New Experimental Advanced DIY Designs.

Rth Junction-Ambient = ~28 K/W !!!!

( Those are really good specs !!!! )
Wow!!!

Click to expand...

not bad, my activ Modul has about 10 K/W Junction-Ambient
-------------------------------------------------------------
spectra for the LCW CQ7P.PC have only group 6

Thanx,bro !

- 10 K/ W !!!! ???? That's F@#$ awesome !!!
( You mean with the fan switched "on ",isn't that so ? )

( But 28 K/W passive is really great also !!! )


-So...For the NW LCW CQ7P.PC....
While changes( power/ spectrum ) between groups regarding phosphor emission are subtle .....
Power differences at blue wls ,are not so subtle .....
Hmm...

Ok ..
You got me convinced on that ,brother ...

I'm going for a ebm-papst fan ,for each heatsink/panel ....

I also like the framework of your new panels ...
Really great ....Low-weight ,minimal ,simple and effective ...
(I like also the "merkaba " placing ...
But this time I want 35 pcs of oslons on the same known heatsink ....
So I can't go "merkaba " easy ....
Still ,led placing , it's acceptable enough,the way it is ...-Rectangular lines'n'rows ...

Ok ..
NW LCW CQ7P.PC..

...






So H4 is further closer to blue,than the rest groups of NW ...
(Lowest value at Cx axis ...~ 0.34)
Also is "closer" to violet ....
That means less green ...( Lowest value at Cy axis,also... ~0.335)

I think i like most the H4 group ....
Too much of blue ,but only 5 of these leds out of 35 ....
So, better be rich at BL 440-445 nm ...
And not too much of green ....
(Lowest possible at 5000 °K )

Very close to "daylight "( 0.333 / 0.333 )

Say ....
If we had a blue die (440-445 nm )and a red phosphor (650-660 nm ).....
Then we would have got a led with coordinates (example ) ....
Cy =0.15 & Cx= 0.4 ...


{ Draw a line from point ~ Cy =0.25 ,Cx =0.15 which is the blue die ..
And point ~Cy=0.27 ,Cx=0.73 which is the red phosphor ..
At which point onto the line ,the light color will be it depends on load and particle size of phosphor ....
More phosphor and small particle ,point will be close to reds (high Cx value ).... }

At any point ,at violet light .....
(Where blue and red ends of spectrum ,"curve " and meet at CIE 1931 standard .......)

Now ..We want to make it more rich ....
Add more reds ,orange ,yellow and a bit of green .....
So we add a green phosphor ...
(Moving towards green side ..Cy values increase ....I.e . 0.15=>0.35..Light color changes, from violet to white ....
Depending on peak value of green phosphor ,previous violet "line " becomes a "triangle shape" now ....
A third point of a green / yellow phosphor wl is added ...Into the triangle ,is all the possible light color shades ,
that can be theoretically ,reproduced by this particular die & phosphors combo ...)


And thus ..we will move upwards ,towards green,yellow and orange ....A bit to the right also ..(at the diagram ,for our example .....)
We end up somewhere at Cy= ~ 0.35 and Cx= ~ 0.45
(generally at that " area ",maybe a bit to the red or yellow or orange ...
Depending on individual needs and plants cultivated ....)

So ...We end up at "Warm White leds" light "area " ....

Who would have thought .....?
That white light leds (blue die + phosphor ) have so many-many-many variations ( " colors " ) ...

Warm white at 2700 K has 72 ( ! ) different shades .....( Oslons SSL )


Trimming a led spectrum, based on that "all-white " school of led growing ,will be tricky ,enough ...
Way more tricky than blending monochromatic leds together ....

If only , there was a spectrum mixing tool for SSLs ..
How nice that would have been ....

- 10 K/ W !!!! ???? That's F@#$ awesome !!!
( You mean with the fan switched "on ",isn't that so ? )

( But 28 K/W passive is really great also !!! )

Click to expand...

fan has to work for this number, more than 20K/W without fan

merkaba is not necessary with all led are white, layout is good.

a real problem will be the driver.. 35 x 3.2V = 112V, there are not much on the market.

