PAR Readings from single CREE LED?

[fry_school101]

Does anyone have the PAR readings of a single CREE LED? I've seen the plots from full setups and I can divide the output by the number of LEDs, but doing it that way doesn't account for the amount of overlap provided by different spacings. Even more helpful would be a reading with optics.

I'm trying to make a guess at some numbers for a setup like the attached pic.

 

[stugray]

I dont have an answer for you, but may I ask what software you used for that diagram.

Stu

 

[fry_school101]

I dont have an answer for you, but may I ask what software you used for that diagram.

Stu

It's just AutoCAD

 

[westmoorenerd]

Wow...I like the diagram.

 

[fry_school101]

Bueller?

For those of you with systems, could you maybe cut a hole in a piece of cardboard and measure from one light?

 

[blasterman789]

Sanjay Joshi did detailed PAR analysis of some various LED fixtures here, and perhaps it will help.

It does confirm that LED with optics do produce significant overlap and un-even PAR ratios. Acceptable if you are aware of it and designing for it, but tends to put another nail in the coffin of the 'spread those Crees around for even lighting' crowd.

For the record, I found a single 3watt RB Cree with a very narrow optic would roast an acropora from 2' away in a couple days. Sorry, I just don't trust PAR meters because they meter the same for fruiting plants or Montipora when either have different spectral requirements.

 

[fry_school101]

Thanks, I've seen that article and many similar ones. I'm trying to establish a baseline and come up with some approximate overlap values that will allow me to design a light that focuses light more in certain areas rather than just cover the tank.

 

[jgrog76]

I just started a similiar thread asking the same information. I am sure if we can figure out what the spread is of a single LED with a chosen optic and distance, we can then design a fixture around that information that will provide an even par distribution. You could also design a fixture that would put out 400+ par in one area, and then 200 or less right next to it. Visually I doubt our eyes could tell the difference. I am thinking the uneven par measured now is due to not choosing the incorrect spacing on the heatsink to match the distance the fixture is from the tank, and the optics chosen. I am beginning to think that distance is a good thing with these lights. It will allow a for more mixing of the light and even par values. LED's lose very little par at the distances we use with the optics which is unique and we should be taking advantage of that.

 

[noahm]

I am thinking that to limit hot spots to a single overlap, you should be using a triangle spacing based on about a 15-25% overlap on each side. Rectangular spacing will create a larger and more intense hot zone of multiple overlaps.

Those are 1 radius spaced 1.7 rad apart. Knowing the emitter/optic lens angle should allow you a ballpark guess as to spacing. I may be dead wrong here, but it makes sense to me. Unlike a sprinkler system, the coverage of a single emitter is pretty even, so the 'head to head' coverage is not necessary, but the triangle spacing should be ideal. Even if you do 'head to head' coverage, you still eliminate a triple overlap point as you will have only single and double overlap areas. The hot zone can only be 3 emitters rather than 4.

Sorry for the poor quality, I have no idea how to post a dwg file in a thread.

 

[jgrog76]

I think what we really need is uneven spacing or mixing. If you space your LED's for that coverage pattern we will not get blue/white mixing and the tank will look spotty. I think that spacing should be thought of in units instead of even spread depending on the cw/rb ratio, and that some emitters will need to be overlapped %50 or more depending on what the color of the next emitter on the heatsink is. For example if you have a 2rb/1cw ratio you will need a pattern that will overlap the cw %50 over the rb coverage area, but the rb coverage you would want just touching edges with each other for even color distribution throughout the tank. Another way to achieve the same would be to put different degree optics on the different colors instead of spacing them out different on the heatsink. Like 40 degree on RB and 80 degree on CW.

 

[noahm]

Yeah, I was just thinking of treating each color array separately and space for full coverage of each one independently. If you are using different efficiencies, then optics may come into play along with spacing, but the triangular pattern should still be the most efficient and least overlap. The final product might not look organized as far as spacing, but each array would be.

