LED Household Bulb Longevity

[alphazeta]

...In the short term, now that high-wattage CFLs are on the market, we just don't need LED lighting.

I'll agree with you on that in regards to indoor lighting. But as for outdoor fixtures, I believe there is a place for LEDs right now. However, I've yet to see a decently designed screw in led bulb. The bulbs with the heatsinked fins are probably the closest but are far from optimal in the cost/benefit ratios.

until then...

 

[2xTrinity]

Persistence of vision maxes out at about 20Hz, otherwise you would see ANY light bulb flicker to some extent.

Incandescent lights will not flicker even when fed with a "noisy" signal as there is not enough time for the filament to "cool off" during the off portions of the cycle. Since light from an incandescent light is only dependent on the temperature, that means that it will not flicker. LEDs on the other hands directly convert electric current to visible light, so are extremely sensitive to the power supply. Personally, I don't believe there is any excuse for an LED light to have visible flickering ever. rectifying and filtering a 60Hz AC input into a fairly flat DC output is trivial, the fact that there are LED bulbs on the market that flicker shows that the companies are not even willing to invest $0.05 for some filter capacitors.

However, the persistance of vision, while true when looking directly at something like a TV screen does apply, that doesn't work with light sources -- try using a 20Hz flickering light bulb to look at spinning fan blades, or running water and you'll see what I mean.

 

[winny]

Persistence of vision maxes out at about 20Hz, otherwise you would see ANY light bulb flicker to some extent.

Beg your pardon? I can easily see if a CRT monitor is set to 60 Hz or 70 Hz. If you by light bulbs mean incandescent, the time constant of them are way to long for the 100/120 Hz flicker to be noticeable. If you run LEDs directly on 50/60 Hz (half-wave rectified or two sets in anti-parallel), you do see the flicker.

 

[Scott Packard]

Today, I visited a newly-opened DIY store. They were nicely stocked with Philips LED household bulbs.

I checked the lifetime (it is written on the packing): All colored ones are rated 20,000 h while both the warm white and the cold white ones are rated 6,000 h only. So we're moving away from the 100,000 h claim ? Good !

I believe the Philips figures are pretty close to reality, since it's a reputable company. And I assume they know how to build converters that don't dramatically reduce the LED lifetime. The difference b/w colored and white lifetime should be caused by the phosphors used in white LEDs.

With these lower figures, one of the big selling points of LED bulbs has dissolved. The 6,000 h for white LED bulbs are really not that spectacular, given that white fluorescent bulbs are rated 8,000 h.

I'm glad to see the rated life come down too. What would be interesting to see is the parameters they use to rate it at 6k hours.
I used to think this was not significant, but for the last several years I've only been getting two usable years out of all my CFLs at home, irrespective of brand. I finally found the parameters for CFLs usable lifetime, and they're more liberal than what I use a CFL for in my home (and something all newspapers and tv here don't mention). The "Energy Star" qualified bulb:

- Reach 80% brightness within 3 minutes (Note: using amalgam [mercury], eventual brightness and lifetime increases while initial brightness decreases)
- Average of 10 samples must be greater than 80% rated initial output at 100 hours, and greater than 40% rated initial output at lifetime rated hours. One sample failure is okay, two is okay with a special letter of explanation. Three and you're not meeting the spec. (5 burned base-up and 5 base-down unless otherwise specified on the package.)
- The lifetime duty-cycle is calculated by running 3 hours on, 20 minutes off. I'm probably using it at 1 minute on, 10 minutes off, which is closer to a "rapid cycle". Does a rapid cycle kill a bulb faster? I don't know. My guess is yes.

http://www.energystar.gov/ia/partners/product_specs/program_reqs/cfls_prog_req.pdf
http://apec.fivevision.com/www/UploadFile/117_168.pdf



As long as it's within those parameters it's considered a good light. Unfortunately, it seems that after two years of use in my house I can't put up with their warmup time and output so I have to toss them. I do not have a place in my house where any bulb is used 3 hours on 20 minutes off.

 

[LEDite]

I've got about 1000 hours on the 10W Triple Cree P4 lamp assembly I built for my porch light.

Each Cree is outputting ~ 165 lumens.

Comparing 8 month-old and current photos, it does not seem to have lost any brightness.

