Need advice on LED mix

[SoFloReefer]

Background Story
So I started building a fixture about two years ago and never finished it. It is on a massive square heat sink that will sit over my 40 breeder. It has three strings:
1. (4) SemiLED UV 700mA max , (4) XP-E Blue 1000mA , (6) XP-E RB 1000mA
2. (14) XP-E RB 1000mA max
3. (8) XP-G Neutral White 1500mA max, (6) XP-G Warm White
These are driven by 3 ELN-60-48D Drivers. Drivers are controlled by Coralux Storm X with 5v to 10v analog converter

Anyways to shorten a long story, I ran a fan on the same power input circuit of the three drivers (used a flyback diode) and ruined all three drivers.

Where I'm at now
The DIY LED market has completely changed in the last 2 years and I am rushing to catch up. I want to run several LDD-1000H drivers powered by a matching Meanwell driver. There are so many options for LEDs today that I'm struggling to decide what to go with. I am going to reuse the LEDs I have but will probably put the SemiLED UVs on their own string. Any advice on what other colors (and channels (red, green, lime, etc) I should add? I have plenty of space on my heat sink and plenty of channels on my controller.

[SoFloReefer]

I am on the fence about adding red, yellow, and green. The neutral and warm white leds hit a lot of those spectrums.

I am thinking of rearranging the LEDs into 5 string like so:
1. (8) Neutral White
2. (6) Warm White
3. (12) Royal Blue
4. (8) Royal Blue, (4) Blue
5. (4) UV 420nm

What do you guys think?

[jedimasterben]

Definitely add lime. You'll have to order them and have them mounted, though, as there is no place on the planet that sells them mounted yet, which seems nuts, but when you think that they were made to use in RGB arrays to enable higher CRI, it figures that customers would want them unmounted already


You can swap the XP-E royals for the Luxeon M - each M will replace six of the XP-E, so power consumption will go from ~70w down to ~46w and you'll have ~22% more flux.

The XP-G are still pretty efficient, and mixing warm and neutral will help with low-CRI of the neutrals. Adding lime will help even more, add ~8x total lime to the array.

That's not a whole lot of violet - particularly if the violet you have are what Rapid sells, they are very low output bins for their price, especially considering the output bins that they sell for some of their commercial contacts. I would consider adding at least a dozen more, and even then you probably can't see that they would be turned on with any other string.

[SoFloReefer]

JediMasterBen,
Thanks for Chimming in. I have been reading your posts and you seem to be extremely knowledgeable about this subject. What is a good source of Violet LEDs, or more specifically, what is a good brand?

What are your thoughts on adding red and green to my mix?

[jedimasterben]

SemiLEDs is a good brand, it's just that Rapid doesn't sell the higher flux bins at this time. LED group buy has much higher flux violets for the same price, but ReefLEDlightng (http://www.reefll.com , a new sponsor here, from the guys behind the awesome Advanced Aquarist articles) sells the proper high-flux SemiLEDs like we all want.

They even have four die violets that emit ~5800mW of flux drawing 14.5 watts, compared to the 'Total Spectrum UV' that Rapid sells that emits only ~1300mW at 9.5 watts, which is 35% as efficient. The ReefLL violets are the most efficient violets on the market, beating even the Luxeon UV and LED group buy's violets, which are both high-flux.

[SoFloReefer]

reefll is a Chinese company. How can I be sure I am getting decent LEDs from them?

[jedimasterben]

SemiLEDs is based in Delaware, although their production is all in Taiwan. Reefll has documentation from SemiLEDs verifying the bins they sell are actually better than what is listed in the datasheet for the LEDs, in addition to using the proper package size and having the gold base. Everything points to them being genuine

[SoFloReefer]

Thanks Jedi. It caught me off guard when I saw Hong Kong shipping but it looks like they are a Russian company with production in China. Some kind of communist alliance.

[jedimasterben]

lol

I just noticed you're in West Palm. I was there on Tuesday evening dropping my wife off at the airport.

[SoFloReefer]

Yup. I make it up to Okeechobee sometimes when I go around the lake on my bike.

I'm thinking of going with the ledgroupby Violets. They have 405nm and 430nm that seem to be much better than the Rapids but not as good as the ReefLEDlab. The ReefLEDlab aren't cheap though and the shipping is really the killer (the wait mostly).

