Minimalistic multichip DIY LED build

[mr.wilson]

As long as they are rated for the same current it is not an electrical issue. However I personaly would want to be able dim those two colors independently of the other. I would also recommend a 620 nm Red over one of a longer wavelenght.

I see this is several lights. First off 1 Watt Emitters are starting to get more rare every day. The trend now is moving to higher wattage emitters like even 10 Watts each. The cost of a difference between a 1 Watt emitter and a 10 Watt emitter is roughly double. So if you compare 100-1 Watt emitters at 30 cents each you get $30 but if you have 10- 10 Watt Emiters at 60 cents each you have $6.00. Now if you with 3 Watt emiters they are just a little more expensive than 1 Watt emiters so 36 of them at 40 cents each costs them $14.40. the only thing holding them back from going with the 10 watt emitters is that all colors are not yet available in the 10 Watt designs.

I find that mixing voltage either makes the 3.4v chips dim or prematurely burns out 2.4v chips. It can also alter the spectrum. With 5 channels, there is no reason to combine green and red. I would mix royal blues with the red, as there are always extra royal blues to spare (dim). I mix the greens with true blue. Again, no great loss if the blues are dimmed to limit the green.

1w LEDs are more efficient (PAR/w) than anything greater. That is what this thread is all about, getting away from 3-5w single emitters. I will admit, I've never heard of 10w single emitters so you may know something I don't know.

Chip reliability and operational cost is more important than saving a few bucks on capital cost.

 

[mr.wilson]

But it's 660nm that's the absorption spectra for chlorophyll C.

Also, (not sure how to multiquote) that seems like a nice controller! May I ask what code you used for the ethernet shield?

There is no chlorophyll C absorption in zooxanthellae. You mean chlorophyll A which has a small peak at 630nm and a significant one at 660nm.

Red light looks good in an LED array if you have lots of colourful corals, but a stark new tank looks pink with red light. It's hard to evaluate an LED array unless you have a fully populated tank to test it with.

 

[asid61]

There is no chlorophyll C absorption in zooxanthellae. You mean chlorophyll A which has a small peak at 630nm and a significant one at 660nm.

Red light looks good in an LED array if you have lots of colourful corals, but a stark new tank looks pink with red light. It's hard to evaluate an LED array unless you have a fully populated tank to test it with.

Sorry, chlorophyll A, my bad.

 

[TropTrea]

I find that mixing voltage either makes the 3.4v chips dim or prematurely burns out 2.4v chips. It can also alter the spectrum. With 5 channels, there is no reason to combine green and red. I would mix royal blues with the red, as there are always extra royal blues to spare (dim). I mix the greens with true blue. Again, no great loss if the blues are dimmed to limit the green..

Mixing boltages is something I have been doing all the time. As long as the current ratings are simular and your not running them beyound the peak it is not an issue. But rememner if you have one chip that is sinking 2 volts and onther sinking 3 colts at 1 Amp then one will be running at 2 wattas while the other is running at 3 watts..

1w LEDs are more efficient (PAR/w) than anything greater. That is what this thread is all about, getting away from 3-5w single emitters. I will admit, I've never heard of 10w single emitters so you may know something I don't know.
..

Yes if look at technology from about 2 years ago this is true you do get more light out of 1 Watt Chip than a 5 Watts chip on a per watt bases. But LED's have come a very long way in the last two years especialy the last 6 months.

Look up the Data Sheet for the XM-L-U2 LED's by Cree. It is pushing close to 150 lumns per watt at 2 Watts, and can out out 950 lumns at 10 Watts. Yes the effectiency drops but this is even true for an 1 watt LED that was mopre effecient at 1/4 watt than it was at 1 watt.

Chip reliability and operational cost is more important than saving a few bucks on capital cost.

