LED DIY (planning, suggestions are welcome)

PetFlora

Well-Known Member
Chiming in, FWIW,

The 410 through 625 are amply covered in the 6500 and 2700, IMHO adding these specific spectrums is redundant, replace them with more WW, but on a separate o/o switch to add once you hit ~ 10 days into flower. None of the strains I have grown respond to 660, which BTW only makes up ~ 5% of mj's PR: 650 is much more useful


Hi all!

I'm planning on making my own LED grow lights, and I would like to have your opinion.

I was thinking about using 100 x 3w leds, with a ratio of something like:

6x 730nm
40x 660nm
20x 625nm
6x 410-420nm
12x 440-450nm
8x warm white 2700k
8x cool white 6500k

What do you guys think?

As for the driver, I've found this one http://www.ebay.com/itm/140727706400 (AC100-240V to 5V DC 60A 350W) ...will it do the job or should I get a 12V one?
 

PICOGRAV

Well-Known Member
They are similar becase they are white. But not identical. The cob's spectrum is wider and less dialed. Slightly more green in it than the xt-e.

I didn't even bring up the cool whites. I stated 5000k cxa vs 5000k xt-e...and the xt-e is 130lm/w vs the cob 124lm/w. read below...




Why are you quoting me for this...considering it is not me making the panel, I only answered his question for what he was planning, not trying to chang his design to do exactly what I do(like you are). If you looked at what I am using, you would know that I would make a white dominate panel myself and supplement the red for the flower boost. But feelmyanger asked about the whites that he already planned on using in his design.
And to expand on that thought I would use both warm and cool white then pure red as well. I think that cool whites with additional pure red are very effective. The addition of warm white would just broaden the red output. But is not total necessary to the design if you do it right, like the apache I have been using have already shown that.


EDIT:
What is the best bidgelux that you are talking about? The one you are using(vero 29) says 120lm/w?
IMO, the hole idea of splitting the spectrum up like people do with the Christmas Panels is to achieve overall high efficiency of the lighting system in two main ways:

1. Using narrow banded LED's to conserve electrical power, as these are generally more efficient then same power whites.

2. Only giving plant light where it can absorb it at higher levels, mostly blues and reds.

The hole system does make sense to me, but whats the point of adding white back into the mix, if the idea of wasted light spectrum is what you are trying to avoid?

Stabilized DC Performance of the Bridgelux 5000K chip is about 120 l/w yes. Here is a comparison of the two chips in real life figures:
Cree Bridge.jpg
 

PetFlora

Well-Known Member
Huh? What could be higher efficiency, one NW or WW led that covers ~ 410-630, or a bunch of individual diodes?

Don't forget that the sun has all the colors, so do quality NW /WW diodes. The plant will take what it needs and leave the rest. What it can't do is manufacture spectrums that it needs, but is not getting when the whole 460-590nms are left out, due to misinterpreting the mj PAR graph

Let me use an analogy of a recent off-road vehicle test I saw on Speed Channel, where a stock original Land Rover type was compared to a $200K+ heavily modified version of the same vehicle. Tiff drove both on the same off-road track, back to back, driving the heavily modified version last, so he had track and vehicle familiarity going into the second run. The difference in time was ~ 5-6 seconds. That's a lot of extra money for a relatively small decrease in time.

What I am trying to point out is, unless you are hell bent on squeezing out a few extra grams, today's crop of NW + WW (that has a vg 650 range) is all you need to grow excellent meds. KISS my brothers
 

tags420

Well-Known Member
IMO, the hole idea of splitting the spectrum up like people do with the Christmas Panels is to achieve overall high efficiency of the lighting system in two main ways:

1. Using narrow banded LED's to conserve electrical power, as these are generally more efficient then same power whites.

2. Only giving plant light where it can absorb it at higher levels, mostly blues and reds.

The hole system does make sense to me, but whats the point of adding white back into the mix, if the idea of wasted light spectrum is what you are trying to avoid?

