ATtiny projects

I found this one at one of my preferred shops. The nice thing is that the output voltage can be tuned from 11.3 to 15.6V. It also doesn't need active cooling and is pretty efficient. Designed to run 24/7 in closed spaces.

This one had been sitting in the back of my head for ages. Now this seems to be ready for a prototype run. No PIR sensor, just couldn't get an affordable small one with integrated lens and amp. Still better than my current CFL bulbs.

My boards were mailed a few days ago, but I just couldn't wait to see them anymore. Tweaking the colors in gerbv helped a bit.

The purple colour of these boards seems to be totally incomprehensible for my scanner.

The purple colour of these boards seems to be totally incomprehensible for my scanner.

The purple colour of these boards seems to be totally incomprehensible for my scanner.

The purple colour of these boards seems to be totally incomprehensible for my scanner.

I finally managed to capture the 'purple-ness' of the boards adequately.

This is what it looks like before the LED catastrophe. I had to remove 16 of them as something was shorting. I suspect that parts of the extremely wide pins (uses as heatsinks as well) were shorting to the ground plane somehow. I clipped them and partially ruined my wire cutter. It wasn't meant to cut such thick stuff. Only 64 more of these to go (LEDs not boards).

This is what it looks like before the LED catastrophe. I had to remove 16 of them as something was shorting. I suspect that parts of the extremely wide pins (uses as heatsinks as well) were shorting to the ground plane somehow. I clipped them and partially ruined my wire cutter. It wasn't meant to cut such thick stuff.

The final board. It takes about 8.5 to 10W, depending on the supply voltage. It is designed to run at 660mA. The LEDs are grouped in sets of 4, which requires about 12V to run at a high enough current. The LED driver takes about 0.7V off, so we're looking at a real 12.5V to 13V on the supply voltage connectors. Each group can be controlled individually, but currently all 8 channels are treated as one. The PWM is generated by a timer compare match interrupt. The data is sent via SPI to the LED driver.

The boards are actually quite purple.

Saving pins by using an ADC channel to read 3 buttons is OK for just one board. But as soon as I connect two of them, the voltage levels are off. This was to be expected, as now two pullups are there, but even after removing one of them... doesn't work. The SMD resistors also vary quite a lot it seems. As I have another pin on the bus, I might later create some sort of message system to let the boards talk to each other (half duplex).