making a new pannel

Hello everyone,

I have red quite a lot on led and I will start to build one in few days.
I was planning to go with oslon SSL80 for hyper red and cree XTE for white and XPE for red, with meanwell driver (700mA) and passive dissipation. the cost was about 400dollars for 130W.
By looking at some grow, I saw the Apache tech, I look a bit on a well known site: Satisled and I found I could make a similar panel of 230W for less than 350dollars including led of 1W with lenses of 15°, driver (350mA) etc...

regarding the efficiency, 1W led are about as efficient as the osram and cree. with the lens I will loses 7% of the photons emitted but I will have 180W of power instead of 130W and for less money. It seems better to me.

Now, I still wonder: How far should I put my panel from the plants with such narrow angles ?
Does the penetration will match the 3W or better ?
I am only going to use white and red having a ratio between the Apache tech 3W:2R and the area 51 4W:2R; but my red will include 630 and 660 with a 1:1 ratio, for flowering it seems to me 660 is necessary. The "green" from the white should give more penetration apparently.

I am not looking to big bushes but to SCROG a bit, do One BUD of indica and indica/sativa and also grow some supercropped sativas.
I saw here some guys growing big plants with apache tech AT120 I guess I could do the same with a DIY one.


By the way my grow room is very small : 3x1,7 feet (90x50cm), and I calculated I needed maximum 150W from the 30W per square foot of Knna and 17W/sqf from the thread of SUPRA using high efficiency cree.
tell me guys if I am on the good path please, because I can't choose between this two configurations.

Sounds ^^^ good to me

....Knna's calculations are dated now , so Supra's stuff seems more appropriate. 17 w/sqft is low, only acceptable with the best (top bin) leds out there. Your efficiency # are incorrect compared to apache, we don't publicly know what leds are inside and they use ACTIVE cooling which will have better lm/per w than your PASSIVE design.

1-1-1 ww/630/600nm is heavy red leaning, but "should" work. People successfully veg with HPS so I don't see an issue. 50 degree lenses will drop your coverage area to the panels dimensions basically.

Height above plants is not a standard issue, read the plant's signs!!! and adjust accordingly

When u r considering cheapest satisleds , why bother using 1W leds? u get there like 100 x 3W leds for about 140$, u would need about 5 of their 650mA Driver for 18pcs - in reality it works easily for 25 x 3W leds thats like 70$, so thats like 210$ for about 200 true W. I built a super cheap satisled panel myself. Looking at their datasheets, dont be fooled they r full of mistakes, so i would trust them as far as i can throw them regarding the lumen outputs ratings and other stuff they state. But i knew that before i decided to do that (if i remember right i calculated from their 660nm datasheet lumen into mW/W Radiation: -> i figured that they r more efficient then Osram oslon 660, ROFLMAO)

I would not bother with anything below 3W. Way too much work with 1W LEDs.

I would also add a fan. It should give you a little bit better efficiency of the LEDs.

1-1 w-R seems ok to me but then keep two W with every 1x630 and 1x660 (guess its what you meant in the first place)

I have two panels, one with secondary lenses/optic on em and one with out, can easily grow with out, just do the scrog you talked about or small auto`s etc. and keep em at about 10" instead of 15-18" as you would with Optic on em (Im talking 3W)

thanks everyone for answering to me

well I think you misundersttod my spectrum, I may not be clear (english is not my mother toungue).
I will go with white and red that's for sure, but my ratio will not be 1/1. I will have a 1:1 ratio for my red 660:630

On my first intention I was going to use 3XTE WW in R3 bin; 2 XTE in 5000K ; 1 XPE-2 in 630nm and 2 oslon SSL 80 660nm.
this was my flowering spectrum. it makes 8leds per modules. and I wanted to use 8 modules for my growroom.
They are 15W each aproximately hence it would make 120W of high efficiency led for 0,45m2 which make 24W/sq foot. not bad?! it's more than SUPRA with the same led.

then I saw apache tech and obviously it works very well. I look at satisled to build the same. (time to do it isn't a problem, I can solder 200 led if It's needed...).
I discover there 1W are not far in efficiency from the Oslon and Cree. And they are much better than the 3W in satisled.
exemple: 50 White 1W for 15$ I have 120lum/W and I will have 50 white 3W for 35$ but they have less effiency (if I remember well only 80lum/watt). In terms of prices the 1W are cheaper and are more efficient.

what the point of using 3W then? PENETRATION? So I thought of using narrow angles like apache tech.

Cool thread, lets see if I can help...

