Sunday, June 2, 2013
Mike said it took about 20 minutes to mill and three toolchanges, so that's some rapid prototyping right there.
The actual board was a prototype I had already designed for milling on the donated Max NC mill that we resuscitated at Noisebridge. Bottom line: the OtherMill frankly did a better job than the $1000+ MAX NC, no doubt because of all the obvious hard work OtherFab folks put in optimizing materials, tooling, feed rates and the software toolchain.
Though it was one-sided, the board was really easy to solder: the holes drilled for the headers were just snug enough to support the pins with enough space to solder between the plastic headers and the board. The milling was deep enough that there was no problem with solder bridges without solder mask. I usually use a flux pen to solder bare copper but I didn't need it; I'm guessing Mike went over the board with a scotch-brite so the copper was clean and easily wetted.
I will go into more detail later on the board and the firmware/software, but it's a straightforward design for an architectural lighting installation. The board uses a ATTINY2313 microcontroller to decode serial commands and send PWM signals to a ULN2808 Darlington driver chip to control 12 volt LED light strips. Each of the three driver channels (for red, green, and blue) can sink about an amp of current and so it can drive several meters of LED strip. The serial input comes from a 75176 differential driver and is thus compatible with RS-485 and RS-422 for long data runs.
I'm particularly proud of the firmware, which does smooth and time-controllable fades between color commands as you can see in this video.
This is an open hardware design and I will be uploading files and firmware to github Real Soon Now for all to share & enjoy.
If you think this fabrication method is as fabulous as I do, go support them on Kickstarter!