Totally OT, LED lighting question

[Bruce Schoenleber]

I recently replaced my Kitchen ceiling florescent fixture with a 18"X48" LED fixture, it uses an LED strip that appears non-replaceable. While the unit is
rated for 30,000-50,000 hours I am skeptical. My question therefore is, what has the greater effect on LED burnout, cycles or burn time?

Any solid state electronics engineers care to authoritatively answer that?:D

Thanks!
-bruce

[david privett]

I have done electrical work for awhile and recently we have been installing a lot of leds . I think the only failure was from a lightning surge I seen. I think long term vibration would cause problems. And why I say that is have you seen cars with leds and one or two be out with the rest doing fine.

[John K Jordan]

I recently replaced my Kitchen ceiling florescent fixture with a 18"X48" LED fixture, it uses an LED strip that appears non-replaceable. While the unit is
rated for 30,000-50,000 hours I am skeptical. My question therefore is, what has the greater effect on LED burnout, cycles or burn time?

Any solid state electronics engineers care to authoritatively answer that?:D

Thanks!
-bruce

In my experience and research the life may related to the heat. Those with poor heat sinks might go quicker. Those with low quality circuitry might also go quicker even though the LED emitters are still fine, same as CFLs. The life ratings on all the fixtures are calculated estimates, not tested times.

My on anecdotal "evidence" is with some I left on all the time. These were porch light fixtures and they did get warm. Both started flickering and failed within four years.

Save the receipt. My new policy is to not buy fixtures with non-replaceable elements.

JKJ, not a solid state engineer or liquid or gaseous state engineer

[Leo Van Der Loo]

I have had Led lights already for a couple of years in use in our house, and have a few that stopped working after a relatively short time, it was not the LEDs but the electronic part within that malfunctioned, heat seems to the biggest problem with that, as it was the lights that had no airflow, closed to the top fixtures that were the ones that I had to replace.

My son here build his new house, and it is totally “off grid” he installed LEDs when building his house, they are outside as well as in every room, this going to be about 5 years now, and he had only one LED light stop working, it was very soon after it was installed, he did get a replacement for it, he also bought several spares, (for just in case).

These lights are all so called pot lights and have lots of ventilation above the ceiling they sit in.

[Bruce Schoenleber]

Talked with an EE friend of mine. His claim is LEDs are uneffected by cycling, unlike florescents. That is the answer I am going with.
Hoping that I will be within the mean of the MTBF :rolleyes:

Thanks for the replies.

[John K Jordan]

Talked with an EE friend of mine. His claim is LEDs are uneffected by cycling, unlike florescents.

That makes sense when you consider that LEDs are often turned on and off many times a second to control the color, temperature, apparent brightness, or even to create a strobe, usually through pulse width modulation. But maybe ask him if he is referring to the LED emitter itself or the whole package including the driving electronics. My understanding is the LED itself is not affected but cycling the fixture/bulb may be affected since that will cycle the electronics that drive the LED. A "great" way to blow a circuit is with multiple heating/cooling cycles - components can break from the thermal expansion itself, and worse, the current surge in poorly designed circuits.

One possibly interesting tidbit about how heat can affect the life of the LED emitter: a discrete LED (the emitter itself) will only withstand a few milliamps of current before it gets too hot and "pops". Much of the circuitry in a fixture is to control this current. Maybe 30 years ago a friend and I experimented with supercooling a single red LED by immersing it in a dewar of liquid nitrogen (minus 321 deg F) and seeing how much current it would take. We set the dewar on my kitchen floor and cranked up the current driving an LED dunked in the liquid. We never could blow it out but it got so bright it made a blindingly bright circle of light on the ceiling 8' above. It was so bright I suspect looking directly at the LED would have caused eye damage. (and a safety warning: use extreme caution when playing with liquid nitrogen!)

For those of us who have been "playing" with electronics for decades the advance in LED technology is astounding, not only in color and brightness but in the cost of both the emitters and the driving electronics (and integrated modules). It will be fun to watch what the next few years will bring.

JKJ

[JohnC Lucas]

I've had LED fixtures in the ceiling of my shop now for 2 years. They run all day most days. they are the 2 light 4 foot fixtures from Costco. I have had on bulb fail. The bulb on the other side works fine. It's the electronics because I swapped the leads on the bulbs and both bulbs will work when connected to the good side. In the kitchen we have 10 individual lights and they were burning out constantly. These were incandescents. I put LED's in their place and they are still running and that was about 2 years ago.

[Ken Fitzgerald]

When we remodeled our kitchen 4 years ago, I started buying and replacing light bulbs with LEDs. I have had a couple expensive 3-way leds (used in reading lamps) fail. I have had a couple regular leds fail. But gradually as my incandescent bulbs fail, I am going to all leds. The combination of new windows, and going to leds has reduced our electrical use according to our service provider by 17-18%.

