Lumen maintenance on LEDs is < 3 years

Lazyman · Jan 12, 2019

I've been reading studies, whitepapers and talking to lighting engineers for years, both in the US and China. The LED industry is trying to hide from you that their lights are disposable after 10K hours of use, as output falls to 90% or lower.

A couple reputable companies have addressed this (see chart from Spectrum King) but most are happy to sell you a panel with a promise that it will turn on for 10 years. They just dont tell you it will be at half it's original brightness. The only fix is to replace the panel.

The lighting engineers all know this, but marketing largely doesn't want it published because it means their expensive panels with a 5 year ROI are disposable, and that ROI won't be met.

Drop That Sound · Jan 13, 2019

I was always against LEDs from the start, my gut feeling has kept me away from spending big bucks on any COB's, or any other current chipped light emitters for that matter. Thankful I made it through the blurple phase in like 2013-2014, with only spending 25 bucks on a cheap 20 watt screw in, to get my pink headache light fix.

Glad I invested into newer digital HID ballasts instead at the time. Even though, my old mags collecting dust will probably still be kicking in 30-40 years in the end times. I'll be underground still using them with a bulk pallet of cheap bulbs, as soon as hyped up expensive LEDs completely take over, and standard HID bulbs become a dime a dozen. The less parts to break down the better right?

COBs and spendy strip\board light panels are all the rage right now, all hyped up with misleading marketing as you say. I could never spend 100's or 1000's even if I made the fixtures myself to save. Maybe water cooling could make them last/stay brighter a bit longer, after all heat is the enemy, but that's for another time.

I'll just play with screw in LEDs, to get my fix for now. At current price of $140 per 60 bulbs (or 930 actual watts with 15.5w bulbs), I'll be at full brightness for ten years if I replace them a few times (less than $300 more, maybe cheaper in due time). Meanwhile the guys that spent a grand or 2 on a 900+ actual watt pre-made fixture, or a few lower wattage ones, will lose out...

I suppose they can change out COB chips, on some fixtures, especially DIY ones. That's cool if you know what you have to do, and the cost to do so.

I don't fall for the whole ROI marketing sales pitch idea. No thanks, I want what I'm paying for, and I want it now! At least for 3 years. For cheap too. Just like the majority of LED tech being manufactured for growing, and hiked up 1000+%. Along with fancy expensive looking fixtures\cooling modules\driver that cost the most.

Douglas.Curtis · Jan 13, 2019

This is incorrect for the new ceramic based LEDs.

Osram is already making chips which lose only 10% over 100k hours. That's 22 YEARS to lose 10% in a flower room. All LED previously were made with a plastic base.

Player2 · Feb 5, 2019

Although I think this link probably deserves its own thread, I'll toss it here.

https://hackaday.com/2019/02/05/what-happened-to-the-100000-hour-led-bulbs/

So yeah, lumen output over time problems... color shift problems... electronic engineered failure time bombs in the circuits...

Of note is the temperature sensitivity. All you homebrewers, and cob/ho overdriven guys especially, need to be water cooling.

Come on growmau5! Get some kits up with gaming computer chip coolers!

Also check out the flashlight guys for dome shaving and polishing.

Drop That Sound · Feb 6, 2019

Or just make a wired socket fixture, buy 8-16w screw ins, and change the bulbs every 3 or so years. In 10 years or so you'll still have spent less than any other fixture, especially pre made commercial ones, but have full output the whole time.

Investing in a DIY cob setup the only parts that are gonna last, and not eventually need replacement, are the heatsinks and framing itself. The fans, cobs, and drivers will eventually fail.

With water cooling though, the Cobs will last way longer i'm sure. Thats the only way I would go really.. i would even find a way to watercool the meanwell drivers too, or simply submerge them in a tub of mineral oil with a radiator lol. Sure it might cost more initially, but in the long run you'll have a nice rig to keep gluing up new chips in the future, if you ever even need too.

Douglas.Curtis · Feb 6, 2019

Player2 said:
So yeah, lumen output over time problems... color shift problems... electronic engineered failure time bombs in the circuits...

The newer ceramic LEDs have significantly slower losses, vs. the plastic based LEDs the article is based on. Time and tech is marching on, and quite quickly.

Player2 said:
Of note is the temperature sensitivity. All you homebrewers, and cob/ho overdriven guys especially, need to be water cooling.