 

[fry_school101]

You would just end up putting the blues at the intersections in a staggered pattern like above, but the original question is still valid because we are looking at an idealized 2D plane cut at some level in the tank. The overlap will change depending on how far you are from the light so looking at circles or even the hotspot diagram in the original post (which is also a circle diagram) only gives you a look at one level. The PAR will drop the deeper you go, but that will be tempered by the value that the LEDs are sharing. I would like to be able to check at several levels since this is going to be used on a display tank with the corals at different heights and not a frag tank on eggcrate. To do this I need to make a plot of the output of one LED and figure out if we can assign a value to the PAR that is getting shared.

If no one has a meter and a LED to test with we may just have to wait and I'll take my own readings after I get ahold of some.

 

[jgrog76]

The other idea I have is to use the optics and height of the fixture dictate the coverage area. In this case the higher up the fixture the better. For instance if you placed the fixture 36 inches above the tank and then used the optics to make each LED cover an area of say 24x24. This would create almost %100 overlap. It should eliminate hot spots and provide even mixing from each emitter. This helps get rid of the 2D plane problem by increasing the distance between the lighted objects and the emitter, and would come close to making it into a point light source. The total par would come from the additive effects of the overlap between each LED.

 

[noahm]

You would just end up putting the blues at the intersections in a staggered pattern like above, but the original question is still valid because we are looking at an idealized 2D plane cut at some level in the tank. The overlap will change depending on how far you are from the light...

Yes. Under any setup, you would probably want to try for just getting full coverage at the highest growth point, and the overlap would just increase as you headed toward the bottom, which would, in theory, be desired to maintain PAR. Narrow optics would allow for a denser array, and higher PAR, the opposite with wider optics. Narrow optics would also create less change in overlap from top to bottom. You could also mix lenses to cover multiple levels effectively. The issue I think we are running into is the fact that a great deal depends on what height you want the fixture, and what PAR you want at a given depth. At any rate, I think you are better off building for 'more than adequate' and dimming according to need.

 

[noahm]

The other idea I have is to use the optics and height of the fixture dictate the coverage area. In this case the higher up the fixture the better. For instance if you placed the fixture 36 inches above the tank and then used the optics to make each LED cover an area of say 24x24. This would create almost %100 overlap. It should eliminate hot spots and provide even mixing from each emitter. This helps get rid of the 2D plane problem by increasing the distance between the lighted objects and the emitter, and would come close to making it into a point light source. The total par would come from the additive effects of the overlap between each LED.

I think I see what you are saying. Basically 'aim' each LED at a single point say 5 feet away, so it is just one 24" Diameter (just an example) footprint for each one that is on the same spot. You would need a slightly concave surface to mount them to, but it should work if the calculations are done correctly for the height. Is this the same as you are thinking?

 

[jgrog76]

Yes that is about what I am thinking. I was going to do without the concave surface though and just deal with the slight oblong shape created. I was figuring that the area of the heatsink is pretty small (10 inches across or 5 inches off center at any point,) that the rest of the errors in beamwidtch, math, rounding, etc would make it close enought. So far with the help of my math gifted brother I have come up with 40 inches for 25 degree optics, and 69 inches for 15 degree optics. These numbers are to cover a 28x28 area with a 10x10 heatsink. These numbers came from drawing a pyramid with a base of 28x28 and finding the point where it was 10x10. The only thing I don't know is how much PAR this would result with per LED, and if my assumtion that PAR total in this case would be a cumalative addition per LED.

 

[fry_school101]

You're going to end up throwing a lot of light outside of the tank unless you angle them and spread them out similar to Santoki's build.

Regardless, we still need to find a reading for a single LED. Please post back if you find one.

 

[ruprecht]

You're going to end up throwing a lot of light outside of the tank unless you angle them and spread them out similar to Santoki's build.

Regardless, we still need to find a reading for a single LED. Please post back if you find one.

PAR is a relative term. Optics and distance (among plenty of other things) will dictate PAR. A LED can't have a static PAR number.

However, if you find it useful it is said that 3 XPEs make 5.0 μMol/m2/s at 1 meter distance in air with 40° optics. This is the warm white light that's heavy in the 500-750 range and light in the 450. PAR is probably close enough to what it should be for the cool whites which have more blue, just divide by 3 for a single bulb. It is internet data so take it at it's face value.

 

[tahiriqbal]

Why not get something like this from flebay!!!!? Item number 170439678002 and draw your own conculsions....