Building New Panels - Sharing the Process

[rrog]

Thanks for that drawing- makes sense. Not sure what I'll do- still looking.

I think I'll simplify and have the fans on a separate power chord. Keep it a bit simpler.

LED Group buy just started shipping today. Cool. These come tinned, but all the vids I see show guys tinning anyway. Should I pre-tin these?

Regarding the wire, just stranded #22 or #20? No special coating / insulation?

 

[tenthirty]

Stranded 20ga would be a better choice.

 

[rrog]

OK stranded 20. I guess there's nothing special about the insulating material. Thanks-

EDIT:

I have to order the drivers, now that LED Power Buy has LEDs in stock. I'd like to just double check the LED specs before I pull the trigger:

10 Cree XM-L per string.
Running at 2A, yielding 6.5W output from each and 3.2 forward Volts each
For 10 LEDs, that's 32V

Per Rives great help, I was looking at the Mean Well LPF-90D driver http://www.meanwell.com/search/LPF-90D/LPF-90D-spec.pdf

25.2-42 volts, 2.15a, 90.3 watts

Does this look OK? I'd be running at 2.15A, not 2A, but I have the overhead as this is rated for a max of 42 V.

 

[xmobotx]

These come tinned, but all the vids I see show guys tinning anyway. Should I pre-tin these?

no; if tinned they are ready

my lumileds from future lighting were nicely tinned; the colored diodes i just got from rapidLED are not tinned ~its got me a little disconcerted

 

[habeeb]

OK stranded 20. I guess there's nothing special about the insulating material. Thanks-

EDIT:

I have to order the drivers, now that LED Power Buy has LEDs in stock. I'd like to just double check the LED specs before I pull the trigger:

10 Cree XM-L per string.
Running at 2A, yielding 6.5W output from each and 3.2 forward Volts each
For 10 LEDs, that's 32V

Per Rives great help, I was looking at the Mean Well LPF-90D driver http://www.meanwell.com/search/LPF-90D/LPF-90D-spec.pdf

25.2-42 volts, 2.15a, 90.3 watts

Does this look OK? I'd be running at 2.15A, not 2A, but I have the overhead as this is rated for a max of 42 V.

looks fine to me,

only thing is mean well to dim has to have a power source to dim,

 

[rrog]

I won't be dimming. I'll just turn on and off strings. Thanks for the check on the driver and Xmobo- thanks for the tinning advice.

 

[luvdemshrooms]

Thanks for that drawing- makes sense. Not sure what I'll do- still looking.

I think I'll simplify and have the fans on a separate power chord. Keep it a bit simpler.

The fan I linked to came with a bracket, so all I had to do was flip the fan so the airflow was correct, drill and tap two 4-40 holes into the bracket (though a locknut would have sufficed), drill two 4-40 clearance holes through the heat sink and attach. Worked great. The fan is a 5 volt fan that came with a 110v adapter. It plugs in with a USB end and can be daisy chained as the fan supplier includes a usb out on the fan as well.

Sorry about the crappy pics. Still can't find my camera.

 

[rrog]

Looks like a great build, to me! Thanks so much for sharing ideas.

 

[cimenk_hel]

Hi guys, and thanks for all the info.

It seems there is some kind of concensus that a 1:1 ratio of warm white to neutral white should cut it for flower? If there is any wild disagreement, please tell us. This setup would keep things nice and simple, especially since if using all XM-L's, both types can be hooked up to same drivers etc.
I have in my notes from this thread and elsewhere that a 2:1 ratio of neutral white to cool white (xm-l again) could be appropriate for vegging. Anyone have a drastically different idea? Like ditching cool whites for veg also (in favor of what?).

And a specific question about neutral whites: I see in one source that was linked to earlier in this thread that these neutral whites are labeled either 4000K or 5000K, would it be a better idea to use the 5000K (it seems to me), and why?

 

[xmobotx]

here's a WW XM-L seems like they can be kinda hard to find on PCB {rapidLED has the XP-G}

as for neutral cree; the ones i got are actually a notch towards cool but; in general you want a color temp anywhere from 5000 to 5500 as they say this is closest to natural sunlight {hence 'daylight' bulbs} mine are 5300 to 5700 {cree presents color temp as a range} rapidLED's neutral is 3700 to 5000

so; in general; i would advise 2700 or 3000 for 'warm' and around 5000 for 'neutral'

hempfield has a plant really paying a lot of attention to a CW diode in his garden giving rise to thinking that it has some strong points for intensity ~i would speculate that it should not be ignored for veg

1:1 n to c would make a good veg light and 1:1 n to w would make a good flower light

would it work? ~hell yes

would it be optimum? well; that remains to be determined

and one could say the same in either direction {2:1 favoring n or c/w in either case}

 

[rrog]

My Crees arrived today. Drivers next week. I went with 1:1 WW and NW. I'll turn off 1 string of WW for veg, increasing the blue ratio.

 

[cimenk_hel]

And also my previous question about Kelvin rationg also applies to the cool whites... there seems to be 6000K and 7000K ones. Anybody got an idea which one has the better spectrum?

 

[2cent13adz]

Hello rrog! I don't get it. In the first post you say this:


If I'm going to do this, I prefer to overkill wherever possible. I'm not looking at cost.


