One Board, One Enclosure, and a Runway

The short version: the redesigned PCB is done and tested, it's now a single board instead of two, and it needed a completely new enclosure built around 3D printing. Component costs keep going up, almost wiping out the savings from the redesign. Next stop is FCC compliance testing, which is turning out to be more interesting than I expected.

It's been since March that I posted an update. In that post I mentioned I was working on a new circuit board to bring component costs down. That redesign is finished and the boards are tested. PCBs take time to fabricate and assemble, so I used the gap to take a break and spend time with my family. I'm back at it now.

Two boards became one

The original design used two PCBs. One sat behind the display and carried the microcontroller, USB ports, and most of the components. The second carried the buttons, rotary encoder, and LED.

In hindsight this was a rookie mistake, and the electrical engineers I hired at the time (yes, it was a team) didn't push back on it either. Two boards means two fabrication runs, a connector between them, more assembly steps, and more ways for things to go wrong. It should have been one board from the start. It is now.

A new enclosure, designed for 3D printing

The single board meant redesigning the enclosure from scratch. This time I designed specifically around 3D printing, because 3D printing is the plan for production, not a stopgap on the way to injection molding. That decision changed how I approached almost every surface.

It also opens up something I couldn't do before: different colors and textures in genuinely small batches.

I kept the rough proportions of the original enclosure but started the geometry over. I went through a lot of iterations.

Here's where I landed (pun intended!). It's simpler than the original and I think it's more elegant, and it's far easier to print and assemble.

Two physical changes worth calling out. The walls are thicker, which makes the device feel better in the hand even though the whole thing is thinner than before. And I moved the battery to improve the center of gravity, so it's much harder to tip over.

The board itself

Beyond consolidating to a single PCB, I upgraded several components. There's also one addition I'm holding back for a future post.

Most importantly: there is now a runway on the back of the board. As far as I know, this is the only PCB in the world with one. Can you guess which airport this runway is at?

About cost

Components and fabrication have gotten meaningfully more expensive since I started this project, enough that the simpler design didn't produce the savings I was aiming for. Consolidating to one board absolutely helped, and it would be much worse without it.

What's next: FCC compliance

I'm confident enough in the design now that I've now started talking to compliance labs. As with every other step in this project, there's a lot to learn, and a few surprises.

PaperMETAR uses a pre-certified ESP32-S3 module, which is the standard way to avoid re-testing the radio itself. My working assumption was that this, plus sticking to Espressif's integration guidelines, would put me on the simpler path: a Supplier's Declaration of Conformity (SDoC) covering the rest of the device.

The digital side of that still holds. What I've learned is that RF exposure is a separate question with its own rules, and it hinges on how far the user is from the radio during normal use. Above 20 centimeters, a device is evaluated as mobile, which is generally a calculation. Below 20 centimeters, it's portable, and that can mean Specific Absorption Rate (SAR) testing. Portable use can also fall outside the conditions of the module's own certification, which is its own complication.

I'm still digging into where PaperMETAR actually lands. There's more nuance here than the 20 centimeter number suggests. Which part of the body matters, since different tissues and volumes have different exposure limits. There are also power-based exclusion thresholds that can take low-power devices out of SAR testing entirely, and PaperMETAR transmits very little. And "normal use" is a key phrase and consideration. PaperMETAR sits on a desk or a shelf across the room and gets glanced at. It is not a device you hold against your head.

I'll report back once I have a definitive answer from the labs. Wish me luck!