Devlog 3: The Design Is Finished, But Hardware Meets Reality

The engineering side of the project is finally finished.

The KiCad project is locked, the Gerber files have been exported, and the BOM and CPL files have been checked against the component library. I went through the board one more time and double-checked the components, footprints, resistors, capacitors, shift registers, and all 128 Hall sensors.

From a design perspective, the board is ready for manufacturing.

The Cost Problem

This was probably the biggest reality check of the project so far.

Because the PCB is larger than 20 × 20 cm, it gets pushed out of the cheaper prototype pricing tier. On top of that, the board contains 128 Hall sensors, 16 shift registers, a large number of LEDs and hundreds of small SMD components.

Once PCB manufacturing, component costs and assembly are combined, the price for producing only a few prototype boards becomes far too high for a personal project.

I spent several hours going through the BOM and trying to bring the cost down. I looked at cheaper sensor alternatives, changed some component selections and tried to avoid additional factory setup costs wherever possible.

Unfortunately, there is only so much that can be optimized when the board itself is this large and contains this many components.

What Happens Next?

At the moment, I am not sure whether the project will continue in its current form.

That does not mean the project is finished. The design files are complete, backed up and ready to return to if I decide to continue with it later.

The biggest thing I learned from this stage was that designing a PCB that works on paper is only part of the problem. Designing something that can actually be manufactured at a reasonable price is a completely different challenge.

This project gave me a much better understanding of Design for Manufacturability (DFM) and, more importantly, the real cost of turning a custom PCB design into physical hardware.

Possible Next Step

If I decide to bring the project back, the most realistic solution would probably be to redesign the board as several smaller modular PCBs instead of one large 200 × 200 mm board.

That could make manufacturing significantly cheaper and would also make individual sections easier to replace, test and improve.

For now, the design is finished and safely backed up.

The next step depends on whether I can find a practical way to manufacture the hardware without making the project unnecessarily expensive.