This sounds to me like a guy who is aiming to build a race-car, rather than go downtown to pick up a pick-up.
Not standard factory-made, something better. Using the best parts only. If so, I get it this far.

Then you tell us it will only be driven at 2/3 speed. And you don't even want a throttle? This I don't get. At all.

If you, for whatever reason, really need to see how the led's perform at 2/3, it's still possible to wire things for speed.
Set things at 3A, add a 100k pot ($5) and you're done. No 10VDC needed.
At first startup, set it in the middle, use a killawatt and turn it up to the desired 6,5w/led.
Then sit back and enjoy the bright light.

Now, sitting there, probably celebrating the new lamp with a decent spliff, how long does it take before you start wondering?
How bright are these led's if fed the full 3A? At this point, you can
a) use your best stoner-smile and go turn the knob, or
b) curse yourself for using racecarparts to build a pick-up, and newer know...

Whatever floats your boat etc. etc., and good luck with the build!

 

[xmobotx]

well 2cent13adz; i have been of the same mind as you on this but now that i have blown 3 diodes on a string running @ precisely the max rating; it seems prudent to consider the manufacturer's tendency to 'under-drive' the diodes

i like your plan too but; in many instances that last little bit of light comes at the expense of disproportionate heat

 

[rives]

What you are talking about is running the chips at their absolute maximum rated current, which is never a good idea with electronics. The cooling gets too dicey and the margin for error is non-existent. I like leaning on things a bit - the diodes in my hybrid fixture are typically run at 400ma, I'm running them at 550ma, but the maximum is 1 amp. What they don't tell you in the specs is that to run them at this level, you will need a fluid-based heat extraction system so that you don't zorch them immediately. A 50% increase in current would only net back Rrog a 30% increase in light - far better to add a couple of LEDs to the mix and get the extra light that way.

Oh, and I've built some pretty radical hot rod engines, but never a full-on race engine. Most of them were good for about a third of the mileage that a mildly modified engine would give you.

 

[xmobotx]

rives the way my reds blew out in that circuit makes me wonder

it was the last 3 {of 10} in the series {negative end} w/ the 1st of the 3 being the blue rated @ 3v .7A among reds rated @ 2.4v .7A {driver is 24v .7A}

seeing it blow i got the impression that the blue stored up some energy then blew and it was the release that blew the 2 reds

can LEDs store energy like a capacitor for instance?

 

[2cent13adz]

@ xmobotx
rrog said "overkill wherever possible", you say "play it safe". That's not the same, is it?

@ rives
No, I'm talking about having the option to go full throttle. Naturally one will need a good cooling system. Didn't rrog order a serious heatsink? And "need a fluid-based heat extraction system"? Please, tell me you're joking.

As for the last few %, my guess is most people see no much difference when turning the knob at the end, and will set for ~90% anyway.

I'd love to talk hotrods and stuff, later? Did any of you have a sock in one of the carb's? (hm, better add one of these... )

 

[xmobotx]

my rationale @ this point {which is new} is that it's better to have a working light than 1 capable of 'full throttle' which does not work

as you said; being able to dial it back would enable one to determine just how hard to drive them

my 4 lumiled rebel ww fixture is plugging away fine @ max rating

FTM; i will defo defer to rives on these matters as i am a hobbyist and though i do happen to understand electrician thingys to a degree; i am not educated in it

 

[rives]

As I see it, the problem with having the potential to go to those current levels is the likelihood of cranking it up, seeing no apparent difference, and walking away and leaving it for 20/4, 18/6, or whatever schedule for weeks on end. When I set up the adjustable drivers on my lights, I use a trim pot to feed the main adjustment pot so that I can keep things at a livable level when it appears to be wide open. The problem is there no good feedback mechanism - it's not like you have to take your foot out of it because there is a corner coming up. You sure as hell aren't going to get expedient feedback from a plant! And aside from the cost of the chips, having the damn fixture die in the middle of a run is pretty unattractive unless you happen to have a spare sitting there. Even if you have another lighting system available, it's going to stress the plants to change the light source over.

And no, I wasn't kidding about the heat extraction system. The only setup that I've seen with GD+'s running at a full amp had them mounted on tubing with a closed loop liquid cooling system. Even if you have a good cooling system, you are still reliant on the ability of the chip to transfer the heat from the internal junction to the sink. As the temperature goes up, you get diminishing returns from the increased current vs the increase in light.

Xm, no, I've never heard of an LED "building a charge" and then releasing it, and can't really dream up a scenario where it would.

 

[2cent13adz]

My starting point was rrog's bold statement "overkill wherever possible", and I liked it. Still curious about his thoughts, as in my first post.

@ xmobotx & rives, I hear you. But you gotta admit, all you're saying is either "play it safe" or "fear of failure". That's ok of course, but it ain't much overkill, is it?

rives, I'm not sure I understand what feedback you're talking about. Can you explain? (no native english) When it comes to the heatsink, luvdemshrooms posted some pic's a few posts back. With this type of heatsink it's no problem to run xml's at 3A. Would you believe me if I told you I've seen it? And if you can't leave the lamp on 24/7 without worrying, you need to order a proper heatsink! (If this remark calls for another smiley to keep things cool, let me know. I fucking hate to use them)