ellen's disorganized thoughts

Every time I send out a new prototype for manufacturing, I spend a lot of time thinking about the possibility of it not working. The reasons for this are twofold: time and energy. It's not a good use of my time to wait for another board revision if the first doesn't work, but the energy I'm concerned with is not my own. Semiconductor manufacturing, and to a lesser extent, PCB manufacturing, is an environmentally destructive process. I refrain from sending off broken designs to prevent the board house from burning coal to create a useless PCB and prevent myself from wasting components with far higher embodied energy than their small size may imply.

I've decided to employ more modularity in my workflow to help address this. My usual approach for designing a prototype has been to create a bespoke PCB design and send it to be manufactured and assembled. If that design doesn't work, I may waste ten microcontrollers on PCBA for that iteration. The approach I'm using now is to have a stock of generic modules on hand for assembling prototypes. Instead of paying JLCPCB to put a microcontroller on each new board, I may add a microcontroller module in a compact LGA footprint. I’m also working on an S2-LP transceiver module and an Si2141-based RF frontend module. Most of my new boards still use some surface-mount components, but the ones that are most expensive and carbon-intensive to manufacture are attached as LGA modules. By using modules, I can assemble a single board, test it, and decide whether to assemble more. LGA modules are also easy to desolder with nothing more than a hot plate, and they're usable for a wide variety of designs. I can even cannibalize one prototype to assemble a new one by moving a module from one board to another. What I give up in miniaturization I gain back many times over in flexibility.

I'd like to begin selling these modules and development boards based on them, and the ability to quickly assemble a new board from a shared stock of modules if I run low on inventory will enable me to be more efficient in how I stock products, which is helpful for a low-volume business. All of these designs are also open-source, allowing users to modify them and create derivatives. By making it easier for an end-user to repair or repurpose a product, I can directly impact the amount of e-waste I generate.

I'd love to see the rest of the industry move in this direction. It empowers makers and feels like the right thing to do.