PDA

View Full Version : Light



Eclip
06-04-2010, 12:11 AM
Lets talk light for a minute.
Now light coming from a globe isnt dependant on the power is uses ie a 25W T5 isnt giving out 25Watts of light this is meaningless as light is measured by the amount of photons (packets of light) per amount of time another measure is lumans output

This equivalent part is what gets me 1Watt isnt the light its putting out its the power its consuming.
most T4 lights are approximatly 60-70 lumans per watt (some new ones up to 90lumans/watt)
most T5 are around 70-80
CFL are around 50-65 but are cheap usually in comparison
MH are aroudd 60-100ish (higher wattagaes are more efficient but cost more to run)
LED are 110-136lumans per watt (these have their own draw backs aswell)

Geneal rule for scrubbing
0.5 actual (not equivalent) fluorescent watts per gallon MINIMUM.
1.0 actual (not equivalent) fluorescent watts per gallon for HIGH filtering.
now watts is what the light unit uses to gennerate its light output which for light is lumans (packets of light per unit time) which is what we are interested in.

So in the guide it says not equivalent fluroesscent watt and as far as i can tell you based it on T5 lighting and that you cant have too much light essentially but cost of running lights and effective scrubbing are two very different things.
Therefore assuming 70-80lumans per watt we are using T5's as the SM100 puts out 7-8000lumans as its 100watt total. (assuming good reflectors so that close to 100% of the light emitted is directed to the screen)

This is for high filtering (1W per square inch)
The screen is 200cubic inches so essentially we have 40 lumans per square inch of screen.
2X100 square inches = 200 inches^2 total
8000/200 = 40 lumans/square inch.
so 40 lumans per squareinch of screen has been found to be high level of scrubbing by your design which speaks for itself in effectivness.

Now lets look at it from a tank volume point of view. tank volume is 100 gallons so thats an easy 70-80 lumans per gallon for good filtering

common sizes of tank and light needs
28 gal needs 2000lumans at 1.5inches away form screen
50 gal needs 3500 lumans at 1.5inches
100gal needs 7000 lumans at 1.5inches.
150gal needs 10500 lumans at 1.5 inches .
500gal needs 35,000 lumans at 1.5inches.

Lets look at Lux or light power per unit area
Now lux (SI) or footcandle (IMperial) is the light power per unit area.
40 lumans on 1 square inch (0.000625m^2) is 64,000 lux (close to 6000 foot candle) as a rough guide to the amount of light you should be aiming for for good filtering. As a guide direct sulight starts at 32,000 lux and goes up to 130,000lux so the light needed is around 1/2 the power of the sun ( this doesnt include infared light which is the power of the sun that we feel its the sun in light intensity)

Lux depends on the area we are illuminating. Decent reflectors are a way to focus the light but they can only do so much so placing the light source close to the screen is a good way to increase effective irridance. look at reflectors that throw the light to the entire screen surface while being as close as possible to the screen to give uniform light distribution. but being in this 1-2inch range is where the aquerium reflectors ad DIY reflectors are going to be able to throw most of the light effectivly.

I hope this helps a little to the knowledge of light needs and hope it helps future development of lighting or design.

rygh
06-04-2010, 10:00 AM
Very good analysis.

And I like the last part about emphasizing Lux.
That is really the critical scientific measurement to use : Lux value at the screen.
Everything else has way too many variables and can be almost meaningless.

Unfortunately, the math does not match what I measure. And I think the problem is that
your assumption of near 100% is pretty far off.
I checked your calculations quickly, and it does seem that a T5 at one inch would be 64,000 lux.
From memory (I need to check again), my measurement was something like 25,000
That was with the best available refector, and an over-driving icecap ballast.
But it was with a 1/8" acrylic shield. Reflectors are far from perfect. Specs are not always what they say.

The second factor, which is excruciatingly hard to quantify, is spectrum.
Algae use a completely different spectrum than we, or our normal light meters, use.
If you match that, you can get a huge increase in efficiency.
Incidentally what I really hope to achieve with LEDs.

