Thursday, April 7, 2011

Credit Where Credit Is Due

When I first started working on home power monitoring, nothing existed for homeowners to understand and manage their home power consumption.

I found this home energy monitoring website, which has been around for awhile now:

His power usage display is spectacular: However, I priced out the components he used, and his Veris A/D converter runs about $1500 last time I checked. Way, way beyond the cost of the elegant Arduino.

Trystan Lea's site is where I first found Arduino and how it can be used for power measurement:
The name of this blog, "openwattmonitor", is meant to show the lineage, as the child of


"The OpenEnergyMonitor project is based on the work of two developers, Trystan Lea and Suneil, both from Wales. "This is a project to develop and build open source energy monitoring and analysis tools for energy efficiency and distributed renewable microgeneration." The project appears to have been launched in the summer of 2009."

It's all happening now, in real time. They have visualized an "energy monitoring and management framework" that encompasses both open hardware and open software in innovative ways I've seen nowhere else and mobilized an impressive force of technical ability to make it happen.

Dave's "Desert Home" blog blazes new trails in measurement of home power usage in the US, and with interfacing the output to Pachube at

Dave's writings describe how he started measuring his home energy consumption and caught his local power company overcharging him. After calling them on it, they replaced the meter and his power bill was affected very favorably.

I bought my stuff from SparkFun:

Home power monitoring is here to stay.

Assembling Arduino and Shields

There are three total PC boards for this project: the Arduino Deumilanove (now called the UNO), the ethernet shield and a protoshield that plugs into the top and holds the circuits required for measurement of the AC voltage and current waveforms. This part of the project won't need an LCD display or control panel - it will be mounted next to the main power feed for the house, in a weathertight plastic box.

Some customers have complained that the ethernet shield sits down too low on the Arduino. They must have fixed that problem, mine sits high enough to clear the USB connector on the Arduino. I'm hoping the shield will work without the need for any hardware modifications. So far, things look pretty good.

I've been considering using an Arduino "Pro" from Sparkfun: because it seems much more durable to solder the arduino and the ethernet shield together. You're certainly guaranteed the ethernet shield won't possibly touch the Arduino that way, and I like the idea of committing a lower-cost, solidly-wired Pro for the project. I'll consider swapping out the Deumilanove for the Pro after I get all the kinks worked out.

I started assembly of the protoshield tonight. First the basic components that came with the shield. Here they are simply placed on the shield, not soldered down. It won't look that bad when it's done....

The red shield will then get the components for the voltage measurement circuit. Test only those components, then assemble the current measurement circuit and test those. The prototyping area on the shield is very, very small, so I want each section to work before moving on to the next one, or it'll be a mess.

I'm thinking about adding temperature sensing to this project. It would be helpful to see how outside air temperature changes compare to energy usage.

Friday, April 1, 2011

Adding the Two Phases

Found a fellow out in Arizona who's come up with his own means for measuring his home AC power consumption:

He's employing a neat trick: "add" the two legs on the input of US 2-phase power feeds by wiring the two current transducers in series.

My original approach was to sample both legs separately, dedicating an A/D input for each leg. That would let me sample only two circuits using an Arduino Deumilanove (which has six total analog inputs...well, two-and-a-half, but that half-circuit wouldn't be useful.) While this is a precise and powerful way to do it, the phase-locked nature of the analog power waveforms allows you to "multiplex" the voltage one on top of the other by connecting one leg of each current sensor together.

The payoff is that you can use only one A/D input to measure power draw on a two-leg, two-phase US-configured power circuit. That would enable the basic Arduino Deumilanove/UNO to measure up to FIVE circuits (A0 for the voltage reference and A1-5 for the current in each circuit to be measured) instead of only TWO circuits. Excellent!

The disadvantage to this approach is that you've got to hack up your current sensor: wiring them in series like this means you ought to cut off one of the leads from the sensors and join them together. And, you can get the connection wrong.

Note to self: do more research to satisfy myself that this actually works.