A lot has happened in the last month. I've been working nonstop on getting everything finished up and ready for the desert -- which is in a few short weeks (!!!) -- and of course we're running into some last minute problems.
There were a few problems with the last round of PCBs, so we updated the design and ordered a new batch. One of the biggest improvement is the two data connectors. Originally we were using connectors that we'd have to manually crimp ends for (240 wires to crimp, to be exact!). We swapped those out with ethernet plugs so we can use ready-made cords. Also, since we only need 3 wires for the data bus, we can use the other 5 wires for power!
I also ordered solder stencils with this round of PCBs and had an assembly line of board building at the last build party.
Capacitive Sensor Hell
The topic has been my personal hell for the past few months. I was able to get the capacitive library working well on its own, and of course, it all blew up as soon as I threw it in the matrix of floor nodes. Now it's returning numbers that are completely unstable and useless.
I have a few theories:
- Interference from neighboring floor node sensors.
- Communicating on the bus at a constant 250kbps is not giving the capacitance library enough time for sensing and processing.
- The new power supply might not be grounded well enough.
We finally picked up the plastic and some high strength silicone glue and started putting it all together. It feels great to finally have a full floor section to load electronics into.
New Bus ProtocolI won't get too much into the details now, but after quite a bit of discussion & debug, we decided to rewrite the bus protocol to make it more efficient. Now we can get the status of all nodes with a single byte per node (previously it took 5+ bytes). More details to come.
Next is to do the final assembly and throw a dance party before Burning Man!