Holy LED light Hood Batman. Solaris.

[GSMguy]

Originally posted by pjf
I understand that the Aquaillumination LED has even higher efficiency.

not according to PFO, where did you get that?

[Philwd]

Pat's comments, IIRC, were to the point Seoul Semi(and by extension AI) must be overdriving the LEDs since CREE doesn't spec them that high and CREE sells to Seoul. Actually Pat didn't come out and say it but cast vague aspersions that way. But Chris from AI came on and pointed out the R&D they have done to keep junction temps below 45C; even pushing the device. All the press releases from Seoul say 100lumens/W which is higher than the stated Luxeon Rebel LED performance. My take is if AI has indeed engineered a 45C max junction temperature solution then the stated Seoul efficiencies are valid.

[GSMguy]

how would seoul know in advance what AI would engineer and then predict stats accordingly?

but my point is that even with the best stats possible the AI is competing vs T5 arguably the cheapest lighting technology available today.

[Philwd]

They didn't. What I meant was even if Seoul is specing an overdriven device the engineering AI has done removes any doubt of reliability concerns. That is what PFO was strongly hinting at. If AI was running at a junction temp of 90C or so then I would have agreed with PFO that there is cause for concern.

I have looked for data from Cree on the device they sell to Seoul but aside from press releases I haven't found anything. I'm even looking through the industry rags I have access to.

And I agree with you on the T5. Another LED generation or 2 plus solving the spread issue then it may be time to jump.

[cclough_KeyDev]

Hi there,

I'm going to stay out of the rest of this thread, but touch on this comment.

Seoul licenses the die from Cree; the rest is Seoul (package, compound, doping, etc.)

LED licensing FYI: http://www.ledsmagazine.com/news/4/8/33/LEDpatents1

We, AI, are 'not' overdriving the LEDs. We are operating them totally within the power and temperature spec of Seoul.

Many have seen the press releases that Seoul's white P4 LED (we use this one) produces 100 lm/W. This is true, when you select the 'U-bin' (which we do). It is the more costly version of the emitter, and like most other products that have a high and low grade are the parts that came through the line and spec'd to the desired performance. The ones that don't pass get stamped with the lower grade.

-Chris

Originally posted by Philwd
Pat's comments, IIRC, were to the point Seoul Semi(and by extension AI) must be overdriving the LEDs since CREE doesn't spec them that high and CREE sells to Seoul. Actually Pat didn't come out and say it but cast vague aspersions that way. But Chris from AI came on and pointed out the R&D they have done to keep junction temps below 45C; even pushing the device. All the press releases from Seoul say 100lumens/W which is higher than the stated Luxeon Rebel LED performance. My take is if AI has indeed engineered a 45C max junction temperature solution then the stated Seoul efficiencies are valid.

[hahnmeister]

But is the 100 lumens/watt a 'produces equivalent to 100 lumens/watt when used in our light' type of figure, or an actual rating? I have seen LED's themselves advertised as '100 lumen/watt' when really they are 25 or 40, and the mfg. is using the application w/ reflector as a relative reading. Its kind of like how power compact curly-q bulbs are marketed... 'equivalent to 60 watt bulb'... (but its really 14 watts). Know what I mean? It seems like for many LED companies, they make these 80-100 lumen/watt claims even though they are really still just 25-50 lumen/watt LED's

[cclough_KeyDev]

LED manufacturers measure the emitter characteristics in an integrating sphere. The measurement is done with a single LED; no additional optics.

Some LED companies that integrate lenses or reflectors may advertise this way, but Seoul, Cree, Luxeon, etc. measure the emitter on its own.

-Chris

Originally posted by hahnmeister
But is the 100 lumens/watt a 'produces equivalent to 100 lumens/watt when used in our light' type of figure, or an actual rating? I have seen LED's themselves advertised as '100 lumen/watt' when really they are 25 or 40, and the mfg. is using the application w/ reflector as a relative reading. Its kind of like how power compact curly-q bulbs are marketed... 'equivalent to 60 watt bulb'... (but its really 14 watts). Know what I mean? It seems like for many LED companies, they make these 80-100 lumen/watt claims even though they are really still just 25-50 lumen/watt LED's

[hahnmeister]

Well then, at 88 watts, and with the superior optics (focus w/o wasting light), the AI unit should make about 2x as much light as the XM 20,000K, if not more... yet it doesnt.

