AVS › AVS Forum › Display Devices › Display Calibration › LED longterm performance?
New Posts  All Forums:Forum Nav:

LED longterm performance?

post #1 of 43
9/22/10 at 1:50pm
Thread Starter 
I have long had the theory that LED lighting techniques would have aging issues in HDTVs. (started with noticing that Local Dimming LCDs will have uneven wear problems thanks to how they operate in a manner similar to plasma, by dimming part of the grid it would cause uneven wear and the backlight would loose uniformity over time) and later I noticed that edge lit sets will also have issues since LEDs dim with use instead of burning out like CCFL.

Then I saw this post by UMR in reference to LED LCDs.

Quote:
Originally Posted by umr View Post

The uniformity of the products black level degrade significantly. They either look like milk is spilled on the display or from the corners. It varies with each sample. The light output also falls significantly with time resulting in dramatically lower contrast with uniformity issues.

Plasma on the other hand appears to be able hold its performance over time much better. I cannot tell a low hours 9G or 8G Pioneer from a high hours unit. The light output is very similar to a new unit along with the blacks. White balance and color can shift, but it is not a strong effect with time in most cases.

Quote:
Originally Posted by umr View Post

I have seen a 75% reduction on LED flat panels after a few thousand hours.

Quote:
Originally Posted by umr View Post

Yep. Panels that once output over 100 fL struggel to put out 40 fL a year later. I suspect the internal temperature of the LED will determine the rate of decay.

I always thout that it would take at least 3 years for there to be noticable changes in a LED LCD, but he is stating that they are occuring after only one year of use. This surprised me to say the least, which is why I am making this thread.

If calibrators could post the numbers from LED LCDs right after the initial calibration and then post follow up numbers after a year of use (without any re-calibration to correct any changes to the PQ) Then we would have a basis for this and follow-ups every year (or even every 3-6 months if changes are found to be significant) would be interesting to provide longterm data to see if LED lifespans are really as bad as UMR is implying. Some CCFL LCDs thrown into the mix would be nice just to see if they have any longterm issues as well. Pictures would be nice if there is an obvious change, but the data from the meters should be enough to go on.

Results so far

Chad B's Post on a LG local dimming LCD, with calibration charts. A small brightness drop off and a large reduction of the color green.

http://www.avsforum.com/avs-vb/showt...0#post19462560
Reply
Reply

AVS Top Picks

post #2 of 43
9/24/10 at 6:49am
most interesting

I just became aware of these reports: we should keep track of this

Thanks for starting the thread

Mark
Reply
Reply
post #3 of 43
9/24/10 at 8:02am
Can this post be sticky'd up top, so we can keep track a bit easier?
Reply
Reply
post #4 of 43
9/24/10 at 8:37am
sticky
Reply
Reply
post #5 of 43
10/7/10 at 12:31pm
Thread Starter 
I'm surprised that this thread isn't getting more attention. Maybe I should have gone with a more sensational title for the topic to grab peoples attention....
Reply
Reply
post #6 of 43
10/14/10 at 1:49pm
Quote:
Originally Posted by PENDRAG0ON View Post

I'm surprised that this thread isn't getting more attention. Maybe I should have gone with a more sensational title for the topic to grab peoples attention....

I think it was stickied too soon. Most people dont check stickies unless their new to a forum or already invested in the topic.
Reply
Reply
post #7 of 43
10/16/10 at 7:39am
It caught my eye. Was considering a 67" LED panel for Christmas. I always do research beforehand, but it's good to know about problems even earlier.
Reply
Reply
post #8 of 43
10/16/10 at 8:28am
Hard data would be nice. The topic is interesting to me as well but so far its just anecdotal evidence from one poster. (though I have no reason to doubt UMR)

In 6 months of 12+ hours a day use my laptops LED screen (that shows clear effects of heat on the uniformity when running intensive programs) measures exactly the same 105fL as it did on the first day.
Reply
Reply
post #9 of 43
10/25/10 at 2:57pm
Here is an interesting paper on calibration and touches on the LED's as a light source.

