Come in and join me....

[SaltyJohn]

Well... I think that both AI and Solaris are using way too much blue for my taste. Let's not forget that even the white LEDs have enough blue spectrum in them. On another note, I can't remember which one, AI or Solaris, claims to be using Seoul Semi's P4 LEDs at 6.500K however if you go over the pdfs at Seoul Semi's page they only produce LEDs up to 6.300K... go figure!? Prolly a typo on their part?

 

Also I am not going to implement that many LEDs, I am looking at approx. 80-100 LEDs (80% white, 20% blues or as close to UVB spectrum I can get them). So yes, counting LEDs, heat sinks, lenses etc, it will approach 1000$ but that is over 80-85% of the costs for my project (my budget is <1500$ for everything, comparing to 4.000$++ for a ready unit)

 

On the power source(s) I am using cheap(er) regulated power supplies, which then I intend to tighly regulate with extra (low budget) circuitry (~3-5$ per array) and of course for sunrise/set a PWM solution is the best way to go in my book. So when I mention trimmings that I do not need, I meant cloud covers, phases of the moon etc.

I need to dim upwards in a short 10-15 minutes period, stay on for 9-10-11 hrs at 100% (or PWM it to the color of my liking to <100%) and then dim down in 10-15 minutes. As far as the controller, I need to keep the clock alive when mains fail, and run a subroutine when the mains come back (dim up/down much faster and then resume at the point that the mains dropped off originally) and some other, 'nice-to-have' or security features that come at no extra cost, but only extra programming on the same chip.

 

For the prototype, of course, since I have ordered small quantities, the cost is -relatively- much higher since shipping is a large % of the entire cost (ex.: 10$ postage for a 2$ micro-controller)

[evilc66]

Here is something interesting that I really didn't pay attention to before. PFO is claiming a 15K color temperature on their fixtures using 25 LEDs (Luxeon K2s). But they are not using all LEDs during daytime lighting operation. They only use 9 white and 6 blue. If they used the highest output version of the original K2 LEDs, thats only 1260lm in white (9*140lm) per foot, and thats driving them at 1.5A. So a 72" fixture (4 modules) is only 5040lm. Granted this is not including blue light (they use royal blue and it's measured in mW and is difficult to convert to lumens), but still. Thats really not that impressive as far as light output.

 

Now I hope everyone realises that light penetration is better with LEDs, and PAR levels are higher which deffinitely help their case. It still doesn't justify the kind of money they are charging.

 

Normally I would be singing the praises of the Luxeon Rebel LEDs right now because you can create higher light density at a given point with the PCBs that are available now, but a single LED star might be more cost effective. It will allow you to spread the total light out a little more evenly over the length of your tank. If you want to kick the light output up considerably, look at the rebels. You can get 540lm on a single star at only 700mA (white only). Plus you can also "adjust" your white color temps by adding a 470nm blue led to give you less of a yellow light. Add a royal blue LED and you can get an even bluer color and a little actinic while you are at it. Nothing else that I have seen on the market can equal that kind of light output in that size package.

[Gomer]

keep in mind that the AI and solaris are using outdated LED technology.

 

a Luxeon white K2 driven at 1500mA will generate 140 lumens witha Vf=3.85. 3.85 x 1.5=5.78watts or 140/5.78= 20 Lumens / watt!!!!

 

The new Cree XR-E Q4 bin (they are up to N2 bins right now) produce 178 lumens at 670mA with a Vf=3.7 (measured data) or 2.48watts or 72 Lumens / watt!!

 

 

so lets put some more numbers to this and lets assume all white LEDs for lumen sake. So a 70" fixture has 4 banks or 100 LEDs.

LEDs=100

Power *=578watts

Lumens produced=14,000

 

 

 

To get the same amount of light with Cree XR-E Q4 bin

LEDs=79

Power *=196watts

Lumens produced=14,062

 

Not only do you need less LEDs, but you don't need to over drive them, you don't generate nearly as much heat, and you can use less power supply.

 

or...

Q4 LED=100 driven at 1570ma (not recommended)

Power=622watts

Lumens produced=34,100

 

 

 

*not wall power, just LED power

[coolwaters]

yeah i hate when LED loses its efficiency when its over 2watts. most of the time its below 1 watt where its 130lm per watt.

 

fyi solaris uses luxeon I LEDs and the efficiency is around 20-40 lumens per watt. (20lm if its over run)

 

the best right now is Cree or for the biggest bang for the buck Edison.