Larry Cobb

 

[VanIsleDSM]

That's because a colour mixed LED fixture with many emitters running from lower current would be very expensive.

Designing LED bulbs that each have their own driver isn't very efficient or cost effective.. the best system is something custom using a larger, more efficient, higher voltage power supply than can feed many LEDs in series.

5mm LEDs have no place in general lighting.. they dim much too quickly.. accent lighting yes.. but not illumination...

The high quality surface mount LEDs have data sheets that show you exactly how much light they will generally lose over the course of their life.. and they hold up exponentially better to time than the 5mm guys.

 

[Juctuc]

Yes..you are right! Many emitter is high price. At the moment at least. Maybe we see these in the future.

Of course with one powersuppy and leds in series.

There is few american companyes which are doing these. In fact CREE is making those by themself now( after they by LLF). They are doing colourmix-downlights with CRI of 92

5mm leds are "old story" in lighting. I live in Finland and we have even streetlight made of these 5mm leds. Well...they will learn they lesson sooner or later. There is many problem with 5mm leds in streetlightning, and the lightoutput is not the biggest problem

 

[2xTrinity]

Yes..you are right! Many emitter is high price. At the moment at least. Maybe we see these in the future.

Of course with one powersuppy and leds in series.

There is few american companyes which are doing these. In fact CREE is making those by themself now( after they by LLF). They are doing colourmix-downlights with CRI of 92

5mm leds are "old story" in lighting. I live in Finland and we have even streetlight made of these 5mm leds. Well...they will learn they lesson sooner or later. There is many problem with 5mm leds in streetlightning, and the lightoutput is not the biggest problem

Street lighting as in traffic signals (most signals where I live are 5mm LEDs) or actual illumination? If the latter, that's crazy, as most street lamps where I live are HIDs producing tens of thousands of lumens each. I believe Cree has built some arrays of XR-Es over in Raleigh as a proof of concept that stack up pretty well, but I don't see how 5mm could come even close to the same amount of output.

 

[Juctuc]

I was really talk about streetlights made with 5mm leds. There is models which have few hundreds of leds in one fixture. Leds are in series. 64 leds in one array. So called direct connect. So it dont need any powersupply, because there is 64 leds in series.

Well...there is no hope with those fixtures..They claim that the servicelife is more than 40 000hrs, they also claim that the CRI is more that 90, and many others claims...

They are also making ledbulbs with e27 base. 64 led inside. It will produce to much heat(4w)for the 5mm led, so i dont think that will last for a 3-4 months without dimming about 50% from original. There is no heatsink or anything..well..hard to use heatsink with 5mm led anyway

 

[spencer]

Re: wow gold6

quote of spammer deleted. Please don't quote spam - Empath

Get this spammer banned,

 

[RODALCO]

There is a trial here in Auckland with LED street lights in certain selected streets.
I will take some photo's and see what happens regarding the reliabilty issues.
I haven't seen them myself yet, just interesting how they compare to the Hi pressure sodiums.

 

[Troubleshooter]

Those Christmas lights are a series of LEDs that light during ONE half cycle of the 60 Hz power. They are dark during the other half cycle, so they flicker very much, especially when you move your eyes. Some of the strings with more than 35 lights light half of the lights on one half-cycle and half of them on the other half cycle.

This knowledge allowed me to make some very simple circuits to flash and/or sequence the lights, using SCRs instead of triacs. I can even make the half strings come on at different times.

I also built some circuits to allow them to be used with dimmers and old sequencers and color organs. I simply hooked two 7 watt nightlight bulbs in series, and then used a cube tap to plug this into the device along with the LED strings.

 

[GregP507]

Re: wow gold6

I think one of the greatest engineering challenges in recent years has been how to shorten the life of an LED emitter. Normally, they are expected to last for decades, but as a business model, longevity is a money-loser for a manufacturer. At the outset of a new technology, the high price and steady adoption-rate can provide a healthy revenue-stream, but as competition cuts in, and the consumer saturation-point comes nearer, those revenue-streams begin to dry up.

