How efficient are blue used in COB ?

Hi everyone,

I need your help on this

I see COB more and more efficients, taking into account the losses from white conversion, how effect are the blue under phosphore ?
80% ?

I wonder what are the most affective monos 450nm blue led, and how they compare to the one used in COB.
It seems everyone agree saying RB is more efficient than white, but Oslons SSL RB is no more than 70% @350mA efficiency isn't it?

Tell me your though on this please

Bump !!

I found Supra saying in another thread
"Steves once offerred the Luxeon M E5G bin, which was the the most efficient LED I have ever seen for sale (63% at 700mA)."

So wait, how can we achieve 65% efficiency on WW COB while the best blue we can buy are not even 70% efficient ?
It seems cree got above the 300lm/W recently with white, why aren't this blue available ?

try running those monos with the same amperage as a tiny cob led and you may understand why

Under the LES the " excitation " LED chips are connected in certain ways ....

Usually a number of diodes wired in series make up a "group ",
while a number of these groups ,are wired in parallel ..

Than means than when you drive a COB at 700 mA -for example - every each one of its diode chips is driven
4 ( or more times ) less ...So while those chips may have about 40-41 % rad. efficiency @ 700mA ,
then when they are driven at 175 mA (or even less ) ,
their rad efficiency may rise up to 80 % (maybe even more at some cases ) ...

But since COB are phosphor conversion white light LED arrays ,there are already plenty of different ways that ,
a part of the initial excitation energy is finally turned into heat ...
(absorption by silicone ,wiring ,chips and phosphor particles ,stokes losses,re excitation losses ,etc )
Usually an amount of 15-20 % in total ...
So,more or less , a COB has (or can have ) a rad efficiency of ~60 or even 65% @ 700mA
( 80 % - 15% = 65 % )

Cheers.

An example ...
Vero 29 3000K 80Ra ....
At 2100 mA ,it has an Vf of 38 V ( typ ) ....

http://www.bridgelux.com/sites/default/files/resource_media/DS33 Vero 29 Array Data Sheet Rev J 20150908.pdf


Let us see inside the LES ...
156 chips ...

Maybe these ones ?

http://www.bridgelux.com/sites/default/files/resource_media/BXFD4545-45mil-x-45mil-or-1.14mm-x-1.14mm-Datasheet-DS-C21.pdf

Say that the 156 diodes are wired like this :
12 x groups (parallel connected ) of 13 chips in series ...


Say that we drive the |Vero 29 at 2100 mA ...
Each group of 13 ,will have an If of 2100 / 12 = 175 mA

At the data sheet we can see at the " Vf vs If " diagram ,that at 175 mA ,each diode has a Vf of 2,92 V (...or so ..) ..

So each group of 13 ( thus the COB itself ) will have a VF of :
13 x 2,92 = ~37 ,96 ..About 38 V ,typical for the Vero 29 At 2100 ma ...

Thus ,probably and quite possibly ,each diode of the Vero 29 is driven at 175 mA ,
when the array is driven at 2100 mA ...Thus the high( final ) efficiency ..


COBs are like fractals ..What the grower does with LEDS / COBs ( driven softer & add more units )in order to increase
the efficiency of the system ,is actually happening inside the COB itself with it's chips ...

@stardustsailor
COBS are like fractals. I like that. Just had a almost Mandelbrot like moment thinking about white LED spectrum lately so COBs are like fractals is good.

Isn't the entire point of COBs is taking a big chunk of inefficient Chinese blue emitting semi-conductor, slice it wafer thin then wire them inhumanly close so they act like 1 big thin semiconductor. Efficiency comes from their design, semi-conductors used, the machine making them, the phosphors used and the company's blends, plus the technical savvy of the company. You could never get something like CREE RoyalBlue XRE's that cool to behave the way the OP wants if he's talking about making some monster grow COB.

FranJan said:

@stardustsailor
COBS are like fractals. I like that. Just had a almost Mandelbrot like moment thinking about white LED spectrum lately so COBs are like fractals is good.

Isn't the entire point of COBs is taking a big chunk of inefficient Chinese blue emitting semi-conductor, slice it wafer thin then wire them inhumanly close so they act like 1 big thin semiconductor. Efficiency comes from their design, semi-conductors used, the machine making them, the phosphors used and the company's blends, plus the technical savvy of the company. You could never get something like CREE RoyalBlue XRE's that cool to behave the way the OP wants if he's talking about making some monster grow COB.

Click to expand...

Yes ...Almost like that ...

