DIY LEDs - The write-up

[S2minute]

Leds

Leds

S2minute, I see you point, I have calculated that using 200W if white led over 48"x24" tank you get maximum 40,000 lux on water surface if you put it approx. 14"-16" over the water surface. That means weak direct sunlight.

These numbers as from wiki:
10,000"“25,000 lux Full daylight (not direct sun)
32,000"“130,000 lux Direct sunlight

But we add also Royal Blue to tank and there is max. 50W of radiant flux of 450-460nm for 100W of Royal Blue and that might possibly bleach out the corals.

This is what i`m going to do. When my new fixture is up and running, i`ll run the leds at the lowest poss setting and if i can still dim up and down from there i`ll do that. I may even add a strip(s) of that charcoal colored fiberglass window screen under the fixture to further block the light intensity if i have to. However first, find a vendor that offers those frags for like 12 for $99 and buy a couple dozen or whatever. Put them in beforehand. Then DON`T remove a layer of screen or dial up your LEDS each week! DON`T. Do it very little,"once a MONTH" and observe! If you OBSERVE your corals and see that they want MORE, then give them alittle more..ALITTLE! Corals don`t adapt to change in a month or two in the wild! It could take years! So if you have all the corals you can handle right off they will ALL adapt at the same time IF you accumulate them SLOWLY.

 

[terahz]

Listen, natural sunlight on a clear day under good conditions is pretty freaken intense and corals on reef flats do just fine. You mean to tell me that the amount of LEDS some of us put over our tanks emit more lumens than Nat. sunlight on a reef? Impossible. It`s all in how you accumulate them. You can`t pack leds close enough to equal lumen output of the Sun on a reef flat.

While I agree that most, if not all of us don't generate nearly as much par as the sun using LEDs, it is absolutely possible to get more than 2000 PAR (average noon, clear skies PAR at 0m elevation) with LEDs even with the now 'outdated' XRE leds. If you put them 1" or less apart and a couple of inches over the surface of the water, drive them at 1A without optics, I'm pretty certain you will get more than 2000 PAR in the first inch or two under the water. I had 12 XRE (RB+NW) LEDs in my 10G initially and was getting around 900 PAR (unadjusted, so more like 1200) just under the surface at 700mA. If these were XMLs at 3A, I'm sure the par meter would be showing an error.

So if you bleach them, your not accumulating them slow enough. I`m i right or wrong? Even if it takes a year or more to bring your LEDs up to the level you want, do it.

Yes! I agree that moving corals under LED for a first time is pretty stressful so most of the bleaching happens due to the change of light not because the LEDs are too strong.

 

[007Bond]

Ok guys, I read the whole thread, and am starting to get the lingo, but am still leaning about leds. One thing I do know is they sound like the best way to break away from metal halide ( heat ) and chillers. for $ vs. $ effectiveness. I need some input on were is best to purchase? ie leds,ps,drivers ect. My tank is 98x36x42deep,the leds will be 9'' above water. how many leds and how many rolls would you suggest ? optics? So I could use some input on a led build. Thanks in advance. Bond

 

[S2minute]

Leds

Leds

While I agree that most, if not all of us don't generate nearly as much par as the sun using LEDs, it is absolutely possible to get more than 2000 PAR (average noon, clear skies PAR at 0m elevation) with LEDs even with the now 'outdated' XRE leds. If you put them 1" or less apart and a couple of inches over the surface of the water, drive them at 1A without optics, I'm pretty certain you will get more than 2000 PAR in the first inch or two under the water. I had 12 XRE (RB+NW) LEDs in my 10G initially and was getting around 900 PAR (unadjusted, so more like 1200) just under the surface at 700mA. If these were XMLs at 3A, I'm sure the par meter would be showing an error.


Yes! I agree that moving corals under LED for a first time is pretty stressful so most of the bleaching happens due to the change of light not because the LEDs are too strong.

