LED as main light

[TheShaws]

Great LED thread. I am fascinated with the concept and hopefully the application down the road. Those of you with the electrical aptitude.. keep it up!

PAR. Not a difficult concept here but often misunderstood. PAR does mean Photosynthetically Active Radiation. However what that means really is the "spectrum" or specifically the wavelenghts of light that influence photosynthesis. This happens to be from Blue to red, in decreasing energy levels and increasing wavelegths. Thus PAR appears white as a mixture of all the colors.

Increasing light intensity will increase the PAR assuming the light source is producing light in the PAR range. PAR in itself is not an intensity but only a measurement of the light level. If my light levels are very low the distribution of wavelengths are not different due to the intensity, there just are not many photons of light in that particular wavelength. PAR is a measurement like Foot candles or Lux. Unlike footcandles or lux, PAR actually only measures from Blue to Red and not violet nor far red, which the others do.

Not sure about the rules here but there is a good thread about light on another board. Here is the link:

http://www.reeffrontiers.com/forums/...?threadid=1516

Also check out Advanced Aquarist
http://www.advancedaquarist.com/issu...02/Feature.htm


The exciting thing about LED's is the light output vs energy input with the almost total absence of heat (lost energy). Look at the potential savings in energy over a 5 yr period not having to run chillers, fan among many other benefits.

[musicsmaker]

TheShaws, thanks. The rules here are basically that if it's reef related information, then it's welcome. Regardless of what site it comes from.

[AndyL]

Ok I'm gonna cover my tookus here - if any of you are getting bright ideas and hunting on ebay - the 20x5mm 16000mcd white LED auction from dynamic electronics ending friday is mine (why bid against each other right?)

I'll post pictures once they're here and mounted on the testbed...

Andy

[olemos]

AndyL are you talking about this

http://cgi.ebay.com/ws/eBayISAPI.dll...category=26207

[DennisRB]

We must consider that LEDs do not tolerate voltage spikes. The chance of them blowing up is much higher than MH and fluro lighting. Spikes from the supply will damage them easily, eg lightning hitting power lines or other disruptions like a car hitting a power pole or wind knocking down power lines etc. Imaging if this happened and took out $1000 of LEDs. Always use a surge protected power board or similar and remember that they can not always work.

[sandisct]

OK... I've just repaired my cracked tank.... hopefully I can fill it up with water safely.... After which i'll run the LEDs array again and post some pics on its brightness with water in the tank....

[sandisct]

I think the shimmering effect that the LEDs create will be fantastic....

[AndyL]

Olemos: nope thats a different one - though I was seriously thinking about that one - the one's I'm bidding on are a slightly higher mcd rating, and don't have to be driven beyond If max to acheive 16,000mcd...

What you get from this LED is average 15,000 mcd with a Forward Current of 20 mA (If you want to boost up the brightness to 18,000 mcd, just increase the Forward Current within it's acceptable range, but beware of sacrifice of white light purity). It's offering you a bright WHITE light source of high light purity for your project.

Farther down:

Absolute max Rating IF : 20 mA

We've learned in the past that you've got to be VERY careful to read the details and check them against themselves when you're buying LEDs - A lot of these ebayers will give mcd ratings acheived beyond the manufacturers max levels.

Andy

[Obi-dad]

I forget who, but someone posted earlier they had two arrays fail - are you using current-limiting resistors?

It is my understanding that if you don't that will kill LEDs. For example, you can't just put 12v on 3 LEDs in series even if the LEDs are rated at 4.3v for the forward voltage. They will be bright for a while but will burn out. When applying a voltage source to LEDs you need to calculate the current going through the LEDs.

[musicsmaker]

DennisRB, good point. I blew up several LED's trying to set up my moonlights before I figured out it was the resistors. FWIW, I have had them running since this thread was going around a year and a half ago. There is a lot of good info in that thread by the way. There are some pictures of corals fluorescing under the blue LED's in there as well.

[robwsup]

You could probably wire in a decent sized capacitor in parallel with your whole LED/resistor bank to protect your bank. Capacitors resist a change in voltage.

An added effect of the resistor would be that you would have a sunrise/sunset effect as the resistor charged or discharged to capacity as power turned on or off.

Kickass! I added something useful to the discussion!

[robwsup]

Actually, I just tested it, and it only adds a sunset effect. Still cool.

[musicsmaker]

Do you know how to make a delay circuit? I think it's pretty simple.

[sfsuphysics]

Another thing to be wary of with ebay auctions is if the pricing seems a tad on the low side scroll down to the bottom to find out how much shippnig is. Most will have low priced items then make shipping completely unreasonable (ie like $13 to ship 20 leds or something).

