CNCreef Asis Pro 824 LED/T5HO Evil Cluster Hybrid

[jedimasterben]

I was actually considering springing for Makers heatsinks, but i need to go home and measure the one I have at home and see if I can fit the Ledil reflectors under the splash shield.

[Tripwyr]

I was actually considering springing for Makers heatsinks, but i need to go home and measure the one I have at home and see if I can fit the Ledil reflectors under the splash shield.

 

Are you concerned about the sagging Evil experienced using an acrylic splash guard under Ledil reflectors?

[jedimasterben]

Optics fit. Tried to put in a pic, but my phone cant open the file to edit the size. Will upload it when I get home, along with a review of the MakersDriver and Luxeon M.

Are you concerned about the sagging Evil experienced using an acrylic splash guard under Ledil reflectors?

I wont have as much heat concentrated into such a small area (well, maybe), so we'll see a bit later.

[MeepNand]

IME, those arrays put out a lot of heat. I have a 100w multichip led over my terrarium elevated 0.1" off the surface of the glass top with no optics and the glass cracked. Consider replacing the guard with some high-heat glass.

[jedimasterben]

IME, those arrays put out a lot of heat. I have a 100w multichip led over my terrarium elevated 0.1" off the surface of the glass top with no optics and the glass cracked. Consider replacing the guard with some high-heat glass.

Or just not using anything lol.

 

So I tested an array today. Threw the 1350B on a 6" Makers heatsink with four of the RB M. Used a Makersdriver with two 700ma LDD. Tomorrow I will repeat these tests, but with 1000ma LDD in place. All measurements taken at 12" from the top of the heatsink (where the splash shield would be, essentially), so about 13" from the LED diodes, with an Apogee SQ-110 PAR sensor. Numbers in parentheses are the 'corrected' values of PAR for the 450nm RB M, which the meter reads about 25% low.

 

No optics:

BXRA only - 100 PAR

4x M only - 160 PAR (200 PAR))

 

Ledil Minnie WWW (74°) / Brooke W (50°)

BXRA only - 400 PAR

4x M only - 605 PAR (756 PAR)

 

Ledil Minnie M (26°) / Brooke S (24°)

BXRA only - 450 PAR

4x M only - 1050 PAR (1312 PAR)

 

 

Holy damn.

 

Pics:

 

 

Ledil Minnie - WWW on left, M on right

 

Ledil Brooke - W on left, S on right

 

 

 

 

Power supply was tuned to 51v to make sure I have enough to supply the LDD with more than enough voltage for the M.

 

This is at 5% lol.

[RK_tek]

Looks killer. Do you think the 6" Makers heat sink is capable of keeping this cool when running near capacity for extended amounts of time?

[jedimasterben]

Absolutely. As long as the fan is running, it was still cool to the touch right next to the LEDs, after I'd let them heat up for 10 minutes or so. This is in addition to running the heatsink upside down on my table, where the fan didn't have open access to pull in air.

 

If I remember correctly, Milad has tested the dissipation at 100-150w per 6" with the included fans.

[Confuse]

Absolutely. As long as the fan is running, it was still cool to the touch right next to the LEDs, after I'd let them heat up for 10 minutes or so. This is in addition to running the heatsink upside down on my table, where the fan didn't have open access to pull in air.

 

If I remember correctly, Milad has tested the dissipation at 100-150w per 6" with the included fans.

 

Would there be enough room for the CB, Cyans, and HV's in a 6" makers heatsink? I was planning to use a 12" heatsink for one cluster.

[dacdaddy]

Looks awesome. You get to put it over a display yet? Wondering how it looks if it's too blue or not. Your pic was w them at 5%. How are you dimming them?

[jedimasterben]

Would there be enough room for the CB, Cyans, and HV's in a 6" makers heatsink? I was planning to use a 12" heatsink for one cluster.

Should be. I would just make sure that the center of the Bridgelux and M PCB (where the LED is) is on the heatsink and not over one of the t-slots. And it all depends on how many others you have. Remember that this cluster is a 1350-B, not a 950-B, with four M.

 

Looks awesome. You get to put it over a display yet? Wondering how it looks if it's too blue or not. Your pic was w them at 5%. How are you dimming them?

