Holy LED light Hood Batman. Solaris.

[GSMguy]

Thanks hahn

[JMaxwell]

Beautiful. Thanks for posting. it will be nice to see the results at all the depths.

[barjam]

Hahn, are you doing one of these for a 6x39 T5 unit by chance as a comparison?

[hahnmeister]

no, never considered it. 6x39 would require a different grid, something like 2'x4' for just the 6" distance. The other problem is 'which one', as in, which T5 unit would I compare? I have a Tek 6x39, but thats not the whole story then is it? ATI, Sfiligoi, or Fauna would have to send me a test unit or I would have to make one from an Icecap retro. I have 3' reflectors, new bulbs, and endcaps I suppose, but lack the 2x39wattT5 ballasts.

I will be doing a T5 reflector comparison though, just waiting on a couple other companies to send me their samples, and waiting to see if Sfiligoi (Stealth samples), Grotech (Photon Lunea samples), ATI (Sunpower/Powermodul samples) and Giesemann (Matrixx samples), and Aquaconnect (Lumimaster), and Current USA want to participate in the 'reflector showdown'.

I think that information will be more useful in the long run. With an understanding of how to sub in/out T5 open air par readings into grid results, as well as how to 'stack' the grid results together (overlapping the grids of say three smaller 2x54wattT5 results or six smaller 1x54wattT5 results to get a 6x54wattT5 result after swapping out the PAR values for each bulb from the reference bulb), this should prove a better tool... only problem is that it will be for 4' bulbs, not 3'.

[jnarowe]

That's about what I had expected and I am glad to see it graphically. Did you use different ranges because of the higher intensity in the MH? It's hard for us nimrods to understand the comparisson because you switched up the colors and made the ranges different sizes.

Now I really want to test my lamps!

[barjam]

no, never considered it.

I can certainly understand that. I just hope someone does a real world T5 evaluation at some point. The closest thing I have found anywhere was your impromptu 6x39 vs halide post a while back.

So your MH was under a luminarc using a real HQI ballast? Thats about as good as it gets for that parcticular bulb, right? I wonder what the results would have been under an electronic ballast using the cheapo reflectors that the retro kits come with (I am thinking the hellolights DE reflectors for example).

[hahnmeister]

Yes, that is about as good as it gets for that bulb. I did get slightly higher readings (4.5% higher only) earlier in the testing about ... that bulb has been sitting around on a shelf for a year, so it might have had to 'break in' again despite only having about 200 hours of use on it.

Setup is Lumenarc DE/mini pendant, PFO HQI ballast, and pheonix bulb with about 200 hours of use. There could be a 5% variation due to it being about 100 hours more than the point Sanjay takes his readings at, but I have seen more than a 5% variation between one bulb and another anyways as far as the bulbs are concerned.

Even if the bulb was 10% lower than what it could be (brand new/100 hours of use) for some reason, and its not even that, it paints a pretty clear picture.

I dont want to go into more other than saying that these are just for these two circumstances. There are reasons in the AI design that I will explain later (in a dedicated thread/article) as to why even though the LED's might be listed as more powerful on paper that this unit should not be used as comparison to the PFO units.

Still, I think it shows what happens when you compare to a halide that is in a higher-end reflector rather than 'some reflector of unknown origin' like Dana's review. The reflector has the potential in these comparisons to hinder or overcome greatly... just like lacking a reflector all together would be an huge blow to the halide's readings, or having a reflector that blocks light or allows it to exit from points other than down into the tank... so there is alot riding on it.

But also, keep in mind, the AI ran at 94 watts with everything full tilt. The HQI pheonix combo runs at 322watts. So the LED's are very nearly 1:1 with the halide, or pretty close to it.

Here is one that makes it even easier to see...