Come on growmau5! Get some kits up with gaming computer chip coolers!

Also check out the flashlight guys for dome shaving and polishing.

I'm with you on the water cooling.

What's the difference in output between a bare chip and using a dome?

Koondense · Feb 6, 2019

Apples and oranges.

Douglas.Curtis · Feb 6, 2019

So, reduced efficiency. Got it. Probably why you don't see lenses on Fluence and other top of the line LEDs?

Koondense · Feb 6, 2019

Douglas.Curtis said:
So, reduced efficiency. Got it. Probably why you don't see lenses on Fluence and other top of the line LEDs?

Sorry, I was not responding to your question Doug

Was more of a comment about the whole thread and addressing all leds as same.

To your question...domes(glass or plastic) retain heat so generally cooling is less efficient with cob leds + lenses. Similar as do glass windows keep the heat inside our homes.

Cheers

f-e · Feb 6, 2019

yep, hotter leds are less efficient, don't last as long, and may be off colour if rated at 25c but run at 85

Light transmission through glass is likely to cause losses around 10%, but some lenses are really thick and heavy. You can expect more. Which makes me trash the idea instantly. Most leds have a perfectly usable beam pattern anyway.

I do have a habit of lifting plants up and hitting the lights when they're so close. So do like a cover. But just a couple of millimeters thick.

It is unfortunate that my leds will take a drive so soon. I didn't see it coming on my first build. I take solace from the fact I use high powered cobs though. It's just a couple of screws, and two wires. If were 100 surface mount led's I would look for cobs that match the ballast.

I have been meaning to find broken lights on eBay. I think people will expect too much for them, but it could be a cheap ballast and heatsink source

popta · Feb 6, 2019

I've been keeping an eye on the LED situation for many years too. It's important to separate the question: "How long do LEDs last?" from the very different question: "What kind of junk will some fly by night grow-light companies try to sell?".

Over in the legal-business world there have long been standards for testing LED lifetime. LM80 is the name of the test used to measure lumen maintenance of LEDs. Testing must be carried out by EPA certified laboratories (because this data is used to check compliance with the EnergyStar program). The test results tell you how many hours it takes for that LED to drop to 90%, 80%, 70% etc. of it's starting brightness when run at a specific power level and temperature.

LED manufacturers publish this LM80 data so it's pretty easy to answer the first question, "how long do LEDs last?". I just looked up the LM80 results for a couple modern, high quality LEDs from Cree:

XPLBWT White LED, 2A @ 85°C / 185°F

L90(13k) = 61,900 hrs
L80(13k) > 75,600 hrs

XP-G3 Blue LED, 0.7A @ 85°C / 185°F

L90(6k) = 30,000 hrs
L80(6k) > 36,300 hrs

XP-E2 Red LED, 1A @ 85°C / 185°F

L90(17k) = 68,900 hrs
L80(17k) > 103,000 hrs

They test these things at very harsh temperatures to give you an idea of the worst case performance, but even under these conditions the time it took to drop to 90% of initial lumens was between 30,000 and 69,000 hours. That's 7 - 15 years @ 12hr/day. So how long can LEDs last? A very long time.

But there's a problem.

High quality LEDs like these are still very expensive, around $3 a piece even when buying >1000 pieces, and you need about 200 of them to match the light output of a 1000w HPS. That's $600 just in raw LED costs. You still need to add a circuit board, frame, heatsinks, ballast, labour to assemble and test it, and some profit margin so there's a reason to actually build the thing.

How many people want to spend around $1500 for a single 1000w equivalent light? Probably not that many. So what's a shady grow-light maker going to do? Use lower quality components. Maybe look for a ripper deal on some bottom barrel LEDs that die fast, cheese out on the ballasts and heatsinks etc.

In other words = Quality LED lights exist, but they're very expensive still. So if you're looking at an LED grow light and it doesn't cost the stars and the moon, you should be asking yourself why that might be, what corners did they cut. Ask the manufacturer what type of LEDs are used in the unit and look up the LM80 data for them yourself.

Ibechillin · Feb 6, 2019

Welcome to the forums Popta! I didnt bother replying to this thread as im already copping enough shit about lighting lately ahaha, thanks.

Douglas.Curtis · Feb 7, 2019

Thank you popta

Great stuff.

Welcome to ICMag!!

