Make by Numbers

XNO LLC is committed to having a positive impact on the world. Therefore, XNO works hard to ensure it doesn’t indirectly harm the world through environmental degradation. Improving this impact doesn’t just mean using recycled packaging or getting rid of straws. For XNO, this requires calculating the company’s projected carbon footprint and working to minimize the most significant sources of carbon emissions. Such a study is called a Life Cycle Assessment, or LCA, and involves calculating the expected impact of each step in production, use, and end-of-life of a product. Since XNO is only developing and selling one product, 4Play, one LCA was able to encapsulate essentially the entire impact of the company (except for the negligible impact of prototyping and coding).

The process itself is fairly straightforward: one simply identifies each of the components of and steps needed to make each product. Then the emissions generated by each of these steps is identified. I used this handbook to make the conversions. The detailed results from the LCA are linked here, and plotted below:

Surprised? I know I was. Only one effect really matters: the manufacturing of PCBs (printed circuit boards). A few other things, like the batteries and packaging, might be worth considering, but everything else is lost in the noise. The total impact of the device was about 3000 mPts (mPts are 1/1000 of a Point, which is 1/1000 of the impact of an “average European”—about 20 Mt of CO2). That means that if XNO sold 3000 4Play devices, it would generate 9,000,000 mPts of emissions. That’s an impact of about 9 people; I’d have made my personal impact on the world 10 times worse than it is now by starting this company. This was unacceptable.

Now that I had some data to guide me, I began digging into the numbers. Why was the impact from the PCBs so much higher than all the other parts of the life cycle—orders of magnitude higher than every other material? This impact has not been well documented, reported, or considered by other engineers. However, I did find a few documents to support the initial numbers I had from the tool I was using to calculate impacts. Here are two of them:

Full analysis of PCB’s manufacturing impact

Presentation on information & communication impacts

Regardless of the exact numbers, however, the conclusion was clear: I needed to reduce the relative impact of the PCBs. Getting rid of every other source of emissions from my company wouldn’t have as large an impact as taking out half the PCBs. They were the priority. So I embarked on a long redesign process to reduce the area of PCB used in each 4Play board. The original design featured four 3" by 3" PCBs for the display, each of which had 16 LEDs to display one level of the board display, like this:

The first thing I realized was rotating the display PCBs to be thin, vertical rows instead of full, horizontal planes would reduce the PCB area to about half of what it was before. At first, this produced a strange, fingered design that looked like this:

The fingers are separated to ensure that players could still point to specific spots on the board, a key part of play that can be hard on some boards, making play more confusing (“where did you go?” “uhhhhhh, here, no no, under that, yeah that one”). I wanted to avoid this awkwardness by keeping it easy for a player to point to a given spot on the board, wherever it was. However, the fingers didn’t really allow this (since they were so close together) but spreading them out would’ve meant larger boards (a 6" by 6" base would offset all of the PCB area I’d just removed), even more plastic, and a more cumbersome design. Something had to change.

I went to friends and family for advice, and ended up deciding to change the input of the boards from a video game style design (the left-right-up-down buttons in the picture above) to an arrangement more specific to tic tac toe: a single button on top of each finger. This button would let users select from the 4 levels on a column by iterating through them. This new input mechanism made pointing out or selecting spots intuitive enough (in my eyes) that I decided users wouldn’t need to physically point to the positions they were moving (since they could just highlight the point itself with the buttons). This is what led me to the box design I have today:

I wasn't entirely sold on this design the first time I arrived at it, and have since played with rotating the device (so that the motherboard is vertical and the 16 columns are horizontal, but still separate), as well as many iterations of the case as I work with manufacturers. The current design is certainly not the final version; I still plan to make some small changes before the molds are manufactured. (That process will happen this summer, and I'll keep you updated here!) Overall, though, this design is settled enough for me to redo the same LCA calculations. Here are the current results (here in spreadsheet form):

That's a significant improvement. Overall, there is about 1/2 the PCB volume as the previous design, and many other parts have been eliminated from the design (though those changes have almost no effect on the final total). The projected impact has been reduced by about a factor of two. XNO’s total impact, if 3000 units are sold, has been reduced by about 100 Mts of CO2, equivalent to the carbon footprint of 5 average Americans or Europeans. This change is profitable for XNO, too. Because of its internal carbon tax ($50/Mt CO2 or ~$1000 per person-worth of impact), XNO is avoiding losing about $5000 in profits with this redesign, a noticeable win for such a tiny company (on top of $5,000-10,000 in saved costs from material cost reductions). This internal carbon tax incentivizes exactly this sort of work: minimizing the impact of those things that are the most harmful for the environment, not necessarily whatever looks best to consumers.

Going through the process of calculating and minimizing the expected impact of XNO was an impressively helpful process. It helped make clear exactly where changes had to be made to minimize impact, and these changes helped spawn a much better design (that's cheaper, easier for XNO to assemble, easier to transport, and easier to use than the previous one). Designs that are strictly constrained by a certain metric often end up being more creative and successful than those that meet loose requirements the first time they're designed. 4Play was no exception. As I worked through various iterations, desperately trying to reduce the use of PCBs in the design, I found myself considering new, more radical ideas than I had when I was just working on making it a fun toy for users. And that ended up making a result that was not only better for XNO and its customers, but the world as well—significantly so. Making products more environmental doesn’t mean making less of them, being less productive, or losing profits. It often means the opposite; when companies use less materials and resources to make the same product, they can make more of them more cheaply. All it requires is taking the time to analyze what factors are actually contributing to the environmental impact of a product, and being willing to throw away earlier designs that don't minimize this impact effectively. This is a process all companies can work through, and immediately make the world a better place, while increasing jobs, productivity, and profits. This is how XNO is working to make sure it has a positive impact on the world.