For our example though some things are different ....
Vero 29 is made out of really cheap to produce ( due to GaonSi tech ) LED dies ,
that feature some new developments ( vertical chips ) ,that actually are not
catecorized as " inefficient Chinese blue emitting semi-conductor " but rather as
" average or even high (some parts of each wafer ) efficiency Chinese blue emitting semi-conductor " .
...

But yeah ...
.....

The whole story / idea ,is more or less ,the way you are describing it ...

Cheers.

Stardustsailor , είστε ένας μεγάλος δάσκαλος .
Ευχαριστώ.

Thank you sds, very nicely explained !
I got all the info I needed.
I just I see 45mil Bridgelux are 340mW @ 350mA it is not bad, but far from the 500-630mW oslon SSL are puting out at 350mA.
But I actually understood when you said Cob are like fractal...
Tell me if this is ture please, using only quick approximations from the datasheet:
Vero is 50% efficient around 350mA, it means the die inside are now @ 30mA approximately. Wish mean Vd around 2,75V. It consume 0,08W and It make 0,25% of the 340mW / 85mW.
So Around 25% of ligth for 8 % of power ?!! understand now...

Phosphor causes about 15% up to 20% losses ?

Will Thayer said:

Stardustsailor , είστε ένας μεγάλος δάσκαλος .
Ευχαριστώ.

Click to expand...

Is it google translating that made such a good job translating ,
or it happens that you are a fellow greek ?
( And as you 'd probably know there's not such thing as "greek nationality " .. )

Thank you ,anyways ...

speedyganga said:

Thank you sds, very nicely explained !
I got all the info I needed.
I just I see 45mil Bridgelux are 340mW @ 350mA it is not bad, but far from the 500-630mW oslon SSL are puting out at 350mA.
But I actually understood when you said Cob are like fractal...
Tell me if this is ture please, using only quick approximations from the datasheet:
Vero is 50% efficient around 350mA, it means the die inside are now @ 30mA approximately. Wish mean Vd around 2,75V. It consume 0,08W and It make 0,25% of the 340mW / 85mW.
So Around 25% of ligth for 8 % of power ?!! understand now...

Phosphor causes about 15% up to 20% losses ?

Click to expand...

Now ,for your example ...

1) Say ,that yes 45mil are the dies ,used in Veros .
2 ) Also say that they use the "average " G2 bin ( 400-420 mW ) .
3) Say lastly ,that @ 30 mA ,each die outputs just 0.15 of the 400mW = 60 mW

Then " LED die " radiant efficiency will be 0,06 W / 2.6 V * 0,03 A = 0,06 / 0,08 = 0,75 = 75%
The COB efficiency ( excitation efficiency - conversion losses ) = 75 - 20 = 55% ..
Yes ...It's amost right ...

It's not just phospor that causes these losses ...
For example ...Some of the blue photons that are not used for excitation ,
are readily absorbed by the silicone material of the LES and cause the increase of it's temperature ...
Most do manage to become "photosynthesis fuel " ,though ...LOL !

But yes more or less ,regarding most COBs ,from the diode junction point to the external surface of a LES,
radiation losses are about 1/5 to 1/4 of the initial radiant output( at the junction point ) ...

The dies used in the manufacturing of the Oslon series ,are amongst THE most efficient
LED dies used in the whole LED industry ...

If not the most efficient LEDs ever ,that actually hit the market
and finally reach to be nicely sited ,above leaf canopies ...

So ,you're comparing two different things ...

Dies used for COBs are selected for their emission wavelength ( around 450 nm )
and not by Vf or output power ...
And usually , or 'most probably ' if you prefer ,
their manufacturing cost is ( or can be ) many times less
than the one of the Oslon chips ...



Cheers .

Did you know that LEDs can have a radiant efficiency over 100% ?



http://www.zmescience.com/research/led-over-100-percent-efficiency-mit-94323/

http://www.science20.com/cool-links/optical_heat_pump_100_percent_efficient_led_turns_heat_light-88230

http://phys.org/news/2012-03-efficiency.html

http://spie.org/x112496.xml

@ Will Thayer

What happens if you mix a greek scientist ,some roses and electronic parts ?
Cyborg Roses Wired with Self-Growing Circuits..
That happens ..


http://www.livescience.com/52872-electronic-plants-created.html

http://www.liu.se/forskning/forskningsnyheter/1.660308?l=en

http://advances.sciencemag.org/content/1/10/e1501136.full

Well ..."They " are busy bees ,lately ....
I've to admit that ....

