DIY LEDs - The write-up

[TheFishMan65]

If they are wide angle then you could get away with fewer. However, one of the reason I like the LEDs is that I plan to point them away from my glass. With wider angles that becomes harder.

There are two things to worry about with LEDsm IMO. Getting enough light and not spotlighting. I think there comes a point where the LEDs may just be too bright - yeah we could turn them down, but then why pay the extra money. IIRC the MC-E are not more efficient than the XP-G it the were I might consider them and turn them way down.

I am not trying to sound nasty, so just in case. I have not seen many tanks pictures of the multi LED starts (or didn't realize it when I did). Why? Maybe they are slow like me wait wait I mean discerning . Maybe they have spotlighting and are working on fixing it. I think we need to hear from them.

 

[hllywd]

Here are the stats from the place where they are selling them-


4 x Cree MCE4WT-A2-WG-M0 on round PCB

Diameter: 44.0mm
Viewing Angle: 110°
Emitting Color: white
Lumen min.: 1720 lm for all dies :eek1:
Lumen max.: 3008 lm for all dies
Kelvin min.: 5000
Kelvin max.: 10000
mA test.: 350 mA each die
mA typ.: 350 mA each die
mA max.: 700 mA each die
V typ.: 12.8 V each die
V max.: 15.6 V each die

The problem I see is that with XR-Es you're already pushing more PAR than most people need. The way I understood, the MC-Es are basically 4 of those on a die, then place four of those on a "star" and you've essentially got 16 emitters in the space you once had 1. You went from overkill to super overkill... in one spot. In a 2' deep tank you'll have a hot spot where where everything will be bleached and the rest of the tank will be dark. I could see they may be useful in deeper tanks to get some penetration... maybe, but you'd still need quite a few for a smooth widespread coverage.

From my experience with a couple of 48 arrays and two more in the works, unless you want a spotlit effect in a deep tank, I don't see an everyday use for them.

 

[Iwishihadgills]

Iwishihadgills.... Not a bad choice but still poor lumen output when compared to XPGs, mind you who gives a damn when surely many of us would probably be toning down their XPGs as they are too bright and loosing few lumen from the efficiency point of view is not really a big issue (someone shoot me NO). Good luck with your purchase.

It would have been nice if Cutter gave their customers the choice of having 2 x MCE and 2 x MCEW on the same PCB with option of individual on board drivers. This way you could have mix the colour quite nicely. I am giving a serious thought to have 4 x Cree MCEW LED, mounted on my future PCB style LED tile project. This way I could drive them at very low current for the sunrise and sunset mode. Once they have done that, I could push them harder to mix with my remaining 10 XPGs and 10 XPE LEDs per tile. Remember currently I have 30 LED per tile from which 15 are XPGs and 15 XPE, however this could change to 5 drivers per tile. This will enable me to have 2 drivers for the whites, 2 drivers for the royal blue and 1 driver for the MCEW (commercial potential)

ahhh you are right. I just look up the XPGs data sheet. Does anyone make those with 4 on one board? That would be even cooler. Also does running leds at a higher ma shorten the lifespan or efficiency ?

 

[Iwishihadgills]

The problem I see is that with XR-Es you're already pushing more PAR than most people need. The way I understood, the MC-Es are basically 4 of those on a die, then place four of those on a "star" and you've essentially got 16 emitters in the space you once had 1. You went from overkill to super overkill... in one spot. In a 2' deep tank you'll have a hot spot where where everything will be bleached and the rest of the tank will be dark. I could see they may be useful in deeper tanks to get some penetration... maybe, but you'd still need quite a few for a smooth widespread coverage.

From my experience with a couple of 48 arrays and two more in the works, unless you want a spotlit effect in a deep tank, I don't see an everyday use for them.

Even without optics which is 110 degree viewing angle. Plus they would be spread out and not just in one spot. Would that still do the spotlight effect?

 

[tahiriqbal]

There is a nice little project done by RentalDeceptionist in the UK and he has used Cree MCE LEDs with great sucess.

A lovely little read for someone who may be interested.

http://www.ultimatereef.net/forums/showthread.php?t=321434

Tahir

 

[terahz]

Ok, something really strange is happening with my LEDs and their PAR readings.

I'm measuring using apogee SQ-225 sensor connected to a fluke 87 multimeter.

So as I was measuring various locations in my tank I noticed that at about 10" underwater the PAR almost tripples! I'm using XP diodes with 25 + 40 degree optics, 1:1 ratio, 600ma whites, 400ma royals and the fixtures are a couple of inches shy of 3ft above the surface of my tank.

So at the surface I get about 200µmol, at 8" deep I get about 125µmol, at 10" deep it jumps to 280µmol again and starts decreasing slowly as I go to the sand bed to end at 240µmol at about 14"

Any ideas? I have a theory but I've never seen anyone report something like that so I'm a bit confused.

THz

 

[tahiriqbal]

Is it happening thought out the length of your lighting fixture @ 10" depth? Maybe your sensor picking up reading from your 25 degree lens at certain point at that depth!!

 

[S&KGray]

Ok, something really strange is happening with my LEDs and their PAR readings.

