Recurse Center Week 8
Package of parts arrived Monday morning, making a long week dedicated to PCB manufacturing. I spent all my time holed up at the apartment, recreating past attempts at a homebrewed process.
I got many steps of the etching process working, but did not make a functional etched PCB by the end of the week as I had planned. I made a last-ditch attempt at masking, the only step totally failing, with a very manual cut-out process. It almost worked. Ordered more parts to hopefully close the gap.
2024/09/30
- Electronics order arrived! I have the correct wire and tool to try wire-wrapping.
- Practiced wire wrapping. Repeated over and over enough to get reliable results, but not fast results.
- Spent a few hours re-learninig KiCAD and made good progress on an ALU layout.
2024/10/01
- Wanted to get the etching process started today.
- Picked up PPE and small drill bits from the hardware store.
- Found a dremel plus kit on Craigslist for pickup, and spent the reset of the day in and out of NJ.
- Read a category theory textbook in commute, but am humbled by the distance to anything I'm familiar with. Going to put this down for a while.
2024/10/02
- Ran a few failed etching tests: it's obvious in hindsight the plan for masking with an inkjet printer would not work.
- Researched dry and wet film photoresist. Found a well-shot video showing the process. Realized I should learn what accuracy is really needed before evaluating different processes beyond the bare minimum.
- Get distracted by photography dark rooms and laminators.
- Feeling conflicted about how isolating my work has become, seeing more days of apartment-bound work ahead.
2024/10/03
- Went back to schematic design.
- Compared ALU schematic to Verilog ALU. Stuck on how to make a modular design without knowing how modules should fit together.
- Attempted a last-ditch mask transfer process with acetone. Still a very poor transfer, barely a shadow of the ink is visible.
- Order pre-sensitized copper clads and dry film photoresist. Will take a few days to arrive, so fully back to simulation and schematic design for the near future.
2024/10/04
- One last idea for masking. Cut a mask by hand from a sticker?
- Don't have a sticker but try layering electrical tape. Free-handed some lines and shapes, no attempt at fine details.
- Etched, and the mask almost worked. Results were mediocre and would not be functional, but larger features worked reliably.
- Revisited KiCAD layout, and for the only part I've tried, the ALU, this accuracy is enough.
- Making a habit of ending Fridays with these summaries.