At 85C theres NO WAY a LM301H is hitting over 3.0μmol/J. If you want efficiency then you want low temps. 25C = 77F, and this is the temp the LEDs are pulsed at when testing. Keeping at 85F, regardless whether HLG posted it or not, will keep efficiencies high. Do not operate at 85C. Lol @HeadHunter has the right idea about keeping temps low. I think others also said similar but it seems it's being argued that the stated 85F is not the recommended. The exact # is less than important aside from staying generally cool. The exact temp is splitting hairs and without a sphere you won't have a clue. I don't have a big sphere to throw a QB in does anyone else? The cooler in temp you keep the boards the greater efficiency they will have, and the hotter they are the less efficient they will be. I'd recommend less than 110F or you could have heat issues in your environment depending on CFM of exhaust ect. I'd recommend 165mA max per mid power chip to try to keep efficiencies somewhat higher during full brightness although the less current you run will increase efficiency and reduce heat even more.
Additionally most L70 tests are done at 25°C. For most any real world scenarios cooler is better for longevity and efficiency. Anytime you push electronics to the edges of operating temps there are substantial degradations in performance and function.
It is measured in Celsius, just like every other data sheet and piece of scientific research out there. 55C is really not that hot and within the acceptable range for these chips specially when you compare it some other ICs like some chips in modern gpus running 80c. The lower you keep the temps the longer the life will be and the closer it is to 55C the closer it will be the performance in test conditions of that bin. Spectrum will vary a little with a change in driving current and temp but not enough to really matter for our application,hell how beat up the walls of our tents are probably have a bigger influence on the light then that though. If your running these panels at recommending spec they should be fine and I would imagine that it would be transistor in the power supply that fails before a diode does, well barring physical damage.
Really the way to cool our Boards would be liquid cooling ..if you wanted the maximum life ..and efficiency .
I used to think that with my old experiments with LEDs using Edison Optos. Honestly you can spend less money buying more boards are running them lower then buying heat sinks or water blocks for only some minor cooling effect. The highest I have seen only my 132s at 1.4A is 95F center of the board and they are just passively cooled by the airflow around the tent. I am actually about to get some 288s to setup a rack for my peppers with some very close spacing 16" (2" for tray, 2" head space on light, leaving 12"). This will replace my hodge podge collection of LEDs and T8s I have been using over the years. Luckily this a down year for me so maybe only have 600-700 starts to care for this winter. If these panels did not run so cool I would not be able to maintain such a close spacing and it up my cell space almost 30% for area in the early stages. As a side note I have always wondered about running them sub zero say -60c or so and comparing draw and output.
you say you are going to run 288 boards close? Am I misunderstanding or are you figuring your heat experience with the 132 boards will allow you to run the 288 boards really close? If you already have run 288 boards to try this please disregard this, but if you haven't yet, I own both 132 and 288 boards and the heat output difference between the two is VERY significant. I run my 4x 132 boards in a 2x4 tent at full power - 75w /board and as you already know the heat output is very minimal even when run at their max on a 24/0 lighting schedule which I do. I have trouble keeping this tent at high enough temps because of this. My other 2x4 tent has 2x 288v2 boards on a hlg-240 driver and at 100% power the heat output is significant. I usually can only run those lights at 75% max power just because of heat, and actually it's still an issue until I get them down to 50% - or less. to run 288 boards close you are most likely going to have to run them so dimmed down, you might as well go down to no more than 60w a board output and run a lot of boards for full canopy coverage all dimmed low. would eliminate the need for heatsinks on them if you do that. 60w or less the 288 boards don't require a heatsink. again, if you have already experimented with the 288 boards already please disregard. I just didn't want you to base you plans thinking the heat output between the boards were similar.
Im sure the maximum operating temp mentioned is just what the chip is capable of withstanding in a most extreme environment. It also probably has a minimum operating temp too. I have some NEW OLD STOCK guitar amplifier tubes still in the original boxes, and all parameters test as new at full voltage at 9 different stages. These tubes when bought new in the mid 60s were guaranteed to operate at full voltage, 24 hours a day, 7 days a week, from -40 degrees F up to +140F, for either 10,000 HOURS, or 2 years.... Whichever came first.....59 Weeks is 10,081 hours or around there. These tubes were screened from the factory, and are insanely expensive. The tubes I use in my guitar amp are Mullard 10M Master Series ECC83/12ax7 Gold Pins. They also used subminiature tubes in guided missiles that could take even bigger extremes in heat, and cold, and shock. Many of them also had gold pins to combat corrosion.
My whole point of the discussion was that I didn't think that the guy running the test in that video a few pages back was running that HLG 288 overly hard based on the heat it was putting out. It does make me wonder if those other boards which appeared to have much beefier heatsinks have something to do with less loss in output and maybe the HLG boards need to either better space the diodes or start providing a better heatsink. I only run mine at 100watts and each set of 2 has a small 6inch fan blowing across the sinks and they produce almost zero noticeble heat.
