tooth's strength would never reveal. Filament printing, whatever it loses in tooth sharpness, keeps considerably more of the toughness a gear tooth's real working life actually depends on, which is precisely the property the resin gear's beautiful finish never had any way of demonstrating before it was put under load.

One figure worth keeping in mind

A resin gear under repeated meshing load can begin showing cracked or chipped teeth after a small fraction of the cycles a comparable filament-printed gear survives without visible damage, a gap that has nothing to do with how accurately either gear was originally printed and everything to do with how each material tolerates the same repeated small impact at the tooth root. The resin gear's dimensional accuracy, in other words, bought exactly nothing toward the property that actually determined how long it lasted in service, since a tooth that meshes ten thousand times before failing and a tooth that fails on its ten thousandth mesh both started out looking, on the bench, like the identical, equally impressive part.

Where the same trap catches other parts

A gear tooth is a particularly clean example of this trap, but it is far from the only place it shows up, since anything printed for how well it holds a fine edge or a precise fit rather than for how it survives being loaded repeatedly is vulnerable to the same mismatch. A snap-fit clip printed in resin for the crispness of its catch feature can fail the same way, cracking after a modest number of insertions where an equivalent filament clip would flex and recover, because a snap fit's real working life is also a repeated small impact and flex cycle rather than a single careful dimensional check. The lesson generalises past gears specifically to any feature whose working life is measured in cycles rather than in a single moment of use, which is a genuinely different design requirement from the one a smooth, accurate surface finish was ever built to satisfy.

What this changes in practice

Choosing a printing process for a load-bearing feature means asking what property that feature actually needs to survive its working life, repeated impact tolerance for a gear tooth, not simply which process produces the sharpest, most accurate version of its shape on a bench inspection. A part's appearance under a careful first look and a part's behaviour under the specific repeated demand it will actually face are two different questions, and a process chosen to answer the first question well has answered nothing at all about the second unless the two happen to align, which for a resin gear's tooth profile they simply do not. The practical habit worth carrying forward is separating, for every printed part, the question of what it needs to look like from the question of what it needs to survive, and choosing the process against whichever of those two questions actually decides whether the part still works after real use rather than whichever one is easier to judge by eye on the day it comes off the machine.

My key error with this

I moved that part to the jetted process because I wanted the surface finish and the feature definition, both of which it delivered, and the result was genuinely the nicest looking component I had made that month. The part existed to hold five or six sensors in a fixed relationship to one another, which means the only property that actually mattered was that it stayed the same shape in every direction under handling and temperature, and the process I had chosen for its appearance does not give a part uniform behaviour through its thickness. The relationships between the sensors drifted in a way that was consistent, repeatable and completely useless, and no amount of recalibration fixed something that was still slowly moving. What replaced the belief is that surface finish is a requirement like any other and deserves to lose to a more important one, and that for a part whose entire job is holding a geometric relationship, the question to ask first is which process gives the most uniform material rather than which gives the best face.