Apache use the best chips available and they are not cree's. They are no longer using 1w chips anymore...now they are using 5w(capable) that are kinda like a better cree xte(but not and I guarantee better). It's a secret cause they are that good...not cause he is trying to pull a fast one. Just trust that apache uses the best chips and one day when he realizes no other companies will want to pay that much for a chip and he will be more open about it. Others will use the cree for cheaper price and still great performance...just not the best. But as soon as someone buys one themselves they can research and blurt it out...I just told him I wouldn't. I like my relationship with them so I keep my word.

Every chip should be looked at as a variable chip. All of which have a maximum efficiency usually way lower than max capacity. And the chips should be ran and thought of as what they are being driven at(the maximum efficiency or close to). Which is why apache's 1w held it's own still to this day with many current 3w/5w chips...they were/are as efficient and power as many the 3w/5w that are only driven at 1w-2w. But now chips have stepped it up and the 3w/5w's efficiency level are better at high drives now so we can use and capitalize on them.

You have the right idea for your DIY panel...copy the best and what do, but improve on things like coverage by using 60* lens, and maybe 660nm like you are doing. The xt-e is a great chip and will make a great grow panel, it's what I would use.


As for hanging height...it actually is set by what will give the canopy the targeted intensity...for the at120(and a51 sgs) 18" gives the canopy ~700-800umol which is what they aim for, and also the right coverage. So assuming your panel will be close in performance with your model(at120/a51sgs) I would start around there if you don't have a par meter.

tags420 said:

Cool thread, lets see if I can help...

Apache use the best chips available and they are not cree's. They are no longer using 1w chips anymore...now they are using 5w(capable) that are kinda like a better cree xte(but not and I guarantee better). It's a secret cause they are that good...not cause he is trying to pull a fast one. Just trust that apache uses the best chips and one day when he realizes no other companies will want to pay that much for a chip and he will be more open about it. Others will use the cree for cheaper price and still great performance...just not the best. But as soon as someone buys one themselves they can research and blurt it out...I just told him I wouldn't. I like my relationship with them so I keep my word.

Every chip should be looked at as a variable chip. All of which have a maximum efficiency usually way lower than max capacity. And the chips should be ran and thought of as what they are being driven at(the maximum efficiency or close to). Which is why apache's 1w held it's own still to this day with many current 3w/5w chips...they were/are as efficient and power as many the 3w/5w that are only driven at 1w-2w. But now chips have stepped it up and the 3w/5w's efficiency level are better at high drives now so we can use and capitalize on them.

You have the right idea for your DIY panel...copy the best and what do, but improve on things like coverage by using 60* lens, and maybe 660nm like you are doing. The xt-e is a great chip and will make a great grow panel, it's what I would use.


As for hanging height...it actually is set by what will give the canopy the targeted intensity...for the at120(and a51 sgs) 18" gives the canopy ~700-800umol which is what they aim for, and also the right coverage. So assuming your panel will be close in performance with your model(at120/a51sgs) I would start around there if you don't have a par meter.

Click to expand...

If you want the best, then you have to pay for it - and usually a lot more than it's worth. But it's a choice you have to make. Personally I would not pay that much extra, for so little more.

thanks you tags420 for information.
I saw your journal, I loved the videos you did

Then I can understand why AT has such efficiency.

in my case my ratio white red is 5:2

Just want to make sure you understand me well here :
I am hesitating between using chepo led from satisled or using cree which are a bit more efficient but much more expensive.
I don't understand why should I use cree if for 400$ I can only have 120W while I can do a 250W pannel for less than 350$ with satisled.
the cheap led will be less efficient but their will be twice more led, so I should get much more lumens...

In case I am "copying an apache tech" with 1W from satisled, I calculate the efficiency in terms of lumen/W it's about 105lum/W
instead of their 120 lum/W
I have 240W of power instead of 120W from apache tech so I will get 25 000 lumen and they will only get 14 400 lumen.
also my panel will be 40x50 so my light will be more spread than their one.
Is that correct? if yes I will win by far?!

in case I am using cree and osram. I will have a panel with a better efficiency and longevity but more expensive and I don't think it will put as much light out than the 240W of cheap led... it would need 200lumen/W which is impossible actually even for cree.
If I am using high efficiency led I will never put lenses on them... loosing 10% of photons will be a waste because you pay a lot for it.

