Self-made LED panel general discussion

[puffenuff]

Yes, but what I learned from TL is that you only need it for a short period immediately after lights out, so it would need to be separate fixture and on a separate timer. Problem is finding small wattage IR bulbs. Reptile lights has one with a remote ballast. Now all that's needed is a cheap light mover. Alas, this won't work with my 8 bulb Quantum BB. Next time I would get 2 @ 4 bulb, unless I bite on some RapidLEDs

Yes, you can use IR when the lights switch off for 15 or so minutes to alter the chemicals into staying in daylight mode if you will. This should trick the plants into thinking they are getting more light, supposedly something like two extra hours. Not sure exactly how it works though, read a couple articles on it back in the day and saw a video of eddie r talking about it too. I don't run mine 12/12, only for a couple hours during mid day and evening.

 

[patrikantonius]

Objectively, a good hps/ballast is proven to grow excellent fat buds. Also proven: they generate a lot of heat. LEDs are sol on less heat, so why not simulate the SDG of hps using leds?

Earlier you were talking about efficiency. Well that's relative. Blue/Red diodes are totally inefficient at producing any other spectrums, and yet those other spectrums are very important. Whites are very efficient at producing a total PAR range, and pretty muich all the 450-650 mj growers need.

In terms of pure yield efficiency (dry grams produced per consumed watt), the best results I've seen, at least for low power installations, are with red & blue LEDs and not HPS. Red & blue LEDs are not efficient at producing large spectrum but isn't that what it's all about, producing the most light where it matters the most ?

 

[PSUAGRO.]

^^^^All of this...ha................Anyways TL knows his stuff and I wouldn't disagree about far red, but Hans last run on http://www.ledgrow.eu/ his panels run the far red constantly and his 153grams from 112w led(not including driver/fan) is documented and pretty impressive...........let me check the panel again/ wanna be sure it runs far red in all phases.........


Edit: Yep the two and only Deep red rebels (Hans calls them farred/ don't know actual nm/ he says it performed better than the "cherry red "714nm leds ) run on all phases veg/early flower/ full flower

Its the two leds in between the xpe blues/ This is the veg/clone phase in the pic...

 

[PetFlora]

^^^^All of this...ha................Anyways TL knows his stuff and I wouldn't disagree about far red, but Hans last run on http://www.ledgrow.eu/ his panels run the far red constantly and his 153grams from 112w led(not including driver/fan) is documented and pretty impressive...........let me check the panel again/ wanna be sure it runs far red in all phases.........


Edit: Yep the two and only Deep red rebels (Hans calls them farred/ don't know actual nm/ he says it performed better than the "cherry red "714nm leds ) run on all phases veg/early flower/ full flower

View attachment 2182717View attachment 2182718 Its the two leds in between the xpe blues/ This is the veg/clone phase in the pic...

Yes you can run them throughout lights on, but when run at the end (and again first learned from TL who goes by another UN on another site, but this reinforces it:

From SkunkSkool

The first and last hour of a day's sunlight is mostly red light because of the scattering effect on blue light. "So at the onset of the dark period much of the phytochrome is in the Pfr form. However, Pfr is unstable and returns to phytochrome Pr in the dark." The red light in sunrise returns the Pr to the Pfr form. "Phytochrome Pfr is the active form and controls flowering and germination. It inhibits flowering of short-day plants (the long night period is required for the conversion of Pfr to Pr) and promotes flowering of long day plants."

 

[jubiare]

I think when we talk about efficiency here, it's about photons usable for growing

 

[jubiare]

Yes, light in the 740nm range seems to be very beneficial, but, as we can see from the radiant curve of the Cree XP-E warm white, only less than 4% of the total light is emitted above 700 nanometers. This means that putting independent far red LEDs seems to be more efficient to cover that part of the spectrum.

mAYBE MORE THAN 4% i THINK? But yes, maybe it's best to use IR emitters, it all depends on how much IR we really need

 

[PetFlora]

I think when we talk about efficiency here, it's about photons usable for growing

Yes that is the common understanding, but...

 

[IlovePlants]

I'm not sure where I'm going with this but I would like to chime in. I think that most of us are actually right on most of our points, but I wanted to elaborate on something. Personally I feel that for the average person it is better to avoid white LEDs for the next 2 years. I'm waiting on efficiency to increase, the best lm/watt reading on any white LED I've seen so far is about 115lm/w which is nice, but for the price of making a 1000w 5800k LED you may as well buy all HPS, grow a few times with them to get some money, before you ever even have enough money for the LED Panel.

