DIY LEDs - The write-up

Status
Not open for further replies.
Grim,

Yes it translates. But nobody knows the math - when you figure it out please let us know:). I think you are probably fine. I think kcress said if you can leave your hand on it your fine. We regularly run 40-50C for electronics at the office.
 
After running straight LED's for almost a year, on my new tank I'm going to mix them up with T5's. I have a IceCap 4 bulb T5 retro, I'm thinking of adding 3 LED's strips using the aluminum channel stock. So it would be a T5 bulb, LED strip, T5 etc. Similar to the new IceCap Reef Illumination fixtures.

My question is, what would you guys suggest as to # of LED's per strip? I was thinking 13. My tank is 48x30x20, that would give me 39 LED's and 4 T5's for total lighting, I think that would be plenty for whatever I want to keep.

My only concern is 13 LED's stretched over about 46" might not blend very well. Any thoughts? I would be using Cree XR-E's.

Don't use optics and there shouldn't be a issue. My setup is spaced out 3" apart and while I haven't checked with water I held a sheet of paper over it and within a few inches you couldn't see a difference in LED's.

What color scheme are you planning?
 
Grim,

Yes it translates. But nobody knows the math - when you figure it out please let us know:). I think you are probably fine. I think kcress said if you can leave your hand on it your fine. We regularly run 40-50C for electronics at the office.

I did monitor the PAR, went from 173 to 171 over that time too so it isn't too hot anyway. I suppose I should play around with the blow dryer and some fans to see what I can find out about temp and output. Gads here we go again:fun2:
 
Don't forget to monitor the spectral output also :) Actually knowing the change and the starting PAR you maybe able to read the junction temperature off of the chart (if there is one).:hmm1:
 
Don't forget to monitor the spectral output also :) Actually knowing the change and the starting PAR you maybe able to read the junction temperature off of the chart (if there is one).:hmm1:

I suspect there is going to be a loss once the LED heats up no matter how much you cool them.
 
I just checked the chart, unless you can detect a 1% change in lux means almost no change in junction temperature. So my guess (and I think I have read it) is the LUX and PAR are two different. Grim got a LUX meter? If you measure the starting temperature, starting LUX and end LUX, the chart can be used to figure out the junction temperature.

Assume a royal blue XRE starting at 25 C. If you have a radiant flux drop of 10% then according to the chart the junction would be at 75 C. So is flux the same as lux or lumens? And does anyone have a way to measure this?
 
Grim now that you have your fixture fired up and running, would you be able to do a spectral diagram for each of your type of LEDs? I quite simply do not believe that the data sheets are a perfect depiction of the spectrum they put out, and have been asking for a spectral diagram for several months.

Pleeeeeeeeeeeeeeeeeeeeeeeeeese
 
Don't use optics and there shouldn't be a issue. My setup is spaced out 3" apart and while I haven't checked with water I held a sheet of paper over it and within a few inches you couldn't see a difference in LED's.

What color scheme are you planning?

Yeah, with it being a shallow tank I wasn't planning on optics at all, 80 at the most.

I will prob go 60/40 on the LED's, with blue being the 60. I run 50/50 now and have my whites dimmed down a bit to give it more blue.
 
Oh, he can really do it. I thought I was just teasing. Never followed the T5 thread or where ever he is posting, I jut know he does good work from reading.
 
hey everyone i see that there are alot of people here who are trying out LEDs and know what they are talking about. I have a 29 gallon Biocube and want to build my own led lighting to build into the existing hood. My question is how many LEDs do i need and how to wire them along with drivers info, power supply etc.
 
Go back to the top of page 150 I wrote a summary there. If you still have question come on back. If there is something unclear please PM and I will try and correct it.
 
First, I rarely use Cree for fixed lighting projects, and I build some pretty big stuff. While Cree leads the efficacy envelope, they are an expensive solution. Bridgelux, Satistronics, etc., are all viable options that get you there a lot cheaper. I've got some 10watt 440nm's coming that a manufacturer is allowing me to test,

OK, I'll bite. One forum newb to another...:wavehand:

I'm looking at a little different kind of solution, one where the Bridgelux might actually be useful. (I've looking to build pendants, for freshwater.) But I'm not seeing the solutions for optics and/or different colors, especially for color mixing. And 440 nm might be interesting.

You seem to have a little experience with this. So tell me/us why we should be going this route - politely?

