4. Electronics production
In the beginning, it was difficult to find some components where we are, specially FR-1 phenolic paper based boards for PCB production. We had to resort to a double sided fiberglass based board. Although it worked well, it took two 1/64″ flat end mill bits to accomplish the work of making the five boards for our FabLab group. First I had to learn several new things to me, among them, the basics of soldering. The failed boards were used for soldering practice.
Other items we had to change were, resistors, capacitors, and crystal. At the beginning of the course it was difficult to get SMD type components on time, so we had to use conventional ones. Neither we had at that time the Atmega microprocessors (we used PIC processors.). Because of this limitation we had to redraw the PCB circuit pattern in a combination of Rhino and Photoshop in order to accomodate the bigger components and save it as .PNG.
Milling of the PCBs in fiberglass based board. This product had a slight curvature. Because of its rigidity, it was difficult to keep it perfectly flat to the sacrifice material, which caused stress and broken end mills. We divided the milling process in four stages; hole drilling, first patern cut, a second rectification pass and external frame cutting to separate the finished PCB from the board material sheet.
After we got the SMD components, I was able to catch up with this assignment. At first we had a low spindle travel speed in the Roland mill which resulted in long waiting times to get our boards made, so I resorted to make the boards by painting the boards which were later removed by a laser, and chemically etched (acid.) The one on the left had the Zener diodes welded on the wrong direction, on the one on the right that was corrected, but I was afraid I damaged the very tiny vias… The final result looks like something out of Mad Max….
Loading the AVRdude with the USB ASP. Looking for processor compatibility:
Later I tested it on the AVR-ISP but an error message appeared. This second ISP iteration didn´t work, probably because of bad welds, and / or damage to the vías. A third iteration was made afterwards in the Roland mill.
I´ve tried to program the Make clean, Make hex and Make fuse on my PC laptop, using an AVRISP,but was very difficult, so I had to ask a friend borrow his Mac and repeat the process of loading the codes on the ISP.