If that is your point where do cut that line.
You can use one 10 Watt Chip to get out 950 Lumns for $8.00
You can use 4 -3 Watt Chips to get 950 lumns for $16.00
You can use 10 one watt chips to get 950 luns for $25.00
You can take 40 1 Watt chips and run them at a lower wattage to 950 lumns for $100.00

Each one down the line is slightly more effecient on a Lumns per Watt bases. As you increase effeciency you increase cost. But as far as reliability goes I would put my money of two 10 watts chips running at 5 watts wach would outlast 12 1 watt chips running at 0.83 Watts each. And with the new design the 10 watt combination is probably putting out at least 20% more light.

As I understoodd this tread the idea of the multi emitter was to be able to put different colored emitters closer together on a single mount. This was to avoid the sdisco shadowing effect. It is also used to get light to penetrate deeper in a tank that is more than 24" tall where with single chips it was nearly imp[ossible to get enough light penetration.

 

[Lassef]

Chlorophyll C absorption in zooxanthellae is common. It is chlorophyll B that not has been reported from corals. Chlorophyll C has a very high peak around 445 nm (there is different forms and at least one other has its peak around 430 nm). There is small peaks around 580 and 625. One investigation has shown that the ratio between ChlC and ChlA changes with depth in Seriatopora hystrix

Sincerely Lasse

 

[1SlickFish]

My gosh I want to build something so bad but t5's and halides are still tried and true.

Talk abut chomping at the bit!? :bounce3::bounce1::bounce2:

 

[mr.wilson]

Yes if look at technology from about 2 years ago this is true you do get more light out of 1 Watt Chip than a 5 Watts chip on a per watt bases. But LED's have come a very long way in the last two years especialy the last 6 months.

Look up the Data Sheet for the XM-L-U2 LED's by Cree. It is pushing close to 150 lumns per watt at 2 Watts, and can out out 950 lumns at 10 Watts. Yes the effectiency drops but this is even true for an 1 watt LED that was mopre effecient at 1/4 watt than it was at 1 watt.

I haven't looked at Cree advancement, as they cater to the task lighting industry. They have high lumen per watt ratings, but this is a poor metric for reef aquaria. The PAR per watt is higher in other brands. Cree emitters also have limited colour options compared to other brands.

There are several criteria to efficacy (operational cost vs. capital cost, quantity of light, quality of light, and even distribution). I still think a three channel 10w multichip (9 x 1.1w chips) is more efficient than a single 10w emitter when you weigh all of the parameters. Epileds are available in a broad range of wavelengths and colour mixing will be thorough with three colour families in one multichip (red, white & blue). You could throw a few greens in with the blues and within each colour group, several wavelengths can be represented.

Mixed colour multichips are a more efficient method of delivering blended spectrum. 10w single emitters create even more pronounced colour banding than conventional 1w single emitter "Lite-brite" arrays. I have some LED fixtures with 2w single emitters and the few red LEDs spotlight corals unless you remove the lenses; however, once the lenses are gone, the intensity is lost.

If that is your point where do cut that line.
You can use one 10 Watt Chip to get out 950 Lumns for $8.00
You can use 4 -3 Watt Chips to get 950 lumns for $16.00
You can use 10 one watt chips to get 950 luns for $25.00
You can take 40 1 Watt chips and run them at a lower wattage to 950 lumns for $100.00

I don't think anyone is advocating running any LEDs at full power. The Epileds used in this thread are 2.5w ea. max and are typically driven at 1w ea, then dimmed down from there to get the desired aesthetic. 100w multichips retail for $35 and give you 9000 lumens. You would have to buy 10 of your 10w single emitters and run them at 100% to get the same number of lumens. That would cost $80 without factoring in a PCB or wiring. Alternatively, you could buy a fancy 5 channel multichip for $100 for superior reef lighting.

Each one down the line is slightly more effecient on a Lumns per Watt bases. As you increase effeciency you increase cost. But as far as reliability goes I would put my money of two 10 watts chips running at 5 watts wach would outlast 12 1 watt chips running at 0.83 Watts each. And with the new design the 10 watt combination is probably putting out at least 20% more light.