When the incandescent light-bulb was introduced by Thomas Edison, the vacuum-tube design was nearly fool-proof. Nearly everyone has heard of the electric light bulb that has been burning in a California firehouse for over a century, and in fact, the same technology used in radio vacuum tubes means that quite often those tubes are still working perfectly after 80-90 years. However, over the ensuing decades, the longevity of incandescent bulbs has decreased over the years. Fluorescent-tube technology is nearly as old, and these as well, have seen their longevity decrease substantially as well. I'm not sure exactly how this planned-obsolescence was achieved, but I suspect it has something to do with allowing the slow infusion of oxygen through the base of the bulb.

Since LED lighting is a robust solid-state technology, I suspect the weak-points will have to be engineered in the driver circuitry, which in itself is quite robust. Possible weak-points are electrolytic capacitors and thermal heat dissipation. If the capacitors are allowed to degrade, and/or the temperature of the circuit is high, it may lead to earlier failure. i suspect it won't be long before we see LED lighting with a service-life similar to that of incandescent or CFL bulbs.

 

[CoveAxe]

Re: wow gold6

I think one of the greatest engineering challenges in recent years has been how to shorten the life of an LED emitter.

No, it isn't. Lifetime and reliability has only increased for emitters. No one is looking at making them fail faster. So far improvements in efficiency/brightness and color, and premature failure of the electronics has been the main driver of replacing LEDs. This will continue to be the trend for a while. Even if LED bulbs became immortal tomorrow, there are so many disposable items that use LEDs that there will be no shortage of markets to sell LEDs to. This is an overblown "problem".

Nearly everyone has heard of the electric light bulb that has been burning in a California firehouse for over a century, and in fact, the same technology used in radio vacuum tubes means that quite often those tubes are still working perfectly after 80-90 years

It has lasted so long because it's not a practical design: early bulbs had very thick filaments. This is good for lifetime, but bad for lighting. It's so dim you can barely read from it in a dark room. I believe they also reduce the voltage going into the bulb itself to increase lifetime as well since they say it has its own power supply. It's about the same brightness as a 4W nightlight, even though it is rated at 30 or 60W. If you reduce the voltage enough on a modern incandescent, it will easily last over a century as well, but again, it will be so dim that it's practically has disappeared. Think about it: If all of these bulbs from 80-90 years ago regularly lasted so long, we should be seeing them everywhere in old buildings. Where are they?

I could also make a very efficient and bright incandescent, but it will only last a few seconds. The current ~1000 hour lifetime for general use bulbs is a general compromise between extremely long life but very dim, and very short life but very bright.

Fluorescent-tube technology is nearly as old, and these as well, have seen their longevity decrease substantially as well.

Do you have any proof of this? I can easily find fluorescent tubes with 75,000 hour average lifetimes. What tubes from the 50s/60s/70s had average lifetimes longer than this?

i suspect it won't be long before we see LED lighting with a service-life similar to that of incandescent or CFL bulbs.

Not exactly. Even as cheap as CFLs have been made, they are still regularly rated at several times longer lifetimes than the incandescents they are replacing. LED has to be at least as good as CFL to replace it. Now since consumers are generally unwilling to pay ~$10 for an LED bulb compared to an incandescent or CFL, bulb lifetime has been decreasing as things like heatsinks and better design have been sacrificed to reduce cost. This will continue until some equilibrium has been reached and then improvements to LEDs will yield better efficiency/lower heat and longer life, or the market will move away from bulbs and go to luminaires with built-in LEDs, which should have very good lifetimes. In either case, there is no way LED is going to have 1000 hour lifetimes: people would just move back to CFL.

 

[Anders Hoveland]

Re: wow gold6

IMO, LEDs at present are not an economical or sane choice as a routine replacement for tungsten bulbs despite Euro regs and greenies in CA trying to push alternatives along.

LEDs can make sense in certain particular commercial applications, where the lamps are left on constantly. Some large department stores have hundreds of lamps, and this can add up to substantial cost savings.

What I have been noticing about many stores is that the availability of LED has resulted in the stores relying more on spotlighting to supplement the traditional fluorescent lighting, to the point that it is now almost a 40% LED to 60% fluorescent ratio.

I agree though, LEDs generally do not make so much economic sense in people's houses, in most instances.