Original plan for me: 175 bow front, 6` long, 2 braces so, 3 sections. 55 leds per section: 48 will be RB XP-E white XP-G in 4 bars of 12 per section the middle bar will have 7 XM-Ls (T6CW and WW), that`s about 1 led per 9.6 sq inches roughly. Fixture will be 12-14" above tank. However, the XP-Gs and XP-Es will all be run at 700MA. The XM-Ls at 1500MA, The leds on the outside edges will have 40 degree lenses IF they hit areas on the bottom, otherwise i`ll use 60s. The XM-Ls i many not use any lenses at all but, if i want to, 60s and 80s are now avail for them. I`ll use 80`s.
I`ll get all the frags i want in there beforehand. Then i`ll run the fixture at it`s lowest setting. If i have to, i`ll use window sereen to further dim the light. THEN only once each month, i`ll make it just a little bit brighter IF it`s time. And i`ll wait. Even if it takes over a year , i`ll wait. Remember LEDS are the new thing for us in lighting and whats the most importent thing to us in this hobby?? ...PATIENCE without that...we doom ourselves to failure. After that you`ll have a tank full of corals all accumulated to the leds at the setting you want. ALSO, if you can, start a seperate frag tank with another LED fixture so any new or "special " frags you want to add can be adjusted to the current light level you have in your main display, then just move um over. At first, all that will take a LONG time to get a decent rotation going but , after, you have a steady flow of LED adjusted frags and corals to do whatever you want with. This is a new thing for us...so you`ll have to take the appropriate steps to not only let your corals adjust but to let you HUSBANDRY adjust to it as well.

 

[SMDBOYZ]

All the items for my build finally arrived 48 xp CW 48 xp RB AND 8 meanwell drivers all on a 60 x 10 in heatsink along with 60 degree optics
Ive read somewhere that 8 drivers wont be a good idea and in reading thru several threads i cant really find what i should use for drivers if the 8 meanwells are a problem can i use 2 other larger drivers and what model would be recommended i need dimming.
Anyone know if rapidled will take back the 8 meanwells.
Would the 8 drivers be ok if i have the blues come on first then a few min later the whites?
Any help is much needed and thanks in advance
Rich

 

[bamf25]

^^^^

You can look at the Meanwell hln series drivers. They have units that will provide upwards of 240 watts, and currents upto 3 + A (I may be wrong on the amps). The only MAJOR difference is you will have to run strings in parallel to use these drivers to full effect. That presents some wiring differences and complications. There are also Thomas Research drivers (nanotuners has them sometimes), they are a high volt driver, but can be hard to get. It is more dangerous to work with high volts SO BE WARNED. They will run long strands in serial.

 

[brassmonkyballs]

All the items for my build finally arrived 48 xp CW 48 xp RB AND 8 meanwell drivers all on a 60 x 10 in heatsink along with 60 degree optics
Ive read somewhere that 8 drivers wont be a good idea and in reading thru several threads i cant really find what i should use for drivers if the 8 meanwells are a problem can i use 2 other larger drivers and what model would be recommended i need dimming.
Anyone know if rapidled will take back the 8 meanwells.
Would the 8 drivers be ok if i have the blues come on first then a few min later the whites?
Any help is much needed and thanks in advance
Rich

You'll have to call them and ask. They are a popular model and always running out of stock. They carry the HLG240 A version and i suppose if you asked them to take them back and you buy those they would be more inclined. But I would take a look at the HLG 185-42....A version for no dimming, B version for 3 in 1 dimming. 42V and 4.4A...a pretty good match for 4 strings of 12. Wattsupply.com carries them...rather lists them but its a special order for them. I have a couple HLGs on the way. maybe Rapid would order them for you too.....

 

[marcer]

I have calculated from following formula:
PAR = h.c/lambda/NA

where:
h = 6.6626E-34 (Planck's constatnt)
c = 299792458 (Speed of Light)
NA = 6.02214E23 (Avagadro's number)
lamdba = 450nm = 4.5E-7 m (wavelength of Royal Blue)

that you can get 3.76 umol.m-1.s-1 (PAR) from 1W of Radiant flux of Royal Blue Leds. Cree XP will give you 0.9W of Radiant Flux for 2.38W (3.4V Uf, I=700ma, 38% efficiency)

That means that theoretically, if you focus all your RB light into your tank, then from 42 RB Cree Leds you can can about 190 PAR at water surface of 50"x24" if you calculate 90% efficiency of optics.

From CW you get more PAR, because white light of leds is created by combining blue phosphorus LED with yellow phosphorus. Yellow phosporus emits light from about 500nm to 650nm (green, yellow, red) where ration beteen W and PAR is better, from 4.17 to 5.34 PAR/W. So you get even more PAR but in green/yellow/red spectrum. So I estimate that about 30-50% of PAR generated by CW is blue PAR and about 50-70% is green/yellow/red PAR.