[AndyL]

Useful discussion again...

So who here has EE knowledge and can come up with some ideas for driving LEDs? My old arrays were driven simply using an ATX power supply for a computer using the 12v power feeds. And as easy as it is to run a potentiometer to simulate sunrise/sunsets - anyone have some ideas as to how to do that digitally?

Obi-dad: It was my arrays that failed - but they did have resistors, and the current draw was within manufacturers specs (I was pulling 23mA - If was rated at 30 if I recall) My thinking now is that the epoxy coating I used to water proof everything probably held the heat from the resistors and the LEDs - resulting in the drastically shortened lifespan.

Andy

[Dr. Drudge]

One popular way to dim leds is to drive them with a PWM'ed (pulse width modulation) current.
In this way you can "dim" the output any degree you like by
adjusting the on/off duty cycle- the leds are "blinked" at varying rates .
Search the web for schematics- usually involves a 555 timer or something similar.

Using a variable resister to limit the current will only dim the led over a small range- the led is either off or on. If you want to dim the output down to , say 50%, you can't use this method.

Also, I dont see how using a capacitor in the above post produces a "sunset effect" .

Cheers

[micaheli]

Okay, the writer of that original PAR site and I have been in contact and he wanted me to post a URL of his when he was doing research on LED's for plant growth.

Here's the URL:
http://fins.actwin.com/aquatic-plant.../msg00601.html

He is also going to log in as me and post some information and possibly get involved in the dicussion. Lets see what he has to say! I'm very interested!

--Micah

[Diego]

Hopefully things have changed since when he wrote that.
Date: Fri, 24 Aug 2001 16:18:26 -0400 (EDT)

Still he doesn't cover multiple LEDs side-by-side or using a reflecting backdrop.

I'd like to stay optimistic.
Maybe someone can test on a nano or a single coral instead of putting 1000s together only to be dissappointed.

Once I get all my electro-lingo in order, I may test in my FOWLR on a low light mushroom or something.

Diego

Originally posted by micaheli
Okay, the writer of that original PAR site and I have been in contact and he wanted me to post a URL of his when he was doing research on LED's for plant growth.

Here's the URL:
http://fins.actwin.com/aquatic-plant.../msg00601.html

He is also going to log in as me and post some information and possibly get involved in the dicussion. Lets see what he has to say! I'm very interested!

--Micah

[micaheli]

I plan on testing something on my nano. Just have to figure out what I'm going to test.

--Micah

[Gomer]

Just a little number comparison to get things rolling. BTW, I am not a reefer, so I might have made some sort of mistake

actually, let me run some basic numbers. Consider a 20g (easier on the numbers. Lets assume 1watt LED= 1watt HW. Lets say you want to run 7.5wpg.

MH
A 150watt HQI setup (reflector/ballast/1 bulb)costs ~300. I was told that HQIs need to be replaced around 1year. 150HQI bulb costs around $75 (ushio 10k). 10 years worth of bulbs will run you $750.

Total for HQI over 10 years=1050


LED
Lets use the Star III (3watt LEDs) You will need 50 Stars. In this number, they are 12.93ea for a total of 646.50.
The current LED ballasts are ~ 80% efficient, so the largest 40watt drivers should work somehow (not sure of wiring) for 10 LEDs. THis means you need 5 @ 63.32 a pop.

Total for LEDover 10 years=963.10

So...over the life of the leds, they are already cheaper..and this isn't even considering additional cooling costs which can be substantial over 10 years for both equipment and electricity

(oh...my numbers for HQI might be a bit off on $, but they should be close enough to make this comparison valid.

oh....go and tac on some $ for the LED heat sinks and you are still good to go

[Diego]

Originally posted by Gomer


So...over the life of the leds, they are already cheaper..

Depends on the 'life of the LEDs'

[Gomer]

Led life (70% power) is 70,000 hours. at 12hr a day, this is almost 16 years

[AndyL]

4g nano-reef with 6 Luxeon emitters:
http://www.nano-reef.com/forums/show...threadid=15832

Andy

[sfsuphysics]

I'm sorry did I miss something? LED ballast? Is that just a generic name for a ac-dc transformer?

[micaheli]

Ballast isn't what it is.. its actually a driver. A fancy term for ac-dc converter.. Yes.. however, they are buffered as well with capacitors so that surges don't affect them as much as well. Other than that, they are fairly simple. The cool part about the xitanium drivers is that some of them are dimmable. You just hook them up to a standard dimmer switch... Which ALSO maens that you can use X10 controls to dim them. So, on a computer, you could have them fade in, and then fade out at night.. how cool would that be.

--Micah