Minimalistic multichip DIY LED build

Hello,

Taking advantage of your experience, i just want to ask all of you, what do you think about 3 dreamchips 90W led in a 53'' x 30'' x 24'' tank, my question is to do it with only 3 or its better have 4 dreamchips?
 
Hello Everyone,

I'm in the process of setting up my 150g 48L"x24W"x30H"....After browsing through the first 30+ pages. Multichips look amazing now reading the back 20+ pages I'm lost as to what chips I should buy as well as the drivers. I work on IT so i mess with computers and heatsinks everyday. I even thought about water cooling the chips :) If some one could help that would be amazing! Goal is to support a mixed reef. I was thinking 2x 100W Cool White + Warm, 2x 50w Royal Blue, Not exactly sure on what to pick or if I should buy the white/blue combo chips :| please help!!
 
Anyone running the LDD solution?
To power 6 LEDs would I need 30 LDDs? I'm planning on controlling them through a jarduino.
Anyone recommend a suitable power supply for this? I'm assuming a 1000w supply would be enough?
 
/Thanks for that, planning on getting the 5up boards, i take it these need to be made by someone or is anyone actively selling them? I already have the dream chips as i was part of the original buy, i have one set up over my mums tank and although its only a softy tank the growth has been unreal.

I take it the LDDs are all powered up in parallel from the power supply? Not sure how many PWM outputs the jardiuno offers may have to run each channel in parallel if there isnt enough (which im pretty sure there isnt).

I'm planning on running these about 700ma per LLD i've had good success with this on my mums tank.
 
I'm using the dream chips with each chip driven by 5x LDD 1000 (ran at 700ma max).

My led strings data is :

Forward voltage/channel: DC 31.5V-36VForward
current/channel: 1400mA
maxRated power: 252W
maxViewing Angle: 140 DegreeDimension: 56mm x 40mm x 4.3mm
Working temp: <70°C

Did a quick calc:

Each led has 5 channels with the data above being the max.

5 x 36v x 0.7mA x 6 leds x 1.25 =945w

So looking at the available power supply a SE-1000-48 would fit the bill, but looking at the data sheet it says the inrush current is 55amps? Won't this trip my breaker? Also when it's up and running does it draw 10 amps at 230v? Does running on 240v here in UK make much of a difference?

Thanks in advance.
 
Have you thought of running two even three PS. That would make them 450/500 or 300/350W PS's. Then if/when a PS fails you won't have total lights out.---Rick
 
Ahhh good idea rick. Would actually work well having two 500w ones like u said for contingency! I will try and get to sleep again now lol :thumbup:
 
61, Do you know what led's they are using to make these? They are not providing any info. fv max A ect.---Rick

Nope. Haven't emailed them either. Just generic leds. You could email them if you are interested in the multichips they sell, and you will need to anyway to specify what leds you want.
fV is probably around 3.6v, and current would be 500ma max.
 
I had folowing the same path, I got my first XM-L before I startef my 5 channels build. I have other reasons to use 5 channels multi chips. At a small area you have 100 LEDs and it blends different wavelengths very well - no disco. With 5 channels you can mix different wavelengths the way you want and you can control them.

Talking about efficiency. Everyone talk about how good XM-L is - but no one tells the true story. It says that it has an efficiency > 140 Lumen / watt - that impress every one. The true is that this is only valid for 700 mA and for certain colours and BINs. At 3000 mA (10 W) is the efficiency of the best XM-L not better than a Epistar running at 700 mA (around 100 lumen/W) At 700 mA -> 146 lumen/W, at 1000 mA ->139 lumen/ Watt, at 1500 mA -> 126 lumen/watt, at 2000 mA -> 115 lumen/watt. This is valid for the best - cool white U2. Cree itself gives no figures for 2500 mA and 3000 mA. For neutral white T5 is the efficiency at 2000 mA -> 100 lumen/watt. 90 CRI - white S6 is the efficiency at 2000 mA only 70 lumen/W. All figures calculated from this

With my multi chip - I have not measured any lumen/watt but I have measured PAR per insert watt and it shows a similar pattern with increasing amperage.

Sincerely Lasse

What you say about the XM-L is true, if running them at a high amperage they are less efficient than running some lower powered LEDs at a much lower current.
You have the choice of low running cost or low set up costs.
For the best bang for your $ I actually prefer the Cree XT-E.

It is not easy to compare the lumen per watt levels of an Epistar LED to that of a Cree as Epistar (so far as I can tell) don't supply data for white LEDs, only the royal blues which are converted to white with a phosphor.
The lumen per watt levels supplied by the finished multi-chip product will depend on the phosphor used (among other factors).

If trying to compare the efficiency of these multi-chip LEDs to a Cree its far easier to ignore the phosphor and stick to comparing the output of royal blue LEDs.

According to the AC-RC supplied data sheet on the Epistar LEDs used in their dream chip, the royals used to make them can produce between 360 and 400mw @ 350ma ( I can't find details of such an efficient chip on the Epistar website, 295mw seems to be the most efficient mentioned there).

The Q flux bin Cree XT-E produces a minimum of 600mw at 350ma, its minimum rating is 50% higher than the max quoted by AC-RC.

At 700ma the Epistar be running at its maximum current (never a good thing) and will produce a max of 720 mw.

