Long life LED products are a bad idea?

Is it worth buying any LED light product with a lifetime greater than 50000 hrs? (there's 9000 hrs in a year).
Surely, in five years time, the efficiency of leds will have approximately doubled, so surely a LED light product that lasts longer than 5 years is a waste of time?....as it would be better to just buy a new led light product in say 5 years time with the more efficient leds in it?...or, say if it was a LED desk lamp, to replace the led lightbulb with one with new leds in it.

In other words, if someone designed a LED desk lamp in which the LEDs lasted for 25 years, and were "Ingrained" into the product (ie there is no led bulb that can be removed and replaced) wouldn't that be a waste of time?..because in 5 years time it would be very inefficient compared to the latest led products on sale.

Page 2 shows predicted efficiency increase of leds...
**broken link removed**

I have it on good authority that the World will come to an end tomorrow so I went out to buy my LEDs today.

If we all held off buying because of predicted future developments, they would be no funding for those developments to take place.

Brian.

I will not pay a fortune today for LED lighting for my home so I will wait until the price becomes more reasonable.
My compact fluorescent light bulbs are efficient, were cheap and last for many years.

I use LEDs in my solar garden lights. Some of my solar garden lights were free from my electrical utility and others cost a whole dollar each.

Thanks, but rather, I am speaking of manufacturers, -is it worth them making LED lamps with huge heatsinks so that they last 35 years?...after all, in 5 years, those leds may be half as efficient as the new LEDs available at the 5 year point of the product life with the customer.

i think it is not going to be done. MTBF assumes no design or process faults under any combination of environmental conditions.

However high quality reputable designers will limit the max temperature rise of the case and thus junction to 85'C using extraordinary methods. For the rest of the world, price competition will cause many escapes.

On my trip back thru Michigan in Jan, I saw many Interstate highway LED failures... About 1%. For car headlights in last couple years , I see slightly less failure rate maybe 1 in 300 cars.with at least one bad LED.

the only good number is a history longer than projected, but a bad design is easy to measure by Tj rise.and HALT testing.

The MTBF theory , in practise is verified with bulk testing in short term and measuring the rate for a required confidence level like 80 or 90%. But this is a bad test , say if they all have a mid term defect that shears the wirebonds after so many thermal stress cycles. Like bending a copper wire so many times till they realize how to improve it by stress reduction.

Price erosion is projected at 50% every 3 years the only way to possibly last 35 years is low duty cycle and very low temp rise and underutilize lumen capacity.

You have to start from where you are. A couple of years ago I changed my 12 50 W halogen down lighters for Philips LEDs. So far they have paid for them selves because of the massive energy saving and with a projected life of a further 15 years. So jump forward 10 years, when I decide to replace them because they are getting dim. How much money will I have to invest to get a worth while return over a reasonable period? Bearing in mind that the cost of energy is likely to escalate.
I think the main problem is that, certainly with down lighters, is that the industry standard for the housing is based on a halogen bulb, and cost peanuts to make. I think that LED downlighters should be sold as complete units, so there would be ample room for cooling and fire proofing. Cheapo downlighters are not fire proof, so need a fire hood (£10). The whole aesthetics of light fittings could and eventually will be changed by the LEDs being built into them.
Frank

According to the guideline bellow, the expected MTBF of a single LED component submitted to standard environmental conditions is something near or even superior to the value that you gave above:

**broken link removed**

This is really a considerable lifecycle, however we must keep in mind that inside a product, there are several diodes in series, so that the overall MTBF is therefore reduced. I’m not aware of the calculation involved, but I presume that in an array of n elements, the MTBF would be reduced to the n-th part.

**broken link removed**

on page 5 of the above , you can see that cree use a heat pipe to cool the led in a 120w low-bay , LED luminaire design.

Do you think its worth using heat pipes, with their extra expense, just to get a few more thousand hours of lifetime?. I mean, why not accept a reduced lifetime, and just replace it after say 3 years, -after all, in 3 years time, the latest LEDs of the day may be twice as efficient as the LEDs of today.