I'm not trying to discourage anyone from using white LEDs, I just feel as though you would be ripping people off by telling them that they need their white light from LED's because they are more efficient. They are not efficient when you look at Dollar/Lumen/Watt/Lifetime. How is a 71 watt Kessil over 300$, that's fucking outrageous, but when you build a pure white version of them yourself they still cost almost as much money, and you spent the time building the units. Check out MajorCoco's thread, I'm sure all of us have. He even states how costly of an experiment it is. So why put growers through all this effort and money when there is already a product, either CFL or Fluorescent, that does the same thing at nearly the same efficiency as some of the lower quality white LED that are available (about 85lm/w).

It's not about saving the environment for me, I worked with an environmental protection agency, as long as you PROPERLY RECYCLE your fluorescent lighting you should do little net harm to the environment. The bulbs I buy are produced in my home country, meaning that I'm not paying to have a giant oil powered monster deliver them from half way a world away. When you buy things from China they don't buy things back from the U.S. or any other country, they have enough land and natural resources to produce what ever they need. Meaning you are contributing to a economically unsound practice of personally dismantling the infrastructure of the global economy.

Don't waste your man hours on this bullshit, I'm sorry about the language, but I must protest! Use the cheapest, local, and most responsible resources. I feel that if you can and feel the need to add white because you are experiencing deficiencies and photo-inhibition, use CFL or fluorescent lighting. I can buy a 40 watt t10 6500k that puts out 3800 lumens of the most PAR efficient neutral light for 4$ per bulb, made in the U.S., and my local recycling plant accepts old fluoros. Winner winner chicken dinner! Within a the next five years, if efficiency keeps increasing, I bet that pure white LEDs will give fluorescent lighting a run for it's money once and for all, but for now stick with what is cheap and what works. Just my opinion, and I hope it helps.
Keep it real,
ILovePlants

 

[patrikantonius]

You threw in the china subject, and there's something that makes me wonder. There is no grow LED light manufacturer who discloses which LEDs they are using. I think this is much more important than any other info that they might give away. Some write that they use "epistar, bridgelux" chips but those chips are the most difficult to find info on. If I were pessimistic, I'd say that this very fact give them the freedom to put whatever they want (understand: the worst quality/cheapest chips that one can find). Some other manufacturers say that they use "cree" chips. This is almost as vague as before. Even if they say using cree xp-e, it is not giving any guarantee whatsoever towards the efficiency of the chips.

I found on the epistar website a 660nm red LED which features 175mW radiant output at 500mA (the voltage isn't given, we can suppose it is around 2.2v like most red LEDs). That gives around 160mW/W radiant ouput.
The Cree XP-E red (625nm) bin N3 (which is used in the ledgrow panels) features 265mW radiant output at 350mA/2.1v which gives 350mW/W. That's more than two times more.
The Osram Oslon SSL deep red (660nm) bin 3T (the best available) features 380mW radiant output at 400mA/2.15v which gives 442mW/W. That's almost three times the epistar's.
The XP-E red bin P3 can even outperform the Oslon SSL.

What I'm trying to say is that theoretically a single N3 Cree XP-E red can outperform two epistars, which I believe are used in most commercial grow panels. Do anyone have the opportunity to compare ? Has the comparison ever been made ? (a single spectrophotometer should be enough to have usable figures, or even a digital camera if the wavelengths/spread angles are similar)

 

[PetFlora]

The above two comments are why I am sticking with my 4ft HO T5 (using aquarium bulbs) until LEDs can match performance & coverage at a similar price. Now, that said I do have a small starter/clone closet where ~ 70 quality LED watts (ie Hans GrowLED.eu), OR a RapidLED solderless kit would be cool

 

[bassclef]

Don't waste your man hours on this bullshit, I'm sorry about the language, but I must protest! Use the cheapest, local, and most responsible resources. I feel that if you can and feel the need to add white because you are experiencing deficiencies and photo-inhibition, use CFL or fluorescent lighting. I can buy a 40 watt t10 6500k that puts out 3800 lumens of the most PAR efficient neutral light for 4$ per bulb, made in the U.S., and my local recycling plant accepts old fluoros. Winner winner chicken dinner! Within a the next five years, if efficiency keeps increasing, I bet that pure white LEDs will give fluorescent lighting a run for it's money once and for all, but for now stick with what is cheap and what works. Just my opinion, and I hope it helps.