At 700ma the Cree XT-E will be run at less than half its rated power and will produce a minimum of 1050 mw

If you take into account that the Epistar figures are for an LED run at a temperature of 25c and the Cree at 85c then things swing even further in favour of the Cree.
There is a graph provided by Epistar showing the drop in output at higher temperatures but it only goes up to 80c, at which point the chip has lost around 10%, so now at around 650mm @700ma.

It looks to me that at the same drive current a high rated XT-E based system is over 50% more efficient than the Dream chip.

At 1A the XT-E will produce 1380mw, on average consuming 3.15 watts, or 438mw per watt.
At 700ma the Epistar will consume 2.45 watts, providing 265mw per watt.

Even at the maximum rated current of 1.5A the XT-E will produce a minimum of 354 mw per watt.
 
A note:
I recently made a thread about the benefits of the Cree XM-L and XM-L2 versus the Luxeon M, and the XM-L2 won. There is also a place to get them for $5 each.

On the subject of Epistar:
not all sellers claiming to use Epistar are doing so. So even though it seems like the Cree is only 50% more powerful, it might be more than that.
I would go with the Epistar datasheet over the ac-rc website.
 
A note:
I recently made a thread about the benefits of the Cree XM-L and XM-L2 versus the Luxeon M, and the XM-L2 won. There is also a place to get them for $5 each.

On the subject of Epistar:
not all sellers claiming to use Epistar are doing so. So even though it seems like the Cree is only 50% more powerful, it might be more than that.
I would go with the Epistar datasheet over the ac-rc website.

except the XML aren't available in royal blue, at least I've not seen one, for whites, I've not read your thread yet, but I still see reasons that either one would be favorable over the other, depending on the specifics of the project.........
 
except the XML aren't available in royal blue, at least I've not seen one, for whites, I've not read your thread yet, but I still see reasons that either one would be favorable over the other, depending on the specifics of the project.........

True. The Luxeon M still win for royal blue.
 
Some over all comparison notes from the prior posts.

As far as the Epstar chips go I have noticed a big differences between different batches I had bought from the same vendor at different times roughly 6 months apart. Upon close examination under magnification I doubtt that these chips were made on the same production line. So I realy think there has to be some buyer be aware warning here.

When comparing chips you especialy in the Blue chips you also need to be aware of the peal wave lenght as well as the band width of the light they are emmiting. I have seen Roay blue chips advertised as peaking anywhere from 450nm to 465nm. I have also seen blue chips that are advertised as peaking anywhere from 460 nm to 480 nm. As an example an XT-E 04 Bin has a range of 450nm to 455nm while an 06 bil has a range between 460 and 465 nm. Very seldom does a venodor specify exactly what bin niomber the particular chip is from.

The other point is the width of the peak. Depending on the chip the peak emitted light might be 460 nm however if were to measure the light emoitted at 455nm and at 465 nm the quantity of light might be 50% of what is emitted at 460nm or it could be as low as 15%. Dependent upon the application a wider or narrower bandwidth would be more effecient. In the case of our Reef applications the wider band width covers more frequencies.

As far as effeciency goes look a a comparison here
XP-G XM-L XPG Wattage XML Wattage XPG-effeciency XML effeciency
350 ma 127 lm 143 lm 1 Watts 1 Watts 127 143
700 ma 228 lm 260 lm 2.1 Watts 2.0 Watts 108 130
1050 ma 318 lm 365 lm 3.3 Amps 3.1 Watts 96 117
1400 ma 400 lm 481 lm 4.6 Amps 4.3 Watts 87 112
1750 ma 575 lm 5.5 Watts 104
2100 ma 663 lm 6.8 Watts 97.5
2450 ma 728 lm 8 Watts 91
2800 ma 806 lm 9.3 Watts 86.7

Yes as you can see the higher the wattage the chip is running at the less effecient it is. But the question remains is is better to run more chips or less chips since our goal is a set amount of light. So if you were looking at a need of 2,000 lm you can aproach it in many different ways.

3 XML chips giving off 666.7 lms each at roughly 7 Watts each for a total of 21 Watts
4 XML chips giving off 500 lms each at roughly 5 Watts each for a total of 20 Watts
5 XML chips giving off 400 lms each at roughly 3.75 Watts each for a total of 18.75 Watts
14 XML chips giving off 142 lms each at roughly 1 Watt each for a total of 14 Watts.

Comparitive build cost would be $49.00 for 3 chips compared to $110 for the 14 chips.


While I only compared one manufacturer of white LED's this could apply to any color family or any manufacturers LED's.
 
Interesting chart Dennis. From what i've seen with to many commercial fixtures, the trouble running fewer chips is spotlighting. They don't blend with the other chips very well. The aquascape ends up resembling a stage show. The sand becomes a mirror image of looking at the underside of the fixture! Better optics in tune with fixture distances to the tank help but i'm still seeing more of it than i find acceptable. Whereas a fixture like the Fluval Sea LED fixture despite it's shortcomings, array of ~ 500+ surface mounted diodes, white, red, royl blue, blue, near uv, from what i've seen of it, blend seamlessly to look like one natural light source spectrum wise. On the higher power/fewer chips end of it IMO, Ecotech Radion doesn't do a bad job either.
In the past year to present i've seen several members of my local reef club switch from LED fixtures back to T5's and Halides. Honestly, i have to say i may've too had i sunk money into the "spotlight" types so prevalent in the marketplace.
Color spotlighting is also a problem i see with the "LED Cannons" which undeniably are great for deep tanks. I'm curious what others here have done to reduce the effect, i'd guess sticking with one type say 12K or 20K high power multi chip?
 
Back
Top