I’ve been eying up the Tweet-A-Watt project for a few weeks now. I think the idea is pretty cool, taking an off-the-shelf power monitoring device (a Kill-A-Watt P3) and tying it to a wireless transmitter in order to get the data from the device to a PC. The tweeting part of the project doesn’t really interest me, but storing power usage data in a database would let me see some long term trends in electricity usage. I might even be able to quell the feeling I have that the “Reduce exterior moisture / Save energy” switch inside of my refrigerator doesn’t do anything.
What I tackled this past weekend was the construction of the wireless reciever (which plugs into my PC via a USB cable) and the transmitter that gets hooked up to the Kill-A-Watt. The entire parts list and assembly instructions are available from ladyada.net. I opted to get the kit out of simplicity. It’s a little pricey, but includes everything you need for the project.
The instructions are pretty good. Provided you are alright with a soldering iron, you can manage just fine – at least up until attaching the power and data wires to four individual pins on the tiny chip found on newer Kill-A-Watts. That part is a little trying. A helping-hands type device is a must have for a project like this and I know that it made things much easier for me. So did a new, super-tiny tip for my soldering iron.
I put the whole thing together according to the instructions and plugged it in. And… nothing happened! I got no display on the LCD and the XBee activity light didn’t blink. The super capacitor takes a while to charge up, but no matter how long I waited, I got no info out of it except for the fact that it would make the standard beeps when the buttons on the front of the unit were pressed. I took it back apart to double check the connections to the tiny chip underneath the LCD display and they seemed fine. I couldn’t see any possible short circuits, but I did read on the forums that it would be important to cut a pathway halfway through the foam on the underside of the display to let the wires out. I also filed a notch in the side of the PCB so that the wires would not be pinched. The LCD has to sit flat in the right location for it to work at all. I double checked the transmitting XBee and discovered that it was not working. I figured I must have fried it as X-CTU couldn’t even communicate with it when plugged into the USB cable. There are instructions about how to revive an XBee that has been put into sleep mode on the forums at ladyada.net. I had to perform the process a few times and it eventually started working again. Upon reassembly, I found that the LCD worked – but there was still no activity light. The LED flashes once when the Kill-A-Watt is plugged in, but that’s it. It works though, so I’m not complaining.
For the moment, the wattcher.py script available from ladyada.net is the extent of the logging I’m doing. By default, it logs to a csv file every five minutes. There is also code for connecting to Twitter and to Google App Engine. Eventually, I’d like to log to MySQL and display the power usage right on my blog. Someday…
Anyway, Tweet-A-Watt is a pretty cool and possibly useful project. It’s on the expensive side if you go the kit route, but no matter how you build it, you end up with parts that can be used on other projects (mainly, the XBee chips and their adapters) if you ultimately decide that this project is not what it should be.