http://www.burosch.de/images/White_P...ibration_E.pdf

The advance of the solid state LEDs as light source for displays is now a fact. However LED
behaves quite differently from the ‘traditional’ UHP or CCFL light sources. A detailed understanding
of this behavior shows that LEDs are more instable with time, temperature and driving current, so
extra precaution in calibration and measurement is needed. While for CCFL and UHP a simple
tristimulus color sensor could yield satisfactory results, to accurately measure and calibrate LEDs a
real spectrometer is necessary. Barco has understood this challenge, and so far is the only
manufacturer in the world utilizing a spectrometer in its product lines based on LED illumination.

Athanasios
Reply
Reply
post #10 of 43
11/4/10 at 2:57pm
Thread Starter 
Still no numbers, is everyone afraid of even looking into this? Whether it is true or not we need someone to do a study of this. Heck, even just maxing the backlight of a year or two old LED LCD and measuring the peak light output would be enough to get started with. (it would tell us a ballpark figure of how much brightness it has lost compared to the hours on the set and give us a starting point)

If this is a real issue we need to track it down and figure out what models are affected and isolate the root cause before too many of these things are sold. (and thus allow the TV makers to fix the flaw) If this does snowball it could turn into a massive lawsuit against a company for false advertisement of lifetime numbers. (50,000 to 70% brightness seems to be common numbers for LED LCDs) Better to catch it early on than years down the road like the Panny black rising. (that too started with one calibrator noticing the issue, but he posted numbers to back it up)

I'll try to bump this up at least once a week since the sticky was removed, just to keep it out there.
And if nothing is found by January, then I'll probably request this topic be moved to the LCD forum so that it can get more exposure, even if it is mostly flame posts from LED fans....
Reply
Reply
post #11 of 43
11/4/10 at 5:18pm
Perhaps umr can chime in with an update or new info... it's seems like he's the only one to post about this so far in terms of hard data.
Reply
Reply
post #12 of 43
11/4/10 at 11:03pm
Thread Starter 
Quote:
Originally Posted by PENDRAG0ON
I would actually prefer edge lit over back lit since the entire grid would age at the same general rate. An LED backlit LCD will have the same potential for uneven wear as a plasma, both rely on a grid of light sources that are proven to dim with use and if I'm going with LCD of any type, I don't want to have to deal with one of plasma's potential flaws in an LCD.

Quote:
Originally Posted by Chronoptimist View Post

Ahahahahahahahahahahaha.

Don't you think someone with one of the XBR8s or similar sets would have noticed a problem by now if that was the case?

All we have is passing comments from one poster with no data. No information on what kind of screen it was. My 7 month old laptop screen that is on for 12+ hours a day, subjected to a lot of heat when running intensive tasks, measures the exact same brightness as it did the day I got it and uniformity is unchanged. There are even two spots that get a lot of heat in particular when running some tasks and while they temporarily lose uniformity as a result, it returns to being uniform when it cools down.

No-one I know with an LED set, even early adopters, has noticed any appreciable drop in brightness or uniformity with their sets.

The technology has been in widespread use for almost three years now, I think the problems would be widely known and documented at this point if it were an issue.

The only set that might have problems is the 6 year old Qualia 005 because it was a very early implementation and was not power efficient at all. (the whole back was a radiator)


If you want quality, you have to pay for it. The set matches all your criteria.

Quote:
Quote:
Originally Posted by Chronoptimist View Post

The only set that might have problems is the 6 year old Qualia 005 because it was a very early implementation and was not power efficient at all. (the whole back was a radiator)


If you want quality, you have to pay for it. The set matches all your criteria.

Funny you should mention the Qualia 005, a while back someone with one of those sets confirmed that he had uneven wear on his set. Oh how I wish I had saved that link....

