Fully Simulating Daylight

[Endevor]

So this project is both for practical purposes to properly light up the saltwater tank for coral and fish, and aesthetic purposes for the viewer.

Currently, my project is run from an Arduino Mega controller. It has 8 channels of 3 LEDs ranging from all over the spectrum. I've classified them into 8 different color types:

Natural White

Cool White

Royal Blue

Blue

Green

Amber

Red

Ultra Violet

 

I've also categorized the day into 10 different phases:

Dawn - Peaks of light before the sun actually rises

Sunrise - Actual sunrise

Morning - Time for sun to reach the middle of the sky

Midday - Sun is right overhead

Afternoon - Sun is starting to lower

Evening - Sun is just about to hit the horizon

Sunset - Actual sunset

Dusk - Peaks of light after the sun sets

Midnight - Typically a deep blue light with maybe a hint of white for moonlight

Darkness - The darkest part of the night

 

My goal in all this is to properly get the feel of each phase and I'm curious if there are PAR or color ratios somewhere I can find online to closely mimic each of these phases correctly. I currently have a general lighting system going through this phase, but the colors seem off. Curious if anyone had some insight on this or knew of some articles.

Thanks

[disaster999]

How many of each LED are you running and what are the intensities of each channel of LED for each phase?

[bfliflet]

I think this might be what you're interested in...

 

http://www.advancedaquarist.com/2013/11/aafeature

[blasterman]

You're going to need a spectrometer

[Mojado]

Two feet deep isn't really helpful here. You're trying to simulate light for animals that are, what, 10 feed deep in the wild.

[Endevor]

How many of each LED are you running and what are the intensities of each channel of LED for each phase?

 

24 LEDs total, 3 LEDs for each channel.

 

Each channel has a range between 0 and 255, I usually have the whites and blues at the highest midday and I leave them about 65, so about 25%... It's only an 8 gallon so I expected not really having to leave them the brightest they can be. Blues typically come a little lower at around 50/255, roughly 20%. The others I have so far been using just to add color for the aesthetics.

 

 

I think this might be what you're interested in...

 

http://www.advancedaquarist.com/2013/11/aafeature

 

That actually does look helpful...

 

You're going to need a spectrometer

 

I will need one though...

 

Two feet deep isn't really helpful here. You're trying to simulate light for animals that are, what, 10 feed deep in the wild.

 

2ft is better than guessing at it. I'm curious what exactly would change at that depth though? My guess would be the spectrum as a whole would decrease some, primarily the green through red hence why it's bluer and darker the deeper you go...

 

I'd love to get something like this:

http://www.apogeeinstruments.com/sq-120-electric-calibration-quantum-sensor/

to attach to the Arduino and adjust the lighting according to the PAR reading instead of by an arbitrary value between 0 and 255. No idea on how to connect/program it however. That one in particular works off 500mV and I don't think the Arduino is sensitive enough.

[BradVincent]

I am glad to see someone else thinking along the same lines as me.

 

1) Depth makes a HUGE difference, both in spectrum and intensity. I think it is perfectly valid to try to simulate 2 feet down. People have good results using straight natural sunlight, equal to 0 feet down, so this should work. This is, however, a vastly different environment than simulating 20' feet down, so this is a big decision to make.

 

2) Most people have an aversion to green, or really everything from 480nm-600nm. This is most of the output from the sun. If you've seen a reef tank in direct sunlight, it looks very yellow, and not as colorful. IMO this is more natural, but many do not like this look. LEDs in this range are not very efficient, and corals probably can't use this light efficiently, so a HUGE part of your cost is adding wavelengths that you may not prefer aethetically and won't help your coral grow.

 

3) Directionality of light is a big difference from realism on a real reef, since the sun moves across the sky. I've thought about cervo-motors moving lights around or building a huge arc of fixed LEDs above my tank and only lighting a few at a time to do this. These are pipe dreams however, and would cost far more than I can afford, and take a way from the look of the tank if I ever could build them.

 

4) Realistic lighting makes for limited viewing, especially if you go deeper than 2 feet. Think about having your lights start ramping up at 7am, go almost full blast at 8am until 4pm, and then ramp down to 5pm plus 14 hours of total darkness.

 

If you want a specrometer. I have an unopened box from http://publiclab.org/wiki/spectrometer. I destroyed a perfectly good webcam trying to build one on my own, and haven't got this to work, so I can't necessarily recommend them, but if you don't have $20,000 budgeted for this, you have few alternatives if you want one.

 

Take a look at some spectral plots. White LEDs and Royal Blues have a HUGE spike around 455nm. Whites then have a gap until about 500nm. Figure out what to add for wavelengths shorter than this blue spike, and what to add to fill this gap up to 500nm and you will have a good start.

[BradVincent]

I am writing a simulator to give me the spectrum at any depth and time. It isn't finished yet, but here are a few preliminary images.

 

 

Here is the spectrum at the surface at noon:

 

 

 

Here is the same thing, only 1 meter down. A lot of the infrared and red are already gone, but the rest hasn't changed much.

 

 

 

Now we are 10 meters down, most of our red is gone.

 

 

Here is 100 meters down. Yellow and red are gone, and ultraviolet and violet start going away too, so it isn't just that shorter wavelengths go deeper. ~470nm goes the deepest, but at this point is very dark. Notice we still have a lot of green.

 

 

 

 

 

Here is what happens at the surface closer to sunset. Rayleigh scattering takes away most of the bluer light, and the spectrum shifts red. Note that the scale has changed so the light is only half as bright for the same height as above:

 

 

 

 

Here is 10 meters down at sunset. Except near the surface, spectrum doesn't change much based time of day (even though everyone makes their reefs go bluer or more actinic). The angle of the sun means more water to go through, means more red is removed. Rayleigh scattering takes away bluer light, and the effects cancel eachother somewhat. The band does get a little narrower, so to be more realistic, turn down UV, Violet, Yellow and Red faster than Blue and Green.

 

 

 

[Endevor]

Interesting charts. Only thing I've been basing my coloring on is looking outside. Been arcing my white and blues during the day, being the highest midday and dimming down in the evening till gone at night. Currently, sunrise is at 7a, sunset at 7p. Reds and Yellow are brightest in the morning and evening to give that amberish look. I have a video posted in the tanks forum with a time lapse of the whole thing. I guess I should probably start ramping down my lights a little sooner to give them more night time. Your results are also pretty interesting. Seems like almost no reds at all really make it deep at any point of day... probably why most lighting only comes with whites and blues... So what I'm getting out of all this is probably focus a lot more on my whites, blues, and greens, arcing them throughout the day and use my red and yellow (amber) as accents for pure aesthetics.