Yeah I know about the driver ....
Tough call that one ....
Although I leave it for later ,the whole driver thing ,
I got some pretty nice suggestions,from a friend ...

here are also some...

http://trpssl.com/driver-search-by-current.html

Those I'll have in mind .
Noted here .
At the cyber-notebook .

Thanx again .

Guod ..
If you had to choose 35 pcs of NW & WW oslons ssl (only )...
Which groups would you choose ?
How you would config the whole thing ?

PetFlora said:

I would agree, IF, you are using only one WHITE, but if using a mix of NW + WW then Merkaba is still a good idea

Click to expand...

Well ,it has kinda a "merkaba " arrangement there ....

Layout and Groups can not be done better for WW & NW LEDs.

Your "approval" on that design ,means a lot .

Ok...
I guess the "spectral configuration" is more than acceptable for starters ....

The IMS construction ,probably,is going to be messy,tricky and time-consuming ....
(But well worth it if it succeeds.....Say 50-75 um thickness of dielectric with λ>1.5 W/m.K....)

...Frame ,probably something like Guod's ...

...Fans 220 V Ac ...(with grills ...)
....
Drivers ,again Guod ,saved the day ....

....
Ehmmm....

Does it worth "dimming-able" the panels ...?

From 350 mA up to 650 mA ?

Oh,it needs a separate 0-10 v power supply ...

Niehhhhh.....

Couldn't a potentiometer had been enough ?

Say 50-75 um thickness of dielectric with λ>1.5 W/m.K....

CooLam® LX
Dielectric Layer Thickness, 17 -22 μm
http://www2.dupont.com/LED_Lighting/en_US/products/coolam/coolam.html

...Frame ,probably something like Guod's ...
ready & testedsame size as your heatsink
http://s1073.beta.photobucket.com/user/guvoo/library/LED-Modul
35 Oslons at 500mA > 56W - no problem for this Heatsink/Fan Combo
at 650mA > 72W - maybe

So about CooLam ,one has to order it in Dupont ? (the whole circuit I mean ...)
Quite informative pdfs to the side ,there .....


As for the http://s1073.beta.photobucket.com/user/guvoo/library/LED-Modul
....
What to say ?
I'm speechless .....
That's what I call tough and solid construction .....
Commercial led grow panel boxes look like ordinary tin buckets,compared to your panel ,Guod .......
(I love the detail of 'golden' (bronze ? )screws at fan grills ...Elegant & classy ... ....)
Anyway ...


I do not think that I'll operate the Oslons more than 500mA ....
So ...I guess ,they'll be ok with a fan ......
Thanx ,again.....


BTW ..
What's your opinion about this " CCC" thing ?
My guess is that ,it will assist (can't say at what grade ...) ,the cooling of leds ...

At least Osram ,says so ....

Regarding the requirements for good
thermal management of the OSLON LEDs,
the copper surface around the thermal pad
should be kept as large as possible. This
serves for distribution and spreading of the
heat and is typically covered with solder
resist.


Edit:
Guod...
Is it any possible way ,utilising a dimmable driver ,
to have a " multipositional " rotary switch ,so to select
Pre-adjusted Current ?
I.e .

0= off
1=350mA
2=400mA
3=450mA
4=500mA

???

I'm sure there has to be a way ,there ....


Also a Fan -thermostat ,"closed-feedback" system ,would be nice ....
So ,heatsink will maintain a stable temp(pre-adjusted or adjustable ) and fan should work ,only when is needed ....
And not the whole time ...(noise reduction /prolonged fan life ? )
I dunno ....Maybe is " too much "...
Or a really " Kick-@$$ " design ....

(Based on Guod's design ,there's plenty of room "under the hood",to accomodate some extra circuits there ...)

But then again ....
Better keep them as cool as possible ....
So fan has to stay on ,all the time ,I guess....

Yeap...
But still I need the leds to be 35 pcs .....
No matter what ...

Also a Fan -thermostat ,"closed-feedback" system ,would be nice ....
So ,heatsink will maintain a stable temp(pre-adjusted or adjustable ) and fan should work ,only when is needed ....

Click to expand...

this was also my first idea, but we deal here with a constant power.
so I did a test over the forward Voltage of the Leds. (Vf falls if temp. goes up)
v

conclusion of this was any speedcontrol of the fan is useless.