And the Sony doesn't match all my criteria, for one it is drastically overpriced for the size. (I'll pay for quality, but only to an extent) Two it is LED based giving it a limited viewing angle. (I'm fine with most CCFL LCd's angles, but I have only seen one LED LCD with acceptable viewing angles, and that was an IPS based LG) And three, my theory on LED uneven wear has all the proof it needs in plasma tvs since they too have uneven wear issues and operate in a very similar manner. This is why I would gladly take a edge lit set over a back lit set if I had to get a LED based TV (but CCFL wins out over both due to proven reliability) I've run this theory by several of the top minds on avs and almost all of them agree with it. The physics behind how LED backlit LCDs work pretty much guarantees that uneven wear will happen at some point, the only question is when.

Quote:
No-one I know with an LED set, even early adopters, has noticed any appreciable drop in brightness or uniformity with their sets.

The technology has been in widespread use for almost three years now, I think the problems would be widely known and documented at this point if it were an issue.

How long did it take people to notice the rising blacks on a Panasonic plasma again? You can't trust peoples eyes, you have to trust meters and hard data. If you know that many people with these sets, then grab a meter and start providing numbers for the LED brightness thread, if you are that convinced it isn't an issue then contribute to the cause of finding the truth. We need a large sample size to know for sure and any help is welcome.

My original theory called for 3 years minimum before the first traces of uneven wear on LED backlit LCDs would start to pop up. UMR's findings surprised even me, I honestly didn't expect anyone to have issues after only one year and I never expected edge lit sets to have any serious issues for an even longer period of time. This is why I want to get to the bottom of this, if there is something about how some TVs are driving these LEDs that is drastically reducing their life, then we need to track it down because there is no way these TV makers are going to do it unless they have to.

Quote:
Quote:
Quote:
Originally Posted by specuvestor
Chrono, the only reason that I thought LED quality for 2010 could be a problem is because the supply chain wasn't ready for it (hence the shortage) and likely need to cut corners. Like wine, I think LED in 2010 is bad harvest and probably should be avoided if possible.

I'm also thinking that corners were cut on several sets to save some money. It isn't uncommon for several models or even several batches in the same model line to get substandard parts, especially when dealing with Samsung who is known for a panel lotto, who isn't to say that some batches got cheaper LEDs? This is just one of many possibilities for UMR's findings that need to be looked into.

But enough of this, let's move this convorsation to the LED brightness thread and keep this topic clean. (I'll copy my previous post, this post as well as Chrono's full post over)

Brought over from another thread to keep it from getting too off topic.
Reply
Reply
post #13 of 43
11/8/10 at 11:57pm
Thread Starter 
First line was a reply from another topic but it leads into the second paragraph.

LED LCD with IPS screens are probably the best bet, assuming they can get over the dimming problem, something I can finally say I have seen in person.

I work at Walmart as a truck unloader, when hauling the freight to the floor I pass by the TV wall some 20-30 times a day so I get to compare some basic things like brightness and viewing angles. On the LED section they have a few Vizio, a few Samsung and a few LG sets. The Vizio 55" local dimming set was always the best with viewing angles of them with the LG's coming in after them and the Sammy's a distant last. The Samsung was the newest set (now about 2 months old) and also the brightest but Samsung has always been hyper bright regardless of tech under the hood so I thought nothing of it, until now. The 55" Vizio display was sold due to an ad mistake and a replacement arrived for it and was put up just 9 hours ago (as of this posting) both were set up the same way (basic out of the box store mode) so they should look very similar right? Wrong, the new Vizio is roughly twice as bright as the Samsung right next to it where as before it was at least 15% dimmer.

The only thing I can think of for a difference that drastic is that they all are indeed getting dimmer with use. So my advice is to avoid store models outright, they are getting driven way too hard and it is drastically accelerating the aging process.
Reply
Reply
post #14 of 43
11/9/10 at 7:36am
Quote:
Originally Posted by PENDRAG0ON View Post

First line was a reply from another topic but it leads into the second paragraph.