The XM 20,000K, last I remember, is only about 20 lumens/watt at best. So 250 w x20 lpw= 5000 lumens. And the AI, at 88 watts, and 100 lpw, would be in the 8800 lumen range. Then, on top of that, the reflector technology of the LED should trump that of even the best lumenarc haldide reflectors (no light bouncing off the reflector back at the bulb, or lost at odd angles)... anyways, this should put the AI unit at at least 10,000 lumens. It doesnt seem like Dana's last tests show this though... they show a slight advantage, but not a huge one. If it was that huge, then this AI unit should be on par with the 250wattDE pheonix bulb.

If this is the case... would the spread on this be good enough to use on my tank that is 30" front-to-back? Could I mount this unit on a light rail perhaps?

[cclough_KeyDev]

Where were lumen comparisons done between our unit and the XM 20,000K? I don't recall seeing that in Dana's review, and we haven't posted any data of the sort.

The point of measurement hasn't really been discussed. If we're looking at peak-to-peak, the LEDs will be quite high in comparison to a reflector based metal halide. If we do an area average, then the two are going to be closer.

Because of the peak intensities with the LEDs, we preform well at depth, but on the flip side because it's localized, our overall average intensities is on par (no pun intended) with the metal halide.

How deep is the tank? 30" is pretty wide for a single one of our lights... Some of our customers are using two fixtures next to each other to avoid dark spots around the edges of their tank.

If you want to discuss this further, feel free to email me. (Some may construe the discussion as a sales pitch, so I'd like to avoid that)

-Chris

Originally posted by hahnmeister
If this is the case... would the spread on this be good enough to use on my tank that is 30" front-to-back? Could I mount this unit on a light rail perhaps?

[hahnmeister]

Sorry, you are right. All the tests were done in PAR, but there is no direct AI vs. XM 20,000K comparison... as only the light distribution chart is given for the LED light...

The comparison includes no 'specific' comparisons. All Dana says is "In other words, the AI LED array outperforms this metal halide lamp and uses 68% less energy."

So I wonder... what bulb would the output (just PAR) of the AI 88 actually match then? A pheonix?

[cclough_KeyDev]

We've done testing with the XM 20,000K; over our scan area we were about 50% brighter, and peak we were over 2x. This was done with an Apogee PAR meter.

-Chris

Originally posted by hahnmeister
Sorry, you are right. All the tests were done in PAR, but there is no direct AI vs. XM 20,000K comparison... as only the light distribution chart is given for the LED light...

[hahnmeister]

How large was the scan area, and at what distance?

[cclough_KeyDev]

I should have known you'd ask!

It was over a 12x12 inch area with both splash guards ~2.5" from the top of the sensor. (Not terribly realistic for the MH, but wanted to keep the same distance)

-Chris

[hahnmeister]

What kind of reflector for the mh?

[pjf]

Originally posted by cclough_KeyDev
If you want to discuss this further, feel free to email me. (Some may construe the discussion as a sales pitch, so I'd like to avoid that)

-Chris

Chris,

Your comments will not be construed as a "sales pitch." Rather, they are welcomed as a "breath of fresh air." You are adding vital information with balance and credibility and we appreciate your professionalism. Thanks very much for your participation in this forum!

pjf

[cclough_KeyDev]

Just a generic spider reflector.

There are so many choices out there, it's hard to determine 'which' fixture to compare to.

I'm open to doing testing on fixtures and providing the comparison data if there are some suggestions out there. (We have an automated test rig that, so we can test other fixtures pretty easily)

-Chris

Originally posted by hahnmeister
What kind of reflector for the mh?

[BeanAnimal]

Chris,

I think the general concensus is that the fixtures need to be tested against decent setups.

A Phoenix 14K in a ReefOptix III or Lumenarc, LumenMax, etc.

The generic spider reflector is a very inefficient reflector.

[Sanjay]

Now if you guys would have come to MACNA and attended my talk you would have seen the comparison of MH in a lumenarc with a Solaris and T5, at 24" and 30" on a 3 ft square grid.

sanjay.

