Building an 'evil cluster' - DIY LED on steroids

[jedimasterben]

Tests at 1000ma on all chips. Again, numbers were recorded using an Apogee SQ-110 sensor 12" away from the heatsink, and with the individual string at 100%.


No optics:
BXRA only - 140 PAR
4x M only - 200 PAR (250 PAR))


Ledil Minnie WWW (74°) / Brooke W (50°)
BXRA only - 500 PAR
4x M only - 815 PAR (1,018 PAR)


Ledil Minnie M (26°) / Brooke S (24°)
BXRA only - 650 PAR
4x M only - 1,325 PAR (1,656 PAR)

So a 42% increase in current (300ma) which resulted in roughly 25% increase in intensity on the BXRA and a whopping 35% intensity for the M, which is almost perfectly in-line with the datasheet.

 

[jedimasterben]

Alright gents (and possibly ladies lurking in the shadows, call me ). Finally got my conduit bent for holding up the lights. Did this mostly to test out the LEDGroupBuy Lumia, but I ended up breaking that chip before I even lit it up :/



Test is done using the 1350B and 4x Luxeon M at 1000ma both channels. Driver is a MakersDriver, heatsink is MakersLED. Tank is a rimless Deep Blue 30, 24" x 24" x 12". Light is suspended exactly 24" above the bottom of the tank.

Pictures were taken with a Canon Rebel T1i with a Tamron f2.8 28-78mm lens. Aperture set to f2.8, auto shutter speed, ISO 800, and white balance set to 'Shade (approx 7000K). Pictures were taken as RAW and imported into Lightroom and software matched the camera's white balance, exported as JPEG. No editing was done on the images.

PAR readings were taken using an Apogee SQ-110 sun-calibrated PAR sensor connected to a multimeter. Both the multimeter and PAR sensor have a +/- 5% reading, so we will assume it is a wash. Numbers given for royal blue are the 'corrected' readings due to the sensor misreading blue light, adjusting 25%, no adjustment for neutral white.

For the Sketchup renderings, all numbers are the readings at the very bottom of the tank, except for the a reading taken dead center at 18" from the bottom of the heatsink. That number is on top of a black line in the middle.

Ok, now that that's out of the way, on to it.




Neutral White BXRA 1350B, no optics

4x royal blue Luxeon M, no optics

Both channels, no optics (no readings, just add the two together)

Neutral white BXRA 1350, Ledil Brooke W (50 degree)

4x Royal blue Luxeon M, Ledil Minnie WWW (63 degree)

Both channels, Ledil reflectors (no readings)

While I was collecting the readings, I had my wife type them into the computer, and when I put the reflectors on the chips and started to give her numbers, even she was like 'whoa, these numbers are MUCH higher, what did you do?' lol

 

[orm]

Excuse a complete noob in DYI led & lighting per se .... why do par reading without water?

Also its a shame that you broke the LEDGroupBuy Lumia as i was excited to hear your comments regarding that chip.

(im a lurker that found my way here after immensely enjoying your comments regarding LED's on reefbuilders)

Regards
Daniel

 

[jedimasterben]

Excuse a complete noob in DYI led & lighting per se .... why do par reading without water?

Also its a shame that you broke the LEDGroupBuy Lumia as i was excited to hear your comments regarding that chip.

(im a lurker that found my way here after immensely enjoying your comments regarding LED's on reefbuilders)

Regards
Daniel

Actually, I managed to get it working last night, I tried again to melt the plastic housing to get to the internal connector and managed to get it! I wrote a review last night, and I'll copy it over to this forum in a bit, will take some work to fix the links.

We do PAR readings in air because results will never be the same twice in water due to any surface agitation, particulate matter, or tannins/stains in the water. Water can also increase PAR up to 25% from surface agitation.

 

[jedimasterben]

After having setback after setback after setback after setback.

So for lenses and reflectors, here is what I used:
BXRA - Ledil Brooke W, 50°
Luxeon M - Ledil Minnie WWW, 63°
Blue Rebels - acrylic 90°
Cyan Rebels - no optics
Hyper violet - acrylic 60° (not to blast LEDGroupBuy, but literally zero of the XT-E optics fit the solderless hyper violet without breaking them)




That being said, I tested each string with the only driver I have that didn't need a power supply, a Meanwell LPC-35-700, which is 48v 700mA max.


