We recently released a brand new enclosure for our Phidget Spatial boards, made of a different plastic and manufactured using a different process than the previous enclosures. Join me as I walk through the journey of making a customized accessory from start to finish.
The First Prototype: Home 3D Printing
Our main goals for the new version of the enclosure are to use a stronger, less brittle plastic and to make something easier to put together. We had one of our engineers make the model in CAD software, then got the green light to move to prototyping.
We entered the prototyping phase with naive optimism about the prospect of 3D printing’s ability to make our designs. We purchased a (relatively) inexpensive extrusion 3D printer from 3D Systems called “Cube” (you can read about our frustrations in our earlier post about 3D printing), but a home 3D printer was not reliable enough to print a full Phidgets enclosure prototype, and an industrial 3D printer was not worth the investment. I wish I could show you our failed prototypes, but they were destroyed long ago. Just imagine a lime-green plastic box with walls that are warped and buckling, resembling the shape of a soggy piece of deep-fried tofu. When we did manage to print an intact enclosure, the tabs used to snap the two halves together broke almost immediately. It was clear that we needed a higher quality printer.
The Second Prototype: Professional CNC Milling
As much as we like to have our own toys for testing, building, and prototyping, we couldn’t justify the cost of a professional 3D printer. So we called up our contact for sourcing in China and asked to get our model 3D printed professionally. When we received the prototype in the mail, it was functional and acceptable, although there was an unexpected difference.
Based on the tracks on the inside of the enclosure, it was clear that it was made using a CNC milling machine, not a 3D printer. This sort of mixup is not a terribly rare experience when getting parts from China. They knew we wanted a plastic prototype for an enclosure, so they ignored our request to have it 3D printed, as they knew they’d achieve the same result for cheaper with CNC milling.
Incidentally, we do have access to a CNC milling machine belonging to a friend of the company, but we didn’t think it was up to the task of making this prototype.
Production Phase: Injection Molding
Since we were satisfied with the prototype, it was time to find a factory in China that would make the mold. We came in contact with a good factory, and began sending design files. We worked out the cost details, but they were unhappy with the enclosure design. There were parts in the snapping mechanism that were too small to be made reliably using a mold.
They suggested a different style of mechanism that would be acceptable for the injection molding process. We were hesitant to take their word for it at this point. Even though they were the experts, a single mold costs several thousand dollars. We had to be sure this new design would work.
Back to the Drawing Board: SLA Printed Prototype
We had to go back to the prototyping phase with this new design, but decided against using the previous method for a number of reasons. The main reason was that the time it takes to communicate with factories in China is inflated due to the time difference. If the factory has a question about your requirements, they’ll send the email when North America is asleep, so you won’t reply until the next day. If there are several details to clarify, communication that would normally take a day could take a week. We had already spent more time than originally expected on this project, so we decided to find someone in our timezone. The other concern was that this new enclosure design had some areas that would probably require smaller drill bits, which would likely increase the cost of fabrication if we stuck with CNC milling.
Soon enough, we found a Canadian 3D printing company that would make our prototype with a process known as stereolithography, or SLA for short. This method of 3D printing involves shining a UV laser onto a thin layer of liquid plastic resin. The resin reacts to the laser and hardens with exposure. The object being printed is moved just below the surface of the liquid, and the process is repeated, layer by layer. Since you can be much more accurate with a laser than you can with an extruded tube of hot plastic, this prototype was the most impressive one yet. We were happy with the design of the tabs, so it was time to move to production.
Making the Mold and the First Production Run
At long last, we received the samples from the first production run. We’re very happy with this new design. While we don’t have immediate plans to purchase molds for other Phidgets at the moment (the cost of a mold is still hard to justify for most of them), we’re glad we went through the learning process of making an enclosure from start to finish.