I've been tracking this and several other related threads for a while. I think you'll find that this group is very open-minded, willing to help, and willing to consider other options when put forwards and justified. Not at all like North Korea. :lol2:
 
I would think that unless there is a VERY good reason regarding the quality of light, the running efficiency alone makes it clear that Cree LEDs are the right choice. What do the 10w Bridgelux lights get, 50 or 60 lumens/watt? That seems to be the average efficiency of these misc brand high output LEDs. This means latest generation of Cree LEDs consumes half the electricity to make the same amount of light, and the next generation of Crees which is due this summer approaches an efficiency which would mean they are consuming 1/3 the electricity to produce the same amount of light...
 
I just checked the chart, unless you can detect a 1% change in lux means almost no change in junction temperature. So my guess (and I think I have read it) is the LUX and PAR are two different. Grim got a LUX meter? If you measure the starting temperature, starting LUX and end LUX, the chart can be used to figure out the junction temperature.

Assume a royal blue XRE starting at 25 C. If you have a radiant flux drop of 10% then according to the chart the junction would be at 75 C. So is flux the same as lux or lumens? And does anyone have a way to measure this?

PAR and LUX are different ways of measuring output. For the purposes of growth potential PAR is what we are concerned with but as far as measuring output under different conditions either will work. I'll play around and see what I can figure as far as the temp. I really don't think I can drive my temps up short of heating up the heatsinks with a torch but I think I can come up with a way to monitor temp and output as the thing comes up to temperature.
 
Grim now that you have your fixture fired up and running, would you be able to do a spectral diagram for each of your type of LEDs? I quite simply do not believe that the data sheets are a perfect depiction of the spectrum they put out, and have been asking for a spectral diagram for several months.

Pleeeeeeeeeeeeeeeeeeeeeeeeeese

You need to be begging sanjay or Dana Riddle, they have the spectrometers:fun2:

I do have a sight spectrometer (you look at the light source through it)that did nail the spectrum of the KZ Fiji Purple lamp that was backed up when someone did post an output graph. I will see if I can figure out a way to get a picture of what it sees but I don't know if my lens is going to pick it up. I can at least tell if the posted graphs are correct.
 
Yeah, with it being a shallow tank I wasn't planning on optics at all, 80 at the most.

I will prob go 60/40 on the LED's, with blue being the 60. I run 50/50 now and have my whites dimmed down a bit to give it more blue.

This is a little more blue looking than in person. I threw some stuff in the bottom of the tank to show the color rendering but it doesn't show up real well.

This is 52 XRE Royal blues, 6 XRE Q5 Cool White, 8 XRE Warm Whites and 18 XPG cool Whites for a 52/32 Blue to white mix

I'll try to get better shots of the color, tank is dry

LEDfixture014.jpg
 
Grim; About your sink temp: The hard part about, "Is it cool enough?", is that any valid comments have to assume that the interface between the star and the sink is a good one.

This means:
1) A nice flat surface on the sink.
2) Very little thermal compound.
3) Reasonably high mounting pressure.
4) Reasonably uniform mounting pressure. In some cases two tightened screws might mean the star goes into a potato chip shape which would be bad.

Of course, with interface issues, a really nice and cool heat sink could always mean smoking hot LEDs that are about to catch fire because they can't effectively pass their heat away to the sink. :)

The fact of the matter is - if you were to keep a HPLED mounted on a 1 sqin star in air cooler than 130F you could run 350mA thru it for at least 10 years of service with no heatsink. So, keeping the heatsink below 150F is more than adequate for over ten years of service. And that is a LOT hotter than you can touch.

As a side note it's been found that the biggest factor in the reduction of light output over time is a function of the air temperature of the air the LEDs are in. That's the air in front of the LED. Any cooling scheme would do well to make sure that air is kept as cool as possible.
 
on a 7.5 gallon tank (12"x12") would 6 LEDs suffice, or should i go or 12? i plan on mostly SPS and zoos but ill most likely have a few SPS up high or a clam on the sandbed

i was considering buying this kit http://www.rapidled.com/servlet/the-61/06-Ultra-Premium-LED/Detail
and i noticed the driver for this kit(6 LED) is the same driver as the kit for the 12 LED kit... so in theory this driver can run up to 12, 3watt LEDs? If this is the case would running 9 LEDS work on the tank?

I would have 5 blues, 4 XP-G cool whites in the pattern,
B W B
W B W
B W B


Thanks,

Mike
 
Last edited:
kcress,
If air cooling the front of the LED is important. What do lenses do? I just ordered mine last night, but thought you might know.

mike,
Read the post near the top of 150. IIRC SPS want a lot of light so I think you better plan on at least 9 and maybe 12. If it supports 12 LEDs it is probably a meanwell. They have a minimum output voltage and IIRC it needs at least 3 LEDs and can drive as many as 12 (or 13 at a lower forward voltage).
 
Status
Not open for further replies.
Back
Top