I also want to mention the value of optics.
With a point source of light, the lux drops as the cube of the distance.
With a linear source of light, it drops as the square.
But a focused beam (laser) drops barely at all.
Basically, with a cheap CFL, no reflector, if you move it from 1" away to 4" away, the lux drops 1/64.
But with a focused beam, say an LED with a 30-deg optic on it, the difference is almost nothing.

Maybe someone can measure the lux of our gold-standard SM100, right at the screen.

SantaMonica
06-04-2010, 10:49 AM
Come over and measure :)

Eclip
06-06-2010, 07:53 PM
Thanks for some solid data on the lux findings you are getting.

There are some funny things with lux in that it is a measure of the light the human eye precieves in that it is weighted toward the yellow/green part of the spectrum. a PAR metre is weighted for what plants see. they are distinctly different as can be seen below in the picture
http://www.americanaquariumproducts.com/images/graphics/par4.jpg

A is what the PAR meter measures and is the wavelegths needed by plants(coral, photosynthesis etc) to provide food for themselves.
B is what the human eye see's and lumans is weighted the same as LUX (what the eye see's)

That is why LED for instance give us very good PAR regulated output now and thus needs to output more light that we are not sensitive too to get similar lumans to CFL since CFL have alot more green/yellow light emitted then LED's.

LED's at 100lumans would give out more PAR then T5 fluro's at 100 lumans with everything else being equal like LUX etc because of the weighting oof the light components. this is why it is easier to make high lumans yellow light as that what its measured on. and why you dont see lumans values for blue and red LED's etc as the weighting is so low on those colours.

Polycarbonate is very good at transmitting light only a few %loss similar with acrylic being a % or more loss.

Algae as far as i can tell form what little data is available is that it makes use of the light spectrum that most plants dont use ie the 500-600nm green yellow part of the spectrum and one of the biggest reason for the push to LED for Display tanks is that is supresses the liklyhood of algae growth as it limits those wavelengths.

What is needed is to find out what type of algae we are actually growing im assuming just the nice green hair algae and then do some experiments on that with different spectrum weighted lights.

This could give a massive increase in efficiency. i have seen somewhere possible SM said that 3200K globes work the best. If you can remember the brand of these globes SM or whoever said it i'll look up some output tables and do some analysis.

last thing is the algae that grows is green hair algae im assuming?

SantaMonica
06-06-2010, 09:33 PM
Green hair is best, but other types work too.

rygh
06-07-2010, 10:00 PM
That "A" curve for what algae needs looks fairly incorrect. Where did you get that?
It seems far too high in the center green regions.

Some good threads with a lot of data (biased, since I started a few)
viewtopic.php?f=3&t=435 (http://www.algaescrubber.net/forums/viewtopic.php?f=3&t=435)
viewtopic.php?f=3&t=560 (http://www.algaescrubber.net/forums/viewtopic.php?f=3&t=560)
viewtopic.php?f=3&t=561 (http://www.algaescrubber.net/forums/viewtopic.php?f=3&t=561)

Several of us have been experimenting a lot with different LED types.
Still no real agreement.
Seem like a combo of deep red + blue + some warm-white is good though.

Eclip
06-08-2010, 07:27 PM
The curve isnt for algae just a comparison from light plants need vs what the human eye sees (from which lux and lumans are deterimened) i believe that the A is just a averaged result form all the photon aceptors in plants (chlorophyll etc etc).

Upon furthur looking it seams as if all green algaes make their food and thus photosynthesise via chlorophyll and thus the non abosrption of green light which gives it its green colour.

the spectrums of absorption for chlorophyll A and B are as follows
http://upload.wikimedia.org/wikipedia/commons/7/73/Chlorophyll_ab_spectra.png
chlorophyll a has approximate absorbance maxima of 430 nm and 662 nm, while chlorophyll b has approximate maxima of 453 nm and 642 nm

So if we can provide light in the 400-500nm as well as 620-680nm region it should be able to get all of its light required. using these two regions will cause the algae to grow the best. Releative equal amounts of these two bands will give the best growth rate. While not so crash hot for DT lighting it would promote good growth.