With the driver I have I can run a 288 V1 at 48v at about 1.1-1.2A was going to try them out and if they do not fit the bill will just use 132 because I can drop them even lower. I am going to attempt to go without a heat sink and just hookup 120mm fans to blow across panels. If that is not enough will mount the panels onto Al plate to make up for the reduction in surface area using the 288 as opposed to 132. I will elaborate on the project some. They are going to be used in a propagation system which will be in a greenhouse used in winter/spring generally the ambient temps are kept at 55-60. Lights would be reverse cycle and adjusted time/intensity to hit a given DLI point. The majority on the stuff I germinate are peppers (3/4) which some of my super hots need 85 to have an even germination rate and are on temp control mats additionally. I ramp up my PPFD from 200 umol @16hrs = 12 ish DLI (if have a different set of lights I put on top of humidity domes for this but would be fine If I can hit that lower with QBs) then bump up about 400 umol @ 12 hrs = 17ish DLI (the reason for 12 is so I can maintain flower if needed in Greenhouse). This continues on until spring catches up. I want to grab the 288s to try them out and run a couple next to 132s, I am not worried if they do not work for this us because I will need some for other purposes. I am also collecting up a lot of 132s this year.
Well first things first, respect to Migro for testing and posting, I like the guy, I've been watching his vids a while. Personally, I would have done it differently, but at least he's testing. It seems he was more or less testing a mish mash of LED to determine general LED intensity reduction, as opposed to what it seemed to me was being debated here, which was comparing the intensity reduction %'s Migro posted between the different manufacturers. Continuing on from the manufacturer to manufacturer comparison standpoint: The only issue I have with the testing was that it didn't seem to control for anything. Turning on a light and letting it run doesn't tell you much except that it might still run after awhile. I think you'd want to control for temp, or intial efficacy (maybe). To me it makes sense that the light operated at 15°F hotter, was the worst performing after 6,000hrs. Also, there's another point of ambiguity, I'm not sure how much dust the splash guard picked up. With a lens like that you've the added possibility of getting dust on both sides, plus the chips too. The guard is easily removed and I'd be curious how much the lens was contributing to the fixtures intensity reduction, I've not seen anyone use them. IMO, if objectivity were important, then I think having all of them running at the same temp would be a good start. I agree that the temp or wattage wasn't overkill, but when comparing the performances of the light engines to each other, the individual temp isn't the issue to me, rather it's the difference in temp from one light engine to the next. They are all phosphor color LED, so temp should effect them all pretty much the same. Some have better heat management than others, but if the heat of the light creating component (LED) is greater on one light over another, the LED longevity will be reduced that much more. I think if you were to control for temp, and then measured light reduction, you'd have more compelling results, at least between the different manufacturers. It seemed there was arguing over the "recommended" temp that a QB should be operated at. Someone said 85F, then it was said 45C (or something), up to 85C. Then someone said L70s were done at 25C. There seemed to be a lot of confusion on temp, and that's why I re-iterated. Whether HLG posts a recommended temp or not is irrelevant. The chips are pulsed at 65mA to keep them 25°C, if they drove them any harder, then the chips would heat up and their Vf would drop (and mess with power consumption) as well as their intensity. When Samsung says 3.0+ it's only achieved by this means of wimpy driving. Any heat will reduce intesity and we don't pulse operate the chips. The figures touted aren't going to be what you're going to get even if you keep the boards at 85F. If a manufacturer recommends a temp, great, but regardless the temp recommended, the cooler the chips the longer the life and greater the efficiency. It seemed people were missing the forest from the trees, seemed they were splitting hairs on the exact manufacturer temp recommendation vs the general characteristics which influence LED operation, as well as the fact that 1 out of 4 lights was 15°F hotter than the other 3 out of 4 lights. If you're comparing 2 LEDs lifespans, you'd want to keep them the same temp regardless who the manufacturer is. If you're not controlling for temp then you'd want to control for something else aside from duration. EDIT: Also, chip binning makes a huge difference too, just because you have a specific chip model it doesn't mean you'll be close to what's advertised. You have to get the correct bins of that model and often times the little guy can't buy the top of the top bins.
I'm not even sure if each of the lights was running at the same wattage. it is kind of odd that he didn't put more effort into this test considering it had to go for an entire year to get results. Honestly, a lot of people probably run their QB's a lil hot esp if you are one that was previously using HID as a heat source, and it was still nice to see that if ran at that temp what light output could be lost. This thread is the first time I've heard about cleaning dust off my QB, I wonder how many are doing this on a consistent basis.
Some info for y'all Output degrades as the die gets hotter. The biggest factor to longevity and efficiency is the efficiency of the internals of the chip itself. There's a number on the data sheets that is the thermal impedance of the die junction to ambient. The better that number is the more efficient the chip is and the better it does at higher temps and in turn higher currents. This is why you need to check your efficiency at the current and temp you will be driving. The LM301H is 25C and 65ma which is very mild for both. Industry standard temp/current for a lot of mid power. Where Samsung LM301 series excels is there temperature differential I mentioned. It has an amazing lack of degradation vs other chips.
I ran across this video by Migro on the drop-off in intensity for some LED's. He tested the QB 288v2. I thought it would be interesting.