A LED lifetime tester

[capwizard]

2 LG and 2 Samsung LED TVs bring-in (different time over this year), said LED screen blinking, TESTed all day worked, I charged the full repair bill by doing nothing, BUT I gave them one year warranty.( those tvs will come back eventually) How can you do? I need build A LED lifetime tester, (A equipment will show in the future of which LEDs will not work.)!, can anyone HELP!!!..

PS: is the LED has ability to recover itself (the internal Diode)?

FD116 LED Lamp lifetime tester
http://www.ledlightsmanufacturemachi...me-tester.html

Lifetime of White LEDs
https://cdn.badcaps-static.com/pdfs/...b8641a8538.pdf

evaluating the Lifetime Behavior of LeD Systems
https://cdn.badcaps-static.com/pdfs/...c7cb64213a.pdf

[mariushm]

Re: A LED lifetime tester

The bond wires that connect the led die to the metal base of the led (the led terminals) could have desoldered themselves.

At some point, with hot/cold cycles, like for example when the led gets hot enough, the bond wires could stretch or move slightly and break the connection, so the led turns off, which means a whole strip of leds could be turned off.

In your particular case, maybe as the TV was brought to you, it was moved from vertical to horizontal position, or something like that, all that shaking and moving that happened may have adjusted those bond wires slightly to make a better connection so you no longer experienced the problem during your test.

Or, maybe you repeated the test with the back case open, which could mean the whole led assembly was better cooled which avoided problems the user experienced.

Or maybe you adjusted the brightness levels changing the environment.

In your place, maybe during test i would have knocked on the led strips with my fingers just to make sure the leds don't have some bond wire issues. which would pop up after returning the TV to customer.

[capwizard]

Re: A LED lifetime tester

Thank you very much, This is A very good idea, I never thought about that, "i would have knocked on the led strips with my fingers just to make sure the leds don't have some bond wire issues."

[Longbow]

Re: A LED lifetime tester

There is no such thing as an LED lifetime tester or one for any other electronic part. Look at the specs and you will see an MTBF number. That means MEAN TIME. Failure of electronic parts is a matter of statistics and cannot be applied to one particular device. (See Heisenberg's uncertainty principle.) Naturally, the obvious cause of problems can be avoided, such as over-current situations. My guess is that the backlight LED's are driven to their maximum possible brightness, which translates into the minimum likely lifetime. Once these devices aquire a track record, service techs will have a better idea of what to expect.

[capwizard]

Re: A LED lifetime tester

LEDs typical lifetimes are 25,000 to 100,000 hours. Now less than one year old TV we had served was LEDs pre-early failure problems, more like Bathtub curve to me, Traditional electronics reliability engineering and failure prediction methodology (FPM) has in its foundation the concept of the life cycle bathtub curve. I will try non traditional method to accomplish this task, use 20kHz to 100MHZ switch power supply for voltage and adjustable high pulse source for current in short period of time let LED be aged to predict LED pre-early failure problem and life expectancy....Will this works?..................

[Longbow]

Re: A LED lifetime tester

Here we have some light bedtime reading on the subject of LED failure modes. Of course without the proper lab conditions, the end user can only speculate but there are still some good points. LED's are not the bullet-proof unbelievable devices most people think they are.

http://www.digikey.com/en/articles/t...ng-led-failure

Another writer has an excellent point about the liklihood of LED system failures based on the quantity of electronics that support them. I like his approach.
http://www.electronics-eetimes.com/n...t-led-lighting

For the most part, in a flat screen, you are dealing with heat and nearby high voltage, 2 things that accelerate failure in most electrical parts. The other 2 big ones are moisture and vibration. A microwave covers all 4 of those, for example. It is also important to keep in mind that LED light output is a function of time under the best of conditions. In the next few years we'll see how they stack up against other technologies like OLED, and atomic enzyme proliferation (AEP). My opinion is still that LED's make great pilot lamps.

[Agent24]

Re: A LED lifetime tester

I call it cheap, overdriven LEDs.

But you are right too. It's glaringly obvious from my own experience that incandescent bulbs and fluorescent lamps outlast CFL and LED lights by miles.

[keeney123]

Re: A LED lifetime tester

My experience with LEDs is they do not like to be turned off and on quickly. Meaning every few minutes like a motion detector would do. This was about 8 years ago when they were just coming out with the high power LED. I am wondering the frequency of the PWM can affect them in a negative way? Perhaps if there was a smoothing cap in parallel it would eliminate this?

[Longbow]

Re: A LED lifetime tester

incandescent bulbs and fluorescent lamps outlast CFL and LED lights by miles.

An interesting comment, since CFL are in fact fluorescent tubes twisted on top of each other. The thing that shortens the life of a CFL is that the lamp heat is fed directly into the switching supply, causing it to broil. On the other hand, folding the tubes back upon themselves increases the light output over a single linear tube, everything else being equal. So in the case of CFL, heat is the killer.