So if you add about 100W of CW/NW light to your 50"x24 tank you add another about 80 PAR of Blue and about 125 PAR of green/yellow/red. So in the end you should have blue 270 PAR. Is that too much for corals? Most of the coral are OK with 400 PAR, some need even 600-800 PAR.

 

[Jimmy54]

WOW ... Quantum Physics for a Reeftank - now we're getting somewhere :crazy1:
What about Heisenbergs' uncertainty principle, and where are those happy, up and down quarks :lolspin: kidding!

PAR? ... important.
PAR/PUR ratio? ... most important.

 

[marcer]

I just wanted to use little bit of scientific approach to show how many Watts/leds you can use to get certain PAR. The measurement seems to be quite inaccurate to me.

btw the formula above is incorrect, but the 3.76 umol/W by RB is correct ratio. Note: W is radiant flux Watts not consumption Watts. LEDs do not have 100% efficiency.

PAR/W = lambda / (c . h . NA) in mol of photons.m-2.s-1

Bottom line: you get about 140 PAR per sq meter (1550 sq. inch) from 100W RB at 700mA (about 37W Radiant lux). If you focus it e.g. into 0.5 sq meter of tank water surface you get 280 PAR. For CW/NW you should get even more PAR but significant amount of PAR is green/yellow/red that is not (as) useful as RB.

 

[bamf25]

Ok lights finally over my tank. This is a 120g tank 48 inches x 24 inches x 25 inches. Lights are about 6 inches over the water, no lenses, there is an acrylic splash gaurd over the leds. Whites at 1000 mA, blues 700 mA

100% whites lights only 13 cools and 13 neutrals.

100% all lights. So the whites plus 4 blues and 35 royal blues.

I tired to make colors as accurate as I could, but it is not easy. I took a all blue, but it reads so blue the photo looks like abstract art. Very little color bleeding despite no lenses, and I did run single color on each bar, but no color banding is evident.

 

[S2minute]

leds

leds

bump

 

[S2minute]

Leds

Leds

I just wanted to use little bit of scientific approach to show how many Watts/leds you can use to get certain PAR. The measurement seems to be quite inaccurate to me.

btw the formula above is incorrect, but the 3.76 umol/W by RB is correct ratio. Note: W is radiant flux Watts not consumption Watts. LEDs do not have 100% efficiency.

PAR/W = lambda / (c . h . NA) in mol of photons.m-2.s-1

Bottom line: you get about 140 PAR per sq meter (1550 sq. inch) from 100W RB at 700mA (about 37W Radiant lux). If you focus it e.g. into 0.5 sq meter of tank water surface you get 280 PAR. For CW/NW you should get even more PAR but significant amount of PAR is green/yellow/red that is not (as) useful as RB.

So you went through all that just to tell us it seems to be inaccurate and incorrect?

and your saying the LEDS don`t give out nearly as much as we thought.

 

[marcer]

So you went through all that just to tell us it seems to be inaccurate and incorrect?

and your saying the LEDS don`t give out nearly as much as we thought.

Sorry, I did not want to offend anybody, I should have rather written little bit, not quite inaccurate. But I have been reading few themes about PAR and I had feeling there there is quite big difference in measured PAR values:

http://www.reefcentral.com/forums/showthread.php?t=1841197
http://www.reefcentral.com/forums/showthread.php?t=1965294
http://www.reefcentral.com/forums/showthread.php?t=1997530

some values area really insane, you can get over 1000 PAR values from 160W only if you focus it to relatively small area like 20"x20". Leds are not perpetum mobile, they can only deliver certain amount of light.

I went through all of this to find out if I should place 42, 56 or 72 RB leds at 700ma over my planned 50"x24" tank. Now I know I will get 190, 250 or 325 PAR only from RB leds. Plus about additonal 100 RB PAR and 150 green/yellow/red PAR from 56 CW/NW at 700mA. So total 440 - 585 PAR at water surface from 230-310 W of Leds. I want to use CAT4101 voltage regulator to dim leds and also adjustable Meanwell power supplies to change maximum current for each group from 500ma to 1000ma.

So I really wonder, why some people with smaller number of LEDs are afraid to bleach out corals. I'd like rather to build light with more leds and then dim it/ lower max. current than to build smaller and not to have enough PAR for SPS in the end.

