200 lumen per Watt Efficacy barrier broken

This is great news:

Cree has demonstrated 208 lumens per watt efficacy (luminous efficiency) in a laboratory LED. I can hardly wait to see these in full production. Still a long way to go for low-cost volume manufacturing, but it's nice to see such great progress in the last few years.

From 70 to 90% energy savings for walkway lighting

I don't have all the details yet, but the information I received from a systems integrator in Canada indicates a big improvement in energy savings by using a wireless control network. The integrator installed new lights along a walkway at (or near) one of the Olympic venues there, and by using LED fixtures (not sure what type of light was replaced, probably some type of HID) they saved 70% or so on energy consumption. I suspect that, in terms of total light emitted, the new fixtures don't put out as much light overall, but they only put light exactly where it's needed. If I'm guessing correctly, this will also reduce wasted light spilling into the sky ("light pollution").

The kicker is when they added some new controls that I worked on. The lights were all powered by a single wiring circuit under ground. The whole circuit is controlled by a single light sensor or timer, so all the lights come on at one time and stay on all night long. The LED dimmer we provided behaves differently, however, and always turns on to 50% brightness (25% energy usage because the eye perceives brightness on "square root of energy" curve). The integrator put motion detectors on each light pole and connected them to our controller. With the additional energy saved when nobody is around, I'm told the energy savings (compared to the original lights) went up to over 90%. That's a lot of kWh each year.

Here's where it gets interesting: we set up all the controllers to self-associate into a pseudo-mesh network. Instead of each motion detector turning it's own light up to 100% when people walk by, the message is automatically repeated along the string of poles, turning all the lights up. After 5 minutes (with no motion at any pole) the lights all dim back down slowly.

It's a little departure from the way we usually build wireless control systems. Typically we want each dimmer to respond to just a few battery-free wireless control switches or maybe one or two solar-powered occupancy sensors. In this case, however, it needs to be more of a free-for-all, where any control point can stimulate a response from all other participants in the system.