I had several unique design problems when I began my entry into manufacturing. My biggest problem was that I was completely new to industrial design and manufacturing; the first CAD drawing in my life was also my first serious attempt at designing a manufacturable product, mc squares. My vision for the product was to democratize the meeting room by breaking up the large whiteboard in the room and distributing smaller dry-erase boards around to the people attending. Once meeting attendees were done sketching their ideas down, they could gather the mc squares and connect them on a larger wall designed to hold the smaller dry-erase boards. Simple enough idea, right? That’s what I thought too.

Given my software design background, I wanted to make certain these little collaboration tools were useful, usable and desirable (these are terms used to commonly describe “user experience” in the digital world). That pursuit led me to purchase a quite expensive, large format resin based 3D printer. The printer allowed me to iterate on my design; testing the size, durability, feel and form of the individual mc squares. But, given the size of the product, the resin printer was not a cost-effective way of producing enough mc squares to test them in the field (or in they meeting room.) We needed to be able to produce a couple dozen squares and mounting brackets at a time to test the magnet strength, the fit between mc squares on the wall, and the functionality of pulling them off the wall without knocking their adjacent siblings off.

Workshop

Initially, we tried creating silicon molds off a resin based 3D print. Max (our resident prototype guru) spent weeks trying to get reproductions that were usable. Again, we struggled to get good working prototypes because of the size of our product, and our tight design tolerances needed for functionality.

[new_gallery ids=”448125,448124″]

That’s wen we discovered a Colorado made 3D extrusion printer, the Lulzbot. We immediately purchased 2 of them on the mere chance they would be the answer for us. The printers weren;t quite big enough to hold a full mc square frame, so we had to carve the frame up in smaller pieces and acetone glue the ABS parts together. I also had to alter the CAD drawings specifically for printing out of a “hobbyist’s” printer. Unlike the resin based printer I owned, there was no support structure material for our Taz3. I could have modified the printer to add a second extrusion head and printed support in PLA, but, in my circumstance, it wasn’t too much work to modify the parts to be suitable for 3D printing.

[new_gallery ids=”448123,448122,448121,448120″]

With the 2 extrusion based 3D printers, and some prototyping of gcode produced by Simplify3D, we were able to produce 5 mc squares a day. I was thrilled with the results. The prints were obviously not as smooth, as the resin based prints, but we produced them for a much cheaper cost ($25 per prototype vs $275 per prototype). And we were able to produce the faster (5 prototypes per day vs 2). Although the $100k + “professional grade” 3D printer produced beautiful parts, I found, given a little up front work and more post-processing, we produced fairly decent parts from a printer costing less than 2% of the resin printer.

In hindsight, I would not have purchased the expensive resin-based printer. It was far less reliable than I would have expected from a professional grade printer, instead I would have used a service bureau for the high-res parts. I expect that my views would change if I had many parts I needed to prototype, instead of a single product. Between both printer technologies I used, there were surprisingly few differences in the production flow. Both required an STL file, both were relatively the same with setting up a file to print, and parts from both machines required a significant amount of post-processing. The tactics were a bit different.

With the resin printer, we needed to “melt” the wax support system off in a specialty oven (this took 2-4 hours) and we needed to make sure the part was supported well in the oven to make sure the heat didn’t distort the part. We then had to thoroughly clean the part, using a sonic cleaner and corn-oil and then several times in a liquid soap bath to remove all oily residue. It was then that we could finish the parts with spray-paint. Quite a process.

With the FDM printer we had to engineer the part a little differently based on our decision to print without support. Basically, this meant we could not have dramatic angles in our features – we modified our angles to no more than 45 degrees of overhang and that worked perfectly. The finishing process was completely different. Instead of having to worry about removing support, we needed to make the part more presentable and remove the telltale visible layering that occurs with extrusion prints. We did this through a repetitive process of spraying the ABS part with a fine mist of acetone and then sanding it down. Each part required 3 iterations of acetone followed by sanding until they looked good enough for us to show the prototypes to users for testing. Spray-paining was not an issue because we just printed the parts in the ABS color that we wanted the part to be.

This entire process of problem solving a product that needs to go to manufacturing has been so enjoyable. Yes, launching a successful kickstarter campaign is rewarding as well, but it’s was the journey of discovering, creating, prototyping, that has been the most rewarding