DIY LEDs - The write-up

[akindbro4u]

I am not sure if I have understood correctly what you are trying to tell me since english is my second language.

What is the optimal size of the lighting for 24"x24" tank to have a nice coverage over the tank coral?

18x18 @ 3" spacing

 

[der_wille_zur_macht]

Is thermal paste/transfer required/recommended if screwing the LEDs down to the heatsink?

Absolutely required 150%. A teeny dab will do, but it needs to be there.

I get the repeater deal. It takes the on off signal from another string of LED's. I could see where it would be useful in a big build. One of those (with a 24v supply) you could make 3 Mean wells run 3 24 LED strings.

Not quite. That "repeater" needs a DC voltage input, which a meanwell can't supply. It also needs a PWM control signal, which a meanwell can't supply. You COULD use the same control signal on that device and one or more meanwells, but that's not a new or unique concept, and you'd still need some "third party" to actually generate the PWM signal. Plus, the "datasheet" includes absolutely no information on logic levels or frequency for the control signal, so you'd be experimenting at best.

The thing I though was interesting was the PWM control. If that means applying the PWM signal is what will turn on the individual channels (which sounds right) one driver could be used to control 3 different strings of LED's at different intervals. I was thinking 12 Blues for dusk/dawn, add in 6 Blues and 6 whites for stage 2 and then 12 white for midday. Just don't know If I am that ambitious:lmao:

Again, the PWM control feature isn't unique - almost all the drivers we're talking about in these threads can take some sort of PWM signal. The only thing "special" about that driver is that it's basically three drivers in one little box. But, as I said above, I'm still very curious to know what's actually going on inside that box, how efficient it is, etc.

From all I can see, "repeater" is just a marketing term they're applying - in reality, it's just three 48v-capable PWM-dimmable drivers in a single package.

18x18 @ 3" spacing

The only thing I'd add to that is that you MUST consider optic choice, height above the aquarium, and aquarium depth in the decision-making process.

 

[HeneryH]

Anyone notice the new Cree XPG spec shows an updated max forward current of 1.5A rather than 1.0A. The note in the spec says this increase is retroactive to all shipped LEDs and is due to increased confidence due to testing. This is important if you want to do the matrix (ie parallel with nodes tied in) method as you can safely run three strings @ 2.5A (833mA each) on the driver and if one LED blows short (leaving only two left at that node) you can have the 2.5 A / 2 (1.25 mA) still be within spec.

 

[der_wille_zur_macht]

Henry, that might reassure people about the potential for a failure to cascade into other failures, but it still leaves the biggest problem you'd have with a parallel configuration - the varying Vf from LED to LED. I don't remember if it was here or in the DIY driver thread but in the last few days three or four people have posted Vfs of individual LEDs they're using, and the variation was large enough that paralleling would make me nervous.

 

[ti-christ]

Henry, that might reassure people about the potential for a failure to cascade into other failures, but it still leaves the biggest problem you'd have with a parallel configuration - the varying Vf from LED to LED. I don't remember if it was here or in the DIY driver thread but in the last few days three or four people have posted Vfs of individual LEDs they're using, and the variation was large enough that paralleling would make me nervous.

There only way to drive multiple chain safely is to use a current mirror. See page 4 on this PDF datasheet. hxxp://www.recom-power.com/pdf/Innoline/RCD-24.pdf

We use that I work to drive an LED backlighting (use in flat screen).

 

[ti-christ]

18x18 @ 3" spacing

Is the3" spacing around the tank or between LEDs. Because, if it's between LEDs, I will be able to put only 36 LEDs. I was thinking to put about 56 to 64 LEDs per section. Is it too much? Is it a bad idea?

I was thinking that by having more LEDs working at the same time but at reduce power, the LEDs will last longer and run at lower temperature while having the same intensity!?!

I understand that the initial cost will be higher but maybe the LEDs will last longer !

For the parts, I am lucky because I am an Electronic Engineer and I can buy part directly from distributor through my company. I can buy Fraen lens for less than a 1$US.

Are the Wide Fraen lens a better choice? I think that it's a 39 degrees optic.

 

[mm95z71]

LED Mix

LED Mix

Sorry for the random post but was curious what people were doing on their Blue/White mix these days. I started with 50/50 and now I am at 2 blue for every 1 white. Just thought the white was washing everything out. Im using Cree XRE Q5 and royal blue. Thanks.

 

[der_wille_zur_macht]

Is the3" spacing around the tank or between LEDs. Because, if it's between LEDs, I will be able to put only 36 LEDs. I was thinking to put about 56 to 64 LEDs per section. Is it too much? Is it a bad idea?