Spectral distribution.
A far as i can tell form Cree's specs the XP-G that is common for LED lighting has some of the peaks covered but to reach the 450-500nm range some royal blues and blues are need to beef up that part of the spectrum and there is no 630+nm led that cree make to boost that part of the spectrum so that mite be difficult but the rest seams to be very well covered. the fluroessence of chlorophyll a is about 670 so soe of the light it acepts it mite be able to break down and make it useful.

grow lights made for hydroponics use a fair amount of red LED's mixed with blues to make the plants grow the best. i need to look at some more LED manufactures to see if any make long wavelength LEDs though

Eclip
06-08-2010, 11:27 PM
Hm reading through 22 pages of that LED thread shows some very good advancement and data found out.

have you increased your flow to your led screen yet rygh and has it helped with getting your screen green as opposed to yellow?
like you said there are ppl using twice the power in LED's that you are using and they are getting green so SM's assumption of flow mite be the crutial thing. afterall flow in a natural reef is 10 fold what we replicate in a tank so maybe that really is the key.

the deep red plus blue spread is looking good. even if you are producing yellow algae it still seams to be filtering very well.

rygh
06-09-2010, 11:58 AM
The trick to making it green instead of yellow was to reduce the lighting.
(Also advice by SM)
Lighting is at 12 hours/day, and there are some duct-tape strips for further reduction.
Increasing flow had no effect.

The 30W of LEDs was just too powerful. Cool!

It is doing fairly well now as the only filter in the system. The skimmer and RDSB have been removed. Extra feeding as well.

The problem I have is that it cannot really out-compete the purple bubble algae in the main tank.
Really seems to be a function of the horizontal design.
So I am building a vertical version. Also needed for a future larger tank.

Gigaah
06-18-2010, 10:01 PM
Manufacturers make any wave length you want. The problem is cost at that point once you wander out of the "main stream" All mainstream blues are going to be between 450-470 which means the blue spike on the white LED's from that manufacturer are also going to be in the same spot since white LED's are blue LED's modified with phosphors.

I've used just warm white 3200k. If you've looked at the graphs Cree has up you can see the spectrum output of them..and all LED's output is roughly the same per given Kelvin rating.I too was looking at and thinking about trying to hit all these chlorophyl absorbtion peaks..but after getting results with the 1 watt warm white only I just dropped it. Trust me I spent weeks pouring over graphs and information on algae its sick.

They reasonably aproximate the output of a low kelvin CFL.I used 1 watt LED's to get the light spread more evenly and at the right intensity. You can also get cheap 1w LED's off ebay that are like 70lumens per watt. roughly 50-70 cents each if you can deal with LED's that are not star mounted. about $1.50-$2 each if they are star mounted. You can buy either in lots of 10-20-50-100 etc and save a great deal of money.

rygh
06-20-2010, 03:02 PM
I had warm-white at first.
I switched to two deep red + 1 blue, and it definitely improved the growth.
But - was it enough to warrant the hassle? Hard to say.

I now need to make 4 x 30W sets for the big tank coming, so some decisions to make.
Hmm.

Gigaah
06-20-2010, 04:32 PM
I'd determine how many lumens those 4x30w will get you. Probably..8000 lumens.

Go straight to ebay.
Here is the cheapest LED solution I've found.
http://cgi.ebay.com/50pcs-1W-Super-Brig ... 2559c5df14 (http://cgi.ebay.com/50pcs-1W-Super-Bright-High-Power-Warm-White-LED-70LM-w-/160419929876?cmd=ViewItem&pt=LH_DefaultDomain_0&hash=item2559c5df14)

I can't find the drivers I purchased but they were $1 each when i purchased 30 of them. they were 1wx3 with 12vdc in. So for 30$ you can power 60 LED's.