These LEDs are particularly designed to work at pulsed mode, on which a big amount of energy is driven to the device, as well a big amount of power is expected to flow on its segment. I don’t consider a good action to reduce the ability of dissipate the instantaneous energy of a component with low mass, due this would quicky raise the substrate temperature to critical levels.

Thing is, the cost of that low bay may well be vanishingly small compared to the cost of getting to the thing to replace it.....

Consider that if a warehouse has to unload and remove a mess of pallet racking and then build a scaffold or hire in a man lift you could be looking at a few DAYS downtime to sort out a faulty fitting, now ask what that would cost Amazon.
Same thing applies in ports and airports for example.
If paying an extra few hundred quid gets another few years out of the thing, that is sometimes money very well spent.

Always when costing products you have to look at which market you are playing in, domestic retrofit is very different from domestic install contracting, is very different from industrial, broadcast, marine, military......

Heat pipes are an entirely sane solution in some applications in some markets.

Regards, Dan.

thanks

Heat pipes are an entirely sane solution in some applications in some markets.

Click to expand...

I agree, but also I believe that when you have heat pipes on leds, then the lifetime determining factor becomes the lifetime of the offline smps that supplies it all...there is surely no use in having a heat-piped LED that lasts 30 years if your power supply only lasts 15 years....I know this wasn't your point, and I am sure we agree on this, its just that shouldn't leds with heatpipes have spec'd up smps's to go with them?....otherwise the heat pipe is wasted money?

Possibly, but only possibly.

Consider that sometimes you are trying to build a very high brightness source, which usually means a LOT of LED in a small space, sometimes you need the source remote from the heatsinks, could be that even with the heat pipe I am only going to get 5 years, especially if ambient is 45 degrees C or so.

There are places (Projection sources, theatrical spot lights) where power density matters to the ability to form a useful beam with optics having a sane size, heat pipes are useful here for moving heat efficiently to somewhere you can accommodate the required heatsinks they are not in themselves a cooling solution.

For really painful cooling problems, look at high power laser diodes, all the thermal issues of LEDs, plus you need to maintain temperature for wavelength tuning, plus the things can be running hundreds of watts.

Like everything heatpipes are a tool in the engineering toolbox, and there is a class of problems they really help with, end of the day MBTF is MBTF.

Regards, Dan.

We need to analyze the question from the perspective of what operation will be chosen, continuous or pulsed. In order to reduce the power consumption of the "lamp" it is appropriate to work the LED array in pulsed mode to make it look brighter with a lower average current.

On this case, the main thermal requirement is not only to efficiently spread the heat over the whole heatsink surface, but also to avoid a huge thermal gradient inside the junction, therefore it is also necessary to increase the thickness of the heatsink right below.

In other words, the decision to reduce the size of the heatsink as you pointed must be technically justified.

Thermal design is the important aspect overlooked in LED Luminaire design.

I've tried selling the concept of pulsed indicators to reduce power, but perceptions of some users wont accept the irritation , besides losses are quadratic with peak to avg ratio from I²ESR so gains in perception are offset by higher losses, unless very slow.
.
Korea's ganged micro heatpipes seem to be the best, with fastest heat velocity and passive air flow . Others for streetlights use silver epoxy and vortex air flow. Each method is subject to patents.

Mil Hdbk883 was dated 1991 before power LEDs were invented and data must be collected after true life time experience not mean failure rates in mass production in short terms.

Cree's design is good and anything less will certainly compromise brand quality and reputation, increase service calls to replace.

power chips are 0.3$/W going towards 0.1$/W in wholesale while highbay basic luminaire engines are 10x this rate in volume. It only makes sense to pay less for quality, if it is small volume and low cost penalty to replace a disposable item. But if it were 1,000 street lamps, you better pay more attention to warranty, risks and due diligence on proven reliability.

