LED lifespan

[BradVincent]

What is the average lifespan of a LED in your tank? Do you feel anything makes a difference?

 

When I first started, everything got both lenses and conformal coating. I've found these actually seem to shorten the life. Any small amount of coating (conformal or hot glue) on or near an actual LED drastically shortens lifespan. I believe lenses may slightly raise the temperature as well. I have yet to deal with corrosion of an uncoated soldered joint except in a few cases where the whole chip corroded (lenses may have prevented this if I used them).

 

Now all my LEDs are bare naked, and I think they last longer (this saves money vs paying for lenses too).

 

Royal Blue LEDs (both Phillips and Cree) have faired the worst, but possibly because I haven't made as much effort to reuse cool blue, green and UV where the whole strand gets removed when 1 LED goes. Royal blues running at 700ma seem to do much better than at higher current. Cree XML/XML2 seem to last longer, even at 3000ma, although I've seen more failures than I thought I would when I switched over from Halides.

 

The LEDs on the AI Sol (commercial fixture) seem to have lasted longer, but even there I had several LEDs fail.

 

I have only lost 1 cheap Chinese LED, which failed within 1 month. I haven't had enough of these for long enough to say they definitely last longer than Crees and Phillips, but it is starting to look that way. I am also running them far short of recommended current.

 

Are manufacturers being overly optimistic about lifespan? Are our tanks overly hostile to LED lights? Even with the failures, I would not save on bulb costs if I switched back to halides, but I might if LEDs where still as expensive as when I switched over in the first place.

[evilc66]

There will always be a chance of failures due to manufacturing issues, but the overall percentage of failures due to this versus production volume is very small (way less than 1%). Heat ultimately is the number one killer of LEDs. Unless you have very specific data on how hot each of your LEDs have become over a period of time, you really won't be able to say without a shadow of a doubt that the LEDs died for any other reason.

 

Really, these are very simple devices with very simple requirements: Keep their current and temperature within spec, and they will live a long life.

 

I have a number of LED fixtures that have been running for years without any failures of any kind, and I use a variety of different brand and series of LEDs. All of them I built myself of course, and I deliberately keep die temperatures way below the manufacturer spec for the sole purpose of helping to keep the LEDs alive as long as physically possible.

[blasterman]

I'm going out on limb and concluding a good deal of random LED failures are caused by the driver and not the LED. LED failures in commercial mainstream lights with proper drivers are rare. Our own LED rigs are prone to solder and wiring inconsistency, and even the tiniest short or jump can cause current regulated drivers to surge and cause all kinds of downstream issues.

 

Lights I've built with fixed voltage ciruits are far, FAR less prone to random LED deaths, even when using cheap LEDs. They basically run forever.

 

LED out-put reduction is usually caused by heat. I've reduced the out-put on many a Cree by running it at max current and poor thermal conditions, but it's a lot tougher with larger LEDs like XML's.

[GHill762]

back when I worked for an electrical supply company (5+ yrs ago) we were told by sylvania at an informational dinner we hosted that their leds would last roughly 14 years if run 24/7, and that power supplies would be the item to need replacing.. they weren't talking about overdriven 3w led's though, they were talking about <1w/led multichip arrays/can lights/signs/etc..

[Sandeep]

One of my pet peeves has always been the often quoted 50,000 hrs that led fixture and bulb manufacturers like to throw around. That is complete marketing bogus. You read about led's failing all the time. They are a complex electronic unit and there is lots that can go wrong, including the power supply as mentioned.

 

I was looking at the bargin bin at my local LFS and they had brand new fixtures, where a entire row of led's in the fixture were not working.

 

I wonder how long they will last, I guess we will find out in a couple of years of real world aquarium use, rather than marketing theoretical hype.

[BradVincent]

I'm going out on limb and concluding a good deal of random LED failures are caused by the driver and not the LED. LED failures in commercial mainstream lights with proper drivers are rare. Our own LED rigs are prone to solder and wiring inconsistency, and even the tiniest short or jump can cause current regulated drivers to surge and cause all kinds of downstream issues.

 

Lights I've built with fixed voltage ciruits are far, FAR less prone to random LED deaths, even when using cheap LEDs. They basically run forever.

 

LED out-put reduction is usually caused by heat. I've reduced the out-put on many a Cree by running it at max current and poor thermal conditions, but it's a lot tougher with larger LEDs like XML's.

 

 

What drivers are you using? I only use meanwell, which aren't cheap, at least not on my wallet.

[blasterman]

I was looking at the bargin bin at my local LFS and they had brand new fixtures, where a entire row of led's in the fixture were not working.

 

The LED lights used in the reef industry are pretty low volume, and the chinese units are prone to failures like this because they are slapped together without a lot of engineering prep. The general lighting industry in general doesn't seemed to be too concerned about LED longevity.

 

Again, the LED's themselves are typically reliable, even the chinese ones. The dead row of LEDs is as I explained a driver or solder issue.

 

It's like we're judging the reliability of computer processors based on hard drive and power supply failures built in china.