[Mike Goetzke]

Even though the life in hours of these LED's is huge if you look at the warranty on the box it no where matches this value. I too remodeled our kitchen a few years ago and replaced the can lights with expensive Sylvania LED's and after about 1-1/2 years they started to fail about one a week. Sylvania replaced them but the replacements didn't even look like the original. Was lucky that I bought these at Menards in that I brought all the bulbs back and they gave a full refund since I had the receipt and one of the boxes.

[Thomas Crawford]

I'm an EE and my company produced IC's for these bulbs for a while. Any early failures would be due to the control circuitry and not the actual LED. The intrinsic reliability of a single LED should be on the order of 100's of years.

[Jim Becker]

I just hung two of those "permanent" LED fixtures in our kitchen. I'm not really worried as the chance is they will out-live me, I suspect.

[Dan Friedrichs]

As others have said, it's the drive electronics (which convert 120VAC to a tightly-regulated DC current) that will likely fail before the LEDs, themselves.

That said, while they make appear to be non-replaceable, it's almost certain that the light fixture manufacturer is just buying the entire LED/driver module from someone else, and you can likely find the entire module as a replacement. I had a lightning strike damage a ceiling fan with an LED fixture, and found the entire LED/driver circuit board for a few dollars on Digikey.

[Curt Harms]

When we remodeled our kitchen 4 years ago, I started buying and replacing light bulbs with LEDs. I have had a couple expensive 3-way leds (used in reading lamps) fail. I have had a couple regular leds fail. But gradually as my incandescent bulbs fail, I am going to all leds. The combination of new windows, and going to leds has reduced our electrical use according to our service provider by 17-18%.

SWMBO has a fondness for 3 way bulbs, it was the only practical way to vary light intensity of a fixture until fairly recently. I'm trying to replace 3 way fixtures with dimmers that replace the switch/lamp holder like this.

https://www.amazon.com/Leviton-6151-Incandescent-Holder-Socket/dp/B001U3Z448/ref=sr_1_3?ie=UTF8&qid=1521123717&sr=8-3&keywords=lamp+dimmer+socket.

I initially bought a different guts-only fixture but it had a short functional life. The Leviton has been good to date. We're running dimmable LED bulbs, the only problem is that the LED bulbs will run at a lower light output than they will start at so turn the knob until the light comes on then reduce the brightness if desired.

[John K Jordan]

...We're running dimmable LED bulbs, the only problem is that the LED bulbs will run at a lower light output than they will start at so turn the knob until the light comes on then reduce the brightness if desired.

I understand that may depend on the electronics in the bulb. Have you tried with others? Apparently the electronics in the bulb itself must sense the output from the dimmer and reinterpret it to drive the LED appropriately, with either current reduction or pulse width modulation. It seems that some dimmers and some bulbs don't communicate or work together well.

A search found several tips. Perhaps this one is useful: http://luxreview.com/article/2016/02/10-things-you-must-know-before-you-dim-led-lamps

JKJ

[Tom Stenzel]

Hello Bruce,

Where I worked we had equipment that had LEDs for a light source. They were powered by a 2400 Hz square wave. So they were turned on to full luminescence and full darkness 2,400 times each second. We had 14 of these and they ran 365.25 (add the leap year, ya know) days a year. After 10 years we had zero failures of the LEDs. Of course you could say it was a cheat, the LEDs had a 50% duty cycle so they actually had been on for only 5 years!

It's safe to say a LED can be turned on and off without any problem. The problems with the lights have been pretty much covered. Heat will kill the LED and the supporting electronics can fizzle out. The screw in LEDs I've bought that died still had working LEDs. Did I take them apart? 'Course I did. I've never seen a LED that needed 25 volts for the forward junction to conduct. But then I haven't paid much attention to the illumination LED market.

If the LEDs don't have a problem with heat dissipation they won't normally just go out. The rated life is reached when the LED dims to about 50% of it's initial output. The small signal LEDs I've seen get too dim to be useful after about 12-15 years continuous use IME. It's a long time.

-Tom

[John K Jordan]

...I've never seen a LED that needed 25 volts for the forward junction to conduct. But then I haven't paid much attention to the illumination LED market.
...

25 volts! That's a new one for me too. Decades ago I bought some expensive LEDs with internal current limiting resisters - with the huge lighting market maybe it is now cheaper to build in some circuitry.

[Curt Harms]

I understand that may depend on the electronics in the bulb. Have you tried with others? Apparently the electronics in the bulb itself must sense the output from the dimmer and reinterpret it to drive the LED appropriately, with either current reduction or pulse width modulation. It seems that some dimmers and some bulbs don't communicate or work together well.