I see your point. But if we're already building LED lamps, why add big, bulky, hot floros? I grow on the micro level and LEDs are much, much smaller. And cooler. And, someone has to experiment with this technology, who is it gonna be? Are we going to just use HPS forever? I think not. In a way, we are holding the retail LED gods accountable by figuring this out. If I can pull 2 ounces from a case with a 50w LED lamp, I think I win the environmental argument.

You threw in the china subject, and there's something that makes me wonder. There is no grow LED light manufacturer who discloses which LEDs they are using. I think this is much more important than any other info that they might give away. Some write that they use "epistar, bridgelux" chips but those chips are the most difficult to find info on.

Which brings me to my next point. Even if they are using quality LEDs, it doesn't matter if the thermal management is no good. The red InGaAIP LEDs lose lumen output exponentially once the junction temperature goes above 40 Celsius. That's not actually very hot. White and blue ThinGans fare a bit better. A large heatsink and good thermal path are a must to get the most out of them.

And we already know how cheaply some of the China units are made--do you think they're giving any thought to the size of the heatsink? Hell no, they just mount a bunch of LEDs too close together on a sheet of aluminum, slap a couple cheapo fans on it and put it into an enclosure with vent holes that are too small. In order to fill a canopy with thick, dense buds, these LEDs need to be working as cool as possible.

Check this out: LEDs Temperature Effect

I actually think water cooling is a great idea. PC builders use it all the time with great success. The water circulates through the heatsink to a large radiator with fan attached. Since we're hitting those peak PAR values, we may as well hit them as hard as we can.

 

[Gastanker]

I've grown very successfully with CFLs reaching up to .65g/kWh which actually beats the the yield Hans has seen with his best LED array - but CFLs are annoying. Don't get me wrong, I love them, and I use them all the time for growing and own likely 100+ bulbs which I all got for super super cheap, like $0.25/bulb, but the constant upkeep with moving bulbs and plants can be a pain if you don't have bukoo time on your hands.

Totally correct about costs - I can run 400w of CFL for under $50 and my 200w of cheap chinese DIY leds are costing me ~$160. But aside from the cost the LEDs are just easier - or at least I imagine. A lot less bulb movement, much fewer individual light sources to worry about, less light depreciation over time... if they turn out to me more efficient that's just icing on the cake.

 

[puffenuff]

The only drawback I've heard of for white leds is that flowering takes a bit longer than with the r/b combos. Not sure if it takes longer to induce or finish, the details from the one report were second hand info, so perhaps it could have been strain or growing condition dependent.

 

[patrikantonius]

I see your point. But if we're already building LED lamps, why add big, bulky, hot floros? I grow on the micro level and LEDs are much, much smaller. And cooler. And, someone has to experiment with this technology, who is it gonna be? Are we going to just use HPS forever? I think not. In a way, we are holding the retail LED gods accountable by figuring this out. If I can pull 2 ounces from a case with a 50w LED lamp, I think I win the environmental argument.
Which brings me to my next point. Even if they are using quality LEDs, it doesn't matter if the thermal management is no good. The red InGaAIP LEDs lose lumen output exponentially once the junction temperature goes above 40 Celsius. That's not actually very hot. White and blue ThinGans fare a bit better. A large heatsink and good thermal path are a must to get the most out of them.

And we already know how cheaply some of the China units are made--do you think they're giving any thought to the size of the heatsink? Hell no, they just mount a bunch of LEDs too close together on a sheet of aluminum, slap a couple cheapo fans on it and put it into an enclosure with vent holes that are too small. In order to fill a canopy with thick, dense buds, these LEDs need to be working as cool as possible.

Check this out: LEDs Temperature Effect

I actually think water cooling is a great idea. PC builders use it all the time with great success. The water circulates through the heatsink to a large radiator with fan attached. Since we're hitting those peak PAR values, we may as well hit them as hard as we can.

Very good points. Most manufacturers specify the output loss vs. temperature rise so it is easy to see how important cooling is, and you are definitely right.
I still think that Crees or Osrams would outperform Chinese LEDs in pretty much any comparable situation, but as we are talking about DIY, thermal management is indeed a very good topic to discuss.
I like your idea of watercooling but watercooling has several downsides, especially for cooling LED panels.
As we discussed earlier, it seems to be optimal to put the LEDs on the largest surface as possible to have an even illumination among the plants and thus optimize light penetration. The bigger the surface, the more difficult it is to cool it with water, or at least it's what I think.
Anyway don't hesitate to share ideas about this because I think it is a great idea.