PAR readings at 28" for ONE heatsink of LEDs (2x BXRA, 6x RB M, 6x Rebel blue, 3x Rebel cyan, 14x HV) in the center of the fixture at only 700mA:
NW - 125
RB - 257
CB - 56
HV - 233


Total 671 PAR


(As usual, Apogee SQ-110 attached to multimeter, and corrected per Apogee's posted PAR curve, 1.0 for NW, 0.8 for RB, 0.9 for CB, 0.6 for HV)


Good lord. This thing is absolutely ridiculous, and look at the PAR from those violets! MAN! I can't wait to get these things loaded over the tank

 

[jedimasterben]

Got it hung and up and running. My grandpa didn't put the eye bolts in the right places, so I was freaking out for a bit when I went to hang it up (my blood sugar also measured around 50, so that explains why I couldn't come up with a solution). My dad suggested just swapping around the non-eye-bolts and they'd line up properly and should hang without twisting. Got home ~30 minutes ago and BOOM!

The light is not perfectly level from front to back, but it is just a couple of degrees off, which I'm fine with.




As for looks, it's pretty **** awesome, but it shows how absolutely terrible I seem to be at planning and execution. Wires still running everywhere, not planning out where the power supply for the fans would go and where their wires would chill out, etc.




I also either didn't do my angle calculations correctly, or the angle of the heatsinks is farther off that I'm thinking it is, as the far left and right edges of the tank are dimmer than the center, and PAR is around 1/3-1/4


I changed the LDD that the arrays run on. PAR would have been astronomical with all 1000mA (700mA for the violet). I am running the four BXRA at 500mA each (in parallel on one 1000mA LDD), the 12 M at 700mA each (on three LDD), the cyan and blue Rebels at 1000mA (each their own LDD), and the HV at 700mA (on two LDD). I ended up only needing two LDD boards and only one 7.3A power supply instead of two, so that is nice.




As it is, with all the LEDs at 100%, it is a nice crisp white with a blue tint. The above pictures are danged friggin terrible. On top of that, I took the old lights down at 5:30 and didn't get these lights on until 9ish, so the corals were all closed and the fish were mostly asleep. Color banding is minimal, but shimmer is as expected with LEDs.




For PAR, measuring all LED strings at the same time, and not correcting, in the center 24" of the tank, it is an even 500 PAR, and the height of the sensor didn't seem to matter much, PAR only drops by 25 or so going from 3" under the water to 9" under. Going off-center from that 24" area, PAR drops off very quickly, going down to around 100-150 on the edges of the tank. The edges are noticeably dimmer, but edge-PAR is still better than most people's lighting


Overall, I'm happy, and it is definitely different and will take some getting used to, but I like it and it looks like a definitively positive change.

 

[ReefLEDLights]

Not Bad...Well done

CCDs do not do tanks well. Even tried old school 35mm. Photoshop worked best but can be misused to represent the true looks.

Should grow coral well and given the parts are a better option than most factory fixtures.

Cheers

Bill

 

[jedimasterben]

Yeah, just used my phone for the pics. I threw them in Photoshop to remove the ridiculous blue tint to them, and it did an ok job, but it'll be oh so much better once I get my hands on a DSLR

 

[_Alex_]

Very nice. And I know all about the low blood sugar thing lol. Been a diabetic for 25 years...

But again nice build! I'm sure you'll get the wires and power supplies all cleaned up soon!


-Alex-

 

[ReefLEDLights]

Yeah, just used my phone for the pics. I threw them in Photoshop to remove the ridiculous blue tint to them, and it did an ok job, but it'll be oh so much better once I get my hands on a DSLR

Even a DSLR has true looks issues. I use a Nikon D200. Some use a Cannon...

A lot of debate on which is better would rather not go there. Nikon Lenses vs Cannon CCD either way they need adjustment for true looks.

Please document the growth...After 18 months you will have weeding issues with some corals.

Bill

 

[kes]

wow, very impressive!

 

[jedimasterben]

Even a DSLR has true looks issues. I use a Nikon D200. Some use a Cannon...

A lot of debate on which is better would rather not go there. Nikon Lenses vs Cannon CCD either way they need adjustment for true looks.

Please document the growth...After 18 months you will have weeding issues with some corals.