Finally figured out the ole autosum button again, so I can provide these numbers:
Total incidental light, LA3: 296,475
Total incidental light, AI88: 81,035

Note that while the AI has light values in the 5-10 range in the corners, the LA3 has light levels in the 80-140 range in the corners, and should be measured on a 3x3' grid, which would jack its number up even more. Otherwise, as it is, the LED unit makes a little more than 1/4 the output of the halide.

So sure, it got trounced by the halide, but not by too much in the end since it was working with a little less than 1/3 the wattage. I would imagine that watt for watt, with a less efficient 20,000K halide bulb, the LED's, with their highest output ones being blue/450-460nm, they would be 1:1 performance-wise with this fixture.

[pjf]

[killagoby]

Nice! I like how you wrote it in Engilsh hahn. lol.

[jnarowe]

OK, so can we translate that to needing three times the LED to equal 1 time the halide?

The reason why I ask, is that when i researched LED for my tank, the numbers were way off the chart in terms of expense to build an LED array that would match my current lamps, in the neighborhood of $100K+. And there was still the question of viable penetration.

[MiddletonMark]

Originally posted by hahnmeister
So sure, it got trounced by the halide, but not by too much in the end since it was working with a little less than 1/3 the wattage. I would imagine that watt for watt, with a less efficient 20,000K halide bulb, the LED's, with their highest output ones being blue/450-460nm, they would be 1:1 performance-wise with this fixture.

I'd also suggest that using a LumenArc reflector probably helped the distribution [esp in the corners, etc] and overall intensity.

While a lot of folks have learned to get quality reflectors, a less efficient reflector would probably give a more similar distribution and lower intensities.

Then again, I'm glad you didn't mimic the product-demos I've seen for LED's ... which often use poor reflectors, 20K bulbs, and generally run a very inefficient MH setup.

Thanks for trying to inject actual data into this discussion.
[that doesn't mean I won't pick it all apart later ]

[JMaxwell]

Hahnmeister,

Excellent charts. No need for the 3-D contours, as you are only working with three parameters. Are you going to push out the data for 12", 18", and 24"? I recall an earlier post where you said that you were testing at those levels. The AI/LED unit might benefit in some situations from being mounted 3"-6" closer to the water, so data on height vs. PAR might be useful.

[hahnmeister]

I have data for 6", 12", 18", and 24".

FWIW, I also hooked up a PFO mini pendant to the grid, and the results werent that bad. The difference between the PFO grid results and the Lumanarc are that the PFO has a smaller area of coverage, sure, but its light levels are that much more intense. It had peak levels at 6" just under 2000.

So lets just put it this way: you need a pretty crappy halide reflector on the halide to match the output of the LED.

[dkle]

So I've been following this thread with great interest; even though I can't afford the LED light at this moment. Thanks for posting the results, since people can argue all of the theoretical points for all they want, but in the end, the numbers don't lie.

My questions for you are:

1. How do you extrapolate from your results comparing between a 90w fixture with a 322w fixture to say that if you have three LEDs then the result will pretty much even 1:1.

2. I just checked the Aquaillumnation fixtures on Reegeek; and the 90w fixture is 12 inches long. The 24'' fixture has double the output. Since the standard MH setup that you used covers 24'', aren't you supposed to test the 24'' fixture to have an unbiased result? Since a big argument on your part was the previous testing used some unknown, not optimized reflectors on the MH setup, I think testing the 24'' LED fixture instead of a 12'' one would make your comparison much more valid.

And if I made some wrong assumption on my part, please let me know. Again thanks for the testing and posting the result.

[8BALL_99]

So if I'm reading right it takes 4 times the leds to get the same amount of light as a good Mh setup? At the same time if you used 3 times the leds there would be no power savings. The other question is if you used 4 times the leds wouldn't they use more power and generate more heat to get the same amount of light?

FWIW I think anyone even considering a 3000.00 light would already have 130.00$ MH reflectors so that part just makes since to me..