I was thinking that by having more LEDs working at the same time but at reduce power, the LEDs will last longer and run at lower temperature while having the same intensity!?!

I understand that the initial cost will be higher but maybe the LEDs will last longer !

For the parts, I am lucky because I am an Electronic Engineer and I can buy part directly from distributor through my company. I can buy Fraen lens for less than a 1$US.

Are the Wide Fraen lens a better choice? I think that it's a 39 degrees optic.

Really, don't get wrapped up about LED to LED spacing - pack them as densely as you need to get the number you want over your tank. The 3" comment I made on the last page was just to put a number out there, to judge the distance from the side of the tank to the "edge" of the array, i.e. the first LED. I suggested that distance should be twice the spacing you end up with between the LEDs.

But spacing and other factors also depend strongly on optics and height above the water - I don't think you've told us height yet?

Sorry for the random post but was curious what people were doing on their Blue/White mix these days. I started with 50/50 and now I am at 2 blue for every 1 white. Just thought the white was washing everything out. Im using Cree XRE Q5 and royal blue. Thanks.

Personal preference and the color bin of the white LEDs you end up with will influence this heavily, so people should list full bin and model numbers if they can when responding.

For instance, right now I have a test rig for my DIY driver set up with 6 x Luxeon Rebels. The bin they're from is way over near the blue end of "cool white" so they practically don't need royal blue supplementation, though I'd probably want it for the pop if not the color.

 

[mm95z71]

Der,
It has been so long since I built this so not sure on the bin. I rebember WG. Is that the bin? I was surprised I went to the 2-1 ratio. Does this sound like too much blue? I did have 30 white and 30 royal blue on a 55 when I started. Probably a bit much. Running 30 and 15 now.

 

[der_wille_zur_macht]

WG is a common bin. It's fairly warm (and a touch yellow/green) so I'm not surprised that you wanted lots of blue with it. Honestly, it's very much personal preference - some people want a 10kk MH lamp and some want a 20kk!

I used to be of the persuasion that it was best to get the bluest whites you could, so you needed less royal blue supplementation. Now I'm wondering if it isn't better to have some slightly warmer whites and just go ahead and use a lot of blue with them, since that'll give you a little wider spectrum and might help with coloration of certain corals. Though that's all wild conjecture on my part.

 

[The Grim Reefer]

Absolutely required 150%. A teeny dab will do, but it needs to be there.



Not quite. That "repeater" needs a DC voltage input, which a meanwell can't supply. It also needs a PWM control signal, which a meanwell can't supply. You COULD use the same control signal on that device and one or more meanwells, but that's not a new or unique concept, and you'd still need some "third party" to actually generate the PWM signal. Plus, the "datasheet" includes absolutely no information on logic levels or frequency for the control signal, so you'd be experimenting at best.

Check out diagram 2. It appears the signal line is coming directly off the supply lines from a constant current driver. No? My wiring diagram reading skills are woefully rusty

http://www.lck-led.com/images/LT-3090-700.pdf

I know it still needs a 12v supply.

 

[fstar25]

Re: DIY LEDs - The write-up

der,
thanks to your advice ive upgraded the 30 led with 2 meanwell 60-48D drivers (how nanotuners normally sells retro kit) to 36 leds with 3 drivers and the are custom machining the heatsink to accomodate having 6 string centered instead of 5. thanks again for the advice.

now i just need advice on string choices. now that ive gone to six i dont know what to do. was going for the following

royal blue-white-50/50-white-blue

any suggestions

Sent from my Droid using Tapatalk

 

[HeneryH]

Anyone notice the new Cree XPG spec shows an updated max forward current of 1.5A rather than 1.0A. The note in the spec says this increase is retroactive to all shipped LEDs and is due to increased confidence due to testing. This is important if you want to do the matrix (ie parallel with nodes tied in) method as you can safely run three strings @ 2.5A (833mA each) on the driver and if one LED blows short (leaving only two left at that node) you can have the 2.5 A / 2 (1.25 mA) still be within spec.

Henry, that might reassure people about the potential for a failure to cascade into other failures, but it still leaves the biggest problem you'd have with a parallel configuration - the varying Vf from LED to LED. I don't remember if it was here or in the DIY driver thread but in the last few days three or four people have posted Vfs of individual LEDs they're using, and the variation was large enough that paralleling would make me nervous.

By definition, the voltages would need to be the same across each node but with perhaps some varying current if the curve is slightly different shape for each LED.

edit - for reference, this is the matrix config we're talking about...