[Gomer]

most of the time its below 1 watt where its 130lm per watt.

 

 

You won't get that high at all with mainstream highpower LEDs, let alone with what is in the solaris/AI

[coolwaters]

Cree Q5 was the best led i heard so far. at 114lm per watt. but as u know it loses its efficiency at higher amps.

we overdrive LEDs cuz i dont want hundreds of them in one spot just to produce a lot of light.

if Solaris didnt over drive the LEDs then the hoods would have 4 times at much LEDs on it....

 

hopefully i'll be getting the soldering iron 2morrow at wal-mart.

sorry for the weeks of delay im been tied up with fixing my friends car and very needy and seemly maintenance friends...

[evilc66]

PFO isn't overdriving their LEDs. They are operating well within their specs.

 

Using the XR-E Q5 or similar LEDs is deffinitely the way to go for larger fixtures, but has big limitations once you start shrinking the size down to nano and pico size tanks. I don't think focusing on any one LED for all applications is the way to go. Bigger is not always better. Anyway, now that the 180lm Rebels are starting to become readilly available, they have the same lm/W as the XR-E Q4's.

 

None of us can really fault PFO or AI for using the LEDs they have. It's financially impossible to keep up with the changing technology. Us as DIYers have the advantage of being able to pick exactly what we need at the time. We can fault them for the cost though

 

It will be interesting to see how things pan out in the near future. Between potential product updates from PFO and AI, and the new competition coming out, the LED fixture world will get interesting quickly.

[coolwaters]

im pretty sure their using Luxeon I on the older models and their rated at 350mA. and PFO isnt running them at that.

[Gomer]

I'm ordering my heatsinks tonight. Once I get them in with thermal epoxy, and can get the dremel to the backside of the optics (need to channel wires since I am not reflow mounting the optics), then I'll get images up. Anticipate pictures in my thread within a week

[evilc66]

Sound cool Gomer.

 

Coolwaters, I was pretty sure they were using K2's on the newer I4 versions. I could be wrong of course. I can see them using Luxeon I's on the originals.

 

PFO is claiming over 80lm/W on the I4 units white LEDs. Seems high.

 

http://www.solarisled.com/FAQTechnical/FAQ...54/Default.aspx

 

It's the 5th question in the FAQ if anyone is interested.

[Gomer]

PFO is claiming over 80lm/W on the I4 units white LEDs. Seems high.

 

 

sure...maybe if they are running at 100mA. Not a chance in hell at normal levels.

[evilc66]

I found it a little comical too. Actually, most of the things that they talk about that involves light is a little comical. I think that their claim of PAR "equivalent to a 400W MH at 24 inches in length" is a little exagerated. I know that PAR levels are higher per watt in LEDs over MH, but I think their claim is a little overboard. Must have been using pretty crappy bulbs!

[coolwaters]

well the K2s is maxed out at 40lm per watt....unless they used the newest K2s...but that just came out a few months ago...and wat the I4 was like last year?

 

lol

"The white LEDs used in the I4 series is above 80 lumens/watt"

i dont believe that...btw does anyone know when they came out?

 

Bulb Watt-Color Lumens Watts lumens/watt

250W 4K Sylvania 20,000 295 67.79661

250W 10K ushio 10500 295 35.59322

250W 20K ushio 5000 295 16.94915

400W-4K Sylvania 36000 465 77.41935

400W-10K Ushio 18500 465 39.78495

400U-20K Ushio 8000 465 17.2043

 

is that correct? if so then dam... my 9 LED array at about 12k is giving me about 50lm per watt. (when i factor in the blue LEDs.)

[evilc66]

Yeah. Make you start to wonder who to believe.

[Gomer]

Keep in mind that Lumens is on a scale relative to our eye sight. You could have a super efficient 380nm light with a horrible lumen/watt value, compared to a very inefficient 500nm light that gives a MUCH MUCH higher lumen/watt value. This is why the high K bulbs have horrible lumen/watt values.

 

so the lesson is, take lumens/watt and lumens ingeneral on a grain of sea salt.

[neanderthalman]

I found it a little comical too. Actually, most of the things that they talk about that involves light is a little comical. I think that their claim of PAR "equivalent to a 400W MH at 24 inches in length" is a little exagerated. I know that PAR levels are higher per watt in LEDs over MH, but I think their claim is a little overboard. Must have been using pretty crappy bulbs!

 

20K MH with a spider reflector.

 

I'd like to see the test performed with a 10 or 14K MH and a Lumenarc III reflector and see how the solaris really stacks up.