It is important to remember that estimates for MTBF are irrelevant for LEDs because the variables for design and process flaws which escape routine testing are high risks. This explains the obvious high failure rates in car head lamps or DTR lights that can only improve by design changes, like thermal design and ESD protection.

https://uk.farnell.com/cci/00c93460101/heat-pipe-200mm-6mm-dia/dp/1373309

..these heat pipes cost only £10 so I wonder why all led lamps above say 10W aren't using heat pipes?
(obviously as discussed they wouldn't be needed if the leds are spread over a wide area etc).

With "lights in difficult places" or "expensive to replace" I would have thought the answer was to heatpipe the leds and then use a power spply mounted near the floor, so its easy to replace, and use high voltage led bank so that i2r losses are low, or have a long life dcdc converter near the luminaire so that you can convert the HVDC to the whatever the leds want in current.

as you know, offline converters can suffer mains transient failures, so the isolation stage would neeed to be near ground level for easy acces, do you agree?

For lights in difficult access areas, the solution ought to be lowering the power density with larger radiators and increasing convection air flow with consideration to animal nests. Gains in thermal conductance can made cheaper with air velocity of the surface without the need for high volume flow rate or more aluminum mass but dust collection is a concern.

The most intelligent low cost bulb is the type with remote phosphor inside the bulb. This reduces density of spot light (eye glare) for visible types.

That being said, I still prefer tri-phosphor tubes with indirect lighting and similar efficacy to some LED tubes at 90 LPW with high CRI quality 88 and no hum, flicker or cascade tube failure issues (independant 1 of 4 or 8 ballast outputs) with 50kh proven MTBF for long term per day. This why Walmart uses this solution in new stores.

Line transient protection is a must, which can be achieved in many ways.

Heat pipe are a solution to moving heat, but you still need a heatsink on the end....

If the pipe costs £10, then my factory gate has just gone up £30 or so, so you try not to need them if there is any sane way to avoid the things.

Agreed that there is little wrong with tri phosphor tubes in a good quality HF fitting, LED is not the answer to every job, sometimes florescent and sometimes MH is the better solution, once in a blue moon, tungsten halogen is the appropriate answer.

Engineering answers are always compromises between conflicting requirements (And plenty of times, the familiar technology winds simply because it is familiar).

Specifiers have long memories, and the early LED lamps seriously over promised and under delivered, if I was specifying a streetlighting or other large project I would be looking for some very serious guarantees on reliability with sufficient penalty clauses to cover the cost of gaining access to replace failed units, there are plenty of horror stories about the early LED lamps.

Regards, Dan.

The following mains transient protectors seem to suggest that transient protection is mostly for "outdoor" led lighting situations
http://www.littelfuse.com/~/media/e...ghting_surge_protection_modules_flyer.pdf.pdf

surely mains transients are a threat to any mains connected LED driver? (indoor or outdoor).....because one can never give a valid lifetime figure for any mains connected led driver, since one doesn't know how much mains transient activity that led driver will see where it is installed.....I mean, even a mains power supply with a lifetime of "20 years", may only last three years if it has bad luck with mains transients?
Also, its interesting that littelfuse's thermally protected varistor based surge protectors don't offer any lifetime guarantee at all in any of their literature...so surely, it is pointless to put heatpipes on leds driven by a mains power supply, because the lifetime will be determined by the mains transients that the power supply will receive.?...and thence the product , even with a heatpipe, may not last long at all, if mains transient activity blows up the power supply after a couple of years

PSU
http://amtex.com.au/power_pdf/HCP650-1000.pdf
..The above psu has an MTBF of 160000 hrs (16 years+) but surely if in a place where there is high mains transient activity then lifetime will be much less?

Power equipments works typically at higher temperature, which is the most significant factor in determining the MTBF, and have a lot of components wich increase the likelihood for individual failure, and even using over rated components, 160 KHours ( ~18 years ) sounds extremely optimistic, even if not subjected to harder electric conditions.

Thanks, regarding temperature, 'lytic caps degrade with time at temperature, but diodes, fets etc do not?....I mean, a fet can go on forever at a temperatre (junction) of 100 degC?