 

[Obi-dad]

That last diagram shows connections between each LED from string to parallel strings - is that correct?

 

[der_wille_zur_macht]

Check out diagram 2. It appears the signal line is coming directly off the supply lines from a constant current driver. No? My wiring diagram reading skills are woefully rusty

http://www.lck-led.com/images/LT-3090-700.pdf

I know it still needs a 12v supply.

Yes, but in other diagrams (and by virtue of mentioning "PWM" signal) the signal is coming from a more traditional PWM source. I really can't grasp how you could connect the output side of a driver to another driver's dimming input, which is another thing that makes me scratch my head. The output of SOME drivers will actually be pulsed to match the PWM input, but on other drivers, the target current is actually changed so the output is a true constant current even when dimmed. And even with drivers that PWM the output, there's almost always some degree of filtering (caps) across the output, so the waveform is almost never a "clean" square wave like I'd assume you'd want on a PWM input.

der,
thanks to your advice ive upgraded the 30 led with 2 meanwell 60-48D drivers (how nanotuners normally sells retro kit) to 36 leds with 3 drivers and the are custom machining the heatsink to accomodate having 6 string centered instead of 5. thanks again for the advice.

now i just need advice on string choices. now that ive gone to six i dont know what to do. was going for the following

royal blue-white-50/50-white-blue

any suggestions

I think your plan is this: put 12 royal blues on one driver. 12 cool whites on the second driver. 6 of each on the third driver. If so, that's what I'd probably do.

By definition, the voltages would need to be the same across each node but with perhaps some varying current if the curve is slightly different shape for each LED.

And that's the problem - a 5% difference in V/I response from one LED to another can translate into a 100% difference in output from the LED! If you're building a luminary to put on a parking lot this doesn't matter. But us reef tank freaks want every single bit of output possible, and I'm not going to risk a design that can have variation that high.

Look at the V/I curve you posted and compare voltage numbers that are 5% apart from eachother and you'll see what I mean. People are measuring several-percent differences between LEDs so it's a real potential, not something I'm making up.
[/QUOTE]

That last diagram shows connections between each LED from string to parallel strings - is that correct?

Yes, it's a "matrix" approach that removes some of the "danger" from a true parallel approach, but still doesn't quite solve the problems.

 

[HeneryH]

Oh well, I already bought -24Ds so we'll see what the variation is.

 

[der_wille_zur_macht]

If you set up a test rig to measure voltage drop across LEDs at an arbitrary current, you can "voltage bin" them yourself and build an array that should have minimal variance - I just wouldn't want to go into a matrix or parallel configuration "blind" without doing that.

 

[ti-christ]

Its not one of your questions, but if you have not ordered already go back and look at page 70 of this thread Starting with post# 1747

Question #3, depends on the cooling capacities of your design ( heat sink ability and active or passive cooling, fans or not)

Thanks for your respone!

I have read again that section but I do not understand what you are trying to tell me regarding the LED ordering !?! Can you be more precise

For the Cooling, I would like to have a passive cooling but active cooling can be used on demand while monitoring the temperature of the heatsink. But I preferer to have a better heatsink with no active cooling if I can do without it!

 

[dnahas]

Thanks for your respone!

I have read again that section but I do not understand what you are trying to tell me regarding the LED ordering !?! Can you be more precise

DWZM said it best

"For the R4 brightness bin, they have OD which is also right on the BBL but one step bluer than the 1A bin available with R5 brightness.

So, if you were on a budget, I'd get the R4 OD, because it's fairly blue, and almost as efficient as an R5. If you want the best, I'd get the R5 1A, since it's almost as blue, but brighter.

In the end, the variations in color between their bins are all fairly small."

My translation get the R5 Bin. As I just fired mine up, a note to others regarding the mix of XPG whites, not XPE's, with XPE RB. der_wille_zur_macht also recommended a 60% RB XPE and 40% White XPG. IMO he is absolutely correct, the XPG is clearly brighter than the XPE.

Thanks again - der_wille_zur_macht!!! You should talk to that guy before he publishes that article on reefkeeping mag. :twitch:

 

[dnahas]

A little off topic here, but wouldn't mind the advice. As I am in the finishing stages, I see my cheap-o rat shack DMM has a limit of 200ma for current so i realize its time to bite the bullet and get a decent meter.

So far my only requirement is one that will measure upto 1500ma for this job, other than that the cheap ones have done the job.

Any recommendations? I saw someone likes the 30$ sears job and I like the price of that!!! Is it worth it to spend a bit more for a fluke or something with additional functions I am likely not to use?