EDIt* here is pretty close. fyi they work at 12v in

http://cgi.ebay.com/20x-Power-Driver-3- ... 4cef8dd12a (http://cgi.ebay.com/20x-Power-Driver-3-Watt-3x1-LED-Light-Lamp-12V-MR16-/330436563242?cmd=ViewItem&pt=LH_DefaultDomain_0&hash=item4cef8dd12a)


Overall i think 1w LED's are more suitable to the application than 3w. They are easier to heatsink too. I have a tank lit with 1w LED. I use a 4' 1" square aluminum tube. On the end of the tube I fixed on a 40mm fan that blows air thru the tube and everything works good.

rygh
06-20-2010, 07:30 PM
Well, I must admit that is really cheap. Wow.
Basically 50W of LEDs for $18. Amazing. Well, $20 shipping though.

The downsides:
* Only 80 lumens/watt. Not bad at all, but not fantastic.
* Totally unknown brand, on ebay. Caveat Emptor.
* Good, but not optimized wavelength.

A lot depends on how long they will be run.
Energy cost and longevity differences usually dominate those initial cost savings.

Gigaah
06-20-2010, 07:42 PM
Yeah..they are definatly not the most efficient LED out there. I have some and are using them. They work. Id unno what more to say. They are the cheapest per lumen. If kept cool enough they should last quite awhile. I guess if you lose one they are pretty cheap:)

I know they could be lower K. If you dig around ebay you might be able to find a better color but I've searched high and low and thats probably the lowest cost you'll get in the warm LED range. The same guy sells a batch of 100 Regular cool white for like 55$ shipped. I'll bet if you message the guy he'll put together a batch of the warms for about that price.

They are not star mounted but I didn't have a problem with that. If you need star mounted your looking at a good 1.50 each if i recall right.

To be honest. Do i think they will last as long as a high quality cree..probably not. Are they ass efficient. No. Do I think given all these they are more cost effective over flourecent given potential longevity? Yes.

rygh
06-21-2010, 01:18 PM
You make some good points.
I can replace them several times for the cost of a Cree.
And sure would save some money if they do work well.

What I might do is mix lots of those, with a few deep reds.

How hard was it dealing without the star mounting?
Is the bottom electrically insulated?
Can you simply bend up the tabs solder wires to those, and glue the base directly to the heat sink?

Did you ever find a data sheet on them?

Eclip
06-24-2010, 11:42 PM
Sorry been away havent had time to check in on my OP. some interesting points. as far as i can tell form reading all the LED posts is that there are only a few of you at different stages of testing.

The very sucessful one that i have found is inkidu with DIY LED's and he used warm whites, deep 680nm red and 3W blues. produced nice green after 3 days on a long screen and that the side with the deep reds was better.

Gigaah you said you got results with the 1W warm whites im assuming that is what you are using now to get green growth? and do you have a spectral graph for them?

the problem with assuming the 8000lumans(lumans are human eye weighted) is that its not acurate for plants and why you find blue led's dont have any lumans/watt rating they have power outputs. The algae uses light that is usually outside our optimum viewing which makes the weighting and luman factor horrible so luman ratings mean nothing.

Up till Now
So from what rygh, inkidu, and gigaah have witness'ed there is growth with just using warm whites which i exect is because of the higher distribution of red then cool whites and there is the blue in its output to allow some growth then the cool whites. inkidu found that warm white mixed with some more blue + deep red (680nm) gives good results as seen from his picture of his scrubber after 3 days. (higher amount of red helped alot not sure if its due tot he dep red or not but probable)
http://www.algaescrubber.net/forums/download/file.php?id=673
This seams to be really good growth

rygh has found that just blue plus red(just red or deep red or a mixture?) works. Do you have any pictures of this i can see. We know this is true as the 50W ebay 8mm LED's that waucedah_joe used which seamed to work really well giving great algae biomass. the specs for those LED pannels are as follows
# LED Bulbs Power: 50watts. (40x Blue LEDs + 72x Red LEDs)
# Red LED wavelength: 620nm-630nm
# Blue LED wavelength: 460nm-470nm
# Red LED Bulb Lumen: 15LM x 72 Red LEDs
# Blue LED Bulb Lumen: 4LM x 40 Blue LEDs

