My past few weeks have been consumed with designing and building a wireless AC dimmer that functions with my wireless Zigbee automation protocol. It has all power circuitry onboard so everything is self-contained in a little AC dongle. It also contains circuitry for wireless bootloading of code onto to the micro-controller, and circuitry for a local dimming slider/knob to plug into it directly.

Controlling all three prototype lighting units with some preliminary message routing framework.

I finished my prototype lighting controller a week or so ago. It is capable of wirelessly controlling 5-channels of LED lights using an automation protocol I’m writing, all encapsulated by the Zigbee wireless protocol. The prototype boxes are in nifty little cases that I cut apart with my fun new Dremel. Stay posted for coming updates.

The boards I designed finally arrived. I was actually able to solder on the 603-sized parts (teeny tiny). And, to my surprise, the board actually works!

Front

Back

Assembled and working

With the light controller working and wireless, I sacrificed a night of sleep to lay it out on a PCB. I give this board maybe a 30% chance of functioning properly, but I’m sure it’ll work great on the inevitable second time around!

light controller 2-layer PCB layout

After going dark for a while, I finally have some work to show. As you may know, the end result of this all is to have LED lighting units on my furniture that are all wirelessly controlled over the ZigBee wireless protocol. I’ve finally gotten to the point where I have a custom LED control circuit using an Atmega168 micro-controller, a TI LED controler, and an XBee ZigBee radio, and an application that can control it.

Here’s a rough video of the result. Basically, I use my application to discover the lighting control over Zigbee wireless. I can then select the node, the lighting channel on the node, and the color to set the lights too. This is very rough prototype just to see the idea in action.

After my months volunteering on the Obama campaign, I was left with an abundance of yard signs.

Since it seemed sacrilegious to just toss them in the dumpster, I decided to make use of them to fill the wall behind my sofa. So using some leftover wood from my coffee table construction, I made a couple blocks with holes.

Taking the metal from the yard signs, I chopped them down into various sets of sizes and gave them a nice coat of paint.

Browsing my local OEM parts store, I was able to find retaining rings, standoffs, clips, and screws to construct holders.

Put it all together using some epoxy and voila! recycling based wall “art”.

Today a 5 meter reel of LEDs arrived. Here are the results.

After building my sketchy micro-controller based AC light dimmer, I noticed that while fading though the dimming levels, it wasn’t a linear response. There was a noticeable “stall” in the fading around the mid-range of values. So to fix this, and waste a perfectly good Thursday night, I had the micro-controller watch the response of the light with a photo-resister and see what the problem was.

Keeping and eye on the light

The resulting response was nowhere near even. So I plotted the values and its derivative.

So to adjust the problem, I altered the distribution of the fade values. By decreasing the density of points around the “delay”, which was around 9000 on a scale of 0 to 16000, I was able to smooth out the fade.

Resulting distribution

Tonight’s experiment was building an AC light dimmer that could be controlled via my little micro-controller. This would basically give me the ability to remote control any standard incandescent lamp, light, or bulb. Or the ability to remotely start a fire in my apartment, depending on my engineering ability. But after a night of building and testing, I can proudly say there were no fires or electrocutions. Yet.