Nope - I never actually attached the optics. It's sitting on the table facing the ceiling and I just set the optics down.

 

The color it makes is pretty stellar - about 14K, and it really wasn't purple.

 

To dim these, I am using a Makersdriver that Milad sent me to review. Potentiometer and 0-10v control for LDD, and it powers the fan, too. The potentiometer still dims them to 0, too.

[jedimasterben]

Also, the Minnie that I have are made to be used with the Cree MT-G, which has a larger diode than the M, and the distribution graphs look different, as well as a different FWHM. When I bought mine, the ones for the M weren't available from Mouser, but now they are (and cheaper than the MT-G ones, too), so I'm considering buying them, but not sure if it'll make an appreciable difference. I'm just concerned with the massive dropoff of intensity with the Minnie WWW.

 

basically, a 10 degree drop in FWHM.

 

EDIT: I guess it is going from 26 degrees to 74 degrees, and a 50% drop for triple the FWHM isn't so bad the more I think about it lol. I just wish there was an in-between size.

 

I'll probably attach the optics to the LEDs and get some true hanging measurements and spread calculations with everything. With the current hanging situation, I was planning on putting the lights 12" over the main tank, which puts them 26" above the sandbed, and almost 40" above the sandbed of the smaller tank, and with that, PAR is much lower. If I want it to be a dedicated clam tank, I'm gonna need a LOT more light than I'm giving it, even with the 26 degree reflectors to make up for the light height.

[Mr. Microscope]

Ledil Minnie M (26°) / Brooke S (24°)

BXRA only - 450 PAR

4x M only - 1050 PAR (1312 PAR)

Fried eggs anyone? Nice!

[jedimasterben]

Oh, that's nothing. That's only at 12" - I want to be able to give that to a clam mantle on the sandbed.

[Amdphenomx4]

Thermal grease works better because the material is in a semi-liquid state.

 

Thermal glue, aka thermal epoxy and the like are rip-offs. I used to make my own with powered aluminum that was 50% metal by weight (as opposed to commercial brands which are less than 10%). I then did a lot of testing and found standard epoxy works just as well because the important factor is how close the star is to the heatsink. With epoxy you can warm it up until it's like warm syrup, and them by pushing the star around you can get a very tight metal to metal fit. From a physics standpoint the distance between the metal of the base of the star and the heatsink is far more important than randomly floating molecules of metal.

 

Thermal tape is obviously a different beast than epoxy.

 

 

Just a small comment on this, thermal epoxy usually has no metal in them as many are made to be non electrically conductive. For instance the adhesive sold by Steve is silicone based. The more permenant adhesive sold by LEDGroupBuy is based on ceramics which are both very good if you want to avoid accidental shorts. On the other hand, there are adhesives like Arctic Silver Adhesive which are 62%-65% silver by weight, but it is slightly conductive. If you want to have an electrically conductive thermal adhesive, you can have metal in it, but if you want non electrically conductive adhesive, you won't get metal. They're also probably much better at heat transfer than the aluminum powder would allow.

Reference

 

Ben', as long as you push the star around until you can feel metal on metal you're good. It's just easier if the glue/epoxy is really, really thin, which you can do by heating it up before hand in warm water.

 

Basic law of squares. For everytime you reduce the distance between the heatsink and star base by one half thermal conductivity goes up 4x.

 

A small correction to this as well. The heat transfer formula is:

 

Rate = k•A•(T₁ - T₂)/d

Where d is the thickness of the material

T₁ is the temperature of the LED star

T₂ is the temperature of the heatsink

k is the coefficient of heat transfer of thermal paste

A is the area the surfaces contact each other

 

If you half the distance you double the transfer rate. If I understand what you're saying, you are following the inverse square law which wouldn't apply in this situation as there is most noticeably no point source of radiation.

Reference

[dacdaddy]

Oh, that's nothing. That's only at 12" - I want to be able to give that to a clam mantle on the sandbed.

Are you switching from 950's to 1350's on the build now? How much more production do you get out of the 1350's?