[ReefLedLab]

Originally Posted by SoFloReefer

The ReefLEDlab aren't cheap though and the shipping is really the killer (the wait mostly).

We have made great efforts to parcel comes up to you as quickly as possible. But if you willing to wait, we can offer the cheapest way of sending by China post. Unfortunately, it is very slow Most of parcels, shipped by cheap China post, is arriving to USA after one month, sometimes after 40-50 days. It's a pity. But VERY cheap! For example - minimal cost of sending parcel to USA through our "two-stage" fast sending is approximately 15USD, and by China post - only 2USD!

So, if you willing to wait, cost of shipping may be very small. To getting cheapest way please contact with us by PM.

[perkint]

Those luxeon M royals do look good. Shame they are so expensive and hard to find. Can't find any of the 3V ones on 20mm stars

Is it just me being a cabbage or are the only lights on the reefLEDlab site multichip ones?

Tim

[karimwassef]

My experience with LEDs has been a little disappointing due to " mix spotlight " issues. The fixtures I've used have different colors and the problem is that they are reflected directly on my rocks.

So I don't see a uniform white light, instead I get a red spotlight, next to a blue and then a green one, white one, etc.... They move around with surface motion, and the resulting disco-ball action is really a distraction.

Maybe the overall is ok from 10 feet away but I'm usually 10" away...

I know there are diffusers and lenses but anything that reduces my light intensity is wasteful to my tank.

At this point, I'm ripping up my LEDs and going back to a MH base with LEDs only in blue/UV (as close to one color only) to add color depth and the actinic like view I used to get with my VHOs (a long time ago).

Is there a good solution to this??

[SoFloReefer]

I've never personally used red and green LEDs. I've only used whites: cool, neutral, and warm, and Blues. I've heard of the "disco ball effect" but never seen it and I've built several fixtures. Maybe the solution for you is to go with whites that cover more of that red and green spectrum you are looking for. Also, put the white emitters directly next to blue ones.

[SoFloReefer]

Originally Posted by perkint

Those luxeon M royals do look good. Shame they are so expensive and hard to find. Can't find any of the 3V ones on 20mm stars

Is it just me being a cabbage or are the only lights on the reefLEDlab site multichip ones?

Tim

I'm glad to see that the other brands have caught up to Cree and even passed them. For my tank though, I can't see getting rid of the XP-E royal blues to save on a little electricity. I am going to be running 50 LEDs on my fixture, all fairly good emitters (high bins), which I think is more than my coral can handle on full blast. It will probably be dimmed at least 50% most of the time. Maybe with time I will be able to acclimate the coral to the intense light.

[jedimasterben]

Originally Posted by perkint

Those luxeon M royals do look good. Shame they are so expensive and hard to find. Can't find any of the 3V ones on 20mm stars

Is it just me being a cabbage or are the only lights on the reefLEDlab site multichip ones?

Tim

Why would you want a 3v model? You'd need a driver that could push 4A to get them to maximum output, which, at best, is rare. The 'standard' 11v Luxeon M needs only 1A to achieve the same output.

[perkint]

Because I'm still playing with ideas and am considering running lower voltage (just about enough to run 4 vero 10s in parallel from one LDD350, so probably under 30V) and have a limited number of channels to play with (I think - again, design is not finalised). Want to run 4 equal panels, so at least 4 of anything I get. And I intend to under drive everything!

If I have the 3V, I can run two per panel (so 8 total, or possibly 3 per panel, 12 total), under driven, all on one channel, to go with the under driven vero 10s (one per panel, 3000K and high CRI), 12 XPE blues (3 per panel) and maybe some UVs (again, not sure what!).

Of course, this is still playing (well, except the vero 10s, bought them and a few spares ) so all plans likely to change. Especially if it means the cost of just one colour (the luxeon Ms) would end up being over $100!!!

Tim

[jedimasterben]

Originally Posted by perkint

Because I'm still playing with ideas and am considering running lower voltage (just about enough to run 4 vero 10s in parallel from one LDD350, so probably under 30V) and have a limited number of channels to play with (I think - again, design is not finalised). Want to run 4 equal panels, so at least 4 of anything I get. And I intend to under drive everything!