Stabilized DC Performance of the Bridgelux 5000K chip is about 120 l/w yes. Here is a comparison of the two chips in real life figures:
The reason people do narrow band is because they see the absorption peaks and think that that if they just use those all the light will be usable compare to the shitty hps spectrum they are used to...and they usually think that green and yellow are wasted. But they are still important and necessary.
Both of those are untrue as white are basically either blue with some red(CW) or red with some blue(WW). And new panel makers don't always know the benefits of white. And the green and yellow that is present is not wasted. Up to 70% of green light can be usable, so I don't understand how adding whites to a narrow banded panel would be negative or any kind of waste, the light is used.

So I was right and the cree xt-e is better, you just proved it your self. Stop bringing up the 6000+k CW...I specifically gave you the top 5000k xt-e(R4 bin 130lm/w in 5000k) against the 5000k cxa3050 (which is better than the bridgelux you are chubbing over). If you found the CW/NW in a 5000k it just makes yours worst.

That comparison in the pic is so false because they are not driving the xt-e properly how someone would efficiently use them...driving at full power then judging, is not accurate. Copare their best vs they others best ...and lm/w shows that without same wattage. All your pic shows is that the wattage's are different...and that you need more...but considering it is a 5w vs a 100w, that was already pretty obvious. The point is how they will work in a system to give you a great grow light. And when you use the right and best components(not cobs) it will be the best light possible. I didn't say the simplest...the best.

I saw you picking a fight with someone in an other thread and was laughing, because it's two new/think they know everything going at it light they know what they were talking about. And I was actually siding with you...but now I see that you were the main problem the whole time. I know you like what you have(probably why you bought it), but it is not the best and no arguing is going to change the facts.

And I still don't know how you think that 100w from a small point source is better for your canopy than spreading 100w of better/more power out over the whole canopy.

P.S. Whole(as in all together, ie...as a whole) is spelled whole with a "W"...not "hole".
 

Spiko

Member
Huh? What could be higher efficiency, one NW or WW led that covers ~ 410-630, or a bunch of individual diodes?

Don't forget that the sun has all the colors, so do quality NW /WW diodes. The plant will take what it needs and leave the rest. What it can't do is manufacture spectrums that it needs, but is not getting when the whole 460-590nms are left out, due to misinterpreting the mj PAR graph

Let me use an analogy of a recent off-road vehicle test I saw on Speed Channel, where a stock original Land Rover type was compared to a $200K+ heavily modified version of the same vehicle. Tiff drove both on the same off-road track, back to back, driving the heavily modified version last, so he had track and vehicle familiarity going into the second run. The difference in time was ~ 5-6 seconds. That's a lot of extra money for a relatively small decrease in time.

What I am trying to point out is, unless you are hell bent on squeezing out a few extra grams, today's crop of NW + WW (that has a vg 650 range) is all you need to grow excellent meds. KISS my brothers
Hi, can you point me in the direction of someone using a DIY light like the one you're talking about?
 

PICOGRAV

Well-Known Member
Seems like I am not the only one bending the truths so to speak...
at.jpg

because it's two new/think they know everything going at it light they know what they were talking about
You might have to think about what you where saying there and try again?

P.S. Whole(as in all together, ie...as a whole) is spelled whole with a "W"...not "hole".
I guess I was just thinking about another type of hole, when I was replying to you...
 

tags420

Well-Known Member
Seems like I am not the only one bending the truths so to speak...
View attachment 2859061
You might have to think about what you where saying there and try again?

I guess I was just thinking about another type of hole, when I was replying to you...
I am not apache, so what ever they say is not me. Nor did I even recommend them in this thread...just that I use them, thus supporting their spectral theory. I tell people all the time their website blows. They also say 14% and 50% lens angles but people don't catch that and think they are talking angles. Plus the site is from 2010. I started growing with them in 2011.
I have shown and given everyone on here the truest info on every AT product I have. The original at120 pulls 158w from the wall...LED's, drivers, and fans. the new at200 pulls 196w. The website maybe wrong for the old models but will be much better soon. And that doesn't exactly matter because the real numbers are better than the website, and the output/w is better than any other unit out and has been for 4 years. And not the at200 is raising it even higher.

