[Kobey]

It's been a while since I posted but I've been busy going off the deep end.

Last time I was here I was hunting specific lights and not finding what I wanted in a price that I could swallow.


Specifically I wanted a grow light for 12" depth that ideally can be dimmable without costing $$$.

It's been 6-months and a lot of road traveled and I have ended up working with a manufacturer and will be making a sample production run of custom lights.

My final 2 configurations are really just a variation of each other one is cheaper to produce the other might be arguably better but the devil's in the details so I'm posting it here for everyone to critique, improve, argue...have at it before I pull the trigger. :wink2:


The 12-LEDs are 1w (12w total) as follows:
Light #1
Cree XT-E cool white 6700k x6
Exotic 410nm x2
Exotic 475nm x2
Exotic 660nm x2


Light #2
Cree XT-E cool white 6700k x5(*)
Exotic 410nm x2
Exotic 455nm x1
Exotic 475nm x1
Exotic 495nm x1
Exotic 660nm x2
(*) there is also a suggestion of using a XT-E warm 3250k in the middle and 4x cool whites.

 

[rebelbuck1993]

It's been a while since I posted but I've been busy going off the deep end.

Last time I was here I was hunting specific lights and not finding what I wanted in a price that I could swallow.


Specifically I wanted a grow light for 12" depth that ideally can be dimmable without costing $$$.

It's been 6-months and a lot of road traveled and I have ended up working with a manufacturer and will be making a sample production run of custom lights.

My final 2 configurations are really just a variation of each other one is cheaper to produce the other might be arguably better but the devil's in the details so I'm posting it here for everyone to critique, improve, argue...have at it before I pull the trigger. :wink2:


The 12-LEDs are 1w (12w total) as follows:
Light #1
Cree XT-E cool white 6700k x6
Exotic 410nm x2
Exotic 475nm x2
Exotic 660nm x2


Light #2
Cree XT-E cool white 6700k x5(*)
Exotic 410nm x2
Exotic 455nm x1
Exotic 475nm x1
Exotic 495nm x1
Exotic 660nm x2
(*) there is also a suggestion of using a XT-E warm 3250k in the middle and 4x cool whites.

based on the color spectrum in the chart provided light 1 will have better growth over all

 

[Kobey]

Light #1 is the current front runner as it's easier to produce, costs marginally less and also has the, again minimal, benefit of being a slightly "brighter" looking light.

 

[jeffkrol]

Neither though I lean to 2 more than one.
both are too "blue" for me..........

Change the CREE's to Luxeons..but besides that:

All personal opinion/preferences though
3 6500k
4 4000k
1 cyan
2 deep red
2 uv a

6060K
CRI of 95
37.1PPF @ 30cm

Only a score of 56.. but that is a compromise to "look"...

 

[Kobey]

It is a very blue light...was trying to keep the green though orange-ish reds lower.

I have easy access to Cree's though maybe I can source Luxeons not sure (that's a Philips brand correct?)

My other option was Light #1 but swapping in a more cyan 495nm for one of the cool blues.
Exotic 410nm x2
Exotic 475nm x1
Exotic 495nm x1
Exotic 660nm x2

Just quickly slapped this together...tonight when I get back I will try to do the calcs with Luxeons and see if they are an option for what I've got to work with.


edit: I should add too that the fewer number of different LEDs used the better...costs and complexity just get reduced. However manufacturing it becomes moot if the thing sucks so up to me to find the balance as it were.

 

[jeffkrol]

Personally yea that is better.. Might just want to "cluster the "RGB" (cyan= green, violet =blue) since any iteration of RGB is "white". You're limited to 4 like this though (1 deep red, 1 cyan,1 violet, and one of your choice.. normal blue maybe).
all whites in the outer ring of 8.

 

[Kobey]

Took your advice Jeff and swapped the 2 outer deep reds into the middle and whites to the outside.
Going to leave the 410nm where they are.

I'm probably being overly optimistic at this but with the extra red and violet I'm hoping this light might actually work for up to 16" depth for low light low tech and can still be used at a 10" depth.

That gives me a very wide range of tanks that this bulb will (again hopefully) work for and the coverage should be roughly 196 square inches per bulb if all my math averages out right.

People were already shaking their heads at me trying to use 4-different leds so not sure how 5 in a more complex pattern will go over...and my market is rather niche on top of that because really how many people use longs or breeders and honestly even do natural planted in 10g all things considered. Forget the fact I'm trying to compete against give or take $10 fluorescents.


As for the Phillips leds they are not an option at this point as we have Crees here and I get to use those first.

Time to find out

 

[jedimasterben]

I still don't understand using dedicated red LEDs. Warm white LEDs with high CRI give around the same radiometric amount of light in red, yet cover significantly more red spectrum, so you're not just getting 660nm, but also a bit above that and quite a bit below it, whereas a typical Chinese 660nm LED has a very tiny FWHM in regards to spectrum. You also don't get awful color banding from a warm white LED, they blend very well with cool white.

