DIY LEDs - The write-up
- Thread starter JohnL
- Start date
- Status
dnahas
Premium Member
Sorry, for those look here http://www.mmmetals.com/pages/extrusion_profiles/aluminum_extrusion_profiles.htm
hllywd
Premier World Traveler
Nice call on the adjustable wrench ReefEnabler, it makes sense to me the shrear strength would be weak. Although I have my whites and heatsinks, but don't have anything assembled, I can't imagine wanting to reconfigure the LED layout so for me it's glue. I figured a 3/4" wood chisel under the star would pop them off fairly easily if I ever needed to replace anything.
Tim
BTW - I'm hopeful the NanoReef GB arrives in the next week or so with my drivers and royal blues so I can begin to join the ranks!:dance:
Tim
BTW - I'm hopeful the NanoReef GB arrives in the next week or so with my drivers and royal blues so I can begin to join the ranks!:dance:
sammy113
New member
Thanks for the adjustable wrench trick Ryan. Yesterday I had to remove an intire row cause the arrangement was exactly the same as the previous row. Luckily I noticed it 10 mins after gluing them but still had a hard time to get them off without damage them. Ended up with a paint scrapper but scratched the heatsink a little.
This helps me to destroy me current lamp cause I'm re-doing it for a new housing
slicker, with lower profile fans and a colored splash guard with the holes for the lenses to peek out 
This helps me to destroy me current lamp cause I'm re-doing it for a new housing
slicker, with lower profile fans and a colored splash guard with the holes for the lenses to peek out Sorry, I do not understand the graphic. However, which bin is better?
der_wille_zur_macht
Team RC
Prepare for more than you ever wanted to know about color. :lol:
Some relevant background:
Pure colors of light can be described as a wavelength. This is a simple physical property that can be measured. However, whitish light is made up of the combination of many wavelengths. We can plot a spectrum of wavelengths with relative intensities to describe the physical properties of the light, but this doesn't really mean anything in terms of how we perceive the actual color of the light (unless you're a geek and can interpret spectral plots well.)
So, other standards have evolved to describe the color of whitish light in more human-interpretable standards. The standard used for LEDs is the CIE1931 colorspace. Basically, it makes a two dimensional plot of whitish light. It's kind of sacrilege to put a colored rendition of the colorspace on a computer monitor since it's bound to be inaccurate, but it will at least give an approximation:
So, the short version of the story is that you can superimpose Cree's binning chart above over that graphic and you'll get a rough idea of the tints of the different color bins. The colors won't be nearly as intense as shown in this image, it's basically oversaturated to show emphasis compared to what you'll get from the LEDs.
The difficulty for us reef keepers is that we're used to other lighting manufacturers (of fluorescent and MH) essentially oversimplifying things when they describe their products as a "kelvin" color.
Kelvin numbers indicate the color a perfect "black body" will radiate at a certain temperature. It's a one dimensional measurement. It's a useful reference in that it proceeds right along the line humans would perceive as "cool to warm" in terms of colors. BUT, it doesn't account for other tints that are often present in white light sources, which is why a 2-d colorspace like the CIE1931 is useful.
So, when trying to compare Kelvin to CIE1931, you can plot the "tintless" kelvin scale as a gentle arc on the CIE1931 space, roughly like this:
The black line near the middle of the chart is the kelvin scale. Typically that line is referred to as the BBL (black body line) again because it plots the color of light a perfect black body will radiate as it increases in temperature.
So now we've got the line plotted, we can put dots on it that correspond to certain kelvin values and compare to regions on the CIE1931 space. But, since the kelvin scale only accounts for a 1-d slice of the 2-d space, we basically need to project perpendicular lines above and below the BBL to create areas on the 2-d space that represent colors that would get classified as certain values on the kelvin scale.
That's why there are those sort-of-vertical lines on the Cree chart with kelvin numbers next to them - those represent the set of values that would get the same kelvin rating, and can be used to compare how "warm" or "cool" colors are when they're not near the BBL itself.
Luckily all of Cree's official color bins are very close to the BBL so none of the colors will have strong tints. But, it's still useful information because it lets you interpret Cree's chart more accurately.
For instance, Cree's 1A color bin is right on the BBL, so we know it won't have any weird tints. It's just beyond 6500k, so we know about how warm/cool it is. Compare that to the 1S bin. It's going to be about as cool, but since it's way above the BBL it'll be tinted. To find out the tint, look at one of the color charts I posted above - if you go above the BBL in about that temp range, you're headed towards a green tint.
Regarding strength of adhesives - I've never seen a datasheet for the thermal epoxies people are using, but the thermal tape's datasheet does show increased bond strength after a pretty short period.
From the binning spec
Some relevant background:
Pure colors of light can be described as a wavelength. This is a simple physical property that can be measured. However, whitish light is made up of the combination of many wavelengths. We can plot a spectrum of wavelengths with relative intensities to describe the physical properties of the light, but this doesn't really mean anything in terms of how we perceive the actual color of the light (unless you're a geek and can interpret spectral plots well.)
So, other standards have evolved to describe the color of whitish light in more human-interpretable standards. The standard used for LEDs is the CIE1931 colorspace. Basically, it makes a two dimensional plot of whitish light. It's kind of sacrilege to put a colored rendition of the colorspace on a computer monitor since it's bound to be inaccurate, but it will at least give an approximation:
So, the short version of the story is that you can superimpose Cree's binning chart above over that graphic and you'll get a rough idea of the tints of the different color bins. The colors won't be nearly as intense as shown in this image, it's basically oversaturated to show emphasis compared to what you'll get from the LEDs.
