LED COB bar project

bmdiyh

Well-Known Member
One more question. If I use 20xCXB3590 for a 32 ft² without lenses. At lets say 12" of the canopy how much coverage would on chip have? Or what if I use lenses? I am asking, because with this number of chips I'll need to rearrange them and maybe make a different design for them to able to cover more evenly the space.
 

SupraSPL

Well-Known Member
Double checked it, yes those figures are correct. That is the 36V class so at 1.4A they are dissipating 49W ea. Cutting edge efficiency for a flowering lamp but currently those COBs are still on order. Jerry hooked me up with a sample pack and I am currently testing them at 700mA (23W) 63.8% efficient. The first time I have tested a flowering lamp that was above 200lm/W, Jerry is the the man!
 
Last edited:

robincnn

Well-Known Member
50% efficiency is Supra's spreadsheet does not mean you power COB with 100 watts and you get 50 watts light an 50 watts heat.
The efficiency is based on theoritical efficiency limit of white phosphoros LED of 333 lumen/ watt or 48.75%

So 100 watts at 50% efficiency would be like 24 watt light and 76 watts heat
Please correct me if i am wrong
 

bmdiyh

Well-Known Member
I played around a bit... still not sure about the coverage of each one of the COB. But do you think this (below) would be a good way to place the 20 COBs in 32 sq ft area?
Screenshot_2.jpg
 

bmdiyh

Well-Known Member
50% efficiency is Supra's spreadsheet does not mean you power COB with 100 watts and you get 50 watts light an 50 watts heat.
The efficiency is based on theoritical efficiency limit of white phosphoros LED of 333 lumen/ watt or 48.75%

So 100 watts at 50% efficiency would be like 24 watt light and 76 watts heat
Please correct me if i am wrong
Well that just sucks ass! Why is that even something of a percentage if it doesn't mean that? I am confused again! :confused:
 

Actionbone

Well-Known Member
Double checked it, yes those figures are correct. That is the 36V class so at 1.4A they are dissipating 49W ea. Cutting edge efficiency for a flowering lamp but currently those COBs are still on order. Jerry hooked me up with a sample pack and I am currently testing them at 700mA (23W) 63.8% efficient. The first time I have tested a flowering lamp that was above 200lm/W, Jerry is the the man!
Efficiency improve very fast! This days! Hps and 3v chips are gone!
 

alesh

Well-Known Member
50% efficiency is Supra's spreadsheet does not mean you power COB with 100 watts and you get 50 watts light an 50 watts heat.
The efficiency is based on theoritical efficiency limit of white phosphoros LED of 333 lumen/ watt or 48.75%

So 100 watts at 50% efficiency would be like 24 watt light and 76 watts heat
Please correct me if i am wrong
Well, you're wrong. 50% efficiency means that 50% of energy is radiated in the form of light and the rest is dissipated as heat.
 

robincnn

Well-Known Member
Well that just sucks ass! Why is that even something of a percentage if it doesn't mean that? I am confused again! :confused:
Sorry i did not mean to confuse you. Please ignore my post. Go with the par watts recommnded on previous page and you will be fine.
 

SupraSPL

Well-Known Member
50% efficiency is Supra's spreadsheet does not mean you power COB with 100 watts and you get 50 watts light an 50 watts heat.
The efficiency is based on theoritical efficiency limit of white phosphoros LED of 333 lumen/ watt or 48.75%
It is based on the theoretical maximum lumens/W for each color curve (LER), but I think you were dividing the LER of a white by the LER of a green (333/683). For a 555nm pure green laser that was 100% efficient (theoretical) we would get 683lm/W and zero heat. For the Cree CXA/B 3000K LER is 325 lumens/W, so a 100% efficient LED of that color curve would create 325 lumens/W and zero heat. At 50% efficiency we would get 162.5lumens and .5W of heat.

As far as the accuracy of the LER, in the case of the 3000K CXA we did get that figure directly from Cree, but Alesh, Mr Flux and SDS also arrive at about the same figure extracted from the PDF curve. The LER, color temp, tint and lumen output will all vary slightly from COB to COB and may shift as the COB ages. SDS did confirm in a lab that the Cree CXA3070 3000K AB bin exceeded its minimum lumen figures as promised.

