LED spectral analysis about to begin. Need your input.

[redfishsc]

We are about to begin doing some spectral analysis of our most commonly used LEDs for DIY arrays. We will also do some PAR testing to compare (roughly) the PAR output of various similar-sized LEDs.


We have available spectral graphs for most of our LEDs already, but we don't have much data that tells us what the resultant spectrum is for the combinations we use over our tanks. I'm sure some folks with ninja math skills could give a solid educated guess on the resultant spectrum if they had the data for the intensity of each LED at each wavelength, etc...

But we want to get some real data. We'll be using a radiometer at a University, and we will drive the LEDs using a benchtop power supply (constant current capable). I'll have a 700mA Buckpuck and a Meanwell ELN60-48 handy just in case.

I have all the LEDs mounted to temporary heatsinks and soldered with quick disconnect fittings.

Here is a list of the LEDs I will be testing. Note that I do not have exact data regarding the tint bin of each of these LEDs but I do have the flux bin, which gets us in the right ball park.

Cree XPG:
---cool white
---neutral white
---warm white

Cree XML
---Cool white

Cree XPE
---Blue
---Royal blue
---Red
---Neutral white

Cree XRE
---Cool white
---Royal blue

Bridgelux
---402 cool white (10 watt capable).

Luxeon
---Rebel neutral white
---Rebel royal blue
---Rebel Cyan

Satistronics
---3 watt "pure white"
---3 watt "red"
---10 watt "pure white"
---10 watt blue 455nm
---20 watt 15,000K cold white


HERE IS WHAT I WOULD LIKE YOUR INPUT ON:



1. What combinations, from the list above, do you think would be the most helpful? (for reef or planted aquaria, we will study them for both applications).

2. Do you have any LED you would like us to analyze that is NOT on the list? If so, PM me and I can arrange to study it for you but you MUST hurry. It must be here before April 28th).

3. What drive current for the 3-watt class do you think is most useful? Most of these are capable of 1,000mA but we will be using 700mA as the standard.

4. We may do some testing with optics, but given our time constraints, we may not be able to use optics at this point. Optics will only serve to show PAR data and footprint changes, but not likely effect spectral analysis, which is our main priority.

5. Any other suggestions are appreciated.

[redfishsc]

By the way, I am keenly interested in testing some of the larger LEDs that are out there, but since I am funding most of the LED expenditures out of my own pocket, I simply could not justify buying anything larger than the Bridgelux 402 and the Satistronic 10w and 20w that I have.

If anyone would like some data on the bigger Bridgelux, the new larger Cree multichips, or import LEDs that go up to insane wattages, PM me. If you can get it to me in time, I should be able to test it.


Also note that I would love to test things like the PAR30 and PAR38 bulbs out there. I have contacted some of the resellers that carry them, and only one expressed interest, but hasn't sent me anything.

[b_s_c1]

I know you say given time constraints in your post but what I would really like to know if there is a shift in the spectrum during the life of the LED's.

As far as combination I think a 25-75%, 50-50%, and 75-25% would be a good base line. I think this will cover most questions and a person would be to take an educated guess on something out of this combination.

[marc111]

Based on what people have been trying I would suggest
1:1 and 2:1 mixes of EPE:XPG and RE:XPG The most relevant will be the newer series as that is what most are buying at the moment.
RB:NW
RB:CW
RB: with both CW and NW
the same with an single blue added to the mix
Also a violet from nanotuners 420 nm led could be added to the mix.
Please pay attention to the beam overlaps as this will significantly affect the measured intensities which is key as both output power and frequency are in the PAR equation.
Note that optics will not affect the spectrum but will have major overlap and PAR effects.

Mark

[redfishsc]

I know you say given time constraints in your post but what I would really like to know if there is a shift in the spectrum during the life of the LED's.

I may be able to test some 1.5 year old cool white XPG against a brand new one, but if they vary much in their tint bin (which is hard to get constant, even from the same supplier) then this will skew the test.

The only way I know to do it is to test a couple LEDs, burn them for a couple years, and re-test them.


As far as combination I think a 25-75%, 50-50%, and 75-25% would be a good base line. I think this will cover most questions and a person would be to take an educated guess on something out of this combination.


Which colors are you talking about? Cool or neutral whites, and blue/royal blue (or combinations thereof)?

[redfishsc]

Based on what people have been trying I would suggest
1:1 and 2:1 mixes of EPE:XPG and RE:XPG The most relevant will be the newer series as that is what most are buying at the moment.
RB:NW
RB:CW
RB: with both CW and NW
the same with an single blue added to the mix
Also a violet from nanotuners 420 nm led could be added to the mix.
Please pay attention to the beam overlaps as this will significantly affect the measured intensities which is key as both output power and frequency are in the PAR equation.
Note that optics will not affect the spectrum but will have major overlap and PAR effects.

Mark


That's mostly where we'll start. I can't get the 420's from Nanotuners, they don't sell them anymore. If anyone has one they want tested, they can send it to me and I can mail it back when we are done.

[wfournier]

I might be able to supply a couple bigger Bridgelux LEDs if interested.