https://www.osapublishing.org/view_article.cfm?gotourl=https://www.osapublishing.org/DirectPDFAccess/357E47B1-B9C0-B349-2A28D3F9D2132436_318668/oe-23-11-A640.pdf?da=1&id=318668&seq=0&mobile=no&org=

https://www.google.gr/url?sa=t&rct=j&q=&esrc=s&source=web&cd=66&cad=rja&uact=8&ved=0ahUKEwjd_qnh0ajJAhWFDywKHReACC44PBAWCEkwBQ&url=http://gooa.las.ac.cn/external/download/980343/3334554/20140825115318474.pdf&usg=AFQjCNHa0_07fbkpN1652-qYzpFwurzekw&bvm=bv.108194040,d.bGg

http://hub.hku.hk/bitstream/10722/209308/1/content.pdf?accept=1
..



Though,they still have some "issues " ....sometimes ...
...

( ...) 1. Introduction
White Light Emitting Diode (LED) technology is currently replacing conventional white
light sources in lighting applications due to their excellent properties of high luminous
efficiencies, customizable spectrum output, low power consumption, good reliability, and long
lifetimes [1]. General white LED lighting is categorized based on Correlated Color Temperature
(CCT) into three general categories: Cool White (CW), with a CCT range of 2500K-4000K,
Neutral White (NW) with a CCT range of 4000K-5000K and Warm White (WW) with a CCT
range of 5000K-6500K [2]. (...)



Cheers.

stardustsailor said:

Is it google translating that made such a good job translating ,
or it happens that you are a fellow greek ?
( And as you 'd probably know there's not such thing as "greek nationality " .. )

Thank you ,anyways ...

Click to expand...

"We are all Greeks. Our laws, our literature, our religion, our arts have their root in Greece."-Percy Bysshe Shelley

"Europe without Greece is like a child without a birth certificate." -Valéry Giscard d'Estaing

"Nothing is more active than thought, for it travels over the universe, and nothing is stronger than necessity for all must submit to it."
Thales



γοογλε τρανσλατε

Cheers,
Will

Steves LEDs now has the Luxeon ES royal blue N4R bin, 70% efficient at 350mA (1W).

For the Cree CXA3590 5000K DB driven at 200mA (13W) I estimate 72% efficiency (234 lumens/W). So the blue diodes in that COB would be running above 80% efficiency at that current. I use that example because the CXA3590 data sheet shows lower current data than the CXB3590. I expect similar or better efficiency from the CXB3590 3500K CD, 4000K CD, 5000K CD and 6500K DB

@REALSTYLES has some of the rare CXB3590 6500K DD, which I suspect deserves the title "Most Efficient LEDs" sold so far. Unfortunately I have been unsuccessful getting the SPD curve or LER from Cree for the 3500K or 6500K so I cannot estimate efficiency

Using "typical" figures, according to the Cree PDF:
@ 700mA (49W), 212 lumens/W

My estimate for lower currents:
@ 350mA (23W) 240 lumens/W
@ 200mA (13W) 260 lumens/W

SupraSPL said:

Steves LEDs now has the Luxeon ES royal blue N4R bin, 70% efficient at 350mA (1W).

Click to expand...

LD CQAR L1 ( 2.8 V @ 700 mA & 1400mW output @ 700mA )
At 350 mA has about 81,5 % radiant efficiency ...
( 1.4 W * 0.55 ) / ( 2.7 V * 0.35 A )


http://www.osram-os.com/osram_os/en/press/press-releases/company-information/2015/osram-significantly-increases-the-efficiency-of-blue-led-chips/index.jsp

http://www.osram-os.com/Graphics/XPic5/00088586_0.pdf/LD CQAR.pdf

SupraSPL said:

For the Cree CXA3590 5000K DB driven at 200mA (13W) I estimate 72% efficiency (234 lumens/W). So the blue diodes in that COB would be running above 80% efficiency at that current. I use that example because the CXA3590 data sheet shows lower current data than the CXB3590. I expect similar or better efficiency from the CXB3590 3500K CD, 4000K CD, 5000K CD and 6500K DB

@REALSTYLES has some of the rare CXB3590 6500K DD, which I suspect deserves the title "Most Efficient LEDs" sold so far. Unfortunately I have been unsuccessful getting the SPD curve or LER from Cree for the 3500K or 6500K so I cannot estimate efficiency

Using "typical" figures, according to the Cree PDF:
@ 700mA (49W), 212 lumens/W

My estimate for lower currents:
@ 350mA (23W) 240 lumens/W
@ 200mA (13W) 260 lumens/W

Click to expand...