Again, they are 2.5w chips running at <1w. The new Cree design has more lumens per watt, but that could very well just be yellow/green light created by phosphors that look bright to the human eye. If you look at the Cree MK-R, it offers 200lm/w, but you won't find a royal blue or true blue anywhere near that level. My point is, these are advancements in phosphors for task lighting, not technological advancements that apply to reef lighting.

As I understoodd this tread the idea of the multi emitter was to be able to put different colored emitters closer together on a single mount. This was to avoid the sdisco shadowing effect. It is also used to get light to penetrate deeper in a tank that is more than 24" tall where with single chips it was nearly imp[ossible to get enough light penetration.

Yes, these are some of the main points of the thread. Low cost and efficiency are two of the other threads of this discussion. Sorry if I come across as argumentative. Your contribution to this thread and time are a great asset... it's just your input is geared toward single emitter applications. It's like having someone post "...but t5's and halides are still tried and true" in a multichip fanatic thread

 

[DrawnToBeaches]

I know this isn't related to a reef tank but i'm looking to build a light for a vivarium rack using some 50w emitters, was wondering if I could run more than 1 50w led on each driver. I've built larger lights with smaller leds, using crees, but I'm guessing this is different. any help is appriciated and links to good cheap leds are welcome, as long as theyre more efficient than t5-ho im happy with em.

 

[Lassef]

There has been very few options to run more than 1 of these multi chips (> 30 LED) to one driver. Power LEDs is best to mount in a daisy chain but few driver has its constant current region higher than 50 V. (these chips has a forward Voltage between 30 and 36 V) However - MeanWell has come with drivers that will fit 50 LEDs multi chip with 1 W emitters (typical max 1600 mA as max current for a complete multi chip) in a daisy chain if you accept to run them at 1400 mA. HLG-185H-C1400B is a dimmable driver that probably will manage 3 to 4 50 LEDs chip with 1 W emitters in a daisy chain (Constant Current Region 71-143 V) With HLG-120H-C1400B you can use from 2 to 3 multi chip in a daisy chain. Not more or less in both cases.

The drawback with these drivers is that DC over 50 - 60 V could cause injuryin in a similare way as AC

</SPAN><OSincerely Lasse<O</O

 

[DrawnToBeaches]

Thank you lasse, also if you don't mind me asking what are the chips you used for the reef with the three cylindrical housings? Colors look great

 

[Lassef]

Thank you lasse, also if you don't mind me asking what are the chips you used for the reef with the three cylindrical housings? Colors look great

Orginal 100 LED 5 channels Dream Chip

Sincerely Lasse

 

[TropTrea]

Again, they are 2.5w chips running at <1w. The new Cree design has more lumens per watt, but that could very well just be yellow/green light created by phosphors that look bright to the human eye. If you look at the Cree MK-R, it offers 200lm/w, but you won't find a royal blue or true blue anywhere near that level. My point is, these are advancements in phosphors for task lighting, not technological advancements that apply to reef lighting.

Actualy it is the higher K rating chips of Cree that are producing the greater LUMN rating. It would be great for some of our mind if all the manufacturers listed there chips in PAR ratings but then in reality PAR is not a true breakdown of what our corals need it is just a little better than LUMs. Even with a par meter you can have a higher PAr from a chip strong in Green light that does little compared to to another chip with a higher light concentration in the 480 nm range. You need to look at specific spectrum points to realy compare.

Yes, these are some of the main points of the thread. Low cost and efficiency are two of the other threads of this discussion. Sorry if I come across as argumentative. Your contribution to this thread and time are a great asset... it's just your input is geared toward single emitter applications. It's like having someone post "...but t5's and halides are still tried and true" in a multichip fanatic thread

Yes I agree but I also disagree. Both single emitter and multi emiter do have there advantages in different situations. The shallower an individuals tank is the more advantage the single emiters are at the taller the tank the more the advantage goes to your multi emiters.