[coolwaters]

Ultimately LEDs have better PAR right?

[The Propagator]

20K MH with a spider reflector.

 

I'd like to see the test performed with a 10 or 14K MH and a Lumenarc III reflector and see how the solaris really stacks up.

 

 

It wouldn't.

 

With all the hype they put into that product pre-production versus the amount that have been actually sold by now I would have to think that some one would have devoted an entire website or at least a review page to them on how great they are. but... NOPE.

Not even Advanced Aquarists on line has have they?

They usually do it with every thing.

[neanderthalman]

Ultimately LEDs have better PAR right?

 

You say that like it's an absolute thing.

 

1W MH wouldn't give you as much PAR as 1000W of LED. Simple as that.

 

Does 150W MH give you more PAR than 150W LED? Don't know. Depends on the LEDs, depends on the bulb, ballast, and reflector for the MH.

 

Can LEDs surpass MH? Possibly. Can they do it economically? Not yet.

[The Propagator]

http://cgi.ebay.com/Clearly-LEDs-Aquarium-...1QQcmdZViewItem

[siwelk]

uh, broken link Prop

[evilc66]

http://cgi.ebay.com/Clearly-LEDs-Aquarium-...1QQcmdZViewItem

 

Looks like 5mm LEDs. It's not going to be worthwhile, even for $60. I wish they had a bigger picture of it.

[The Propagator]

Email and ask for one...

[coolwaters]

http://cgi.ebay.com/Clearly-LEDs-Aquarium-...1QQcmdZViewItem

 

u think u can grow anything with 12w-14w of light?

the max lumens efficiency on thos 5mm LEDs is 40-45lm

 

if it were more compact it would make a good light for a small refuge.

[SaltyJohn]

Cree Q5 was the best led i heard so far. at 114lm per watt.

 

Coolwater, how do you figure 114lm/w? That would be awesome!

 

From my math, the Q5 (their latest bin) is rate at min 107 -max 114 (@350mA) and at maximum If the LED gives out ~225% (from their graph) of 114lm ie~256.5 lm at (3,7V * 1A =) 3,7Watt, so the 'new', overdriving math tells me:

256.5 lumens / 3,7 Watt = 69.32 lm/W which is still great, but still, very far from 114lm/W

 

Am I misreading any info on their data sheets? Can you assist me with this one?

 

well the K2s is maxed out at 40lm per watt....unless they used the newest K2s...but that just came out a few months ago...and wat the I4 was like last year?

 

"The white LEDs used in the I4 series is above 80 lumens/watt"

i dont believe that...btw does anyone know when they came out?

 

I am pretty sure they are not using the K2, but rather the Rebels (or most probably Seoul Semi's P4 bellow, much like Aqua Illumination's units are) which are rated 100lm/w (@350mA) and at 180lm @ 700mA. Overdriven again, at 1000mA they will give almost 215/220lm (according to their graphs) @ 3,55V, which makes for a (3,55 V * 1A = 3,55W) 220lm/3,55W=62.85lm/W. Now it can be still overdriven up to 3,99V (according to Philips' data sheets always) which might approximate the 80lm/W they claim, but with a really shorten lifespan!

 

Last, Seoul Semi's P4 -the U-bin at least that I have tried- is also rated at 100lm/w (@350mA) and at 170lm @ 700mA. Overdriven again, at 1000mA they will give almost 240-245lm (according to their graphs) @ 3.8V, which makes for a (3,8V * 1A = 3,8W) 245lm/3,8W=64.47lm/W. It can obviously also be still overdriven, to push more lumens out per watt, but life expectancy goes bye-bye again.

 

Unfortunately for us, no manufacturer has really taken any leaps from competition, so we are left to decide on minor details which LEDs we should be best using for our projects. I am basing my decision making on optical viewing angle, thermal dissipation, and of course product availability and lets not forget co$t!!!

 

For example, the Cree LED has a typical viewing angle of 90 degrees, but the P4 has 132 degrees and the Rebel has 160. This favors the Rebel in my book, since the lm/w that it puts out over the 160 degrees, when focused with a proper lens, will give me more lumens per square in/cm of tank area, which is basically what I am after...

 

However I can source Seoul's P4 LEDs cheaper in my part of the world...

 

Heat sinking properly these babies so I can approach overdriving, higher lumen output adds substantial cost as well...

 

And on, and on she goes...

 

So...yes... it is a painful exercise, but much has already been invested in this one, and I am not cutting my losses short anytime soon!