If I have the 3V, I can run two per panel (so 8 total, or possibly 3 per panel, 12 total), under driven, all on one channel, to go with the under driven vero 10s (one per panel, 3000K and high CRI), 12 XPE blues (3 per panel) and maybe some UVs (again, not sure what!).

Of course, this is still playing (well, except the vero 10s, bought them and a few spares ) so all plans likely to change. Especially if it means the cost of just one colour (the luxeon Ms) would end up being over $100!!!

Tim

The 3v M is still going to be ~$12 and will emit ~1800mW flux at 1A - which is only ~300mW more than the $3 royal blue Rebel ES at the same current, much less the newer, more efficient Luxeon T, which are the same diodes that the M is based on.

You no longer need low kelvin to have low CRI in the Vero line - the 90CRI minimum line now includes 4000K and 5000K.

I would use either a pair of 12v Luxeon M or a pair of Luxeon K royal blue with the Vero, depending on what current you plan on running them at. The 90CRI 4000K Vero 10 emits ~910 hot lumens at 350mA, you'll want a minimum of ~13,000mW of royal blue to bring the color temperature into the 12-14K range, which is two 12v M at 1A, two Luxeon K 8up at 500mA (which can also be run up to an amp). Or if you're sticking with '3w' LEDs, then you'd need a minimum of 8x when run at an amp, and if you're underdriving them say to 700mA, then you'll need ~11x.

You don't need to underdrive LEDs to keep their lifespan high - you just need to keep them cool. I've ran LEDs at their maximum current (and some far above) and as long as they're kept cool, they'll make their L70 rating (or exceed it).

[karimwassef]

Maybe you can run LEDs hard as long as you keep them cool, but the light output to power consumption curve yields diminishing returns.

[jedimasterben]

At a certain point, of course - but I don't understand purchasing LEDs that are made to be run with enormous output (and have relatively high cost) to hinder them with low current to slightly exceed '3w' class chips, in turn taking many more LEDs to do the job.

The lumen maintenance to 70% of the M is 50K hours when run at 700mA and keeping the die at less than 135C, which, to be frank, should only be achieved with the poorest of cooling. The lumen maintenance data available for the Rebel ES (which has poorer thermal transfer than the M or the T) also shows that above 120C, 500mA and 700mA show no difference in L70, and at 105C, 500mA, 700mA, and 1A all show no difference.

In the past, LEDs were much less efficient at higher currents, and worse still at higher temperatures. LED tech has come a long way in just a few short years, thermal transfer rate has increased, heat tolerance has increased, and efficiency at higher currents has increased significantly, as well as seeing a decrease in forward voltage. Pushing, say, an XR-E, which was near the beginning of LEDs over our tanks, at an amp was really pushing it hard, especially if your cooling wasn't up to snuff - efficiency was ~26% at 25C for the top bin (935mW output, 3.6W input) and ~23% efficiency moving to 85C . Nowadays, pushing a Luxeon T to an amp is nothing for the chip, efficiency remains high at a test temperature of 85C (over 56% for top bin, 1620mW output, 2.86W input).

I'm not saying that if you buy these chips you HAVE to run them to an amp (or whatever maximum current), but with tech where it is now, it is unnecessary to underdrive to remain efficient and sail past the alternatives.

[perkint]

Already bought the veros - got the 3000K 97 CRI ones, which are the coolest 97 CRI they do (I think!).

There are two reasons for looking to under drive them, and that doesn't include the little bonus of increased efficiency (which, as you say, is not massive). Mainly keeping heat production low as they will be going into an enclosed hood (and I hope to have the fans off most of the time), and, when I upgrade in a year or so, I'm hoping I can just swap a few jumpers round on O2Surplus' wonderful driver board, swap a few LDDs and then spread out the 4 clusters to light a much bigger tank

That hope in mind (& yes, I know I'm quite likely to simply build a new light at that stage, as it gives me chance to play again, with the new LEDs which will be out by then! But it's still part of the brief I'm following for now...) I want to build 4 units the same. And I need to keep an eye on light spread to avoid rainbow affect. Shouldn't be too bad, but was (slightly) noticeable just using my test panel (3 XML (1 each of WW, NW & CW), 3 XTE RB and 3 XPE B). Hopefully be reduced when using 4 panels. But does make me think it might be more noticable if using smaller number of light sources (eg 8up luxeon Ks) rather than lots of sources mixed together on 20mm stars.