Lost in a SOG · Feb 7, 2019

As we type the big agripharma companies are all going, or have gone, LED precisely because the lifespan is so much better on the latest gen osram, cree, samsung etc diodes than hps especially with good thermal management.

Also you can get the right spectrum specific for a crop so less junk wasted photons plant cant fully utilise.

popta · Feb 9, 2019

Thanks for the welcome guys. I personally haven't taken the LED plunge yet. I've been watching their efficiency and lifespan steadily march forward for more than ten years now, but it seems like there's been no progress at all with respect to the high initial cost. That still seems like a problem to me. I don't necessarily *want* a costly light that makes up for the initial expense by lasting 10+ years when the efficiency of new LED models is still steadily increasing. 2.2 μmol/w does sound pretty attractive, but in 5 years aren't I going to want to dump it and buy that hot new 3.2 μmol/w panel? Yeah I know me, of course I am

Given how long it's taken to get to this point though I wouldn't be surprised if it's another seven or eight years before the initial price for this level of performance comes down to a more competitive level and who knows if I'll even live that long so maybe you just have to jump in a at some point.

The efficiency and lifespan of the red LEDs in particular is getting very compelling. Some are over 3μmol/w these days, very attractive. Maybe I should dip my toes by getting some red supplementary lighting. Like this kind thing http://www.hyperiongrowlights.com/products/inter-light/

popta · Feb 9, 2019

Can you edit a post here? That link should say http at the front, not https.

Ibechillin · Feb 9, 2019

You can edit and private message etc after you have 50 posts.

aridbud · Feb 9, 2019

Good info. For us, it costs almost the same as HID lighting which we had until last 8-9 years. We replace lights every other....wait for a sale, then jump on it rotating LEDs. Each generation of LED's are getting better. Knowing wattage/chip-set helps.

Electric utilities are less, for which we're grateful using LED....heat build up isn't an issue.

As others mentioned....do your research before purchasing.

Lapides · Feb 9, 2019

Drop That Sound said:
I was always against LEDs from the start, my gut feeling has kept me away from spending big bucks on any COB's, or any other current chipped light emitters for that matter. Thankful I made it through the blurple phase in like 2013-2014, with only spending 25 bucks on a cheap 20 watt screw in, to get my pink headache light fix.

Glad I invested into newer digital HID ballasts instead at the time. Even though, my old mags collecting dust will probably still be kicking in 30-40 years in the end times. I'll be underground still using them with a bulk pallet of cheap bulbs, as soon as hyped up expensive LEDs completely take over, and standard HID bulbs become a dime a dozen. The less parts to break down the better right?

COBs and spendy strip\board light panels are all the rage right now, all hyped up with misleading marketing as you say. I could never spend 100's or 1000's even if I made the fixtures myself to save. Maybe water cooling could make them last/stay brighter a bit longer, after all heat is the enemy, but that's for another time.

I'll just play with screw in LEDs, to get my fix for now. At current price of $140 per 60 bulbs (or 930 actual watts with 15.5w bulbs), I'll be at full brightness for ten years if I replace them a few times (less than $300 more, maybe cheaper in due time). Meanwhile the guys that spent a grand or 2 on a 900+ actual watt pre-made fixture, or a few lower wattage ones, will lose out...

I suppose they can change out COB chips, on some fixtures, especially DIY ones. That's cool if you know what you have to do, and the cost to do so.

I don't fall for the whole ROI marketing sales pitch idea. No thanks, I want what I'm paying for, and I want it now! At least for 3 years. For cheap too. Just like the majority of LED tech being manufactured for growing, and hiked up 1000+%. Along with fancy expensive looking fixtures\cooling modules\driver that cost the most.

yesum · Feb 9, 2019

I use 14 watt house lighting screw in LED's with the bulb part cut off. If they wear out in 3 years no biggie. I can grow plants 20 inches high with them which is good enough for me.

Lumen maintenance on LEDs is < 3 years

Overkill is under-rated.

Attachments

Well-known member

Autistic Diplomat in Training

Member

Well-known member

Autistic Diplomat in Training

Well-known member

Autistic Diplomat in Training

Well-known member

Well-known member

Member

Masochist Educator

Autistic Diplomat in Training

GrassSnakeGenetics

Member

Member

Masochist Educator

automeister

Rosin Junky and Certified Worm Wrangler

Well-known member