This graph shows the 'typical' spectral quality of the 70CRI versions of the XT-E. Getting an 80 or 90 CRI XT-E will shift the peak slightly and will increase amber, as well.

 

[Kobey]

Thing is I dont really want the oranges and lower reds and at only 12" depth I dont need a pile of CRI. Yes they make a tank look warmer but for a grow light they are not needed and 10k should still be pleanty "white" with the other LEDs bringing out color in the tank.

Or I will be completely wrong and have manufactured the 2nd worst looking algae growing bulb ever. Lol.

 

[jedimasterben]

Wait, so these are straight up grow lights? 660nm and a touch of violet/blue.

 

[Kobey]

Yes thats my goal. A specific grow light for 12" deep planted tanks that doesnt look like total bunk to our eyes.

 

[jedimasterben]

Well, then you're asking for two very different things. If you want a light that grows pretty well and doesn't look like garbage, then use a cool/warm white mix with a couple violet. If you want a light whose primary function is maximum growth, then it won't really look great.

 

[Kobey]

Yes but still given a 12w-led configuration like I have with the parameters I've set myself there is a better middle ground than anything else out there. I mean sure you can go with a Current+Pro and dim it down to 20% but that's an equally impractical solution if you ask me.

Also I'm betting the vast majority of 12" deep tanks do not sit in sit in darkened specifically controlled environments. They are in the middle of living rooms, bedrooms, what have you and typically have plenty of other ambient light around them including direct and indirect sunlight to more than pick up the missing amber / warmer colors of this bulb.

Really I hope it comes out as a dimmable 10k bulb with better reds that's a little on the warmer side compared to your average 10k fluorescent.

Having said that I am here for insight

What would you like to see given 12x1w LEDs that promote better growth at 12"?

 

[Kobey]

So it seems to be just about ready to go!
We are going to do a sample run but based on the suggestions here (thanks all!) and combined with what I proposed and the experience of others here there was a couple changes. Mainly the LEDs will be increased from 1w to 2w and swapping 2 whites for testing.

It's not been fully approved but everyone seems confident we're are going to create 6 bulbs to start with in the following configuration.

12x 2w LEDs:
4x cool white 6700k
2x neutral white 4500k
2x 410nm violet
2x 660nm red
1x 475 blue
1x 495 cyan

The bulb itself is dimmable and the cost between 1w-2w on the sample is somewhat irrelevant so 24w total gives more results. That is also the reason for the spread of different LEDs, so we can cover up various combinations after the fact and test real light levels to see how it all comes together before making a choice on production.

I'm really rather excited I might actually have this thing in my hand by the end of the month!

 

[jedimasterben]

Mainly the LEDs will be increased from 1w to 2w and swapping 2 whites for testing.

Is the heatsink increasing in size for the doubling in wattage? If not, then you'll want to leave it at 12w total.

 

[Kobey]

Is the heatsink increasing in size for the doubling in wattage? If not, then you'll want to leave it at 12w total.

The fixture itself is the heat sink you can see the ribs of it in the other photos better. It's actually overkill for 12w and will do 24w.

The bulb itself is 4-3/4" wide and 5-1/4" tall overall. Think floodlight type footprint.

 

[jedimasterben]

The fixture itself is the heat sink you can see the ribs of it in the other photos better. It's actually overkill for 12w and will do 24w.

The bulb itself is 4-3/4" wide and 5-1/4" tall overall. Think floodlight type footprint.

Yes, I have used and helped develop several of this style before, and I can tell you that the heatsinks used in them vary greatly in mass and range from almost satisfactory to extremely poor, even for 'just' 12w of dissipation.

 

[Kobey]

Don't know the actual makeup of the heat sink but will find out now that you asked...I would be surprised though if it was inadequate in any way and while the 50,000-hour lifespan is a marketing thing, if the lights died after a year they wouldn't be in production at all at this point.

 

[jedimasterben]

Don't know the actual makeup of the heat sink but will find out now that you asked...I would be surprised though if it was inadequate in any way and while the 50,000-hour lifespan is a marketing thing, if the lights died after a year they wouldn't be in production at all at this point.

The 50K hour rating is something that is pulled from the datasheets of Cree, Philips, etc brands when run at test current and below their upper thermal limits. Epistar, SemiLEDs, EPI, etc all seem to list theirs much lower and with significantly lower thermal threshholds, so any Chinese vendor that claims 50K hours, I wouldn't believe them for a second.

 

[Kobey]

True enough but even if the lifespan was only 5-years at 12-hours a day, less than 1/2 the claimed rating it would be satisfactory and competitive with fluorescents cost. Plus if heat was an issue they would not survive anywhere near that.

Guess I will be taking heat readings and leaving one of them on 20-hours-a-day until it dies just to find out.