DIY initial thoughts

LOL Holy Shit Trunche0n I'm sorry. I didn't even realize it was your thread. LOL Sorry, stoner moment.

I think it's damn hard to select the proper components when building a panel. It's what primarily holds me back from building my own. IMO once you put CREE in your mix other companies offerings begin to show their limitations, especially when running at the same mA. At 700mA I bet those CREE CWs just dominate everything on your panel, as in the XTE CWs will be responsible for any bleaching that happens since their blue output is pretty significant. So the decision is to drive the XTE CWs at a lower current or maybe use a XPE/XPG instead is there and are the kind of decisions stoners like me best leave alone. LOL. Best Wishes Trunche0n.

Ok - For now I am concentrating in dialing in the flower panel as My CFL Veg is going well and i'll have something to flower before I have something to start a fresh veg with.

To recap I am at:
22WW : 12(660) : 8(630) : 6CW for my flower spectrum

45% Cree XT-E Warm White Bin Q4 2700k min cri 80 XTEAWT-00-0000-00000HCE8
25% Luxeon ES Red 660nm Bin EX6 LXM3-PD01-0300
17.5% Cree XPE 630nm Bin P3 XPeReD-L1-0000-00801
12.5% Cree XT-E Cold White Bin R5 6200k min cri=70 XTEAWT-00-0000-00000BH51

Now I 'Think' I understand you Fran where you are saying the cree's pack way more of a punch regarding the lumen output..

BUT instead of turn down the crees below 700ma i would prefer find a more powerful red 630 and 660 so I can make this work - If that means running 1000ma for the reds so be it.

This osram Oslon SSL 80 LRCP7P-JRJT (JS) looks a little beefier and may be able to match the Crees??? is this a possibility?

I really cant find a more powerful 660nm - can anyone help?!

Ok So to mach the crees luminosity I could run the 5w LZ1-00R205 ledengine led's but havent read anything on these and I am in no mans land regarding my ratio

Certainly out of my depth now

*blows the horn of gondor* will try to answer all Qs so far best i can...Supra probably has some more to say.

Trunche0n, your choice of colours looks solid to me except still too much blue from those CW personally. I would switch them to neutrals maybe, like Pet said.

I can't comment on the XPE coz i never used them. but remember lumen output is not the same as radiometric output. You could conceptualize it as lumens are "biased" towards what the human eye is best at seeing, whereas radiometric output is the amount (measured in power) of actual light waves that are emitted. Of course, with whites, they specify in lumens coz they are dicks

^^ This is why white LEDs like the cree's look like they shit on the red LEDs.

Fran, RE: running at 700ma - Personally i run my reds harder, but the most sensible option is simply to use a larger proportion of red LED (because of efficiency losses when driven hard). As far as certain LEDs bleaching while others are ineffectual, i would be equally attentive to beam angle in this case, in conjunction with more diodes.

Trunch, Oslon SSL 80's are really nice - look to stardustsailor's thread for his build with them. They also do a standard red, warm white, cool white, etc so you could potentially look into using them exclusively.

I run Ledengin LZ-1 myself at 1.05A. (660nm, and some protesting 460nm heh...blue LEDs become less efficient as current, rather than junction temp, goes up) I really like them, though the beam angle is narrower than i would like personally. Should be some pics of them up in the next couple days.

Also pretty sure there is an R3 bin of the LZ1s.

My gf is frowning at me so enough

Hope this helps

I would go with the plan as you have it now. You are not going to find a better deep red than the Phillips Luxeon ES and I can say that I run mine at 700ma. Cree and Philipps are on par with each other and who has the best of what usually seems to be mostly a matter of who put out the last upgrade. The CW will look to your eye to blow everything else away; your eye would see the cheapest chinese crap cool white as brighter than the best 660nm. Your eye sees blue better than any other portion of the spectrum and sees red rather poorly. Lumens are a useful unit of measure for lighting an area for humans but your plants don't "see" things the same way and lumens is really almost useless in discussing light for plants. You've got a basic good plan, everyone would tweak a little differently but after a while, you have to kill the engineer and make something or you just end up with a lot of discussion and no light.