CREE LED's are royal blue 450-465nm
blue 465-485nm
red 620-630nm

From some of my own calcs the 50W pannel can be replaced with 30W of cree now that 50W pannel did a 12x12inch square with another inch around the outside with ok growth leading to a 13"x13" square or between 144 and 169sqinches of area. one sided ofc. this works out to be only 0.32W per sqinch for good growth. It must have been hitting the right wavelengths or at least getting very close to promote green algae growth.
The problem i see with the cree LED's are that the blue is a fraction on the low side and may need some blue mixed in to cover a better blue band in comparison but this mite be good or bad its hard to tell without experimenting with both. what is obvious is that it needs the red+blue. The broard distrubution of light from the warm whites may help but is the extra light needed well im not sure but i guess having small amounts of light across the whole range can only help with the other pigments and compounds eg. carrotoniods etc.

What i think needs fine tuning is the amount of red to blue and experimentation with the deep red to red and blue to royal blue combination. From some calcs of my own i have found that with cree LED's that XPE royal blue to red should be in a ratio of 3:5 to@700mA to get very similar to the 50W grow pannel at only 30W. Essentially the SM100 could be powered with 30W or slightly less of LED's if the ratio can be found! we are close but not quite there i dont think, the grow pannel is effective and we cna mimmick but im not sure if its quite the right ratios almost i think. replicating the grow pannel would cost around $100 plus heatsinks.

Does anyone know a good place to order well priced heatsinks from? im in australia but there was mention of using just square tube with a fan blowing through it?

Sorry for the long winded post but i thort i should do a little summing up since reading all the nice development thraeds about LED's. i will start to tinking and check in the coming month or so...

Gigaah
06-25-2010, 01:57 AM
I do not have a spectral graph for those LED's. However I've looked at enough spectral graphs to see most knock offs pretty closely follow cree's phosphors and havn't seen any significant deviation from cree warm white to cheapo no name warm whites.

I'm not currently useing them for algae scrubber. Those LED's were used as my experiment. The LED's were intended for and are not in use as tank lights instead. Warmwhite + blue + ultraviolet FTW

rygh
06-25-2010, 01:27 PM
There are some old pics here:
viewtopic.php?f=3&t=509&start=10 (http://www.algaescrubber.net/forums/viewtopic.php?f=3&t=509&start=10)
But I will try to take some new ones. About the same, but darker green.

I have built a heat sink with a piece of cheap aluminum angle iron. Home-depot stuff.
See details here.
viewtopic.php?f=3&t=435&start=0 (http://www.algaescrubber.net/forums/viewtopic.php?f=3&t=435&start=0)
It worked OK, but not fantastic.
For the second rev, I ended up using these. You can put quite a few watts on them.
http://www.allelectronics.com/make-a-st ... OID/1.html (http://www.allelectronics.com/make-a-store/item/HS-128/HEAVY-DUTY-ALUMINUM-HEATSINK-TRAPEZOID/1.html)
For my next rev, I will probably scour some local stores (silicon valley - lucky me) for surplus CPU heat sinks.

Question for Gigaah:
How hard was it dealing with bare LEDs, without the star mounting?
Is the bottom electrically insulated?
Can you simply bend up the tabs solder wires to those, and glue the base directly to the heat sink?

rygh
06-27-2010, 09:29 PM
As requested, some quick pics of LED based growth.
About 5 days worth. With my horizontal design, seems to need to be cleaned more often.

SantaMonica
06-27-2010, 11:45 PM
Looks really good.

Gigaah
06-28-2010, 12:57 AM
bend solder tabs up and have at it. I usually insulate under the solder tabs as well. With some thin plastics(like the stuff most consumer packaging is made of).

rygh
06-28-2010, 10:19 AM
Sounds good. Sure would save a lot of money.