Bill

I usually shoot with a Canon T1i when I can steal it from my sister




Ok, so to change my wife's disposition on the light, it needs to be covered like the old lights were:

So, if I'm going to do that, I'm also going to ditch the MakersLED heatsinks. They're very nice, but there's no point in having them hidden.

I'm also more than likely going to actually go with less LEDs. My 12x Luxeon M royals at 700mA are equal to ~8x M at 1000mA, so that will reduce complexity and reduce the number of drivers that I need, and reduce the amount of CB needed. They'll only be a bit less efficient at the higher current (35% more light, 48% more power, 597mW per watt versus 545mW per watt) and I'll still be able to cool them properly, so I'm not worried about the extra current. Heck, I may not even need to run them at full power, the PAR is nuts as it is.


Some renderings I've been working on in Adobe Illustrator. The top layer is the LED layout with the outline of the tank, and the colored circles represent two things. The larger, black-outlined circles are the LED spread at the bottom of the tank when the light is suspended 18" above the rim. The smaller, white-outlined circles are the LED spread where the light hits the top of the tank.

NW BXRA, Ledil Brooke W (50°)

RB M, Ledil Minnie WWW (63°)

CB Rebel, standard 60°

HV LGB, Carclo Ripple Wide (44°)

And something, I'm considering testing out, putting Carclo elliptical lenses on the HV, would be 47° x 8° spread.

This is what the light spread looks like from them on a white LED:

Would keep almost 100% of their light in the tank. Bill, what are your thoughts on them?

 

[ReefLEDLights]

I usually shoot with a Canon T1i when I can steal it from my sister




Ok, so to change my wife's disposition on the light, it needs to be covered like the old lights were:

So, if I'm going to do that, I'm also going to ditch the MakersLED heatsinks. They're very nice, but there's no point in having them hidden.

I'm also more than likely going to actually go with less LEDs. My 12x Luxeon M royals at 700mA are equal to ~8x M at 1000mA, so that will reduce complexity and reduce the number of drivers that I need, and reduce the amount of CB needed. They'll only be a bit less efficient at the higher current (35% more light, 48% more power, 597mW per watt versus 545mW per watt) and I'll still be able to cool them properly, so I'm not worried about the extra current. Heck, I may not even need to run them at full power, the PAR is nuts as it is.


Some renderings I've been working on in Adobe Illustrator. The top layer is the LED layout with the outline of the tank, and the colored circles represent two things. The larger, black-outlined circles are the LED spread at the bottom of the tank when the light is suspended 18" above the rim. The smaller, white-outlined circles are the LED spread where the light hits the top of the tank.

NW BXRA, Ledil Brooke W (50°)

RB M, Ledil Minnie WWW (63°)

CB Rebel, standard 60°

HV LGB, Carclo Ripple Wide (44°)

And something, I'm considering testing out, putting Carclo elliptical lenses on the HV, would be 47° x 8° spread.

This is what the light spread looks like from them on a white LED:

Would keep almost 100% of their light in the tank. Bill, what are your thoughts on them?

To be honest I'm most familiar with the Cree X Series LED

Based on the spectrum you chose I think you have a winner.

Its very close to my personal builds. Slight differences in looks but you got the growth covered.

The White LED Pic looks like the spread of Carclo 10197 Optics details I'm a huge fan of force multipliers and optics rock.
'
By tweaking the height and angle of a fixture or multiple fixtures one can focus most of the light on the reef and minimize spill on the glass.

The Makers Heatsink is just OK...I'm Old School and prefer the size and mass of the Heatsink USA T Slot Heatsink. The Makers is a tad on the thin size.

Thermal Management is key and mass and surface area are key to keeping LEDs cool. The cooler the LEDs are the brighter they are given their drive current.

Just food for thought but seen a lot of DIY LEDS and can say

Well Done!!!

Bill

 

[jedimasterben]

To be honest I'm most familiar with the Cree X Series LED

Based on the spectrum you chose I think you have a winner.

Its very close to my personal builds. Slight differences in looks but you got the growth covered.

The White LED Pic looks like the spread of Carclo 10197 Optics details I'm a huge fan of force multipliers and optics rock.
'
By tweaking the height and angle of a fixture or multiple fixtures one can focus most of the light on the reef and minimize spill on the glass.