As far as figuring heatsink...since you are going with "best" brand you should be fine with about 16in2/watt. I wouldn't go thinner on the aluminum face than 1/4". If you have some scrap angle aluminum, you could make some ghetto fins and use thermal adhesive to attach them and get your square inches of cooling fairly cheaply. Or you might get lucky on e-bay or a recycle electronics place for some cheap readymade heat sinks.

Truncheon you are correct it is complicated to sort out the percentage of blue but it is really the first thing to key on when building your spectrum IMO. I use a spreadsheet to get the job done which is possible thanks to the specified bins. KNNA recommended 15-20% and specified actual output watts. I will run your numbers and get back to you regarding the approx percentage blue.

FJ the XP-E P3 bin does excellent at 700mA. It is 42.2% efficient at 700mA (@ 50c). I was so impressed by that I changed my original design which was relying on warm whites to provide the 630nm red. I calculated the oslon 630nm JS bin to be only 24% efficient at 700mA (60c).

As far as sourcing heatsink - 2 things i forgot

1. You could consider using a copper plate if you go with the "base plate" style design. If you want to give yourself a headache, you could worry about the differing thermal expansion coefficients of copper and Alu (stars) and the long term wear of constantly expanding and contracting at different rates as lamp heats up and cools down.

2. i get loads of really cheap heatsinks on ebay.

I was hoping to use these! each rated for 215w tdp
View attachment 2614248View attachment 2614249
if i can figure ou a way of attaching (thermally efficient transfer) these well to aluminium sheets maybe one per panel I figured it was do. They are old stock now so 10 dollars each and monster coolers!

You are a legend! thank you :thumb:

SupraSPL said:

Truncheon you are correct it is complicated to sort out the percentage of blue but it is really the first thing to key on when building your spectrum IMO. I use a spreadsheet to get the job done which is possible thanks to the specified bins. KNNA recommended 15-20% and specified actual output watts. I will run your numbers and get back to you regarding the approx percentage blue.

Click to expand...

a rough calculation for the amount of blue
cool white(6500K) = 28% blue; 25% red
warm white(2700) = 8% blue; 47% red
and 0% blue for all reds

a percentage calculation over
22 parts ww
6 parts cw
20 parts reds

gives about 8-9% blue
and this is too much?

Guod with reality check....

I have this in my threads, but think you will find it most enlightening, even though written in 5/2012

Quote:

Quote:
Originally Posted by Phaeton View Post
The spectrum of light the plant uses efficiently changes with the intensity.
At low light levels red is used most then blue, just like the chlorophyl charts.
As the light gets brighter more blue is used, and more. The break point is reached about 50% of max leaf capacity then green starts coming on. As the intensity continues up red and blue remain steady but green use continues to grow until at maximum it is almost half of all the light energy being used.

This is one reason why, even with the rather shitty, lopsided spectrum produced by HPS, one can still get good results with them:


I'm glad we've finally gotten some good studies on green light over the past several years, as has been mentioned previously (link) by a few of us.

While 'every lumen (or rather, PPFD) is sacred', I'm of the camp that would prefer a higher level of (adjustable) full-intensity, multi-spectrum (i.e. 'white') light incorporated into the main fixture, for that very reason.

And with the recent increases in the efficiency of neutral whites, there's no reason why you can't get perfectly good results with just a two-channel, adjustable led fixture (neutral white, and red), supplementing with the aforementioned only as needed.

For reference, here are the LUXEON (Rebel) Neutral White and CREE (XP-E) Whites - relative spectral distribution:

This image has been resized. Click this bar to view the full image. The original image is sized 801x530.


Now, let's look at all of them superimposed over the PRC (can't get this chart to C/P)

Quote:
Originally Posted by Phaeton View Post
UVB and near infrared are used for non chlorophyll purposes on healthy well lit plants.