 

I have strings of 3watt LEDs on the ceilings of some nightclubs that are past 50k hours 24/7 , and some bridgelux approaching 40k hours with zero dimming.

[GHill762]

One of my pet peeves has always been the often quoted 50,000 hrs that led fixture and bulb manufacturers like to throw around. That is complete marketing bogus. You read about led's failing all the time. They are a complex electronic unit and there is lots that can go wrong, including the power supply as mentioned.

 

I was looking at the bargin bin at my local LFS and they had brand new fixtures, where a entire row of led's in the fixture were not working.

 

I wonder how long they will last, I guess we will find out in a couple of years of real world aquarium use, rather than marketing theoretical hype.

well.. the 50000 hrs doesn't mean it won't fail, burn up, etc. it means if it's run properly and kept cool they should easily outlast the driver and run for many years..

 

let's say an average car can go 200k miles (totally made up number), many cars are totalled, blow up an engine, destroy a transmission, etc long before than number (many times because of the way we abuse them), there are also some cars that will make it to 400k miles and outlast the life expectancy..

[jedimasterben]

One of my pet peeves has always been the often quoted 50,000 hrs that led fixture and bulb manufacturers like to throw around. That is complete marketing bogus. You read about led's failing all the time. They are a complex electronic unit and there is lots that can go wrong, including the power supply as mentioned.

 

I was looking at the bargin bin at my local LFS and they had brand new fixtures, where a entire row of led's in the fixture were not working.

 

I wonder how long they will last, I guess we will find out in a couple of years of real world aquarium use, rather than marketing theoretical hype.

I'm sure if you contact Philips, Cree, and Bridgelux, they'd be more than happy to hear your concerns about their longevity data.

[Gooburz]

It is all about the conditions you have the lights in. I have PAR38 bulbs and I take them out about once a month and clean and dust the heat sinc because guess what even with a cover on I have no clue how but salt creep still gets up there. Depending on how hot you run your apartment/home can effect the lights, do you have any air flow?

 

Every product out there known to man has a warranty, but those all apply to ideal conditions. You can't buy a set of tires that are rated for 50,000 miles and get made when they need to be replaced after 10,000 because you drive like a moron or live in horrible conditions.

 

And no matter what manufacture you buy from you will get faulty products, its gonna happen. So if you expect things to be perfect and last exactly how long they say you are in a dream world.

[Reefwiser]

This is something that needs more research. We need to see long term what LED's will do. MH and T5 will put out light for a long time.

But the beneficial spectrum is short lived on them. Causing us to replace bulbs every 6 months to keep the spectrum at its most effective level.

As we track their lighting with PAR meters over the coarse of years we will be able to see how long LED light are effective.

[jedimasterben]

This is something that needs more research. We need to see long term what LED's will do. MH and T5 will put out light for a long time.

But the beneficial spectrum is short lived on them. Causing us to replace bulbs every 6 months to keep the spectrum at its most effective level.

As we track their lighting with PAR meters over the coarse of years we will be able to see how long LED light are effective.

Look at the manufacturer ratings. Most are rated for 70% of their luminous or radiant output after 50-80,000 hours at specific current and thermal ranges. There is no real phosphor degradation like there is with fluorescent or metal halide bulbs.

[MOJOEJOEJOE]

I sort of look at my build like this(im sure others do too).... I had pc`s prior to this build 3+ years ago and I remember replacing my bulbs every year at about $30 each. I had 4 bulbs so that was $120/ year. I dont own fancy gadgets for measuring intensity but my corals dont lie. If the light was bad the corals would eventually let me know. That being said, Im not sure that anyone can say every led will last for 10 years or 50k hours but you should get much more out of it than a MH,T5or PC.

[GHill762]

in most cases you'll want to upgrade to new technology long before your leds go kaput anyway..

[blasterman]

This is something that needs more research. We need to see long term what LED's will do. MH and T5 will put out light for a long time.

 

No, it doesn't need more research. LEDs as a commercial lighting standard already have enough research. The reefing industry is just a niche' industry with a lot of weird opinions. Different physics don't apply to LEDs because some guy making $9.00 in a reef shop says so.

 

Ask any facilities engineer about T5 -vs- T8 and they will tell you the exact lumen loss per month and how often they have to change the bulbs to reach spec. Given them the specs of a Cree ceiling fixture and they'll have the math worked out to the foot lumens in a few hours. They'll also tell you they hate T5 because the bulbs don't last as long as standard burn T8s. Reefers are so enamored over T5, but in the mainstream lighting industry T5 isn't that popular because of longevity issues.

 

LED's as a mainstream lighting technology are already mature and displacing MH and T8 as fast as their budgets allow. There is no 'wait and see yet'. Because somebody likes the color of their phoenix over a 120watt Chinese fixture doesn't change the fact you have billion dollar companies moving to mainstream LED fixtures and have the cost factors down to the penny.

 

MTF rates of LEDs as a whole are likely way less than 1% as Evil says. If you get a failure rate higher in a fixture off Ebay it doesn't warrant discussion. If I can build a basic fixture with spare parts that can run 50K hours with no light loss imagaine what Philips and Cree can do.