[jedimasterben]

At 700ma, the 950 throws out at least 1400lm, the 1350 throws 2050+. The reason I used the 1350 was because I had already broken off four of the M and this would basically mimic the light I was planning to have over the little tank before I decided it should be a dedicated clam tank. I'm gonna do some more testing, and not sure if I'll put on another 1350 and four more M, then hang the light at 12" (which actually puts it in the way of viewing the end of the big tank), or if I'll keep the same hang height, go with four 950 and twelve M like the big tank, or what. I've got lots of drivers and LEDs that I can play with, so we'll see where it takes me!

[jedimasterben]

Intensity graphs from the datasheets of the two Minnies.

 

 

The inside line is the Minnie for the M, outside is the MT-G Minnie (both the wide reflectors). Doesn't look like too big of a difference, other than the MT-G having a wider footprint of having the highest intensity.

 

EDIT: and the graph for the medium Minnies are even closer. The line stopping at 45 degrees is the Luxeon M Minnie.

[dacdaddy]

At 700ma, the 950 throws out at least 1400lm, the 1350 throws 2050+. The reason I used the 1350 was because I had already broken off four of the M and this would basically mimic the light I was planning to have over the little tank before I decided it should be a dedicated clam tank. I'm gonna do some more testing, and not sure if I'll put on another 1350 and four more M, then hang the light at 12" (which actually puts it in the way of viewing the end of the big tank), or if I'll keep the same hang height, go with four 950 and twelve M like the big tank, or what. I've got lots of drivers and LEDs that I can play with, so we'll see where it takes me!

I was just looking and they say it can take only 6-10mA to kill someone. Is it because we are switching this to DC current that we don't risk exploding our hearts, or because of resistors in the drivers or what?

 

With the 950, you like 3 M's huh. Further back we compared the M's as 4.5:1 to the rebel ES. I forget what we compared the 950 to as far as 3w NW chips. Any idea?

 

Btw, awesome work w the graphs!

[Paleoreef103]

I was just looking and they say it can take only 6-10mA to kill someone. Is it because we are switching this to DC current that we don't risk exploding our hearts, or because of resistors in the drivers or what?

 

With the 950, you like 3 M's huh. Further back we compared the M's as 4.5:1 to the rebel ES. I forget what we compared the 950 to as far as 3w NW chips. Any idea?

 

Btw, awesome work w the graphs!

Rebel ES @ 700mA throw out 180lm. If a 950 throws out 1400lm @ 700 then you're looking at ~7.5-8 Rebel ES per 950.

[McJosh13]

The low voltage is what makes it safer. Voltage is what provides the potential for current to run through your heart. The amount of current that runs through something (i.e your body) is a function of voltage and resistance of that object, Current = Voltage/Resistance. A low voltage source cannot provide enough potential to run a current through your body to your heart. Worry about voltage not current

[evilc66]

I was just looking and they say it can take only 6-10mA to kill someone. Is it because we are switching this to DC current that we don't risk exploding our hearts, or because of resistors in the drivers or what?

 

With the 950, you like 3 M's huh. Further back we compared the M's as 4.5:1 to the rebel ES. I forget what we compared the 950 to as far as 3w NW chips. Any idea?

 

Btw, awesome work w the graphs!

While a low current can kill you, you need a high voltage to break your skin (a dielectric/insulator). You also need the current path to cross your heart to do the most damage. Both AC and DC can kill you. With the systems that we deal with though, the voltage is relatively low, and the drivers have a lot of safety factors built in to them to help protect themselves, the LEDs, and you. Also, a GFCI is added backup.

[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.

And think of voltage as pressure that pushes current along. More voltage = more pressure = 'faster' current pushed through you.

[dacdaddy]

The PAR measurements in air, whats the general rule of thumb how they translate to water? Half? Or maybe closer to waters index of refraction 1.33?

[jedimasterben]

Water attenuation will depend on numerous factors, but light can refract in water and increase intensity by up to 50% with surface agitation. This is why all PAR measurements for commercial fixtures and all are taken in air.

[bpaston]

These lights are beasts!!!! If you are reading along with Jedimasterben's build, go ahead and build one you will have no regrets. I can not wait to get this thing above a tank a start seeing what it will do growth wise.