I'm measuring using apogee SQ-225 sensor connected to a fluke 87 multimeter.

So as I was measuring various locations in my tank I noticed that at about 10" underwater the PAR almost tripples! I'm using XP diodes with 25 + 40 degree optics, 1:1 ratio, 600ma whites, 400ma royals and the fixtures are a couple of inches shy of 3ft above the surface of my tank.

So at the surface I get about 200µmol, at 8" deep I get about 125µmol, at 10" deep it jumps to 280µmol again and starts decreasing slowly as I go to the sand bed to end at 240µmol at about 14"

Any ideas? I have a theory but I've never seen anyone report something like that so I'm a bit confused.

THz

Maybe that depth is where the beams from the surrounding emitters start to overlap, multiplying the PAR by however many are overlapping.

 

[terahz]

Is it happening thought out the length of your lighting fixture @ 10" depth? Maybe your sensor picking up reading from your 25 degree lens at certain point at that depth!!

Yes, it is almost everywhere at that depth.

Maybe that depth is where the beams from the surrounding emitters start to overlap, multiplying the PAR by however many are overlapping.

Well the emitters are 2.6" apart, and most are with 40 degree lenses, so at 10" underwater that's almost 4ft. I'd think they'd overlap by that distance.

 

[tahiriqbal]

Overlapping is most likely the cause for this to happen.

 

[Taqpol]

Does anybody know where I can order Cree XR-E Neutral Whites mounted to the same star board that RapidLED uses?

Someone a few pages back flamed the majority of LED users saying we should all be using Neutral instead of Cool white LEDs and that Cree was really bad. I don't know what his problem was, but it did get me thinking. Some of the complaints about a 1:1 CW/RB mix is that it can be a little too purple and that the reds in corals don't pop as much. I sort of agree with this, and just for kicks I wanted to try putting two neutral whites in place of cool whites for every twenty four LEDs just to try and get a better red spectrum. They would be ran without optics for better spread and so that I don't get cones of warm light.

Has anyone else had this thought?

 

[hllywd]

Even without optics which is 110 degree viewing angle. Plus they would be spread out and not just in one spot. Would that still do the spotlight effect?

I don't have anything scientific, but IMO the four MC-Es will act almost like one point source. I've shown singles and my arrays on the ceiling and while the optic may be 110°, the light is not as bright at the outer edges as the center. Simply put, yes I think they will give you a spotlight effect... and really bright if the tank isn't fairly deep.

Maybe a couple of these hung a few feet above a 2' cube with some type of optic to narrow the beam and keep most of the light in the tank would work, but you'd still need a bunch of blues to cool it down... (maybe they come in RB)

If you read through this thread, I believe these were discussed, and dismissed for some various reasons some time back.

 

[widmer]

Scary? Each one of the four emiters produces 430 lumens at 350ma. That is four times your efficiency of normal leds people are using. So in theory you would need less of these to light your tank then your normal xp-g's. Yeah the heat will be more but i dont think to much that a better fan can't handle. $63/4= $15.75 per 430 lumens per emitter only driven at 350ma. And the lumens put off by each emitter is 430-480 lumens, 430 is the minimum output. Each mc-e emitter from rapidled is $25 which is already a good price.

It's a fun idea to have so much light coming from such a small footprint, and it is an excellent price as compared to the normal cost of a single MC-E, but...

It is certainly not more efficient than XP-Gs. Remember, watts = volts x amps.

So when running the MC-E @ 350 mA, you have 430 lumens/(0.350 A x 12.8V) = 96 lumens/watt

Meanwhile, when running an XP-G @ 350 mA, you have 139 lumens/(0.350 mA x 3.0V) = 132 lumens/watt

This means that the XP-G is actually 1.4 times as efficient as the MCE at turning electricity into light.

 

[westmoorenerd]

I've been really contemplating using the 3-up XP-Gs from LED Supply for $22 since it puts out about as much light as an MC-E, but with less electricity. No real need to have 4 MC-Es on a single board unless you're using them for headlights or flashlights....or need to penetrate through 5ft of water.

 

[westmoorenerd]

Just found this well written DIY....a little underpowered for a reef, but good reading none the less: http://aquariums.seaspraydesigns.biz/my-200-gallon-tank/diy-cree-xp-g-r5-led-aquarium-light-%E2%80%93-part-iii/

 

[jtma508]

Someone a few pages back flamed the majority of LED users saying we should all be using Neutral instead of Cool white LEDs and that Cree was really bad. I don't know what his problem was, but it did get me thinking. Some of the complaints about a 1:1 CW/RB mix is that it can be a little too purple and that the reds in corals don't pop as much. I sort of agree with this, and just for kicks I wanted to try putting two neutral whites in place of cool whites for every twenty four LEDs just to try and get a better red spectrum. They would be ran without optics for better spread and so that I don't get cones of warm light.

Has anyone else had this thought?

I'm not at a point where I can start a build quite yet but I've been questioning the obsession with CW for quite some time. Given that MH is the 'gold standard' for lighting, I've been studying spectral plots of 'typical' MH setups. If you compare those with spectral plots of the LEDs people are using (typically CW + RB) the difference is abundantly clear. The CW/RB combo is decidedly short on the green-yellow through red wavelengths found in MH spectra. Of the three white CCT's, CW is the most deficient. My sense is that we should be using a mix of neutral and warm whites to best mimic the MH spectrum. We have plenty of PAR.

