Chapter 10. Refine your design and iterate

…when we design something that can be used by those with disabilities, we often make it better for everyone.

- Donald Norman

About four months passed while I fed cartridge after cartridge of plastic into the 3D printer. I ended up with suitcases full of prototypes of different designs I tried, and I am happy I labeled every one of them and filed them away for posterity. Strangely I kept coming back to a close variation of my original design: the “half-hexagon”, as I got into the habit of calling it. With the limited design knowledge I had started out with, I came close to the final design with the first iterations. Did I just get fixated on my original designs, or did I hit s bull’s-eye early? I do not know.

I looked at this $5,000 of 3D printer materials as another investment in my education. Trying out hundreds of different shapes and sizes in such a quick turnaround fashion helped me learn a lot about CAD, 3D printers and how probable my design might be.

Wasting time on silicone jelly molds

After a couple of months—the general product design was beginning to settle—the changes I was making were becoming increasingly subtle. The next challenge was to do some testing. For that, I needed enough parts to complete an entire installation of my product in a pond.

From a local store, I bought plastics supplies that enabled me to build a silicone mold from which I then was able to make many similar parts quickly. How this works is, you mix two liquids and pour them into a container, covering your original piece of plastic. Mixing the liquids causes them to set to a firm, jelly like substance within about a half hour.

With the jelly firmed up, I carefully cut the original piece of plastic out of it and what remained was a jelly mold from which I hoped to produce many pieces quickly. At least, quicker and cheaper, than it would be if I tried to make many pieces with the 3D printer.

I tried making pieces out of different materials. Some were transparent; others looked a little more like plastic injected parts. Overall, though, the whole jelly mold method of making many pieces was a waste of time and money.

Between the significant loss of accuracy, the difficulty in extracting usable parts from the jelly mold and the fragile copies that emerged, I abandoned this method of producing many test parts after a month of experimentation. It was one of only two times in this journey that I got so frustrated, I thought my journey would end right there. I remember one day in particular. Hunkered down on my garage floor with some of the resulting pieces in my hands, yet another duplicate piece fell apart in my hands under pressure, I got that hot feeling of impending disaster at the back of my neck. From the mostly useless plastic pieces I extrapolated, falsely I might add, the eventual plastic injected pieces would also be poor.

Technical problems fight hardest right before they are about to surrender, so I did not give up—and thank goodness, too—my short-lived negative conclusions on that day would turn out to be flat wrong.

Figure 5 – a more complex, “improved” design

Figure 6 – the half-hexagons attached to the leg assembly

Figure 7 – metal pins would hold the structure together

Back to the 3D printer

After a lot of experimentation, I came up with a design that I thought would allow me to construct a strong honeycomb from individual pieces. This new design is illustrated in Figure 5 on page 49. I liked this new design for a couple of reasons. In theory at least, it had a lot of “pokeoke” — Japanese for making a product foolproof which I will go into in more detail later — forcing it to be assembled the correct way only. With this new design, it wouldn’t matter whether the customer installed the piece upside-down or not, because it was symmetrical along a horizontal plane. Secondly, the “plugs” at the end of the piece would fit into another half-hexagon one way only: the right way, and only into the right holes. The idea was, when the customer ultimately assembled the pieces, they would insert metal pins (see Figure 7 on page 62) into the holes like the bar you see in a door hinge on which the two metal flaps hang. The design promised to be a lot stronger than the original design, so I was vigorously patting myself on the back until I learned about a serious design flaw: It was a plastic injection molding nightmare.

As designed, the mold for that part alone would have to have many moving parts that would slide out of the way after the injected plastic cooled, including at least seven thin metal rods that would serve to sit where the holes would be.

The design had far too many complexities to be cost effective, which I was about to discover.

Making your product only work one way – the right way

Recap: The Japanese have a term for it: Pokeoke. It is a way of designing something so that it can only be used the right way. In my first iteration of my product, user testing uncovered two significant but not critical ways of assembling the product incorrectly. This presents to the customer the opportunity to install the product incorrectly, thereby increasing the possibility of dissatisfaction.

Take the time to design into your product features that make it possible to assemble, install and use the right way only. At the very least, make it obvious that the product is not being used properly if and when a customer assembles, installs or uses it incorrectly.

For example: imagine you have designed a plastic, belt-attachable mobile phone carrier. It attaches to the belt that holds your customer’s pants up. After the first few hundred units came off the production line, you discovered that 20% of your customers mistakenly install the unit upside-down, thereby increasing the likelihood that a customer’s mobile phone will accidentally fall out of the carrier. To solve the problem in the second iteration of your product, you add a swivel to the top part of the unit such that gravity will pull the bottom of the unit towards the ground. A completely fictional example of course, but one that illustrates how a product can be designed to limit, correct or prevent incorrect use of a product.

Take a look at a book titled Universal Principles of Design by William Lidwell. It is expensive, but it is worth it. It explains 100 different such design issues that will help you make a better product. Unfortunately, I only got hold of the book after I had committed to a particular design for the first run of my product. I recommend you read the book from cover to cover before you commit your own designs to manufacturing.

One little piggy goes to the market

Fortuitously timed, there was a two-day molding and injection equipment and services conference close to my home in Seattle. I loaded up my favorite prototypes and headed down to the conference. I showed my wares to about ten or twelve locally based plastic injection vendors, and three of those vendors in particular held my attention. One had good international contacts and a proven record of accomplishment of quality and cost-effective outsourcing to China; they also seemed small enough to focus on a tiny, one-person start-up like mine. The second company seemed less equipped to work with small companies, and although they clearly had the expertise, it was often hard to get their attention. The third company, Cascade Plastics in Fife, Washington, had the expertise I was looking for and from the type of questions they asked, they appeared to have had experience dealing with entrepreneurs. I was looking for a company that could essentially “catch the ball” when, due to my own ignorance of molds and plastic injection, I would make mistakes. Probably the most compelling aspect of Cascade Plastics was their guarantee contractually that the whole product would snap together in the way I needed it to. Two of their sales engineers went over my designs and gave me some very good advice on where I could reduce the complexity and cost of any subsequent mold making. They then took what I had produced and made dozens of further improvements to the CAD models I had created before beginning the mold creation process itself.

One key characteristic of any mold is whether it has undercuts or not. I go into this in detail in the next chapter, but worth touching on here because it has a huge impact on every aspect of production. Imagine you are in your kitchen making a jelly mold. You pour the warm liquid into your mold, place it in the fridge and some time later, you remove it from the fridge and extract the wobbly jelly from the mold. You will notice that such jelly molds are made so that the jelly pops out of the mold easily. That is, there are no features or extrusions within the mold itself that might block the jelly from sliding out, whether it’s one of those old-fashioned jelly molds (Figure 8) or it’s one of Scooby-Doo.

Figure 8 – a typical jelly mold

In every relevant sense, basic molds made for plastic injection molding must obey the same rules as jelly molds. You can add such undercuts to a plastic injection mold but it will make a significant difference in the complexity and price of the mold; “sliders” of many types will be added to the mold that move out of the way once the plastic hardens, allowing the piece of plastic to pop out without obstruction.

• Molds without undercuts are much simpler and cheaper to make than molds with undercuts.

Cost of this stage: $5,000. Costs so far: $30,800