You have been a member since aug 2013(new) and when you first came on you blew up the led section about your little project. Cool DIY but not a revolutionary thing that will change or raise the bar in the industry. And you kept talking about the lumens and how bright they were to you. I have never tried to bring down your decisions specs, they are what they are...only shown you that there are better options.

If you want to call me an asshole...don't be a pussy, just do it.
I was nice at first, and said nothing about you or what you are doing. Just about actual cob info...then you took it personally because that is what you use. But just because you like and use them doesn't change facts about what is truly best. And when I showed you real info you tried to disagree by using opinion, not facts.
 

lax123

Well-Known Member
feelmyanger, whats your Budget? i will smoke some and then present you a cob alternative. because if i built my DIY Panel again id go cob now.i think stuff like going 100W sole cob is fail, id go for e.g. 5x 20W instead, smoothens disadvantages of cob and 3W -1 much better coverage 2- no freaking 10000x 3Watters to solder...disadvantage -lack of lenses, as i just discovered yesterday (got lenses for my Panel) some freaking cheap lenses make a huuuge difference, atleast visually, got no "meter"
 

lax123

Well-Known Member
pico, regarding the pic: i think you just cant go and do something like 120V * 1,9Amps that would be considered 2xx VA but not W... its not dc but AC...or am i wrong? and that high Amps value could be because of like "initital current" for powering up, like a normal bulp 60W that sucks much more for a Brief Moment. but that 120W = 1000W hps id think thats bs
 

PICOGRAV

Well-Known Member
pico, regarding the pic: i think you just cant go and do something like 120V * 1,9Amps that would be considered 2xx VA but not W... its not dc but AC...or am i wrong? and that high Amps value could be because of like "initital current" for powering up, like a normal bulp 60W that sucks much more for a Brief Moment. but that 120W = 1000W hps id think thats bs
I know Lax123, you are correct there, W = PF × A × V, with out knowing the Power Factor I can not tell the true wattage of the panel. Switch mode PSUs, don't really have an inrush of current like that, and the only other things in the panel using power, would be the two fans, the two LED drivers and the fan power supply.

Even if you calculate it with a very bad PF of say .90 you would still be using about 205 watts, and I don't think those little fans, and the power supply losses would add up to almost 85 watts?

Does anybody know why we have to calculate for the power factor ? ;)
 

lax123

Well-Known Member
and now i looked at the other pic, the calcs u made for that led comparison. pico you calculated for the worst Cree Scenario while driving them @ max current where they have their lowest Efficiency. recalcing for 350mA...you would Need 95W of cree to compete with 112W of the other cob -> cree more efficient...in that case. hehe so every one can put the Facts as he pleases ;-) on the other Hand you would have to buy 95 crees and Drivers in that Scenario...dont think that would be perfectly cost efficient...
 

multipass

Active Member
Hey, got a couple electrical questions...

If you put a LED PCB (Multi-chip or Star) with LEDs onto a heatsink.. Is the heatsink safe to touch?
Is only the +/- on the PCBs electrically conductive? or is the whole PCB?

I'm thinking about trying a few multichips on heatsinks... or I'm also trying to find some more empty LED Flood Bodies from non-working LEDs(still cant find any).

thx =D
 

lax123

Well-Known Member
thats why i asked the OP if he wants to get them on pcb...because then you dont have to worry about ZAAPPP, only +- is "charged", the Emitters "unconductive" back is attached to the pcb core n stuff to distribute the heat...
 

PICOGRAV

Well-Known Member
and now i looked at the other pic, the calcs u made for that led comparison. pico you calculated for the worst Cree Scenario while driving them @ max current where they have their lowest Efficiency. recalcing for 350mA...you would Need 95W of cree to compete with 112W of the other cob -> cree more efficient...in that case. hehe so every one can put the Facts as he pleases ;-) on the other Hand you would have to buy 95 crees and Drivers in that Scenario...dont think that would be perfectly cost efficient...
It was a 105 watts of Cree if I remember correctly and the comparison is still fair, I took the output of the bridgelux at its peak power and compared it to Cree's output at its peak power, how is that not a good comparison?