The Makers Heatsink is just OK...I'm Old School and prefer the size and mass of the Heatsink USA T Slot Heatsink. The Makers is a tad on the thin size.

Thermal Management is key and mass and surface area are key to keeping LEDs cool. The cooler the LEDs are the brighter they are given their drive current.

Just food for thought but seen a lot of DIY LEDS and can say

Well Done!!!

Bill

I've actually got several plate heatsinks from Heatsink USA here, they're 4.85" x 12" (along with a couple of the 1.813" x ~14"). As far as mass goes, the Makers beats them out, as the entire thing is aluminum - but bigger is always better (to a point, anyway).

And yeah, the one separate pic is the 10197. What are your thoughts on it, I know you sell that particular one. Would it be a good alternative, in this case, to the Ripple wide? It would keep more of the light in, basically focused on the two rocks - the SPS rock and the clam rock. I don't so much mind the light spill on the glass - once my algae scrubber begins getting its good growth on, I won't have anything wanting to grow on the glass, it'll all want to grow on the scrubber. It'll be an algae-on-algae brawl!

 

[johnanddawn]

I recently replaced my standard Cree optics with the carclo ripple wides and really like the improvement
I had less light splash on the glass with the standard optics (par meter and algae growth confirm this) but the ripple wides boosted my overall par and improved crossover between LEDs. Of course the real test is with the corals and mine showed obvious improvement after the change. My sps went through a mini growth spurt almost immediately.
I have not used the ellipticals but looking at your diagrams they may be worth a try

It does look like the center region of your tank will have some "tunneling" near the surface but blended by the time it hits the bottom. But with your two pillar scape it will not be noticeable and will give you a lower light area. And obviously surface agitation is going to completely change your theoretical pattern

Nice build

 

[jedimasterben]

The only reason for higher PAR in your setup is because they are a narrower angle - 99% of all setups use 60° lenses, the Carclo wide are around 42-47° FWHM depending on what LED you put them on. Only problem with doing that is that you will create more spotlighting if you leave it at the same height as the 60 degree lenses, and if the LEDs have any space between them...

[video=youtube;O5CXbIWYS3c]http://www.youtube.com/watch?v=O5CXbIWYS3c[/video]





And even without surface agitation, the water will refract the light and make the light cones all willy-nilly. Kinda like how it is right now - the very ends do not get hardly any light, it seems, but I measure over 130 PAR (uncorrected).

 

[ReefLEDLights]

The 10197 as in the pic will reduce the spill on the glass and improve PAR on the sand bed.

I like them and use them on the LEDs closest to the glass on my personal tank.

That said they are not as popular as the Carclo ripple wide. Either way you have the flexibility to adjust and slightly tilt your fixture to optimize the optics.

My tank is 30" Deep and the lights are 16" off the water and dont have any disco effect. As far as efficiency the Carclo are the best we tested including spread or PAR on the edges.

There is no standard to optics. The 60 optics you mention is that max spread or most of the light. The Carclo ripple wide have a max spread of 60 degrees with most of the light within 45 degrees.

Bill

 

[jedimasterben]

Lenses are rated by their FWHM, so the 10209 would be a 44 degree lens on an XP-E. XT-E would probably be 47 degrees or so. And then 70% of the light emitted is contained in the inner half of the cone of light.

Now, that's not to say that all lenses sold as 60 degree will be 60 degrees FWHM. Most acrylic lenses are mass produced in China now, so tolerances are very lax, and a 60 degree lens may measure out to be 45-75 degrees FWHM.

I think I'll give the elliptical lenses a shot. If anything, it's a cool concept to try out

 

[jedimasterben]

Big update coming... light is finished with a new twist on it!

 

[jedimasterben]

LED only

T5 only

Everything is still a bit angry for not having light for around 24 hours, but they're less mad now, it seems.

Final array:

2x Bridgelux BXRA-40E950-B-00 with Ledil Brooke W
8x Philips Luxeon M RB with Ledil Minnie WWW
8x Philips Rebel blue with generic 90 degreen lenses
20x LEDGroupBuy 430nm hyper violet with 60 degree Carclo lenses

All but the HV run on Meanwell LDD-1000H, the HV on LDD-700H.

T5 tubes are ATI Coral Plus run on an HEP SD218-40 UNI dimming ballast. Will probably swap for ATI Blue Plus because it is so white with these on that it looks pretty terrible.