Week 5 - Electronic Design

Chintan Vaishnav


To design a circuit board with micro-controller that operates an LED with the help of a button.

Schematic Design in Eagle

The schematic design was straight forward with the help of the tutorial in fab academy.

Schematic (Front)

Board Design in Eagle

The board design was challenging and tedious. It was very hard to overcome crossover points even after trying all the tricks. Ultimately, a combination of rearranging the components and autorouting traces with routing grid set to 10 mil worked well. Auto routing got about 75% of traces routed. The final three or four that needed routing were super tricky as well.

I tried the following tricks:

  • Changed the Grid->Alt to "0.0125" for finer control over movements of components and traces.
  • If I hit a wall in overcoming crossover points, ripped a large number of traces, moved components around, and auto routed traces.
  • Selectively rotated components.
  • Found the closest VCC and GND pins to minimize routes.
  • Upon running DCR, Eagle complained about several errors (18 in total) about clearnances being small. I adjusted tracks using ALT+Move, and got it down to 11. At this point, the clearances looked reasonably good, and I decided to go ahead with milling the board. In all, editing the board took several hours of iterations before I had a board that was satisfactory.

    The following two Eagle tutorials were very helpful:
    Eagle Tutorial 1 by Jeremy Blum
    Eagle Tutorial 2 by Jeremy Blum

    Board in Eagle
    Board in Eagle

    Exporting and Milling the Board

    Exporting the board (traces and outer edges) using Eagle and GIMP was easy. An important step here was to invert the images (using "image->invert" on GIMP) so that copper to be removed from the board is in white.

    I had to mill the board on Modela twice. First time around, the 1/64 drill bit must have gotten pushed up as I tightened the screws. Consequently, the traces came out light. Next time, I adjusted the drill bit, but also deepened the z-axis of the path to -0.15 (instead of -0.1). The board came out perfectly.

    Stuffing the Board

    I faced two challenges whil stuffing the board:

  • It was really hard to solder the 20 MHz resonator. It has three pins at the very bottom, and required heating up three pins with solder tacks on them simultaneously., and then quickly tacking the resonator. The first time I did it, I messed up several times, and ended up ripping a trace, requiring me to mill my PCB again and start stuffing all over again. The second time, I soldered the resonator with care and was able to do it cleanly.
  • Damaged Board Trace
    Damaged Board Trace due to difficulty in soldering resonator

  • The tracks on this board were very close to each other. I ended up bridging them a few times and had to be really careful as I removed the bridges with a braid.
  • Final Board
    Final Board

    Hopefully, the board will work next week.