Building New Panels - Sharing the Process

[rrog]

For the Cree XM-L they run 10W at 3A, but based on Rives good guidance I am running them at their recommended ~2A, dropping output to 6.5W. So I added more LEDs to get my 300W panel. 50 total. I would rather buy more LEDs than run them at their max. That's what I meant by overkill. Also the effective wattage is ~ 75W PSF of canopy. That's likely overkill also.

 

[2cent13adz]

rrog!

If you are happy running things @ 2A, I'm happy! Still can't see why... what is worst case scenario? Them led's will only last eight years, not ten?

 

[rives]

I think that Rrog's statement referred more to building the fixture with top end components and trying to kill every snake prior to it rearing it's ugly head. He is new to the whole idea of DIY LEDs (I suggested it to him) and I believe is coming from a minimal electronic background. If you've followed the development of commercial fixtures, you are undoubtedly familiar with the wild claims made by importers, minimal to non-existent heat sinks, and running the chips at a small fraction of their nominal rating. I want his first experience with this stuff to be positive, and if he wants to push the envelope a bit, to do it after he acquires a bit more background. This fixture already is packing some serious power for it's intended application.

By feedback, I meant that with most mechanical toys, you have noise, vibration, etc giving you some sense of when you are pushing the limits - that is obviously impossible here.

By the way, welcome to ICmag!

 

[xmobotx]

Xm, no, I've never heard of an LED "building a charge" and then releasing it, and can't really dream up a scenario where it would.

thats what i thought too ~perplexing that the reds blew

 

[habeeb]

here's my version of overkill..

-oversized heatsink
-1 extra fan
-running chips at lower watt, thus increasing chip count by 2
-copper stars / heatsink
-air cooled room




2cent, the thing your not getting.. is the higher you drive the chip the lower the output. look at any chart and see highest lumens comes at lower amp ratings.. in this game it's not "go big or go home" , it's about finding a happy medium between chip count / output. look at this chart to see you only get less the more you put in, add some of those numbers up and you could end up with a light that puts out 3/5 the light you could be having, of course at the expense of having to have more chips, but that's a good thing spreading the light out.. :

 

[tenthirty]

2cent, the thing your not getting.. is the higher you drive the chip the lower the output. look at any chart and see highest lumens comes at lower amp ratings.. in this game it's not "go big or go home" , it's about finding a happy medium between chip count / output. look at this chart to see you only get less the more you put in, add some of those numbers up and you could end up with a light that puts out 3/5 the light you could be having, of course at the expense of having to have more chips, but that's a good thing spreading the light out.. :

IMHO, the plant just like a diode runs more efficiently at a moderate level of light.
Being a biological system, the plant will adapt to whatever light is presented.
If light is in abundance the need for efficiency is decreased. No???

By feedback, I meant that with most mechanical toys, you have noise, vibration, etc giving you some sense of when you are pushing the limits - that is obviously impossible here.

Sometimes you can hear them sizzle and pop. ;-)

 

[2cent13adz]

@ rives: Then I didn't get it from start! If building 'your way' keeps him in his comfort-zone, that may be a good thing. Agree about factory-made stuff. The feedback thing; you get the temp of the heatsink. If it's cool, so are you. And TY for the welcome!

@habeeb: Eh, well the only thing I didn't get was op's wish to run slow. Interesting chart, did you do it? If so, can I ask which led was used? And at what temp? Looking at the graph in the xml pdf, it's pretty much linear at a constant temp. Would be nice to be able to compare. I wonder about the drift in CCT, is this behaviour normal among 'all' led's? Temp-related? Never gave that a thought... Anyway, I'm sure we can agree on this: When it comes to heatsinks, the game really is "go big or go home".

 

[rives]

Sometimes you can hear them sizzle and pop. ;-)

Yep, usually just prior to letting out that all-important smoke!

The feedback thing; you get the temp of the heatsink. If it's cool, so are you.

True to a point. The junction temperature of the chip far exceeds what makes it to the heatsink. The rate at which the chip's "pipeline" to the sink is limited in how quickly the heat can be conducted away. Then you either permanently damage the chip or let the above smoke out.

 

[2cent13adz]

How do I edit a post?

habeeb, graphs to look for in the pdf:
Relative Flux vs. Current (TJ = 25 °C)
Relative Flux vs. Junction Temperature (IF = 700 mA)

Then give me your thoughts of what makes things inefficient, current or temp?

 

[rives]

I think that you have to have 50 posts to edit, as well as sending more than one private message.

 

[2cent13adz]

rives, ty for info. this useless post makes seven... I'll have to be careful not to waste the one pm... rrog, sorry for cluttering your thread with noob-stuff, ok?

 

[rives]

You can delete the one pm that you use, and use the "space" again, I think. Kind of a goofy setup.

*edit* Hell, thinking about it, maybe you can only receive one and have to delete it to get any more.

 

[rrog]

Always good to have dialog. Thanks guys.