[hahnmeister]

dkle,
The total incidental light output is how I calculate 'total output'.
"Total incidental light, LA3: 296,475
Total incidental light, AI88: 81,035"

So I take the sum of all of the 576 points on the grid together. This is a bit like how lighting mfg's calculate 'total light output' for a bulb, by taking the sum of all points. The total light ouptut for the AI is about 27% of what the LA3 is. The actual wattage of the AI was 94 watts, and the HQI was 322 watts... or the AI is 29% the wattage of the halide in this case. Also take into account that there is a fudge-factor here of the pheonix having a variable output throughout the year (and the LED will change slowly over time) and it will eventually drop to say, 75% of its output after a year... so the efficiency per watt of the two units isnt that far apart.

Lets say a person used an EVC 20,000K though, for more actinic. Then every point on that grid would be multiplied by the ratio of about 81/88 (values taken from Sanjay's info archive for PPFD). And then the total light output would also be multiplied by that... or 272890.

So I was careful in that 1:1 statement because I qualified it by saying that IF A PERSON PREFERRED A BLUER TANK... as in, LED's vs. a 20,000K bulb. Then watt for watt, the LED's would compete rather well. LED's are even better at making blue light than daylight, and halides get worse and worse at it. The pheonix is an 'oddball', almost monochromatic blue halide... similar to many LED's out there in spectrum, but with an output that beats a ushio 10,000K. So as soon as you start looking for another bluer bulb (yeah yeah, the aquaconnect 14,000K...lol, okay smart-@r$$es...lol) like a 20,000K, the halide can lose its edge from the par/watt standpoint.

Sure, one could argue that the LED's should be higher number to compete (but keep in mind that the cost breakdown that I gave before would have to be augmented even further then). I think the 'precident' has been set by some other mfg's that their units are to be compared to 250watt halides, and these Seoul Semiconductor LED's are supposed to be the top of the line for 2007. The other reason for this comparison is because it is what Dana Riddle did in his test, and the numbers given for the halide in comparison didnt seem as competitive as they should be to me.

[hahnmeister]

8BALL_99, yes, if the LED numbers were increased, to match the output more (as well as wattage), then the heat output would be about the same as well.

The only respect where the two technologies arent in line with each other at present would be cost. To match the output and wattage of a halide, you would spend 3-4x as much on the LED system over 5 years, as per the breakdown I gave earlier.

[dkle]

Thanks for the clarification. Great info.

[crazzy]

[QUOTE]Originally posted by hahnmeister
[B]Lol, all you guys are too funny. All we need to add is a 6-pack per person to turn this thread into a full blown roast!

Things have changed... in a turn of luck, it turns out that I will be getting that 'testbed' sample from AI after all. It was mentioned that since I asked to test it first, I would get it first after all, and then perhaps send it directly to Sanjay (or back and then to Sanjay, either way). Anyways, Its pretty nuts for me until about the 19th, but as of that night, I will have the lumenarc all polished up and the grid dusted off & ready to go for some PAR testing. So look for some hard data to get posted around the 21st.

Has the unit now been sent to Sanjay?

[cclough_KeyDev]

Not yet - just got it back from hahn today. I'll probably ship out to Sanjay next week.

-Chris

[QUOTE]Originally posted by crazzy
[B]

Originally posted by hahnmeister
Lol, all you guys are too funny. All we need to add is a 6-pack per person to turn this thread into a full blown roast!

Things have changed... in a turn of luck, it turns out that I will be getting that 'testbed' sample from AI after all. It was mentioned that since I asked to test it first, I would get it first after all, and then perhaps send it directly to Sanjay (or back and then to Sanjay, either way). Anyways, Its pretty nuts for me until about the 19th, but as of that night, I will have the lumenarc all polished up and the grid dusted off & ready to go for some PAR testing. So look for some hard data to get posted around the 21st.

Has the unit now been sent to Sanjay?

[killagoby]

Phew! Thank god I like a bluer tank. That's one of the reasons I went with Solaris. One of my LFS's had one hooked up and the Frogspan in that tank looked radioactive!

