Testing
So you have a device! We know what it can do—but how long can it do it for? And under what conditions? Once we began making functional prototypes, new questions emerged about the longevity of the product. Yes, one can calculate, for example, how long a battery should last according to its datasheet, or look up the specifications on individual materials and parts, or use an oscilloscope to determine exactly how much charge is used per activation. But for full assemblies it’s impossible to make confidant claims without testing a device to its limits.
Longevity
Pushing our buttons
Our second test-rig button-pusher, “Carl 2.0” was a slapped-together four-bar linkage powered by a cam. Simple, but reliable and efficient!
To respond to some initial questions about the longevity of our devices, we built our first button pusher: “Carl” (video below). The original rig, made from laser-cut econowood, tested a single device nearly 30,000 times every 24 hours, and could run on an external power supply indefinitely. An Arduino with a bluetooth shield would timestamp each press for future debugging. For style points, we water-jetted a second pusher (above) out of aluminum, alleviating concerns that laser-cut econowood would fatigue or wear over time.
Rotation Iteration
Over time, the design of our device changed to incorporate a rotating bezel: a much more difficult feature to stress-test than mere buttons. Carl required an upgrade, and new, more ambitious goals. For this challenge, the test rig should be able to:
1) Rotate a bezel of modifiable proportions 6 times, and then return the dial to its starting position;
2) Press the button once per position, and log all presses.
After some simple sketches in Fusion 360, and research into the mathematics of Geneva drive geartrains, an idea began to take shape, and we got to laser cutting:
Once the theory behind the device had been tested and confirmed, wooden gears were replaced with delrin versions to reduce friction and improve longevity; the original button-presser was cannibalized and included; a DC motor was mounted to drive the system; and a “modifiable gear adapter” (clear) was added to the Geneva drive. This adapter would change often to accommodate differently sized devices, but at this point in the device history was used to test only the device’s hidden internal electronics.
The final step was to add a counter to make tracking the number of press and rotation cycles easy. To our great satisfaction, the final version of Carl V2 demonstrated that our devices, on a single coin cell charge, could function for over 35,000 rotations and presses. (Video available at top of this page).
Ruggedness
keeping cool under pressure
Longevity means nothing if a device breaks in the field. This has meant addressing all kinds of environmental factors. Premiere among them: water ingress.
The IPX7 rating for waterproofness requires a device be able to resume operations following thirty minutes of submersion in water at a depth of one meter. And so, informally…
IP rating must be verified by certified testing facilities, but it’s good to at least get an idea of a device’s capabilities with the occasional depth test. And yes, rest assured: it survived!