LED LCD with IPS screens are probably the best bet, assuming they can get over the dimming problem, something I can finally say I have seen in person.

I work at Walmart as a truck unloader, when hauling the freight to the floor I pass by the TV wall some 20-30 times a day so I get to compare some basic things like brightness and viewing angles. On the LED section they have a few Vizio, a few Samsung and a few LG sets. The Vizio 55" local dimming set was always the best with viewing angles of them with the LG's coming in after them and the Sammy's a distant last. The Samsung was the newest set (now about 2 months old) and also the brightest but Samsung has always been hyper bright regardless of tech under the hood so I thought nothing of it, until now. The 55" Vizio display was sold due to an ad mistake and a replacement arrived for it and was put up just 9 hours ago (as of this posting) both were set up the same way (basic out of the box store mode) so they should look very similar right? Wrong, the new Vizio is roughly twice as bright as the Samsung right next to it where as before it was at least 15% dimmer.

The only thing I can think of for a difference that drastic is that they all are indeed getting dimmer with use. So my advice is to avoid store models outright, they are getting driven way too hard and it is drastically accelerating the aging process.

Is this really empirical proof of anything? Chronoptimist's anecdotes were dismissed because of

"You can't trust peoples eyes, you have to trust meters and hard data. If you know that many people with these sets, then grab a meter and start providing numbers for the LED brightness thread, if you are that convinced it isn't an issue then contribute to the cause of finding the truth. We need a large sample size to know for sure and any help is welcome."

But a Walmart truck driver who didn't personally set up both sets to know if they were unchanged -- every day -- is a good example. This is an interesting topic if real data and pissed off users start making observations, but so far it seems a lot like FUD.
Reply
Reply
post #15 of 43
11/9/10 at 7:46am
Data from an LG 55LHX local dimming full LED set I calibrated twice:
1st calibration in January 2010, Day mode was calibrated to 69 fL measured with a Chroma5 profiled off an Eye One Beamer, an older version of the i1Pro (Beamer was within calibration date).
2nd calibration after 10+ months, same calibrated Day mode measured 64 fL with the Chroma5 profiled off an i1Pro (also within calibration date). There was a lack of green in the grayscale.
There might be a small difference in the 2 Eye One spectros, but it should be minimal seeing as both were NIST recertified by X-Rite.
Reply
Reply
post #16 of 43
11/9/10 at 10:56am
Thread Starter 
Quote:
Originally Posted by RetroElectro View Post

Is this really empirical proof of anything? Chronoptimist's anecdotes were dismissed because of

"You can't trust peoples eyes, you have to trust meters and hard data. If you know that many people with these sets, then grab a meter and start providing numbers for the LED brightness thread, if you are that convinced it isn't an issue then contribute to the cause of finding the truth. We need a large sample size to know for sure and any help is welcome."

But a Walmart truck driver who didn't personally set up both sets to know if they were unchanged -- every day -- is a good example. This is an interesting topic if real data and pissed off users start making observations, but so far it seems a lot like FUD.

I can't explain away two sets set up identically in the exact same spot where the newer one is now the brightest set on the wall vs the older set being one of the dimmer LEDs. A small difference in brightness can be explained due to unit varrience, a new set being more than twice as bright as the old one is kinda hard to explain away. I'm leaving it at that.

And I didn't offer it as the end all be all to prove this as fact, just offering this as something I observed that I can't explain any other way other than that they had all dimmed with use due to how hard they are driven in a store enviroment. Now we have 2 people who have seen this in person, neither offering any proof of it other than what their eyes have seen.
Reply
Reply
post #17 of 43
11/9/10 at 11:03am
Thread Starter 
Quote:
Originally Posted by Chad B View Post

Data from an LG 55LHX local dimming full LED set I calibrated twice:
1st calibration in January 2010, Day mode was calibrated to 69 fL measured with a Chroma5 profiled off an Eye One Beamer, an older version of the i1Pro (Beamer was within calibration date).
2nd calibration after 10+ months, same calibrated Day mode measured 64 fL with the Chroma5 profiled off an i1Pro (also within calibration date). There was a lack of green in the grayscale.
There might be a small difference in the 2 Eye One spectros, but it should be minimal seeing as both were NIST recertified by X-Rite.