Es verdad, amigo. And the best, most economical sources of IR for most growers are still simple, good ol' halogen lights, and Reptisun fluoros for UV-B, at present. (the dangers - and potential litigious nightmare - of incorporating expensive UV-B LEDs into a fixture notwithstanding)

Besides, without complete, independent control of the latter two (i.e. independent of the activity and photoperiod of the 'main' LED fixture), the grower's control over their desired photomorphological changes becomes rather tenuous at best.

If one is going to go that route, then both UV- and high-intensity, blue-mediated light damage (since that is what it is) should be adjustable - both in intensity, as well as photoperiodicity and duration.

As the higher-energy end of the spectrum isn't really a 'finishing/maturing' as much as it is a degradation (i.e. blue and UV pass through clear trichomes just fine; it's only when they become cloudy that they show any significant absorption of that energy, and quickly turn from cloudy to amber - at which point one should watch 'em like a hawk to keep your product from degrading too soon and ruining the desired effect), it should always be incorporated judiciously at first, and in small doses - until the effect on that particular cut is well-established, after which it can then be predicted with a 'fair' level of accuracy.


(note: neutral white I called 'normal white' here for some reason. Wonder what I was smokin' at the time...<whistles>)

As one can see, the CREE Neutral White (I call it 'Goldilocks', because it's almost 'just right' ) has a RSPD that still allows nearly ~25% of its total power in the blue range (and plants only really 'need' ~8-10%), and more that 1/3 of which (i.e. the area under the curve) is over ~580nm or so (which has a Photosynthetic RS of over 90%!) - which is much better than even your typical 'Enhanced HPS'.

Couple that with strong white light (green-response chlorophyll extending throughout and deep into leaf structures, with a net effect at or near that of the (mostly) surface-level blue and reds), which also takes care of most of the ~660nm+ you actually need for photomorphogenesis - and you can get by with 630nm reds just fine.

(i.e. 630nm red is ~95% of the PSR of 660nm, AND they currently still have ~20-30% greater radiometric efficiency - as well as being cheaper than the deep reds - so there's more 'bang for the buck'):

As one can see, the CREE Neutral White (I call it 'Goldilocks', because it's almost 'just right' ) has a RSPD that still allows nearly ~25% of its total power in the blue range

Click to expand...

regardless of manufacturer..
Radiometric flux of...

..warm white
2700K about 8% blue
3000K about 10% blue
3500K about 13% blue

..neutral white
4000K about 16% blue
4500K about 19% blue
5000K about 22% blue (would not call this neutral white)

at 5600K we get 25% blue

For fans of the neutral white, check out the numbers on the new XM-L2, very impressive. I only calculated up to 2amps since it would be cheaper and more efficient to use XT-E at that point. The 5000K T6 XML-2 is 48.6% efficient at 1 amp (50c). If I understand correctly, this means the best Cree blue chips are ~70% efficient at 350mA.

Hey supra very Very impressive,
As a bonus these are available mounted on stars in europe which is great. the t3 looks healthy enough in the 630 range. Would prob need to bump up the blue with a sparing amount of U2 but could you recommend a 660 red to keep up with the intensity of the xm-l2?

Im assuming you mean, when you say "keep up with its intensity", run at around the same wattage? (at 2A?) rather than wattage + beam angle? or both?

The only that springs to mind is i think there's a LedEngin LZ3, which is rated for something ridiculous like 10 watts?

prepare for it to cost.

indeed: 30 dolars for a 10w 90 degree ledengin LZ4!!

silasraven said:

prepare for it to cost.

Click to expand...

Well I was hoping there might be some 660nm that had 120 beam angle and maybe could "keep up" running the xml at 1A? Then again the xt-e at 700ma doesnt seem comparably less in output regarding the difference in cost (half the price and probably half the heat)

Everytime some one posts a new led find I keep on thinking "yes! this will be a more bang for buck way of doing things than what I am currently thinking of".... probably due to my snail pace grasp of running the math....

SnotBoogie said:

Im assuming you mean, when you say "keep up with its intensity", run at around the same wattage? (at 2A?) rather than wattage + beam angle? or both?

The only that springs to mind is i think there's a LedEngin LZ3, which is rated for something ridiculous like 10 watts?

Click to expand...