The difficulty for us reef keepers is that we're used to other lighting manufacturers (of fluorescent and MH) essentially oversimplifying things when they describe their products as a "kelvin" color.
Kelvin numbers indicate the color a perfect "black body" will radiate at a certain temperature. It's a one dimensional measurement. It's a useful reference in that it proceeds right along the line humans would perceive as "cool to warm" in terms of colors. BUT, it doesn't account for other tints that are often present in white light sources, which is why a 2-d colorspace like the CIE1931 is useful.
So, when trying to compare Kelvin to CIE1931, you can plot the "tintless" kelvin scale as a gentle arc on the CIE1931 space, roughly like this:
The black line near the middle of the chart is the kelvin scale. Typically that line is referred to as the BBL (black body line) again because it plots the color of light a perfect black body will radiate as it increases in temperature.
So now we've got the line plotted, we can put dots on it that correspond to certain kelvin values and compare to regions on the CIE1931 space. But, since the kelvin scale only accounts for a 1-d slice of the 2-d space, we basically need to project perpendicular lines above and below the BBL to create areas on the 2-d space that represent colors that would get classified as certain values on the kelvin scale.
That's why there are those sort-of-vertical lines on the Cree chart with kelvin numbers next to them - those represent the set of values that would get the same kelvin rating, and can be used to compare how "warm" or "cool" colors are when they're not near the BBL itself.
Luckily all of Cree's official color bins are very close to the BBL so none of the colors will have strong tints. But, it's still useful information because it lets you interpret Cree's chart more accurately.
For instance, Cree's 1A color bin is right on the BBL, so we know it won't have any weird tints. It's just beyond 6500k, so we know about how warm/cool it is. Compare that to the 1S bin. It's going to be about as cool, but since it's way above the BBL it'll be tinted. To find out the tint, look at one of the color charts I posted above - if you go above the BBL in about that temp range, you're headed towards a green tint.
Regarding strength of adhesives - I've never seen a datasheet for the thermal epoxies people are using, but the thermal tape's datasheet does show increased bond strength after a pretty short period.
akindbro4u
New member
ar 120+ led's you never know what I might come up with. I way want to light, say a clam with a different color bin, or more blue, or maybe even a different color all together. I tend to orver think, and things on this tank. I'm in no hurry, so it gives me all kinds of time to build in possible changes later. I'm of the belief that we ned to go with the flow, and be willing to change as our reef's need.
der_wille_zur_macht
Team RC
akindbro4u, my "solution" to flexibility is going to be a "base" array on angle or slivers of heatsinks, plus several smaller spotlights (4 - 6 LEDs on a CPU heatsink) on long leads, clamped to the array with thumbscrews or some other type of clamp that's easy to reposition. That'll let me alter the position, color, and angle of the "highlights" without having to rebuild.
I have to give credit to the Japanese spotlight-style tanks for inspiration to do that.
I have to give credit to the Japanese spotlight-style tanks for inspiration to do that.
Now, it's clear and crystal :lol2:
akindbro4u
New member
depends. How many leds per row were you thinking?
akindbro4u
New member
HeneryH
Henery Hawk
I'll probably be around 5 x 11 or 6 x 11.
Your linear pattern could be staggered for each of my 5 or 6 rows pretty easily but I was wondering if anyone had a more creative way.
A few pages back (maybe pg 97, see below) but it really didn't come close to the 60:40 I will be looking for.
TheFishMan65
New member
Heat Sink
Heat Sink
Well I am testing the heat sink I got:
It has been on for about 60 minutes. Using an infrared thermometer I can read mid 80 (Fahrenhiet) on the inside and mid 90 if I turn if off and measrue around the LED on the outside. This is with 1 amp.
Heat Sink
Well I am testing the heat sink I got:
It has been on for about 60 minutes. Using an infrared thermometer I can read mid 80 (Fahrenhiet) on the inside and mid 90 if I turn if off and measrue around the LED on the outside. This is with 1 amp.
TheFishMan65
New member
Well close to 2 hours and upper 90s (maybe 100, but I think it was on an LED kind of bright to see what you are aiming at). In side has not changed much. I forgot to mention that ambient is about 70.
Neogenocide
New member
HenryH,
That is my design that you posted. I've got 1 of the two fixtures done and it's growing on me. That was designed on 10"x10" heatsinkusa.com stock. It runs fairly warm, but without being enclosed, no fan is needed. I do however plan on installing a low rpm, low cfm, low db fan to help keep them even cooler. As far as color, it bluer than anything I've run before and like it so far. You could easily up the LED count to get the 60/40 count your looking for.
Here is a cruddy pic taken with a point and shoot of 1 led fixture, 1 175w MH Pendant. LED's about 6" off the water, no optics.
Scott
That is my design that you posted. I've got 1 of the two fixtures done and it's growing on me. That was designed on 10"x10" heatsinkusa.com stock. It runs fairly warm, but without being enclosed, no fan is needed. I do however plan on installing a low rpm, low cfm, low db fan to help keep them even cooler. As far as color, it bluer than anything I've run before and like it so far. You could easily up the LED count to get the 60/40 count your looking for.
Here is a cruddy pic taken with a point and shoot of 1 led fixture, 1 175w MH Pendant. LED's about 6" off the water, no optics.
Scott
Last edited:
chuckstyl5
Member
When I plug my meanwells in without anything hooked up to DIM- and DIM+ not all of them fire up dimmly but when i apply voltage they all work ... Any suggestions??? FYI I have all the blue DIM-/+ spliced together and same with the white
- Status