@slipon: I am actually growing Jack the cleaner2 by TGA and jesus OG; and I hate indoor autoflower, I think they are to small and the quality is not very good. but in outdoor damn, they are the way to go!
the problem with naked led (with no lens) is that you need to grow SCROG or make small indica one bud.
I had a 250W HPS and I am trying to find a decent penetration like this with led. 30 to 40cm is it possible to achieve with 3W naked?


@lax123:
then I am sure you ask yourself why bother with high efficiency led from cree when for 1$ a watt satisled do the job.
Are you satisfied with it? lenses or no lens? What is the penetration they can achieve?

@PSUAGRO: 50° will be to narrow then... thank that's great info. but when looking at the apache tech, it's true that the center recieve much more light but by puting the panel far from the plants you get a bit more coverage and it's 14° lenses...

@SMD:right I will add to 12V fan, It should do the job. but I already have a huge and massive heatsink: 40x50cm with 32 "ailettes" of 4cm high. if I remember well it can dissipate something like 230W with the 100cm2/W of KNNA.
but any way, it's a good advice, I will use fans to supplement it. Thank you
I hope you understand what I mean. and thanks everyone to contribute, it is very helpful.

Hi speedy,
I approve the choice to add a fan, Cheap leds produce too much heat
Remember that lumens are not necessarily a good way to compare
Let me ask you a question: In case you would use cree leds, why did you choose the XT-E ?

lax123 said:

I built a super cheap satisled panel myself. Looking at their datasheets, dont be fooled they r full of mistakes, so i would trust them as far as i can throw them regarding the lumen outputs ratings and other stuff they state. But i knew that before i decided to do that (if i remember right i calculated from their 660nm datasheet lumen into mW/W Radiation: -> i figured that they r more efficient then Osram oslon 660, ROFLMAO)

Click to expand...

How do you calculate the mW/W ? Have you got the spectrum of the 660nm cheap led?

How do you calculate the mW/W ? Have you got the spectrum of the 660nm cheap led

Click to expand...

Im im not mistaken its something like this:


LED: Vf 2,4V @ 750mA @ about 72,5 lm -> 1,8W for 72,5lm thats 40,27lm/W

for monochromatic 660nm Vlambda curve is about 0,061 and multiplicator 683lm/W

(40,27lm/W)/(0,061*683lm/W)= 966 mW

I think oslons have something like 48% efficiency for 1800mW that would be like 864 mW

@lax123:
then I am sure you ask yourself why bother with high efficiency led from cree when for 1$ a watt satisled do the job.
Are you satisfied with it? lenses or no lens? What is the penetration they can achieve?

Click to expand...

Yes thats what i asked myself. I have lenses coming 60°. Cant say more than that i have nothing to compare to and no measurement stuff. I dont know if they will do the job. Plants r growing, cant say how good or bad, im a bloddy first timer, so if my results suck chances r its not because of the light

Skyled:
I didn't say that but I can understand what he means I think:
I thought when you compare two led of approximately the same spectrum, lumens can be taken into account. I am comparing a oslon which has a wide spectrum tending towards 645nm so it will get more lumens than a simple led having a spectrum from 650 to 670. Here I guess the advantage is more for the cheap led.
Anyway I looked at the red 660nm from satisled and they are not as efficient as oslon... far from that.

I choose to go for XTE because their price enable me to have more leds than if I was taking XML2. I can have 2x more lumens with them and for the same price... It just required more work. This was also the SUPRA's logic
But thinking like this you have better going with cheap bridgelux from epistar outputing 220lumens at 700mA against XTE at 140lumen at 700mA but at twice the price.
That's where I am stocked...
yes I know lumens means nothing when the spectrum is not the same. but KNNA said photons is more important than spectrum because of quanta theory. an HPS has a bad spectrum but give so much photons that plants love it.

how did u calculate that?

The same as you, what is strange is that I did not find the same thing... I had 782mW. So I must have done a mistake. Sorry
But oslon SSL are not 864mW but something like 950mW and they also can be driven at 1000mA instead of 700mA which is quite interesting if you go with XTE.

edit: wait a minute, your right, 960mW for oslon is at 1000mA. the cheap led from satsiled are better then...

But then explain to me why you say that these leds make a lot of heat... if they are as efficient as osram and cree?

Of what I saw, some people wants to prove that leds are better than HPS. I can understand that and I think some have to show it. But on the next hand when you don't want to prove you can do more than 1gr/W you don't need to take this led.
because 240W of cheap led will do better than 100W of very high efficient led... You will just pay more electricity but invest a lot.
Am I wrong?