For 42 and 76 XP-E I would use string of six, but for 56 I would need to use strings of seven. But Cat4101 has limit 25V, minus 0.5Vf, so max. 3.5V on one Led. Can you reply what is your measured voltage drop on XP-E RB? Cree specifies exactly 3.5V in their datasheets.

 

[cedus]

Dear all,

I am currently running a 53 gallons reef tank with a size of 27 inches x 20 inches x 24 inches (height).
It is filled mainly with SPS on the top and medium level and LPS and soft at the bottom.
I have been satisfied with my 250W XM 10000k 2x24W Osram67 and its high PAR but I think it's time now to change for something dissipating less heat and lighting less my floor !

I am ready to DIY my own led setup. I think I will build a 20 inches x 12 inches fixture.

I come here to have your advice of the number of LEDS I should take and their type . Do you think 36 should be enough or should I go for 48 ? What mix should I go for ?
20 XP-E Royal Blue, 16 XP-G5 ? Any other suggestion, better idea ?

Many thanks for your help

 

[TheFishMan65]

Good question. I would go 48 with dimmable drivers. 36 might work, but it would be a shame to find out the hard way it as not.

 

[perikaruppan]

hi,
i had already done a fixture with XREs for my 130 Gal (4'6X2'X2') using 84 LEDs (1:1, RB:CW).

Now planning to remod it to the XPG's(CW/NW) and XPE (RB's/Blues)

I currently have:
XPGWHT-L1-0000-00H51 - 50 Nos
XPGWHT-L1-0000-00FE4 - 10 Nos
XPE Royal Blues - XPEROY-L1-0000-00A01 - 80 Nos
XPE Blues- XPEBLU-L1-0000-00Y01 - 10 Nos
All ordered and delivered from Cutter... with 75 Nos of 60 Degrees and 75 Nos of 40 Degree Lens.

Too much reading too much confusion.

I was thinking of doing a 90 LED system with the 2:3 Ratio of CW:RBs.

Was thinking of the row spacing to be at 3" and column spacing of 2"... but if i do that, taking into consideration the fixture to be of 39"x12" i get about 20 Columns and 5 Rows, that makes it a 100 LED build... i feel that will be too much light.

Planning to drive the XPGs at 1A and XPE's at 750mA. will be using CAT drivers.

So kindly help me out ...
1. What should be the spacing?
2. If they were 1:1 arranging the LEDs were not a problem, with 2:3 how do i stack up the LED's, just throw them where i like ? or is there a norm? can do the straight line arrangement or Diamond arrangement?
3. Or should i do 3 modules of 30 LEDs each and place them where i like?, if 30 LEDs, how should they stack up with regards to spacing and arrangement?

 

[kcress]

Ok lights finally over my tank. This is a 120g tank 48 inches x 24 inches x 25 inches. Lights are about 6 inches over the water, no lenses, there is an acrylic splash gaurd over the leds. Whites at 1000 mA, blues 700 mA

100% whites lights only 13 cools and 13 neutrals.

100% all lights. So the whites plus 4 blues and 35 royal blues.

I tired to make colors as accurate as I could, but it is not easy. I took a all blue, but it reads so blue the photo looks like abstract art. Very little color bleeding despite no lenses, and I did run single color on each bar, but no color banding is evident.

Very nice. Thanks for putting these up. One more data point into our data banks.

 

[Obi-dad]

PAR = h.c/lambda/NA

where:
h = 6.6626E-34 (Planck's constatnt)
c = 299792458 (Speed of Light)
NA = 6.02214E23 (Avagadro's number)
lamdba = 450nm = 4.5E-7 m (wavelength of Royal Blue)

I am not a physicist or mathematician, but this formula is not correct - there is nothing in there to account for the number of LEDs. You get the same number with this formula for 1 LED as you do 10,000 LEDs.

It looks impressive to throw out physics terms like Planck's constant, but is not a valid formula.

 

[TheFishMan65]

I was looking for information on Color Shift and came across this. It deals with lower power LEDs. So I don't know how well it translates to high power LEDs.

It shows that the shift is fairly small for current 3nm and most of it happens near to 0 current range.

It also shows a 5-10% variation in voltage for a given current.

I thought some folks might find it interesting.