No ,most efficient LEDs ( thus actual dies ) are probably the Oslon ( as a series ) ...
Most efficient COBs ,may be the rare CXB3590 6500K DD...

Anyway ..Details...


Cheers.

With 71,4 % (max ) efficient 660nm red diodes * and 81,5 %(max ) efficient 450 nm blue ones ,
no wonder why the monochromatic LED lighting fixtures ,
still are (or can be ) the most efficient ones (quantum-wise )...
Specially the Zelions HL...

https://www.sylvania.com/en-us/products/display-optic-specialty-lighting/Pages/ZELION-HL-Horticulture-LED-Fixtures.aspx

http://www.osram-os.com/media/resource/HIRES/546291/538626/15-gb.pdf

* 0.45 W / ( 1,8 V * 0,35 A ) = 71,4 %
http://www.osram-os.com/Graphics/XPic4/00158091_0.pdf/LH CP7P (EnglishDeutsch).pdf


If only ,Osram wanted to get serious with it's COBs & LED grow lights ....

Things right now ,would 've been way different ...

But ,hey ...

Stubborn germans ,once were " stuck " at the " Red & Blue " idea ...

When they did found out about the green light and it's importance,
moved to the
" Red & Blue & EQ white or red & EQ white ** " idea ....

**EQ white : Blue excitation dies + green phosphor = 198 lm/ W Cyan " White "
http://www.osram-os.com/Graphics/XPic6/00088569_0.pdf/LUW CRDP (EQ white) (EnglishDeutsch).pdf
194 lm / ( 2,8V *0,35 A ) = 197,959 lm /W


.......

Nowdays,they are hoppin' to the " Red (& Blue) & EQ white & FR " wagon ...

Somebody please ,just tell 'em to stop for a sec ....

Say ..



What about a monster 150 Watt Soleriq 3000°K 80Ra ?
Or , at least the " Red (& Blue) & EQ white & FR " idea ,in a COB package *** ?

Just sayin '....



Cheers.



*** Since the phospor of the EQ white is layered on top of the blue excitation die ...
Then there's not much of a trouble ,sticking in a board about 200 diodes in total ,
Mainly consisting of of 660nm Hyper reds -hey ,maybe even some 625 nm ones,too !!! - ,
and then have some 450 nm blue ones ,and some EQ whites as also some FR ones ...
Cover them with a nice optic silicone layer and ...
The "Horticultural COB " ..
The Soleriq HL !!!!


Maybe even ... Separate channels ?
One common ground and 4 contacts .....
- 450 nm
- EQW
- 625 nm + 660 nm
- 730 nm
...
F**K ,yeah !
Now,I like it even more !
I can see it comin ' ....

" Hey dude, did ya hear the latest news ?
It took just a COB for the germans ,
to do what the swedish did with an odd plastic-pet-cage -looking thing ,
called LX something ......

Wow ...

LOL !
I trust that is doable ...

....
....By the germans ,for sure it is ! ....

But ... ....

Nice, guys, learning lot of things again.
Osram seems really good, I have a couple of hyper red SSL I could mix with the 4000K vero 13. Could be a killer mix.
It is just that 10mm stars are a pain in the ass to solder...

speedyganga said:

Nice, guys, learning lot of things again.
Osram seems really good, I have a couple of hyper red SSL I could mix with the 4000K vero 13. Could be a killer mix.
It is just that 10mm stars are a pain in the ass to solder...

Click to expand...

I will be adding some Hyper red ssl 150 on 20mm stars to my CXB3590 3500k fixtures. I also have some ssl royal blues that I have not decided what to do with.

Cheers,
Will

stardustsailor said:

No ,most efficient LEDs ( thus actual dies ) are probably the Oslon ( as a series ) ...
Most efficient COBs ,may be the rare CXB3590 6500K DD...
Anyway ..Details...

Cheers.

Click to expand...

Very interesting potential but I wouldnt give it a title since it looks like that diode is still in the lab. The model that is currently currently available is 72% @ 350mA IF you can source the top bin. If I understand correctly OSRAM reels are sold in groups of bins so you have to buy a reel of 600 to find out what bin it is? Mouser is selling them as the low bin 1120 mW for $2-$3 ea.

Same problem with the deep red, I know of no way to secure the top bins , you have to buy reels and play the "lottery" to see what bin you might get? At 350mA the top bin deep red (4T) is 57.6% and cost $5/PAR W. I have some 3Ts and they are nice for a mono.