The only other deficiencies (relative to MH) is the missing peaks at around 420nm (true actinic) and the missing cyan wavelengths around 500nm. Again, relative to MH spectra.

 

[westmoorenerd]

jtma...I feel the same way as you do, I tried getting some custom 3-up LED stars made with CW, NW, and WW, but it didn't work out. I was also trying to get some 3-ups with 2 RB and 1 B.

 

[jtma508]

If you compare the aforementioned Cree XP-G white spectral plots to T5 spectra (as opposed to MH) the issue becomes even more clear. Since T5's use phosphors (like white LEDs) we see the same kind of 'multi-banding'. Look at the UVL ActinicWhite and AquaSun, for example. http://www.advancedaquarist.com/2009/3/review

We see the signature peaks at ~435nm, ~480nm, ~540nm, ~605nm

Now compare the outputs of the different XP-G whites: http://www.cree.com/products/pdf/xlampxp-g.pdf

The RB's have a peak at ~460nm and the Blues at ~475nm

Compared to T5's our use of RB puts us a bit above the actinic component that ranges from ~420nm to ~450nm (that ~435nm peak). This is probably why we keep seeing "Does anyone know where I can get 420nm LED" threads. IMO this has translated into people believing they need UVA when what they truly need is more output in the violet (~420nm) spectrum.

On the XP-G spectral plot you can see that a combination of NW and WW LEDs would give you the peaks at ~540nm and ~605nm respectively. But we're left sucking wind at that ~480nm (cyan) peak. Does that matter? I don't know. But we should work to find out.

We know we have the PAR with LEDs. But there are people complaining about loss of color in corals. Seems like (all things being equal) that would have to rest in spectral issues. Keep in mind, when we measure PAR we are measuring a range of photosynthetic active wavelengths. So, if we measured a PAR of 300 for a combination of emitters that had outputs of 25% at 435nm, 25% at 480nm, 25% at 540nm and 25% at 605nm would that be the same as a PAR of 300 produced by a group of emitters where 100% of the output was at 540nm? It's only my hunch but I don't believe zooxanthellae are quite that indescriminant. It seems more like that they have very specialized (evolved) spectral requirements. Anywhere along the PAR spectrum is unlikely to suffice, IMO.

If you look at MH spectra http://www.advancedaquarist.com/issues/aug2004/review17.htm (10K shown here) you can see the peaks at ~420nm, ~450nm, ~530nm, ~575nm and ~600nm (we'll ignore the IR peaks --- we hate those). Clearly, a combination of WW and NW will cover the 575 and 600 peaks better than CW. RB/B along with the whites will cover that ~450nm peak. We're again short on the ~530nm and ~420nm peaks.

We're limited by what the LED manufacturers produce. But that doesn't mean that the available wavelengths align sufficiently with what the organisms we're trying to support want and need. We do know that T5 and MH work really well supporting reefs. IMO, I believe we need to work toward making our LED builds replicate the conditions created by lighting that we KNOW works. Keep my PAR example in mind.

 

[The Grim Reefer]

As long as you have enough Red (630nm +) Blue (450 to 490nm) and green (490 to 560nm) all the other colors will fill in. The magic is getting the balance and blending worked out. Getting enough violet might be an issue, I don't have my tank up yet so I have no idea if the royal blues are going to have enough bleed into that range without adding some UV Leds or not.

I am not sure but I think just using natural whites rather than a mix of warm and cool white might be the best way to boost the red spectrum without getting a spot effect. Especially if you are using optics

 

[jtma508]

I'm pretty convinced UV is a bad idea. Indigo/violet in the 420nm neighborhood absolutely but below that not so good. A paper published by Banaszak, Santos, LaJeunesse and Lesser in 2006 (http://homes.bio.psu.edu/people/faculty/lajeunesse/PDF/2006Banaszak_etal._JEMBE.pdf) studied the production of UV 'sunscreens' in zooxanthellae --- mycosporine-like amino acids (MAAs). These protectants are colorless and the researches found that zooxanthellae burn-up 19% of their total energy budget to produce MAAs. That is energy that should be used for growth and reproduction. There's been mixed concensus on whether UV exposure (280-400nm) is linked to bleaching.

My biochemistry recall is pretty rusty at this point, but naturally occuring photosynthetic processes are pretty wavelength-specific. The proteins involved absorb relatively narrow bands of light and evolve to take advantage of the ambient conditions over the long-term. For example, there are as many as 11 different MAA's found in shallow water corals each absorbing a narrow UV sub-band.

I don't think I'd mix WW and CW. I think NW and WW would make a better mix. As I noodle about with my design and research I'm leaning toward 4 independently controllable channels (via RKE ALC) with B, RB, NW and WW. But I'd really like to find some good (and affordable) 420nm 3w LEDs to work in there as well. And perhaps the blue channel should really be a mix of blue and cyan to cover that trough at ~480nm.