 

[habeeb]

@habeeb: Eh, well the only thing I didn't get was op's wish to run slow. Interesting chart, did you do it? If so, can I ask which led was used? And at what temp? Looking at the graph in the xml pdf, it's pretty much linear at a constant temp. Would be nice to be able to compare. I wonder about the drift in CCT, is this behaviour normal among 'all' led's? Temp-related? Never gave that a thought... Anyway, I'm sure we can agree on this: When it comes to heatsinks, the game really is "go big or go home".

no, did not do it, found it

XM-L2 T6 5000K bin is all I know, here's link

was powered on to get readings, was not run for a matter of time

CCT, not sure on much, from what I know lower you power, lower the color, higher you power higher it shifts. I don't know how that plays when actually using one for hours and it's rises in temp..

heatsinks, I would have to say there is opinion involved, being there is many factors on a heatsink ( LED spacing / mass to consider / watts used / future re-use / fans or even none / what's available .. ) I mean no one is gonna buy a 8 inch wide sink for 10 watts for a cloning project.. but yes in ease a big sink is desirable..

check out my sink in my albums, it's 'oversized" I would say, but I wanted to re-use it for flower if I ever needed, so in this aspect I went over as I know I would use it later, it also "just fit" where everything came together wanting fans / absolute performance / spacing size...




I am more interested that you mention someone running 3amp XM-L. we are hear to learn more, so I wonder if you might tell us more, on there setup, and color combination used for XM-L and would hope later pics if possible. I am wondering watts a square foot, color of chips used, and for flower or veg..

problem I see being, every chip changes color, even if you look at the XM-L to XM-L2. so the game of LED is in a constant change until they reach a "limit" and only squeeze out 5-10 lumens each chip..

 

[rrog]

For some reason I thought I was going to get one of those battery LED testers with this LED order. Not the case. I guess Amazon or eBay for the little dealios? I'll run there now and see.

OK so I ordered up a little 2 AA battery LED tester, some alligator clips to make more grounds, roach clips- whatever needs a little action.

Also a roll of white and a roll of black stranded 20AWG hook up wire. Keep my + and - wires straight...

 

[2cent13adz]

This is a random c/p from my brain, made earlier today:

Never claimed to be Einstein, but trying can't be wrong. If nothing else, it's kinda fun. You reading here might suggest you share that feeling. I'm always hunting for knowledge, not followers.

To make a test setup for finding out how a led performs at different loads, one would want everything but the current the same at all times. If temps are going wild, the test is useless. I can't do this test myself, but Cree can. So I use their numbers. Doing so, I find the main enemy to battle is heat, not amp's.

I'm not stupid enough not to know where the heat is coming from. Still, if we can beat the heat we can use the juice.

This in turn makes me think the most important part of any lamp like this is the cooling system. This is where ppl try to cut corners, when it actually is the place to spend the most. A good heatsink will last forever, led's are like any electronics; new model every month. Maybe this is why I seem to take the risk of a few blown led's in any experiment lightly. All ledlamp building is experimental!

 

[2cent13adz]

This is now:

@ habeeb, very interesting link. The guy doing the test state in his op "I allowed for the jig to cool as the test currents increased so the LED temp ranges from 25c on up as it got warmer, probably not 85c at any point but certainly warm." More of the same if you scroll down. According to the xml pdf, you lose some 12-15% right there. Some guys asking for performance at 4A, that's the spirit! They are my people!

As for me talking 'bout a 3amp lamp, are we down to "pic's, or it didn't happen?" Hehe, I just might bite. Let rrog decide if this is the place to do it, after all this ain't the 2cent-buildalamp-thread. Once I get started, can go on til doomsday...

 

[rives]

2cent - Why don't you go ahead and start a thread? You will find that almost everyone that frequents this part of ICMag likes to kick things around and examine each other's take on things, and it's obvious that you have information to share. Also, you will quickly get the post needed for "full" access to things.

And I agree with your viewpoint on heatsinks, but not necessarily with "all led lamp building is experimental". Many of the builders on here are more focused on the benefits for the plants over trying to stay on the bleeding edge of LED fixture development. Some just want something that works better than commercial units, has reduced power and running costs, and works well in their environment.

 

[xmobotx]

@ this point i have put together 3 'fixtures' {though i do intend to combine them into 1} w/ SMT diodes

now the 1st is lumiled rebel ww running @ 3v .7A {they are rated to 1A} running problem free now for about a month

the other 2 are running right up to the max rated spec

now; the cree fixture is running 3v 3A for 9 watts >it is rated for 10 watts & i have had no problem there

but the nasa bars are running right @ the max spec of the 660s 2.4v .7A{which may have been OK if i werent mixing a couple 1A rated royal blues} BUT I did have a problem w/ that circuit {catastrophic failure}

i guess you could say i am experimenting

i think i may just go w/ an additional diode on the crees and the nasa bars

 

[2cent13adz]

rives, the thing about experimental is, we are all doing it. We just do different experiments, to achieve different goals. No matter if you're pursuing a low bill or the overkill race-lamp. Don't forget those focused on spectrum, making small adjustments hoping to improve things. Every lamp looks different, custom sized and using a variety of led's. What else do you need to call it experimental?

Starting threads and such, I'm not sure. Kinda enjoy a low profile. Haven't posted anywhere since OG went down, and not too keen on spreading pic's around these days. I'll think it over.
---Guru Meditating---