Failure is always an option

By Jonathan Downing

At LUNAR we solve difficult problems every single day. Nobody ever comes to us saying, “Hi, we have this really easy thing for you to do. It is totally straight forward and should be really simple.” As a result, we are constantly faced with challenges that nobody anywhere has ever surmounted. We often solve problems like these with prototyping, which in my opinion is by far the most important thing we do at LUNAR. Creating the most cost effective, fast-to-produce prototype to test out an idea is a skill that LUNAR excels at, and one which deserves a bit more recognition. This knowledge and ability saves our clients large sums of money and time whilst simultaneously increasing the quality of the final product.

The critical question we ask before starting any prototype is: what do I need to learn from this prototype? Knowing what we’re trying to prove usually dictates the degree of fidelity of our prototypes, which in turn drastically affects what prototyping method we use to create them. As long as a prototype teaches you what you asked of it, in most cases it does not matter how it is made.

Useful prototypes can often be quite simple. I once came to a large client meeting at a global medical device company with spray painted cardboard boxes, string and ½” PVC pipe. I could see some questioning faces in the room, but, yes, this is what they paid the big bucks for. 20 minutes later, as we reviewed the 3rd possible subassembly configuration for their new device by simply rearranging the boxes (hanging them from different lengths of string tied to the PVC pipe framework), it was clear that they loved the interactivity of the model and what it quickly taught them about the issues the team faced. The meeting was a huge success.

Prototypes can also be very complex. When we are trying to make design decisions based on functional performance we call such models “works like” prototypes. They can look like hell but must work like the real thing. For the Novint Falcon 3D game input device, we needed to test several factors with one prototype, so we created this hyper-adjustable beauty.

Yes, it looks like it was created by a badly trained, inebriated monkey but it helped us immensely to test out cable types, spring tensions, gearing ratios, linkage sizes, workspace spatial resolution, motor types, base stability, electronic control and communication, and much more. Testing all these separately just would not have been true to the dynamics of the actual product (that “fidelity” thing again) so we had to design all this flexibility into one prototype. This is by far my favorite prototype I have helped create since I arrived at LUNAR.

With all the learnings from this first prototype we created the next works-like model, which included all the final geometry and decisions from the first model converted into parts that could be easily manufactured and assembled, seen below on the left. One of the things this model taught us was that in order to make the “G” shaped part (G-link) easier to assemble through the opening of the new one-piece chassis, the hole would have to be larger. A couple of minutes with the Dremel in the workshop opening up the chassis hole took care of everything, giving us another valuable reminder why we never skip a chance to prototype. This discovery and others allowed us to create the “Looks-Like Works-Like” final prototype you see on the below right, which is nearly indistinguishable by the untrained eye from the final product.

Lego is a particularly interesting and versatile prototyping medium, and not just because it’s a nostalgic throwback to the days of being a kid playing with the colorful blocks while watching the A-Team. Using Lego is a great way to create simple block layouts of components inside of a device. The following images show one of the many prototypes LUNAR created to explore the layout of a laptop docking station; the left image shows a concept for the cable management at the rear of the dock and the right image shows the final design for Belkin.

One concept we try to convey to clients is that “Failure is an always an option”. (I think Adam Savage from Mythbusters has that on a T-shirt. I should get one of those.) Fortunately, a prototype that does not work can be even more valuable than one that does work, since we can usually learn more from an unsuccessful concept. If you can see and understand why it didn’t work, then you can use that new knowledge to inform your next prototype , dramatically improving its chance of success. Again, this comes back to solving difficult problems. If the problems were all easy we would not need to do as many prototypes and the prototypes would always work. But then I would love my job a whole lot less.