[Longbow]

Re: A LED lifetime tester

My experience with LEDs is they do not like to be turned off and on quickly. Meaning every few minutes like a motion detector would do. This was about 8 years ago when they were just coming out with the high power LED. I am wondering the frequency of the PWM can affect them in a negative way? Perhaps if there was a smoothing cap in parallel it would eliminate this?

Excellent points. It seems that LED driver designers have not thought enough about LED physics to realize that they do have the equivalent of a "filament" found in both inandescent lamps and fluorescent tubes. The longest lasting fluorescent tube arrangement uses a programmed start ballast which warms up the filaments on the tube slowly before hitting it with the high voltage. Some people can't tolerate the 3 second wait, so they select the instant start ballast instead. I think we will see something similar come out for the high power LED's.

Now on to a topic a little removed...LED drivers themselves. Next gen LED general lighting tubes will contain only LED's and have a separate LED ballast. I believe that the separate ballasts will be developed along the same lines as the older FL tubes and eventually supply a very clean d.c. signal to the tubes. That is not the case with the combo units available now that have rudimentary switching supplies inside the tubes. Those internal supplies produce a pretty ugly drive signal with plenty of RFI which exceeds FCC guidelines. The result is optical flickering and RF noise in your electronic products. Nobody is sure yet, but LED persistence is near or at 0. A good d.c. supply is mandatory.

Early LED driver IC's use PWM as a dimming method, but it is definitely wrong to use a simple PWM drive signal to limit average LED current. More sophisticated drivers actually provide a constant current and yes, we would like to see the thing ramp up and down slowly to save the tiny gold bonding wires inside. Some people equate PWM drive method with a switchcing power supply of any kind. They are not the same thing. An LED wants a stable constant current and in my personal opinion the lifetimes will be much longer with gentle turn-on and turn-off times.

[Agent24]

Re: A LED lifetime tester

Indeed, but standard straight lamps like a T6 seem to last forever. Simpler design probably, and like you say, lower heat concentration.

I got some free CFLs in a promotion, some cheapies. About half failed faster than incandescent. The others maybe lasted about the same length of time. One might have lasted longer but I think that was a Philips I bought myself.

I've gone back to incandescent, although this time I'm using Halogens which are supposed to be slightly more efficient than standard bulbs. They don't have to warm up, they cost peanuts, and they seem to last longer. I'm not interested in buying LED lamps unless I have to.

[Trevtron]

Re: A LED lifetime tester

Some interesting points raised here.
Re. PIR switching reducing lifetime- perhaps it is the driver circuit rather than the actual LED's that are failing. I have pulled apart various LED 'bulbs', and invariably it's the driver that has failed where excessive has not been a factor.
Once upon a time, LED's were deliberately pulsed to increase perceived brightness (ever held an old LED calculator near an AM radio?) and this didn't seem to kill them.
But regarding LED failures in recent TV's, I am sure the makers will have done their homework. The LED's will be being pushed close to their limits with the knowledge that, statistically, some will fail. But the cost saving on a large production run will hopefully (for them) outweigh that of an occasional warranty claim.
I was never a fan of CFL's, they never lived up to the hype IMO.
I have some LED bulbs around the house, mainly those that are on for longer periods of time, and am happy with them. Except the 'filament' types that flicker at 50Hz!

[Agent24]

Re: A LED lifetime tester

2 LEDs burnt up in the first LED bulb I had experience with (an unknown brand one - Orbit Lighting or something) - someone else's bedside lamp who complained about the heat output from incandescent.

It failed again after repair with another 2 LEDs burning up, this time I think the PCB may have been trashed from the burning (will attempt repair again though)

I tentatively bought a Panasonic on sale as a replacement. I must admit I don't have high expectations.

[RJARRRPCGP]

Re: A LED lifetime tester

I've gone back to incandescent, although this time I'm using Halogens which are supposed to be slightly more efficient than standard bulbs. They don't have to warm up, they cost peanuts, and they seem to last longer.

I had bad luck with halogens! I saw some halogens in the past and they seem to blow often! I got one in 2014, IIRC and that was out in just a few months!

And I think I saw another halogen do the same kind of thing, IIRC... They seemed to be no better than a 1920s incandescent! (Those with the meek 1,000 hour rating) Watch what manufacturer you get!

[capwizard]

Re: A LED lifetime tester

Once upon a time, there have red apples and green apples, the seven Dwarfs need the blue apples to make A Snow white, the witch finally use gallium nitride made a blue poison apple, Dwarfs use blue apples make a lots of Snow whites, but Snow whites lifespan are too short.

The witch help Dwarfs improved new recipes base on blue apples + blue apples (GaN-on-GaN) they hope Snow whites 2.0 ( LED 2.0 ) will last forever and ever. If not they will replace with Laser Snow whites (laser diodes).. ...................... ................... .................THE END...