A search found several tips. Perhaps this one is useful: http://luxreview.com/article/2016/02/10-things-you-must-know-before-you-dim-led-lamps

JKJ

I haven't really spent time on it. The dimmer does 'go dimmer' with an incandescent lamp so I'm sure the LED lamp is the issue rather than the dimmer. This application doesn't require low light output and I expect LED lamps will get better at variable output as R&D continues.

[Jerome Stanek]

I haven't really spent time on it. The dimmer does 'go dimmer' with an incandescent lamp so I'm sure the LED lamp is the issue rather than the dimmer. This application doesn't require low light output and I expect LED lamps will get better at variable output as R&D continues.

Are these LED's made for a dimmer.

[Jim Becker]

The LED fixture/lamp power supply must be dimmable and the dimmer switch must be designed to support LED. Older dimmers cannot work properly with LED (and CFL in many cases, too)

[Jason Roehl]

The LED fixture/lamp power supply must be dimmable and the dimmer switch must be designed to support LED. Older dimmers cannot work properly with LED (and CFL in many cases, too)

Some older dimmers do work with dimmable LED bulbs. It can be a crapshoot, though. At work, I have hundreds of LED bulbs on some commercial Lutron dimmers that came out long before LEDs, but they're pricey dimmers to begin with ($130+ per).

[Matt Meiser]

The ~2002 vintage light/fan controllers we have in several rooms in our house dim the Costco LEDs we have. We've had one failure, very early and the replacement in that spot has continued so it was a fluke I'm sure. We converted 90% to LED, the rest being 4' florescents I'll eventually convert and some flood lights on the second story soffits that I can't reach and we rarely turn on anyway. Open fixtures, closed fixtures, ceiling fans, recessed cans, bulbs facing up, down, sideways. Most of ours came from Costco but some from Lowes too, usually the Utilitech ones.

My mom has had some more trouble with those jelly-jar style fixtures. I might buy the plastic globes for her and drill a few holes top and bottom for airflow. The bugs get in anyway so not going to make a difference there.

[Art Mann]

I am coming a little late to the party, but I have been responsible for life testing LEDs for the automotive industry, both interior and exterior. The first thing you have to deal with is what we called infant mortality. It isn't possible to test LEDs, or any other electronic element for that matter, adequately enough to strain out all manufacturing failures. If you install a whole lot of assemblies, some will die quite soon due to this phenomenon. After that, LEDs tend to last a very, very long time. We have switched LED tail and brake lights on and off hundreds of thousands of times and they seldom fail. The same is true of instrument clusters. As John pointed out, LEDs are dimmend with pulse width modulation and they are cycled on and off millions of times in just seconds. The single biggest root cause of failure is heat. All automotive electronic designs are life cycle tested to failure in the lab over a wide variety of environmental conditions. The big killer of LEDs is heat.

In the automotive world, most of the testing is done at 14VDC. There is very little active electronics involved. In many cases, there is only a single current limit resistor and those almost never fail. House hold lighting requires some type of power supply and that is probably where most of the failures will occur, as many have already mentioned.

[Jim Koepke]

LEDs are here to stay. They may take on a few different configurations before it all shakes out.

As to the pulse width driving of lights, it is interesting when driving at night to see the flashing streams of a taillight when my eyes move back and forth to check the side mirrors. Some people do not notice this effect.

On the other side of things when incandescent lamps were on the way out one of my purchases was boxes of 60W bulbs. Hopefully enough were purchased to last a long time in the one place they are needed, our well shed. During the cold months when freezing temps are expected a 60W light in the well house is enough to keep things from freezing.

With LEDs possibly appearing in appliances an old problem will likely go unnoticed. If the door switch on a refrigerator failed stuck on, the system couldn't keep up with the heat produced by an incandescent lamp and the refrigerator wouldn't get cold. It would also run all the time.

jtk

[John K Jordan]

...when incandescent lamps were on the way out one of my purchases was boxes of 60W bulbs. ...

There are a couple of places I want to use incandescants so I bought a stock a few years back from Amazon. I like to purchase 130v bulbs instead of 120v.

[Matt Meiser]

If you install a whole lot of assemblies, some will die quite soon due to this phenomenon. After that, LEDs tend to last a very, very long time. We have switched LED tail and brake lights on and off hundreds of thousands of times and they seldom fail.

Well, not everyone has this down, or at least not 7 years ago. We had an LED taillight fail on a 2011 model. I paid $350 for the assembly despite numerous online reports of the failure. A few months later the manufacturer agreed to pay for any that failed and we got a rebate. Ironically in between we got rear-ended and both assemblies were destroyed. Adjuster and I made a deal to use a "recycled" rear bumper, including all the reverse sensors and such, and trunk lid the shop found at a local junkyard, err recycling operation, in a matching color (one of those tri-coat paints that's hard to match) in trade for replacing both taillight assemblies with new parts. I probably won on both accounts.