Eclip
06-28-2010, 11:44 PM
looking good rygh i assume each algae pannel is 30W worth of LED's on those screens?

rygh
06-29-2010, 10:59 AM
To be specific: It is 30W total for all 3 screens.
And only on for 12 hours/day, or it tends to fry the algae.

The build thread is here.
viewtopic.php?f=3&t=509 (http://www.algaescrubber.net/forums/viewtopic.php?f=3&t=509)

Good is relative though. It certainly grows algae, and keeps N/P to 0.
But it is not efficient enough to win over the bubble algae.
I am no longer a fan of horizontal designs.

Gigaah
06-29-2010, 12:16 PM
Hey Rygh can you give me a quick idea as to the lumens or atleast wattage of LED you used over what sq in area?

rygh
06-29-2010, 12:31 PM
It is 30W over a 195 sq in single sided screen.
Lumens does not really work well, since it is mostly red + blue, but noted below.

8 x Deep Red LedEngin. LZ1-00R205
- Running 700 mA (low power), 2.9Vf = 2W each
- About 100 lumens, but not very accurate for deep red.
- About 400mW radiant flux.
- Peak wavelength around 660 nm.
4 x Blue LedEngin. LZ1-00B205
- Running 700 mA (low power), 3.4Vf = 2.4W each
- About 50 lumens.
- Peak wavelength around 450 nm.
3 x Warm White Luxeon K2 TFFC Star 05027-PWW4-140
- 140 Lumen
- spread wavelength

Eclip
06-29-2010, 07:11 PM
thinking about designs and water flow. Would having flow enter the pipe in the middle effect flow distrubution much?

Gigaah
06-30-2010, 12:52 AM
Yes, It would have an impact. Obviously you'd have two seperate screens seperated by 2-3 inches(due to the "T" piece). I don't know for sure if it would be helpful or makes matters worse to be honest tho.

rygh
06-30-2010, 10:25 AM
Better would be to have water enter from both ends.
Same benefits, with no "T" in the middle.
Also, dual pipes can be smaller.
Of course, more pipes, and things to disconnect for cleaning.

Gigaah
06-30-2010, 12:44 PM
Interesting idea. water in both ends. Have you tried this/seen the results rygh?

If I can use roughly the same wattage per sq inch and use just warmwhite I could light a 10x15 vertical double sided scrubber for $100 "heatsink" and all using my favorite 1w LED's. My patented square aluminum with a fan blowing thru the tube. I can get two 42" 3" square aluminum tubes 1/8" wall thickness off ebay for $36. 80mm fans are a perfect fit, one per side. It would do the job. My tank has a couple 4 foot aluminum tubes with 50 1w LED's on each. The tube is 1" sq 4ft long 1/16 wall thickness with a 40mm fan strapped to it drawing air thru it.

rygh
06-30-2010, 01:49 PM
No, never tried dual-inlet. I did not want to deal with all the extra pipes.

Very interesting heat sink idea.
I usually just increase the heat sink size until no fan is needed, but your idea has the advantage of pulling the hot
air away from the tank as well.

Gigaah
06-30-2010, 02:18 PM
It does that too. But it really came about as a cost saving measure.

Eclip
06-30-2010, 07:20 PM
Good idea on the heatsink. i think we are over heatsinking stuff normally tho there are some simple calculations to work out the thermal resistance needed of a heatsink. The use of fans greatly increases the the thermal resistance properties turing a 1C/W into .4 etc

The warms white LEDs do work at growing algae though so i mite give them a try at first and see what the result is for a vertical scrubber

rygh
07-01-2010, 11:11 AM
For the curious:
Heat sinks are rated in Degrees Celsius per Watt.
LED junction temperature in Celsius = <LED wattage> * <Total System Degrees Celsius per Watt> + <Room temperature in Celsius>
Total System Degrees Celsius per Watt = Heat sink C/W + LED bonding C/W + LED Star PCB C/W + Thermal adhesive C/W
To make it worse, heat sinks are very different depending on air flow. Can be many orders of magnitude.