In that view I consider the tank height of roughly 24" as the point where it is basicly a toss up between the two and considering build cost I would still go with single emiters. I have run 250 + watts with single emiters on a 120 gallon tank without the need for cooling fans. Can you do that with multi chips? On this and some other threads I see some individuals spending more on chip cooling than the rest of there lighting system.

Single emiters generaly have life expectancies over 20,000 hours when run at max power. Considering we only run the lights 12 hours per day max that means roughly 4.5 years. In 4.5 years will your multi emitters be absolute? The way LED lighting is making changes every day I would say a system 2 years old in either single or multi chip will be absolute.

 

[DrawnToBeaches]

Lasse are the dream chips the $150 from acrc? Do they have a cheaper version cause those are sick, figure I would only need one for a nano

 

[TropTrea]

Lasse are the dream chips the $150 from acrc? Do they have a cheaper version cause those are sick, figure I would only need one for a nano

On a 25 gallon tube thank one of these would be extrem over killl. You would probably end up running it at about 25% of full power, and have roughly $400 invested after the cooling drivers etc. Then there would only a single point light source so you may end up unhappy with the extrem shadows.

For 25 gallons you only need around 60 watts total of LED's at the most if you keep things effecient. I prefer at least 4 light sources hitting any given point to keep shadows from being extreme. So in your case you only need 4- 15 watt light sources. Since there are no 15 watt multichips I know of you would probably go with 4-25 watt multichips running them at roughly 60% of full power. There are some 20 Watt multichips but the problem with them is finding the right spectrum combination considering you will be running them at a higher precentage of full power.

A much easier solution in my mind here are the single emitter chips, for your application. Running roughly 4 neutral whites at about 2.5 watts and then running about 8 Royal Blues, 4 True Blues and 4 Near UV's, at about 1 1/2 Watts each. This is simular to what I run on 40 gallon breeder tanks with the exception of how much power I'm running the chips at. Total build cost would be just a little over $100.00.

 

[noplay180]

I have a 130" long 30" front to back and 24" tall looking to start on my lighting any suggestions also 8 would like to keep a variety of corals

 

[Lassef]

Lasse are the dream chips the $150 from acrc? Do they have a cheaper version cause those are sick, figure I would only need one for a nano

Yes it is from that vendor. I run mine at an input of around 170 W (max 250 W) and use one driver/channel. There was some 5 channels 50 LEDs chip manufactured last autumn (max 120 W in input) - I do not know if they will produce more of them later on. The 100 W Lumía 5.1 is an other option (it is a 50 LEDs solution). I have not tested that chip. Personally, I prefer one more white channel instead of the red/turquoise channel in that chip but I have not seen it in action so I can´t judge if it is good or bad. The AC-RC version and the Lumia is both based on 45 mil emitters (max input 2.5 W).

I´m very pleased with my chips but it is absolutely not for free. I estimate that each cannon cost me around 350 $.

I have three to a 120 gallon tank.

Sincerely Lasse

 

[DrawnToBeaches]

I don't see it as overkill especially for a reef, I've been running 18 xml over a 50 cube fresh water, near 80% power, thing is I have them 36" from the substrate, i would do something similar with a cannon, the extreme shadowing I havent noticed with my cannons. also your estimation of cost is extremely off, I have a pcu cooler I got for about $40 I have the perfect drivers laying around which cost me about 8 each, and the power supply I would need is only $45, then the Led is 150, the cost isnt that bad to get a full spectrum powerhouse of a cannon for any column tank or cube. if you have experienced hard shadows you either place the light way too close to the water surface or you have a layout that caused it. I look forward to trying something like the dream chip but I may wait for a few months for the price to come down a little. would two 50w RB and one WW suffice for a cube reef? I'll be running them at about 75% with dimmers might add violet to it later.