[pjf]

Originally posted by cclough_KeyDev
Not yet - just got it back from hahn today. I'll probably ship out to Sanjay next week.

-Chris

Thanks to Chris and Aquaillumination for your generosity in letting your units be tested and reviewed!

[divernm]

hahn, great data, I would like to ask a favor of you if I may. and this thought process was set off by AI's website (their 5x5 inch square analysis), I have been running my led light for over four months now, and have noticed a couple of things, first my corals are growing faster, second the light is just as bright at the corals on the led side as it is on the mh side (under water lux meter) for the same height in my tank. directly under the lights. what I am not getting is overflow, I use a spider reflector with my 150 watt xm mh, hqi, 10,000k and two 96 watt pc one .420 and one .455 my tank is 28 inches deep and 23 wide, a lot of light on the metal halide side of my tank hits the wall directly behind the tank and the floors around the tank for about two feet on the three viewing sides . I have the Mh system 9 inches off the water any closer and the heat coming off the front of the system makes it hard to keep my tank cool (no chiller), the led side has no over flow on the three viewing sides and very little on the wall side about two inches worth, above the shadow the live rock leaves on the wall.
I can not get a good feel for how much of the light that the metal halide fixture puts out actually gets to the corals and how much goes out the tank through the glass and into the room missing my coral completely. the metal halide may put out three times the light but how much is used.
could you place the metal halide 29 inches up from the test fixture and see how much light hits a 24 inch by 18 inch square directly under the light and then see how much light hits the same area from the led placed at 20 inches up if you could do me this favor I would greatly appreciate it because as an engineer I know that you do not get anything for nothing, and I can not wrap my arms around why the led light is working so much better then the mh.

thanks in advance

Mike

[hahnmeister]

My tank is 48x30x21h, right side is a ushio 14,000K (no T5s on), and left is the AI. So yeah, I would need 600-some watts of LED to light my tank at least. Right now, I get away with less than 500.

The light from a halide is not wasted as long as it enters the tank before it spreads out too much. So a large reflector that you can position closer to the water and still have a nice spread is a good thing.

What you ask for with the grid from a larger distance is possible, but it doesnt tell you much. All it tells you about is how the reflector converges/diverges (or neigher) with its light. The thing is, the 'big glass box' acts as a giant reflector anyways. This has its advantages over a light that just illuminates the coral directly with parallel rays. Your corals will most likely grow towards the dominant light source (Phototropism) and not develop pigments over as much of their tissue because more of the coral is in shadow. For softies, this is not a problem because they can adjust themselves (as well as some LPS), but not so with SPS. The light that enters the tank, yet bounces off the side walls provides light to areas of the coral otherwise in shade.

The other reason measuring light output over 18" away isnt much use is because the index of refraction of the water. The spread of any light source will be augmented by the water surface. Light rays that enter the water at more than 45 degrees will most likely not even enter the water, and light rays that enter at 45 degrees will 'bend' to 30 degrees underwater. So a light that covers a 48x48 grid if measured 30" above an open air grid will have its output concentrated even if over water.... maybe only covering a 30x30 area (depends on the height of the fixture above the water and the depth of the water).

What I am saying is that there are too many factors for one to calculate what you ask with an open air test. You just have to look at some tanks that use them. The best suggestion would be to use the grid results that are at a distance closest to whatever distance you plan on mounting the fixture above the water. This will this will tell you more than a 30" distance grid test because by using the size of the source, you can determine how much light is going to go into the tank at the surface (what makes the most difference).

[paveking1]

I have just read this thread interesting thread . I have 2 solaris lights on my tank . They are a great light once you get past the price . I run one of there older models 250k up front& 400k on back of tank .Iget great growth & color all the way to the bottom of my tank . Heres a pick of my cap at 11 months old under this setup ,it was a 3" frag when i got it it is 14" round now . I have no chiller & water is at 78 at night & 79 in day .