I've grown very successfully with CFLs reaching up to .65g/kWh which actually beats the the yield Hans has seen with his best LED array - but CFLs are annoying. Don't get me wrong, I love them, and I use them all the time for growing and own likely 100+ bulbs which I all got for super super cheap, like $0.25/bulb, but the constant upkeep with moving bulbs and plants can be a pain if you don't have bukoo time on your hands.

Hans made almost 1.6g/kWh with his LEDs on his best attempt; I believe LEDs can be much more efficient than anything else when used wisely.

 

[bassclef]

I still think that Crees or Osrams would outperform Chinese LEDs in pretty much any comparable situation, but as we are talking about DIY, thermal management is indeed a very good topic to discuss.

Agreed. If the Chinese LEDs would just realease the freakin' datasheets, then we could see how to best use them. But I'm afraid there aren't any.

I like your idea of watercooling but watercooling has several downsides, especially for cooling LED panels.
As we discussed earlier, it seems to be optimal to put the LEDs on the largest surface as possible to have an even illumination among the plants and thus optimize light penetration. The bigger the surface, the more difficult it is to cool it with water, or at least it's what I think.
Anyway don't hesitate to share ideas about this because I think it is a great idea.

I know, air cooling is just easier. I visualize copper pipes with LEDs attached somehow running through the growroom. Water would simply flow through the pipes to a radiator or chiller. The tricky part would be finding a flat surface to mount the lights to. You could drive them harder and keep them cooler this way but I wonder if the effort would be worth it.

 

[jubiare]

if you spread the leds nicely around the right heatsink (enough dissipating area, you calculate this by summing up the whole area, you count all sides/fins), you don't need to actively cooling. And anyway cree, osram, and/or rebel you wouldn't need to space them apart less than an inch, that is silly/unnecessary. Best is two inches apart

For peace of mind, 21inches for every watt burned is ideal. At the moment I have my leds running lke 16inches fr every watt burned, still undecided what to do, if putting a fan or, or stick some extra finned heatsink on. It could easily be ok as is, but you want to be sure! One important thing about junction temperature is, apart from longevity, once you run them too hot the nanometer are altered too. no good

 

[FranJan]

So what about RGB white LEDs instead of phosphorous white? No energy wasted on conversion, though heat management is probably a bitch. Check these out and with a little binning you could probably base an entire panel around these.

CREE Multi-Color XLAMP MCE 100LM RGBW COOL
LedEngin Multi-Color RGBA Full Color
Illumitex ABEO

There's some interesting numbers here I think. I love the fact that Cree sticks a 7000K in with theirs. LedEngin's got that amber that everyone's lovin' lately and can be binned quite well for growing. And Illumitex looks like it wants a chunk of the market. Still there's a lot of numbers involved with these and they just may not be ready for prime time.

And how come when you go to Cree's website and you search for "gardening" or "horticulture" you get zero results. Is Cree trying to tell us something .

 

[Gastanker]

Hans made almost 1.6g/kWh with his LEDs on his best attempt; I believe LEDs can be much more efficient than anything else when used wisely.

kWh? Really?

Fucking brilliant.

 

[IlovePlants]

RGB white seems like the best step towards great white LEDs. Are there any spectral analysis on the spectrum blend? ei. As differing wavelengths approach each other they collide and alter each others wavelength, so does the 525 nm green of the Illumitex actually fill in the spectral gap (since the rest is simply absorbed by chlorophyll) or is it too narrow only allowing for minor mid band absorption? I personally like the look of the Led Engine to fill in the spectral gap. I think this type of LED is "The Great White Hope", and that's all you fans of boxing or James Earl Jones. Very nice post FranJan, just what I've been looking for!
Sincerely,
ILovePlants

 

[bassclef]

So what about RGB white LEDs instead of phosphorous white? No energy wasted on conversion, though heat management is probably a bitch. Check these out and with a little binning you could probably base an entire panel around these.

CREE Multi-Color XLAMP MCE 100LM RGBW COOL
LED Engine Multi-Color RGBA Full Color
Illumitex ABEO

The Crees and LED Engines are not efficient enough at first glance. Yikes, and expensive too. You could get much more efficiency buying separate LEDs. I like the Illumitex a lot, but at $450 for a 70ish watt light it would take some convincing (although it's probably worlds better than a Blackstar). Their 660nm red is the same light output as Osram's GD+ version at a lower current (if I'm reading that right). Good light, that one. If Illumitex dumbed it down a bit and upped their marketing game, they could probably clean up.

Also, I don't like the idea of having so much far red (730nm) in the light. I think we get enough using whites. I'd rather replace an equal division of wavelengths with more red that the plant can use for flowering, say 5R:2W:1B as a starting ratio.