[BeanAnimal]

Some of us did

[aclos3]

Given the prices that LED setups demand, they should be compared against a MH setup that costs around the same amount. If it is not possible to make a MH setup that costs the same, at least give the MH the best reflectors. A $20 bent piece of aluminium is hardly a fair comparison against a $1000-$4000 setup... come on!

[pjf]

Even at 88 lumens per watt, the Solaris LED is more efficient than most mid-wattage (175-400 watts) metal halide lamps. The mean light output of nearly all metal halide lamps in this survey fall well below 88 lumens per watt: http://www.lrc.rpi.edu/programs/nlpi...pdf/table1.pdf. The only MH lamp whose mean light output reaches 88 lumens per watt is the 400-watt Venture pulse-start model ED28 (35200 lumens / 400 watts = 88 lumens/watt). Metal halide lamps with probe-starts or lower wattage (below 400 watts) are even less efficient.

[pjf]

Sanjay,

Thanks again for testing my refugium LED fixture a while back. As you can tell, aquarist LED technology has come a long ways since then.

Cheers,
pjf

[BeanAnimal]

Now THAT is a relevant study! Good grief where do you find this stuff PJF. Honestly, how can you cut and paste things like that and keep a straight face. I can find MH bulbs with 30 lumens per watt too! I can also find MH and T5 bulbs that exceed 90 lumens per watt, some of them over 100. These parameters can be gleaned directly from the manufacturers data sheets. GE, USHIO, Phillips, Sylvania, Osram, etc.

More bold catch phrases and cherry picked data. This entire arguement has become silly, as you refuse to acknowledge the relevant facts in favor of pasting misleading information to back your opinion.

[BeanAnimal]

Originally posted by pjf
Sanjay,

Thanks again for testing my refugium LED fixture a while back. As you can tell, aquarist LED technology has come a long ways since then.

Cheers,
pjf

As he can tell? Priceless!

[hahnmeister]

Originally posted by aclos3
Given the prices that LED setups demand, they should be compared against a MH setup that costs around the same amount. If it is not possible to make a MH setup that costs the same, at least give the MH the best reflectors. A $20 bent piece of aluminium is hardly a fair comparison against a $1000-$4000 setup... come on!

Thats what I have always argued too. I mean, if you are looking at cutting edge lighting, halide vs T5 vs LED, you arent going to be comparing top shelf LED's to bottom rung halides. This comparison to one of the lowest output bulbs on the market, and a crappy reflector just doesnt do much to sell people... no matter how popular the bulb might be. If Im going to look at an LED setup, I would be comparing it to a good halide... but something bluer (since if Im considering LED's, Im not looking for a daylight spectrum light...yet)... like a pheonix, or EVC 20,000K if comparing to a 250watt... and HQI no doubt. And if Im looking at throwing down a grand or so on LED's, I would be springing for lumenarcs (DE pendants I suppose). With some of the newer models coming out which are being compared more to 400 watters... then perhaps sweetening the pot by comparing to Aquaconnect 14,000K 400's would be a good basis for comparison. Thats a royal whoop@$$ bulb... 400wattHQI, SE, and blue. If I was considering LED's, I would more than likely not be interested in the more daylight halides... and, if nothing else, all I would have to do is substitute the results by the percentage of higher output according to sanjay's graph.

So really, the only thing that is needed is a comparison, with actual PAR readings, at say.. 12", 18", and 24" away, of ANY halide bulb in a lumenarc or lumenarc mini... since anyone can simply substitute the output of the bulb picked with any other bulb by multiplying every number on a grid by the ratio of the tested bulb to another bulb in Sanjay's tests.

So that would be the test I would like to see...

A pheonix 14,000K on a HQI ballast in a lumenarc DE, or for that matter, any bulb in a lumenarc, with readings taken at 12", 18", and 24" (a 3' square to a 4' square testing grid). This would provide real data that nobody could argue with anymore. Heck, if its easier, any bulb in a lumenarc mini pendant would be a realistic comparison... as we can substitute PPFD's from Sanjay accordingly... but for those who dont get this, I suppose trying to compare the LED's to a halide that has a similar output would make more sense.