 

[DrawnToBeaches]

Hey lasse, Your tank is awesome, those cannons make it look incredible. very well done on it i hope some day i might make something similar just a little smaller =]

 

[Lassef]

Hey lasse, Your tank is awesome, those cannons make it look incredible. very well done on it i hope some day i might make something similar just a little smaller =]

Thank you. My humble opine is that you can´t get to much of light but you have to slowly adapt the corals and have a high grade of water movement. I have mainly softies - but light for SPS - and my corals are growing a lot. I have a high intensity light only for 4 - 5 hours and a total light period of 12 hours - the normal high light period at a reef in nature. My figures is 300 - 500 PAR as max at the bottom (60 cm of water).

Sincerely Lasse

 

[mr.wilson]

Actualy it is the higher K rating chips of Cree that are producing the greater LUMN rating. It would be great for some of our mind if all the manufacturers listed there chips in PAR ratings but then in reality PAR is not a true breakdown of what our corals need it is just a little better than LUMs. Even with a par meter you can have a higher PAr from a chip strong in Green light that does little compared to to another chip with a higher light concentration in the 480 nm range. You need to look at specific spectrum points to realy compare.

I agree PAR (µmols) are not an ideal metric, but it's a decent indicator, and much better than lumens. Of course, one needs to look at the quality of the light via spectrograph, and quantity of said light via quantum meter (PAR). This will give you a rough idea of PUR, providing you know what each individual coral requites. Again not a magic number, but as close as we are going to get.

Yes I agree but I also disagree. Both single emitter and multi emiter do have there advantages in different situations. The shallower an individuals tank is the more advantage the single emiters are at the taller the tank the more the advantage goes to your multi emiters.

In that view I consider the tank height of roughly 24" as the point where it is basicly a toss up between the two and considering build cost I would still go with single emiters. I have run 250 + watts with single emiters on a 120 gallon tank without the need for cooling fans. Can you do that with multi chips? On this and some other threads I see some individuals spending more on chip cooling than the rest of there lighting system.

The application where a single emitter would be advantageous is a nano tank with a lid, as clearance is minimal for cooling and illumination angles. A 50-100w multichip would have to be run without a lens (instead, a flat glass shield to keep clean/dry) to cover the tank adequately without hard shadows. A grouping of 10w 3 channel multichips would avoid the pronounced colour banding you get with single emitters.

If you don't mind a giant heat sink, you can cool a 500w multichip without fans The common limitations of coolers are fan noise and heat sink height. Price is not a problem if you are buying online.

Single emiters generaly have life expectancies over 20,000 hours when run at max power. Considering we only run the lights 12 hours per day max that means roughly 4.5 years. In 4.5 years will your multi emitters be absolute? The way LED lighting is making changes every day I would say a system 2 years old in either single or multi chip will be absolute.

I agree 100%. I have been working on a multichip fixture for over two years and I still haven't caught up with the newest innovations. When PFO was selling fixtures for $4000, you had to amortize the cost over ten years. Now that we are looking at $500 fixtures, 2-3 years is acceptable.

Most hobbyists spend $5000-$30000 on their complete reefs. When lighting was in the 4 digits, it was a careful decision, but now it matters little whether it is $500 or $750 per fixture, especially when your tank only needs one. Capital cost isn't a major concern outside of this DIY forum.

Efficiency was a big deal when we were using three 400w metal halides and another 440w of T5. Dropping down to three 100w multichips was a huge electrical savings even without looking at chillers, evaporation and bulb replacement. Now that many have switched over to LED, it matters little whether the fixtures are 100w or 150w each. We are talking about a few dollars a month in operating cost. Okay, so efficiency isn't a major consideration either.

Longevity is the other issue to look at. All of these lights we are discussing will be in the dump or our garages in 2-3 years, so whether it's 20,000 or 50,000hrs life span, it's moot. So let's take longevity off the list of parameters in selecting a good light fixture.

So what's left...

1) It has to be new
2) It has to look cool
3) The cost has to be low enough not to set off alarms with wife/girlfriend (or both)
4) The fixture has to involve the user in some way through building or programming
5) For most consumers it has to have the nod of acceptance form the forum members and reef bloggers