Sparkfun seems to have added  sweet little RFM12B wireless tranceiver modules to their store. They look like a great alternative to the individual transmitters and receivers I have been using in my designs. They operate at 1.8-3.6V, meaning they’re slightly incompatible with my 5v designs. I’ll be ordering a few to play with and post results, but unfortunately my time has been limited lately (this job thing is totally overrated).

Automatically monitoring home energy usage is the new green thing, and will hopefully some day not only be limited to nerds (like me). In my case my energy is not really all that green – living in a 1920’s New England colonial-style house, we are powered by with nice greasy heating oil.

Having built out the basic remoting infrastructure and the receiver and remote sensor devices, I had to find a way to monitor oil usage. While an oil-flow meter was out of reach for a first attempt, I noticed that the burner had an LED on the front which turns itself on when the burner is on. So I took the light sensor board, wrapped a little piece of paper around the LED and the light sensor to keep them stuck together and had it send the status of the LED to my computer every 30 seconds to be stored in a mysql database for further analysis.

This project has two awesome goals. Firstly, to measure how much oil I’m using (duh). But further to test how the automation solution copes with continuous data points being sent — a long term test or soak test as we software engineers call it.

Here’s the first week’s data showing the boiler on time, missed data time (gap) as well as the day’s average, high, and low temperatures:

Some interesting things jump out:

• There are periods of up to an hour at a time during which no data is received (7.5% of every day’s data on average). There seems to be no pattern to the times – I am blaming RF interference.
• At first glance, there is no correlation between outside temperature and oil consumption. There are, however, two factors not in this chart, namely sunshine and wind. They make a huge difference as my house gets hit by the sun all afternoon which reduced the need to heat.
• I contributed 18 gallons of oil to the planet’s demise last week (the burner is rated 1.2 gallons/hour)

Take-Aways:

We need to find a more error-tolerant transmission scheme as interference is to blame for significant data loss. The firmware’s automatic re-send feature (re-sends commands twice instead of just transmitting once) was disabled in this test. However, it might also be an excellent reason to investigate the use of transceivers (these look awesome) to allow more sophisticated error handling at the protocol level.

CatrPillr is an open-source based home-automation library based on cheap electronic parts, Arduinos, and the software to bring it all together. Wireless remote sensors and actuators report data measurements from around the house and act on commands sent to them. Applications range from energy monitoring to full house automation such as controlling lights and garage doors from an iPhone. The complete software and hardware is available via an open source license and easily adapted to new remote sensors.

CatrPillr currently doesn’t provide a turn-key solution but blueprints, circuits, ideas, and software which are the foundation of hopefully many devices to come. Knowledge of software, basic electronics, and AVR-micro controllers is currently required to put these things together.

Welcome to CatrPillr! The site just went up!