I dont really understand what u r calculating.

idk but oslon datasheet tells me @ 350mA about 48% Efficiency. So lets say even if its worse driven higher it has 48%. I calced for satis running @ 1,8W, so with 48% @ 1,8W that would be 0,864mW. Thats less what i calced for satisled, and their datasheet is 4 years old....

But then explain to me why you say that these leds make a lot of heat... if they are as efficient as osram and cree?

Click to expand...

Vs
lax123:

Looking at their datasheets, dont be fooled they r full of mistakes, so i would trust them as far as i can throw them regarding the lumen outputs ratings and other stuff they state.

Click to expand...

All i want are some cheap smokes
And as im a first timer: i read somewhere newbiemistake #1: spending too much money without having aquired the proper skills. Looking at my poor plants i know that true

You will just pay more electricity but invest a lot.
Am I wrong?

Click to expand...

At some point, maybe a year or longer the "expensive" led will get cheaper than the cheap led because it consumes less power for the same output. Strongly depends on what u pay for electricity and how cheap u can get the expensive leds.


I think what we would need for satisled is CaliJoe working his magic with his fancy equipment

Thank you very much lax123. you helped me a lot.

I am not a newbie grower in fact, I am just new on this forum, but I have a bit of skills under 250W HPS and outdoor growing.

In my area the summer is hot sometimes more than 40°C so inside is aboute 30°C, I am growing in a small cab and I just want to harvest 100gr per session. I was able to do it with almost no effort with the HPS, but the summer, temp gets too hot, and I don't want to bother anymore with huge fans etc...

That's why I am trying to go with led. my total budget is about 400$ with this I calculate I can have a 120W high efficiency led panel made of cree and oslon
or I can have a 300W panel made of simple cheap led from satisled.
I guess I can do better than 0,7gr/W with satisled panel so a 200W would be sufficient and I will save lot of money. SUPRA was able to grow good plant with old osram golden dragon, and now cheap led are as efficient as old cree and osram. So I can imagine it should work.

I know about the satisled datasheets, in the same time, they say on their site: the datasheets are old and no longer true because the brightness of our led is evolving with the rest of the led market. that means they should actually be better than the datasheets.
I don't thinks satisled leds are necessary bad, they just make them in huge quantity, and anyway XML or XPE2 are not expensive to make. we pay the engineering studies behind it. satisled and other chinese manufactures are copying hence we don't pay any studies.

I will look at this famous CalieJoe then

Do you mind posting some of your plants and a photo of your pannel? are u using only red/white or also some blue?

Check these babies ....



S.L.H. fems into 10 lt pots ( peat-perlite ) and under ~260 W of pure Oslon SSL 80 light power ( Hyper Red & 3 different whites -no blues )....
Entering the 4th week into flowering ( 12 / 12 ) ...
Osram Rules !

SDS I thought will not come... I asked you some questions on your WW 4 ever and I didn't had answer.
Any way I am glad You came, because I do love your interventions.

What is the actual ratio you use? your plants are beautiful indeed. Also I notice the penetration with 80° angle is very good, how tall are the plants?

I am interested in osram oslon but their price is 4$ and I can't offer more than 100W panel if I go with only osram.
I red your opinion on WW oslon SSL80 and you convince me, but prices/quality/efficacity is right now going to the cree XTE according to me.

I thought the best led in 660 where the hyper red oslon SSL. but Lax123 just kill my statement in a few calculations. knowing that this led are only 1,2$ I would more easily go with it. what do you think?

I am higly thinking of going with 3 XTE WW 3000K, 2 XTE 6500 or 5000K I don't know yet; 2 XPE-2 red 630nm ; and 1 660nm from satisled
This spectrum should be good for flowering.
in veg I will have 2 XTE 6500 or 5000K ; 1 XPE 2 red 630.
and I will reach the flowering spectrum by adding progressively (dimmer) the WW and the red 630; so that I can govern the stretch with the blue red ratio...
When the stretch is finish I can put the 660 to make the bud more mature and with a better density.

I using a transfo, and a regulator as the alimentation. It enables me to drive each color separately. hence I can dim the WW, the 630nm and the 660 in fonction of the plants. my spectrum will hence be adapt to all kind of plants. Cool?! maybe just dreaming ^^

From my experience do it yourself projects usually end up being more expensive than what was bargained for. Usually there is something you didn't know you needed a tool etc. I hope things work out for you but I also wouldn't be surprised if what you created ended up burning up and you are out all that money. I hope that is not the case.

Thanks for encouragements !!
well I beg everything will be fine...