Why are COBs better, and whats so great about the 3590?

[MeGaKiLlErMaN]

you mean ooohmoles?

No I meant PPFD, People do tend to get that confused with the PPF/oOoOoOOOooMoles. Yes they are related closely but when you look at less than 500PPFD from a DE then you start to realize how COBS are better over all..830PPFD compared to 440 you know?

 

[CobKits]

in retrospect to answer the OPs question why 3590s are all that:
-proven performance
-well known trusted name
-great holder and reflector options
-one of the more efficient 36V chips out there (was probably at the top until new vero)

i predict that 50V will be he new 36V soon. why not? little difference in design strategy

 

[GreenSanta]

are we gonna get 50 000 + hours with 3590s on passive heatsinks? That's the real question. If we do, we are all laughing. My last few digital ballast have all given up in less than 2-3 years of use. HIDs are more expensive than people think.

 

[JorgeGonzales]

are we gonna get 50 000 + hours with 3590s on passive heatsinks? That's the real question. If we do, we are all laughing. My last few digital ballast have all given up in less than 2-3 years of use. HIDs are more expensive than people think.

Cree tests at 85C, which is much higher than anybody here aims for as a temperature, and 50,000 hours isn't a problem.

 

[welight]

Cree tests at 85C, which is much higher than anybody here aims for as a temperature, and 50,000 hours isn't a problem.

Cree actually bin at 85C, not test, that is to say that the bins are yielded/selected for output based on both 25 and 85 C, the output number drops of course at 85 V 25, the hotter they run the less light they produce. This is referred to as hot binning. The reality is the moment you fire an led it reachs 25 C almost instantly so it has little value as a spec in real world applications, this applies to pretty much every led cree manufacture and most of Crees competitors have gone the hot bin route, because it reflects real world apps, not sure did I say real world enough times
Cheers
Mark

 

[PhotonFUD]

are we gonna get 50 000 + hours with 3590s on passive heatsinks? That's the real question. If we do, we are all laughing. My last few digital ballast have all given up in less than 2-3 years of use. HIDs are more expensive than people think.

Under driving and keeping them cool, you will probably never ever see them go out in your lifetime. Some of us are running at 10% nominal or even less so they should last almost forever.

led is essentially a nano size arc welder with some fire resistant material on it.

 

[JorgeGonzales]

Cree actually bin at 85C, not test, that is to say that the bins are yielded/selected for output based on both 25 and 85 C, the output number drops of course at 85 V 25, the hotter they run the less light they produce. This is referred to as hot binning. The reality is the moment you fire an led it reachs 25 C almost instantly so it has little value as a spec in real world applications, this applies to pretty much every led cree manufacture and most of Crees competitors have gone the hot bin route, because it reflects real world apps, not sure did I say real world enough times
Cheers
Mark

I was referring to lifetime testing, and specifically the 85C L90 portions of Cree's LM-80 results. They usually end up over 50,000 hours, at currents that are in the ballpark of DIY and commercial builds. Higher currents and/or temperatures are more hit and miss:

http://www.cree.com/~/media/Files/Cree/LED Components and Modules/XLamp/XLamp Application Notes/LM80_Results.pdf

 

[welight]

Data I have from Cree(internal) shows CXB3590 L70 at over 105,000 hours at 105C based on 2100ma operation
L70 means at the referred hours ie 105000 hours, the led will retain 70% of its flux and colour stability from the day you first fired it. If you run them at say nominal 60C at 1400ma, you can conclude you would be into the 200000 hour range, L90 is around 20% less than this however since Cree only have around 7000 hours of actual run time data and based on Cree data on other leds with over 20000 hours of real test data, I have little doubt L90 will match L70, but thats just my view
Cheers
Mark

 

[Rahz]

I've read that because the lumen output doesn't follow an expected curve and deviates over time that the longer they run tests the higher the LM end up being, with some samples showing LM70 at over 200,000 hours.

Can anyone explain the LM80_Results under Reported TM-21 Lifetimes? Specifically why they generally show the 70 80 and 90LM being the same. i.e. 44,400 44,440 44,440

 

[PhotonFUD]

I've read that because the lumen output doesn't follow an expected curve and deviates over time that the longer they run tests the higher the LM end up being, with some samples showing LM70 at over 200,000 hours.

Can anyone explain the LM80_Results under Reported TM-21 Lifetimes? Specifically why they generally show the 70 80 and 90LM being the same. i.e. 44,400 44,440 44,440

View attachment 3778702

I think it is because they had no failures with the given test parameters.

 

[CobKits]

L70 means at the referred hours ie 105000 hours, the led will retain 70% of its flux and colour stability from the day you first fired it. If you run them at say nominal 60C at 1400ma, you can conclude you would be into the 200000 hour range, L90 is around 20% less than this

thats pretty damn flat, L90 out thru 150,000+ hours... i hope i still have 90% of my vitality 37 years from now.....

 

[CobKits]

No I meant PPFD, People do tend to get that confused with the PPF/oOoOoOOOooMoles. Yes they are related closely but when you look at less than 500PPFD from a DE then you start to realize how COBS are better over all..830PPFD compared to 440 you know?

ppf =ooooooohmoles
ppfd=oooooooohmoles/sq m.

 

[MeGaKiLlErMaN]

ppf =ooooooohmoles
ppfd=oooooooohmoles/sq m.

Yes but that. The first is one measurement at a given height. The second is the average over an area at a given height.

 

[wietefras]

PPF is the total amount of light coming from the COB or fixture. There is no height involved.

PPFD is a spot measurement. There are lots of parameters which can influence a spot measurement. One of which is the height, or more precise the distance and angle to the lightsource. Also the optics mounted on the COB or wether the walls are reflective or not.

PPFD is basically something you should measure yourself in your grow room.

PPF is what the manufacturer should provide in the specs.

 

[Smoke-A-Cola]

So would you guys say efficiency is more important than spectrum?

 

[PhotonFUD]

So would you guys say efficiency is more important than spectrum?

Why not both? They aren't mutually exclusive anymore.

 

[Rahz]

I think it is because they had no failures with the given test parameters.

But the LM is based on percentage of light over time so without failures the LM70 should be longer than the LM90.

 

[Airwalker16]

theres a slew of new chips that are more efficient and cheaper than the 3590, and we expect that since they are available in similar color temps and CRIs that they will produce the same as the crees, but we need a few more months to prove that. for example vero 4000k is not the same spectrum as cree 4000k, etc.

heres a test i did last night with 3590 at middle of the pack despite being the most expensive chip in the test.

All the chips below it are less than $20...

View attachment 3776888

Where you labeled CT 2828 did you mean GT?

 

[CobKits]

yeah typo. see cobkits thread for list of actual part numbers tested

 

[CobKits]

Why not both? They aren't mutually exclusive anymore.

they kind of are. for a given chip model, efficiency measured in lumens per watt will always be inversely related to QER