Here is a little right up I did on COBs the other day:

"I will explain to you a few of the concepts of operation I have formulated in my own mind. Firstly the article you linked me to is interesting and does raise a few good points, like the 3 watt chips used in the panels where way under efficient compared to 1 watters. There are two sides to this story, the first being, these chips suck but this is only looking at the bloggers facts at face value tho, I have looked myself into a lot of panel designs and one of the common practices manufactures will do, is run high watt chips well under the rated power. Now from their point of view, you could advertise a system with 30 x 3 watt diodes but never actually say its a 90 watt panel. You could then, some what hiding the fact that you are doing this by using a shitty PSU and connecting fans to the circuit bringing your over all power usage up to about the 90 watt mark.

I don't know what panel this guy tested on how but I know there are very good diodes available in a wide range of wattage's from less then one watt up to 10 even 20 watts. Poor design and questionable marking tactics has given a bad name to LED grow panels in general. I can assure you there are 3 watt chips that would be equal or better then to 3 x 1 watters.


The COB design really his a massive step into upping the power and control over LED light. I don't if you used to follow computer chip manufacturing back in the day, early 2000's but this is a lot like what is going on with these new LED chips. As they make diodes bigger and bigger things start to get out of hand with regard to current flow and utilizing different materials, things really would start to get out of control trying to push 10 amps through a single diode. COB are a master peace of power distribution on the micro level, they are using technology similar to computer chip dies, different packaging technics is allowing for higher heat dissipation, better bonding, higher temperatures and ultimatum high efficiency.


In the video you can see every stage of the chips illumination, we start by seeing the parallel diode banks energizing, when one bank is saturated it consumes the voltage drop and you have another bank energize. Power increases and then you final have a uniform voltage across the die, the chip works in the lossy state for quite a large in-cress in voltage, the diodes at this power are only working with them self's, the light as many point sources as there are diodes at this stage. When you get to the upper end of this state, this is where you see the diodes excite the color coating, the light starts to sharpen off almost, things are starting to increase in temperature around the diodes, this is the power level in which the diode start to couple with the plastic (I think it may be some sort of phosphor coating?). At this point the uniform coating and each individual LED is coupled to each other acting as one, your only limits will now be the failure temperature of the diodes or coating. Efficacy wise, think about hundreds of people pulling one load up a hill, as apposed to each person caring a portion of it individually, the individuals are highly effected by small changes in the environment, but using a single load and letting everyone work together gives the load much more momentum.


No one has to think what I say is correct, I could be very well talking out of my ass but always what I try to imagine is how to relate things you can observe into your everyday life concepts. I know in my head I understand this but weather this all can make scene in your head is another thing, I hope its helpful."
 

lax123

Well-Known Member
what about reliability and redundancy, if one "Piece" of the cob Fails is it = boom 100W gone? if you had 95x cree instead, if one Fails = who gives a shit...?on the other hand i think you d have to account for the costs, if 95 cree cost like 500$ while one cob costs 100$...
 

bbspills

Well-Known Member
Well if you had 95 cree's wired in series, then if 1 fails, they all fail.

That was my main issue with my DS100 light, that I had LED's burn out multiple times, which made the light useless.

At least with the COBs (assuming you power them individually), then if one fails you only have to replace that one, while the rest work.
 

lax123

Well-Known Member
Well if you had 95 cree's wired in series, then if 1 fails, they all fail.
??? lol quite the opposite to what i just said ;-) i guess you didnt use constant current Drivers, did you? edit: ah i see DS100...maybe they were in parallel not in series?
 

PICOGRAV

Well-Known Member
what about reliability and redundancy, if one "Piece" of the cob Fails is it = boom 100W gone? if you had 95x cree instead, if one Fails = who gives a shit...?
There is a hundred more parts plus the interconnections to worry about, plus any series string will go out of one LED in that string fails, also good luck changing a single chip like that if you have soldered them in could turn into a bitch, with the Vero chips no solder is needed, just screws and clip in power connector.

Its also easier IMO to manage a single source of heat. Your analogy is like saying, don't buy a Micro processor because if one transistor fails your done, build it with individual transistors... no one has time for that.
 
Top