So for people learning about LED, lumens really do confuse the issue but we are stuck using them for now because LED manufacturers do not give us the output in Watts or photons in most cases. When we look at the efficiency we get a better picture of how awesome any particular light is or isnt. For example:

a 50% efficient green LED dissipating 1W emits .5W of visible range photons (341.5 lumens)
a 50% efficient 3000K LED dissipating 1W emits .5W of visible range photons (162.5 lumens)
 
Last edited:

SupraSPL

Well-Known Member
Efficiency is a great starting point but there is more to the story for example:

a 50% efficient 5000K LED dissipating 1W emits 2.24 umol/s
a 50% efficient 3000K LED dissipating 1W emits 2.33 umol/s
a 50% efficient deep red LED dissipating 1W emits 2.77 umol/s


Some LED geek trivia if you have ever wondered what umol/s means, 2.77 umol/s = 1 668 000 000 000 000 000 photons/s or 1.668 quintillion photons emitted each second for each dissipation W. That is the same rate at which the federal reserve will have to print dollars during the next financial crisis (kidding...maybe)
 
Last edited:

robincnn

Well-Known Member
It is based on the theoretical maximum lumens/W for each color curve (LER), but I think you were dividing the LER of a white by the LER of a green (333/683). For pure green we get 683lm/W from a 100% efficient LED. Warm white tends to be around 325lm/W. So for the Cree CXA/B 3000K LER is 325 lumens/W, that means a 100% efficient LED of that color curve would create 325 lumens/W and zero heat. At 50% efficiency we would get 162.5lumens and .5W of heat.

As far as the accuracy of the LER, in the case of the 3000K CXA we did get that figure directly from Cree, but Alesh, Mr Flux and SDS also arrive at about the same figure extracted from the PDF curve. The LER, color temp, tint and lumen output will all vary slightly from COB to COB and may shift as the COB ages. SDS did confirm in a lab that the Cree CXA3070 3000K AB bin exceeded its minimum lumen figures as promised.

So for people learning about LED, lumens really do confuse the issue but we are stuck using them for now because LED manufacturers do not give us the output in Watts or photons in most cases. When we look at the efficiency we get a better picture of how awesome any particular light is or isnt. For example:

a 50% efficient green LED dissipating 1W emits .5W of visible range photons (341.5 lumens)
a 50% efficient 3000K LED dissipating 1W emits .5W of visible range photons (162.5 lumens)
Thanks. Makes sense now.
1 watt green 100% efficient 683 lumens, no heat
1 watt 3000k 100% effieicnt 325 lumens, no heat
683-325=358 lumens did not just disappear. They are diffrent as lumens are based on human eye color sensitivity. Eye is more sensitive to green than a cct of blue/red
I will have to do all my passive cooling design calculations again.
Now I have more respect for cree when i see their efficiency.
 
Last edited:

testiclees

Well-Known Member
Efficiency is a great starting point but there is more to the story for example:

a 50% efficient 5000K LED dissipating 1W emits 2.24 umol/s
a 50% efficient 3000K LED dissipating 1W emits 2.33 umol/s
a 50% efficient deep red LED dissipating 1W emits 2.77 umol/s


Some LED geek trivia if you have ever wondered what umol/s means, 2.77 umol/s = 1 668 000 000 000 000 000 photons/s or 1.668 quintillion photons emitted each second for each dissipation W. That is the same rate at which the federal reserve will have to print dollars during the next financial crisis (kidding...maybe)
Double checked it, yes those figures are correct. That is the 36V class so at 1.4A they are dissipating 49W ea. Cutting edge efficiency for a flowering lamp but currently those COBs are still on order. Jerry hooked me up with a sample pack and I am currently testing them at 700mA (23W) 63.8% efficient. The first time I have tested a flowering lamp that was above 200lm/W, Jerry is the the man!
Wow that is impressive performance. I will hold off on the 3000k cb and try to acquire those. Do you have any impressions yet from the sample pack?

