Because it's the weekend, I'm focusing on DIY stuff. Sunday is the day to let one's inner idealist out. Why not shine some light on grass-roots power monitoring efforts?
First, let's take a look at Dan Stahlke's home power monitor. He chose the TED (The Energy Detective) approach, but went 100% DIY. He started with home-brew breaker panel inductive current sensors. Tying the sensors to a control box, he embedded the intelligence to compute load, display a usage line graph, and communicate with a PC. He uses an Atmel ATmega168 microcontroller. While the project is interesting, I'm beginning to see a trend in what people learn from these types of DIY investigations: "...the moral of the story is that in our house the hot water heater, refrigerator, and laundry completely dominate the power usage. The compact flourescent lights and computers don't really even make a dent."
I won't bother quoting further from Dan's project write-up--it's short, so just read it on his website so that you can also check out his nifty charts.
From my point of view, the most interesting thing about Dan's solution is the granularity of the data he gathers. Because he is sampling the load so frequently and with such precision, he is able to draw very specific conclusions about his power usage through ad hoc non-intrusive load monitoring (NILM). I don't think any of the existing commercial solutions can match his aggressive 9,615 samples per second ADC sample rate.
Edward Cheung, NASA engineer by day, developed a similar current-sensing solution back in 1999. Unlike Dan's solution, Ed's ties in with his house's demand control system using X-10. The microcontroller is a Microchip PIC 16C74. Obviously, the solution is dated, but it is also possibly the first documented instance of a working home area network (HAN) that utilizes a load monitoring (power metering) node. If you know of something earlier, please leave a link in the comments section.
Why is all this relevant? Besides being interesting, the Cheung solution might play a role in the current patent dispute related to the StatSignal HAN patent I will mention in tomorrow's post. But remember: I'm not an intellectual property expert...I just play one on the Web.
First, let's take a look at Dan Stahlke's home power monitor. He chose the TED (The Energy Detective) approach, but went 100% DIY. He started with home-brew breaker panel inductive current sensors. Tying the sensors to a control box, he embedded the intelligence to compute load, display a usage line graph, and communicate with a PC. He uses an Atmel ATmega168 microcontroller. While the project is interesting, I'm beginning to see a trend in what people learn from these types of DIY investigations: "...the moral of the story is that in our house the hot water heater, refrigerator, and laundry completely dominate the power usage. The compact flourescent lights and computers don't really even make a dent."
I won't bother quoting further from Dan's project write-up--it's short, so just read it on his website so that you can also check out his nifty charts.
From my point of view, the most interesting thing about Dan's solution is the granularity of the data he gathers. Because he is sampling the load so frequently and with such precision, he is able to draw very specific conclusions about his power usage through ad hoc non-intrusive load monitoring (NILM). I don't think any of the existing commercial solutions can match his aggressive 9,615 samples per second ADC sample rate.
Edward Cheung, NASA engineer by day, developed a similar current-sensing solution back in 1999. Unlike Dan's solution, Ed's ties in with his house's demand control system using X-10. The microcontroller is a Microchip PIC 16C74. Obviously, the solution is dated, but it is also possibly the first documented instance of a working home area network (HAN) that utilizes a load monitoring (power metering) node. If you know of something earlier, please leave a link in the comments section.
Why is all this relevant? Besides being interesting, the Cheung solution might play a role in the current patent dispute related to the StatSignal HAN patent I will mention in tomorrow's post. But remember: I'm not an intellectual property expert...I just play one on the Web.
Leave a comment