Co-op work at Apeiron Labs. The use of the part was to act as a hard stop on a structural aluminum tube inside the vehicle. Additionally, this was intended as a proof of concept; any part that needed to attach to tube could use the same mounting geometry. The hard stop had to prevent a sliding mass moving under [REDACTED] Gs; failure to do so would result in damage to the vehicle. The most difficult constraint was clearance. COTs shaft collars existed, but there was not enough radial room around the tube, the tube was thin-walled, and the load it had to hold was large. Clamping a thin wall hard without crushing it, in almost no space, points pretty directly at a collet.

Hardstop assembly with the clamping nut

A collet works here because the clamping force comes from a taper rather than from a set screw or a band. Tightening the nut drives tapered segments inward, so the load spreads around the full circumference of the tube instead of concentrating at a point. That is what lets you get a lot of holding force onto a thin wall without denting it. The other benefit is space; since all the geometry as coaxial, and could fit within each other, the limiting factor in how thin the collet could be was actually the hoop stress on the outer piece rather than any geometric limitation.

Cutaway showing the collet segments

I went through a lot of variations, on geometry and on materials, to find one that maximized the amount of clamping force. The “final” version uses aluminum for the inner part and steel for the outer one. The reason for this is that aluminum-aluminum connections have an extraordinarily high coefficient of friction compared to nearly all other materials. Using aluminum meant that I could reduce the clamping force. Galling is the usually the reason people advise against this, but since the hard stop should never actually slide against the tube, there is minimal risk of galling. The outer part is steel (4140) since it has to be as thin as possible (due to space constraints), and a stronger material could withstand the same stress with thinner geometry.

I also built a variation around a modified commercial off-the-shelf part, purely to get something clamping on a tube sooner. Starting from a COTS collet meant for sealing tubing meant I could test the clamping concept without waiting on custom parts.

Section through the COTS-based prototyping variation

Section of the variation on the tube

I ran FEA in Onshape’s simulation to check factor of safety and to predict where stress was going to concentrate, which was worth doing before committing to metal parts.

von Mises stress in the Onshape simulation

Before any of that, I printed the geometry in FDM to check that the thing was viable at all: that the segments moved the way I expected, that it assembled, and that it actually gripped a tube.

FDM printed prototype clamped on an aluminum tube

The FEA tells you what should happen, and the printed part tells you it assembles, but neither tells you how much clamping force you actually get. So I designed a fixture to measure it. It drops into an arbor press we already had, which saved building a load frame from scratch, and reads force through a 500 kg load cell with an HX711 ADC/amplifier. I calibrated it against known weights so the readings meant something.

Test fixture CAD in the arbor press

Section through the test fixture

I designed and printed the fixture and did the electronics for it: load cell, amplifier, and an Arduino reading it out on a breadboard.

Test fixture set up in the arbor press with the load cell electronics

The design ended up drawn up and ready to quote.