I will have a look at the Luxeon K and Ts

Tim

[Meddler]

Originally Posted by SoFloReefer

I've never personally used red and green LEDs. I've only used whites: cool, neutral, and warm, and Blues. I've heard of the "disco ball effect" but never seen it and I've built several fixtures. Maybe the solution for you is to go with whites that cover more of that red and green spectrum you are looking for. Also, put the white emitters directly next to blue ones.

There was an article a year or 2 ago covering light-penetrating spectrums for reefs. Without going into details, the verdict was that blue and violet are the highest absorbed colors. 6500K bulbs gave a high-saturation wl of 444.6nm and 5500k bulbs put out the highest wl at 525.2nm. Considering that 400-440nm is the highest rate of absorption for corals, and not wanting to neglect the other spectrums (but wanting to focus where it si needed), I would expect to see lights with a (roughly) following layout:

12 LED Puck:

3x 6500K White (covers lower blue spectrum)
1x 5500K White (covers green spectrum)
2x 460nm Royal Blue (covers mid-blue spectrum)
2x 440nm Bright Blue (covers high blue spectrum)
2x 420 Pure Violet (covers mid-violet spectrum)
1x 400nm Ultra Violet (covers high violet/UV spectrum)
1x 660nm Deep Red (covers low red spectrums)

Certainly, these could change. Some want to add green for coloration. The green spectrum will be made up in the 5500K, which puts out a dominant wavelength of 525.4nm. Some may want more UV for that effect. In the end, I would think that this should be the basis of a light, and we would ADD more LEDs to it to bring out the effect that we want. This way, the corals get their basic requirements, and we get the look we want out of our tank. That being said, that would require a more expensive fixture, as the added LEDs would only be for show, and would provide minimal, albeit some, positive effect on coral growth and health.

[ReefLedLab]

Originally Posted by Meddler

3x 6500K White (covers lower blue spectrum)
1x 5500K White (covers green spectrum)
2x 460nm Royal Blue (covers mid-blue spectrum)
2x 440nm Bright Blue (covers high blue spectrum)
2x 420 Pure Violet (covers mid-violet spectrum)
1x 400nm Ultra Violet (covers high violet/UV spectrum)
1x 660nm Deep Red (covers low red spectrums)

I think, this changes will be useful:
1. Royal blue should have wavelength not more that 450nm, preferably 445-450nm. Longer wavelength is not good for fluorescence.
2. PC Amber should be added to increase visual appearance of "warm" not-fluorescent colors. If you will add PC Amber, please remove from set 5500K LED.
3. 440nm LED looks as very dark blue, i.e. darker than royal blue, but not as dark as violet. Will be better remove this LED and increase 420nm (410-420nm is the best option) up to 4pcs.
4. Please beware with 400nm LED. Some type of it may have significant amount of radiation near to 380nm, this radiation may cause some fluorescence of small particles in water, it is looks very bad. I advice you to use 400-410nm, not shorter wavelength.
5. 660nm LED should be used with care. Too much of radiation with this wavelength may cause unwanted algae. Best options is 630-645nm LED, if you can't find it, please use 620-630nm LED.
6. Blue LED 470-475nm should be added in any case. This wavelength is corrects the lack of this part of spectrum in all types of white LEDs.

There are also several niceties, that may be important for you, please familiarise: www.advancedaquarist.com/2014/4/lighting

[perkint]

So, you would suggest a basis like this?

3x 6500K White
2x 445-450nm Royal Blue
1x PC Amber
4x 410-420nm Pure Violet
?x 400-410 UV
1x 630-645nm Red
?x 470-475nm Blue

Tim

[ReefLedLab]

Mostly yes.

You may choose one of this options:
1. No one 400-410nm LEDs and 4pcs 410-420nm LEDs.
2. Only 1pcs of 400-410nm LEDs and 3pcs 410-420nm LEDs.

Regard to blue LED. Only one LED will enough for this set.

Also I would like to ask you again to read article on AA. That article contain many important points, that should know all DIYers to avoid mistake in construction of own LED fixture.