EXAMPLE:
Ambient temperature of a normal hot room = 80 deg F = 27C
For longer life, and optimal light output, best to keep the junction temperature below 80C.
My LedEngine Deep red.
Watts = 2
LED package = 5.5 C/W
Star MCPCB = 2 C/W
Ignore thermal compound and such.
So 80 = 2 * (5.5 + 2 + <heat sink>) - 27
Solve for heat sink = 3.5 C/W
That is actually a tough requirement. It takes a decent heat sink to do that, especially with no air flow.

However, junction temp can go up to 120 without any serious problems. Just shorter life.
Solve for heat sink = 6.2 C/W. Much easier.

Gigaah
07-01-2010, 12:47 PM
I'm not sure if the used watts of the LED is a perfect indicator. I'm assuming that a 3w led that puts out 100lm runs hotter than a 3w LED that puts out 300 lumens simply because more of the energy is turned into light as opposed to heat. Am I wrong on this? I'm also pretty sure that when you run a 3w LED at 2 watts your running at far more than 2/3rds the light and far less than 2/3rds the heat. I know when you overdrive an LED you certainly don't get 50% more light for 50% more wattage simply because of the junction temp. I know there are some great threads about this on nano-reef.com. thats where Evilcc hangs and a few other LED and electrical dudes.

rygh
07-01-2010, 03:17 PM
You are correct.

With older LEDs, I remember the efficiency was on the order of 10% or so, so 90% heat.
I checked the LedEngine Deep Reds, and they actually had some of the data.
About 15% efficient at max 1.5A current.
Maybe 18% at the low power I am running.
I cannot find any data on the white ones.

So yes, that 2W should be less. Probably 1.7, maybe as low as 1.3 or so.

Eclip
07-01-2010, 09:08 PM
essentially most white led's are blue (royal blue) LED's with phosphor coatings the cree XPE royal blue LED's at 350mA use 1.12W and output 425mW of light so thats about 38% efficient so only really really putitng out around .7W of heat. when driven at 1A uses 3.5W and outputs around 950mA so is only 27% efficient.

harder you drive thigns the less efficient they become and more heat they produce. so it makes sense to dirve at lower currents and use more LED's then cost becomes a factor.

The heatsinks i have been looking at are usually very good although once you load the heatsinks up with a fair amount of wattage they may struggle but from conrad heatsinks some of these while being more expensive then others have heat resistances less then 1 with the ones being made for forced airflow being alot lower around .4-.2. They are fairly expensive though as far as heatsinks go but using 6-8 LED on a heat sink compared with those nice trapozoid heatsinks rygh mentioned mite make them practiable

http://www.conradheatsinks.com/products/flat_hd.html

Eclip
07-01-2010, 10:41 PM
Also i was looking at the centre entrance flow as a T piece using alot of glue then cutting the full length and only using 1 screen if the Tpiece is any good with an approiate amount of glue would be strong

rygh
07-02-2010, 11:02 AM
Wow, 38% efficient is really good.
Of course, the phosphor absorption and re-emission is probably not 100% efficient. But I really have no idea what is is.
So If it is that efficient, that 2W goes to 1.24, so the target would be about 5.6 C/W
Which is good, since it is on par with the LED/PCB, which is around 7.
So your fancy heat sink of .84 C/W free air would support up to 6 LEDs.
FYI : I mounted 3 on each of those cheapo trapezoid ones.

Agree that a T should be plenty strong.
Suggest that you can leave a small 1/2" section with no slot, right at the T, then cut a small notch in the screen.
Partly for strength. But also to deflect the water that is coming directly out of the T.

Eclip
07-03-2010, 07:10 AM
good point on the T piece would add to rigidity and give better flow distrubution. i have reworked some of my calcs and have found that 1 fancy heat sink is fine for 15 XPE and 2 XPG all running at .7A about 35 watts of energy