Did Jerry give an ETA for the next batch?

Thanks S, you've been my guiding (led) light.
 
Last edited:

Actionbone

Well-Known Member
It is based on the theoretical maximum lumens/W for each color curve (LER), but I think you were dividing the LER of a white by the LER of a green (333/683). For a 555nm pure green laser that was 100% efficient (theoretical) we would get 683lm/W and zero heat. For the Cree CXA/B 3000K LER is 325 lumens/W, so a 100% efficient LED of that color curve would create 325 lumens/W and zero heat. At 50% efficiency we would get 162.5lumens and .5W of heat.

As far as the accuracy of the LER, in the case of the 3000K CXA we did get that figure directly from Cree, but Alesh, Mr Flux and SDS also arrive at about the same figure extracted from the PDF curve. The LER, color temp, tint and lumen output will all vary slightly from COB to COB and may shift as the COB ages. SDS did confirm in a lab that the Cree CXA3070 3000K AB bin exceeded its minimum lumen figures as promised.

So for people learning about LED, lumens really do confuse the issue but we are stuck using them for now because LED manufacturers do not give us the output in Watts or photons in most cases. When we look at the efficiency we get a better picture of how awesome any particular light is or isnt. For example:

a 50% efficient green LED dissipating 1W emits .5W of visible range photons (341.5 lumens)
a 50% efficient 3000K LED dissipating 1W emits .5W of visible range photons (162.5 lumens)
awesome job @SupraSPL you must open a account on PayPal so we can donate you some cash!! Your hard work teaching us every day is invaluable!
 

testiclees

Well-Known Member
awesome job @SupraSPL you must open a account on PayPal so we can donate you some cash!! Your hard work teaching us every day is invaluable!
bro he does deserve it. A few months ago i was bragging about my mars900! Today i have an optic 360 and im working on a 50%+ efficient cxb build. Mostly thanks to Supraspl but also with support and insight from many of the bright folks here on the led diy threads.

Its been a great experience joining up here !
 

churchhaze

Well-Known Member
Yes, spot on. It might be even better to compare photon output (µmol/s) than energy output (W) but this is good enough.
Neither PAR W nor umol/s on its own tells the complete story. I know that all research shows that photon flux density is what matters for photosynthetic activity, and not the power of the radiation, however this misses a very important part. (which i'm sure you realize)

Lets say you have a 35% efficient "grow lamp" where all photons emitted are 699nm. It's true that this 699nm lamp must have a higher PPF than a similar 401nm lamp with 35% efficiency, but this isn't enough to conclude that PPF matters more than W.

A 699nm lamp is going to do a bad job of growing plants regardless of it's PPF. If we could add energy to some of those 699nm photons somehow, it would shift those particular photons toward 400nm. This means adding power. Having the right spectrum is why efficiency matters and not just umol/J. If we could make some of those red photons blue, it would cost energy, but PPF would stay the same. Unfortunately, we need some of these blue photons.

Since we need so little blue, the spectral distribution of a good grow lamp will be heavier on the red side, and thus what I said above becomes a mitigated problem, but you still need blue (and more than one spectrum), so it's hard to outright dismiss PAR W while only looking at PPF.
 
Last edited:

Greengenes707

Well-Known Member
Church...Reductio ad absurdum

@bmdiyh
I run 24 3070 over 32+ft^2 with great results and have my lights 20+ inches with or without reflectors. And get an average instantaneous PPFD of 900µmols with very even coverage and little drop off at all.
Your design looks good.
 

Greengenes707

Well-Known Member
Read more clearly.
I did a few times...still not following. Other than taking things to the extreme ends to show a point.
What am I missing that you're trying to show?\

EDIT:
And of course you edit before I can respond


Chaz used to use the example of a 555nm not growing very good, but would have a lm/w out the ass. My response was...do you really think someone is going to grow with a pure 555nm light?
The concept of pure reds is more likely to happen. But also one would have to be VERY uneducated in growing to think that is the cats meow. So I guess I agree...but still call it a Reductio ad absurdum
 
Top