Reference:

gallium nitride
https://en.wikipedia.org/wiki/Gallium_nitride

Nobel Prize For Blue LED Invention - Physics Nobel Prize Goes To Scientists Who Perfected LED Light
https://www.youtube.com/watch?v=TbUmURyBPO0

Soraa's LEDs are made with an active material layer of gallium nitride and a gallium nitride substrate. Having a single material leads to LEDs that can take more current and thus produce more light on a package of a given size. It also means that there's less wasted heat, which can degrade the life of LED lighting.
https://www.cnet.com/news/startup-so...-with-led-2-0/

GaN-on-GaN technology is ‘LED 2.0' says blue LED inventor
http://luxreview.com/article/2013/11...e-led-inventor
Nakamura described GaN on GaN as ‘LED 2.0', explaining that it has the advantage of having very few crystal defects compared to normal LEDs. ‘We can operate these at high current, high temperature, and reliability is very good,' he said.

GaN on silicon or SiC?
http://www.edn.com/design/wireless-n...ilicon-or-SiC-

Substrates for GaN Technology


Low-cost laser on the way
http://taf.fi/en/2012/09/12/next-generation-of-leds/

Shuji Nakamura: GaN-on-GaN Will Become the Next Generation LED Technology
http://www.ledinside.com/interview/2...led_technology

Laser diodes are lighting's future, says Nobel Prize winner Nakamura
http://luxreview.com/article/2014/11...inner-nakamura
There are already a couple of laser diode-equipped cars in existence. The BMW i8, which launched this summer, is the first car to use laser headlights developed by Osram....the laser diode's efficiency is ten times higher than that of the LED headlamp.

[capwizard]

Re: A LED lifetime tester

"An LED wants a stable constant current and in my personal opinion the lifetimes will be much longer with gentle turn-on and turn-off times."

I agree with you 100%, because the way they drive the LED lights are strobing, oscillating to save the energy, the LED lights are constant blinking for bird's eye.

Birds need at least 100 frames per second to see TV images as a moving picture. Having a bird in the same room as a TV with a lower frame rate than that will be very stressful for the bird.

[jiroy]

Re: A LED lifetime tester

Away from LED's , i think this was a pure case of weary outlets and /or cheap Chinese Electrical Multi-Outlets , or even an abnormal load of different electrical units on one source . So yes , they will come back to you if i was right , because the problem is still out there . In case they didn't return , then you can assume you were considered a failure repairer , lol . Once i've sold a brand new Computer and the customer showed up three days later , complaining from hang ups and sudden rebooting . Knowing it was brand new , i guessed there was something wrong at his house , and like expected , it was one hell of a sight ..Some 8 to nine i don't know what electrical and electronic equipements were hanging from one outlet source .
I told the guy , fix this this and you won't see me again ... And i never saw him again .

[michael dranfield]

Re: A LED lifetime tester

I had bad luck with halogens! I saw some halogens in the past and they seem to blow often! I got one in 2014, IIRC and that was out in just a few months!

And I think I saw another halogen do the same kind of thing, IIRC... They seemed to be no better than a 1920s incandescent! (Those with the meek 1,000 hour rating) Watch what manufacturer you get!

The problem with incandescent bulbs in the filament resistance is much lower when cold, if you were to make a circuit with a triac that switches the bulb when the mains AC cycle is at zero your bulb would last much longer ,also modern bulbs are no longer gas filled, they are evacuated to save on cost.
A war time coiled carbon filament would last you much longer !

[Longbow]

Re: A LED lifetime tester

I was never a fan of CFL's, they never lived up to the hype IMO.
I have some LED bulbs around the house, mainly those that are on for longer periods of time, and am happy with them. Except the 'filament' types that flicker at 50Hz!

Hard to be excited about a product that has an inherently short lifespan. As you noted, the problem is heat from the tube in close proximity with the electronics. The whole point of CFL's is to get a lot of light output in a small package and they do that job well. Plus, you don't have to put up with the wimpy pale yellow color of low wattage incandescents. Folding the tubes back upon themselves causes the UV inside one part of the tube to excite the phosphor on several other sections of the tube for increased tube length/output efficiency. If you separated the power supply from the CFL tube itself, you would see a big increase in lifespan because the supply would not have to run hot.

[Longbow]

Re: A LED lifetime tester

Indeed, but standard straight lamps like a T6 seem to last forever. Simpler design probably, and like you say, lower heat concentration.

By now we actually have a mature fluorescent tube technology and electronic ballast track record. Running the lamps at 30KHz or greater eliminates flicker because of long phosphor persistence. To date, nobody knows much about LED persistence (if it even exists) and there are no specs on LED tube flicker.

Every once in a while I have to replace an electronic ballast but those things are built to last. Tubes usually fail due to filament burn out so they can't start. Interestingly, the smaller the FL tube diameter, the higher the length to output efficiency, but they run hotter because of smaller surface area. T5 is currently used for shielded accent lighting or grow lamps because they are too bright for direct lighting. I think the T8 will be around for quite a while because it is a good compromise between size and output..and because there are so many T8 fixtures in use. T8 LED tubes will be a good product as soon as the ballasts are provided as separate units and are cost effective.