When the incandescent light-bulb was introduced by Thomas Edison, the vacuum-tube design was nearly fool-proof. Nearly everyone has heard of the electric light bulb that has been burning in a California firehouse for over a century, and in fact, the same technology used in radio vacuum tubes means that quite often those tubes are still working perfectly after 80-90 years. However, over the ensuing decades, the longevity of incandescent bulbs has decreased over the years. Fluorescent-tube technology is nearly as old, and these as well, have seen their longevity decrease substantially as well. I'm not sure exactly how this planned-obsolescence was achieved, but I suspect it has something to do with allowing the slow infusion of oxygen through the base of the bulb.

There is some level of truth to this, but another factor is there is often a tradeoff between lifespan and efficiency. In an incandescent bulb, a slightly longer thicker filament will last much longer but glow at a slightly lower color temperature, substantially cutting into efficiency. It can be cheaper to have to frequently change bulbs than paying to use 20% more electricity, in many instances.

With metal halide lamps, despite the bulb having a long lifespan, it is often recommended to change them out every 1 to 2 years to save money, because the lumen maintenance of metal halide is so poor and businesses often resort to using higher wattages than they need because many of their 6-year-old bulbs are so dim.

Some cheap fluorescent tubes can have up to 30% lumen depreciation after only 2000 hours of use, due to degradation of the phosphor coating. Obviously in this type of situation, just because the lamp is still working does not necessarily mean it should not be replaced.

 

[lunas]

Re: wow gold6

You are forgetting the 22 year life span and conversely the life of a cfl is rated in 6 hours of burn time per day not 24/7

24/7 for 22 years would be 192,720 hours
6 hours per day is 48,180 hours of burn over 22 years

The biggest reason for premature burnout is prolonged heating with cool down periods they last a lot longer.

 

[Anders Hoveland]

Re: wow gold6

The biggest reason for premature burnout is prolonged heating with cool down periods they last a lot longer.

They also do not last so long if they are frequently turned on and off. I am sure there is an optimal middle ground in between. (talking about CFL here just so no one gets confused)

 

[lunas]

Re: wow gold6

Yes and if you give it time between the on and off cycle it improves longevity greatly.

This is why i have a 1 hour delay on my aquarium between on and off i do 3 on cycles 5 hours on 1 hour off then 5 hours on 1 off 4 hours on then 8 hours off.

 

[SemiMan]

Re: wow gold6

You are forgetting the 22 year life span and conversely the life of a cfl is rated in 6 hours of burn time per day not 24/7

24/7 for 22 years would be 192,720 hours
6 hours per day is 48,180 hours of burn over 22 years

The biggest reason for premature burnout is prolonged heating with cool down periods they last a lot longer.

Are we still talking CFL?

Maximum life for a CFL (pretty much any fluorescent) is continuous burning. This will result in the maximum "on" time possible.

 

[SemiMan]

Re: wow gold6

I think one of the greatest engineering challenges in recent years has been how to shorten the life of an LED emitter. Normally, they are expected to last for decades, but as a business model, longevity is a money-loser for a manufacturer. At the outset of a new technology, the high price and steady adoption-rate can provide a healthy revenue-stream, but as competition cuts in, and the consumer saturation-point comes nearer, those revenue-streams begin to dry up.

When the incandescent light-bulb was introduced by Thomas Edison, the vacuum-tube design was nearly fool-proof. Nearly everyone has heard of the electric light bulb that has been burning in a California firehouse for over a century, and in fact, the same technology used in radio vacuum tubes means that quite often those tubes are still working perfectly after 80-90 years. However, over the ensuing decades, the longevity of incandescent bulbs has decreased over the years. Fluorescent-tube technology is nearly as old, and these as well, have seen their longevity decrease substantially as well. I'm not sure exactly how this planned-obsolescence was achieved, but I suspect it has something to do with allowing the slow infusion of oxygen through the base of the bulb.

Since LED lighting is a robust solid-state technology, I suspect the weak-points will have to be engineered in the driver circuitry, which in itself is quite robust. Possible weak-points are electrolytic capacitors and thermal heat dissipation. If the capacitors are allowed to degrade, and/or the temperature of the circuit is high, it may lead to earlier failure. i suspect it won't be long before we see LED lighting with a service-life similar to that of incandescent or CFL bulbs.

I once came across a Wikipedia page that said "This page deleted, thereby greatly increasing its accuracy". The same would apply to this post.

There is absolutely nothing that is factual in what you wrote. Literally nothing.