Are there concrete characteristics to a SAD lamp? Could I achieve similar results if I fill a room with led bulbs?
back
4 comments
When I was researching them (several years ago now) the only key thing that everyone agreed on is that they have to be really honkin' bright, with 10000 lux at 2 feet being the standard. You'll need even brighter bulbs for your scheme, since you'll be placing them further away. It's an inverse square relationship, so at 6 feet you'll need a 9x as bright bulb.
Inverse square applies to a light emanating perfectly spherically that isn't bouncing from anything else. If your room doesn't have dark walls, you will need much less than the inverse square law would suggest.
Similarly, if you apply the "inverse square" law to sound waves as if they are emanating in a perfect sphere from our mouths, it's pretty easy to show that you'll never hear anybody very well, even shouting at the top of their lungs, from even a few dozen feet away. However, between the fact that our voices are partially directional and the fact that we pretty much always have a floor of some kind and are often in rooms that provide even greater constraints, we are rarely in situations where inverse square actually applies to our audio.
Some people like to use blue LEDs. I have 36 1W blue LEDs (0.2W each of actual light output) shining up at my white ceiling and it seems to work just fine.
My target brightness was 0.2W per square metre. That's based on a classic melatonin suppression study. More energy did not cause a significant amount of extra melatonin suppression. Unfortunately my brightness distribution is rather poor.
The blue light is on for 12 minutes every 2 hours during the day.
Seems pretty easy to duplicate with LED bulbs with these characteristics:
CRI: 85 or above
Color Temp: 4000K
About 70-80W of CFL, so that might be around 60W of LED. Using omnidirectional bulbs, perhaps 7-8 mounted near each other all shining to get the same 10,000LUX intensity at your face.
You should look at the spectral power distribution. CRI (color rendering index) score and CCT (correlated color temperature) are pretty uninformative metrics.
The CCT just tells you (approximately) how the relative excitation of short vs. medium/long cones in the eye from a particular lamp compares to that of a black body radiator, and the color rendering index tells you how whether a particular set of paint chips looks roughly the same under your lamp as it would look under either daylight or an incandescent bulb (or some value based on linearly interpolating between the two, depending on color temperature). To judge color differences, the CRI takes cone responses and multiplies by a matrix (chromatic adaptation transform), and then uses a standard color difference formula.
Most of the so-called “full spectrum lamps” have a spectrum nothing at all like sunlight. It’s not at all clear what kind of spectrum would help fight seasonal depression, but I suspect you want a bit of near ultraviolet radiation, which wouldn’t be measured by either CCT or CRI.
The color temperature used should probably be 5700K, as that closely resembles sunlight. Sunlight-like light can negatively affect your circadian rhythm (body always thinks it's day time), however.
Thanks for the reply. I'm using about 250W of 5050 led ribbon to decorate my house for the holidays. The strips I've purchased look like they will blend into my crown moulding seamlessly. I think I will try and use an arduino to rig them up to slowly work their way to full brightness 15 minutes before my morning alarm. Something tells me I will wake up before the alarm goes off...
You need a full spectrum light bulb, regular light bulbs won't do anything for you.
You need 10'000 [insert correct unit here]* and to be quite close (about 20 to 30 centimeters) for 30 minutes.
It should be done in the morning, or else it could affect your sleep pattern.
* I'm not sure if it's lux or lumen or candela, sorry