So it slightly dimmed over 10 months and the green shifted somewhat? Not a major difference like what UMR reported, but we still don't know how those sets were calibrated or what brands there were or if they were edge lit or back lit....

You wouldn't happen to have calibration reports that you could post for that LG would you?
Reply
Reply
post #18 of 43
11/9/10 at 4:21pm
Yeah, the green reduction was actually pretty significant. The white balance was visibly purplish at the low end after 10 months.
The first attachment is how the TV performed in January after the 1st cal.
The second is how it performed 10+ months later with the same settings before I touched it up. I am taking it by faith that the settings were not changed at all. At a glance they looked like they had not, but I didn't go into the 10 pt adjustment and check all those. The customer didn't want them locked so he could easily change the 240Hz settings.

 

Mark LHX Expert 1 (day).pdf 233.39453125k . file

 

LG 55LHX before touch up ISF day.pdf 172.4208984375k . file
Reply
Reply
post #19 of 43
11/9/10 at 4:35pm
Thread Starter 
Quote:
Originally Posted by Chad B View Post

Yeah, the green reduction was actually pretty significant. The white balance was visibly purplish at the low end after 10 months.
The first attachment is how the TV performed in January after the 1st cal.
The second is how it performed 10+ months later with the same settings before I touched it up. I am taking it by faith that the settings were not changed at all. At a glance they looked like they had not, but I didn't go into the 10 pt adjustment and check all those. The customer didn't want them locked so he could easily change the 240Hz settings.

I added a link to this post in the first post under a "results" subsection.

Thank you for providing the charts.
Reply
Reply
post #20 of 43
11/10/10 at 2:10pm
a Notebook Review.com thread
Quote:''i have found that with the blue led and yellow phosphorus backlit LCd,there are severe discoloration after two years of continuous use.There is a clear yellow shift in the colour displayed by the LCd."

Reply
Reply
post #21 of 43
11/11/10 at 8:13am
Just got a newer local dimmer and had to sell some other stuff for cash. My old Dell 2405 CCFL has turned yellowish after a lot of use with the backlight on low for most of its time and I have had that for about 5 years or so. If this were to happen to my new LED I would be disappointed though I did get it for a great price. I will definitely be watching for any changes. Honestly, I give my TV regular workouts on 100 for Contrast and Backlight with a nice screen saver every so often. I personally feel you should break in your TV and drive it high once in a while. After my Dell went out I know that having any part dimmed for too long even on CCFL's in probably not the best idea. Nor would it be good to max it out for a long time but it seems as if your TV should be on the higher end of its capabilities if you want its light output to last. But this seems to be saying that operation will dim them for whatever reason. Not sure I buy the large numbers he is selling. That doesn't sound good at all. I hope this is wrong. My black levels hope this is wrong.
Reply
Reply
post #22 of 43
11/17/10 at 1:58pm
Quote:
Originally Posted by Chad B View Post

Data from an LG 55LHX local dimming full LED set I calibrated twice:
1st calibration in January 2010, Day mode was calibrated to 69 fL measured with a Chroma5 profiled off an Eye One Beamer, an older version of the i1Pro (Beamer was within calibration date).
2nd calibration after 10+ months, same calibrated Day mode measured 64 fL with the Chroma5 profiled off an i1Pro (also within calibration date). There was a lack of green in the grayscale.
There might be a small difference in the 2 Eye One spectros, but it should be minimal seeing as both were NIST recertified by X-Rite.

You may want to post your 55LHX findings in the 55LHX thread and your web site.

-nony
Reply
Reply
post #23 of 43
11/23/10 at 8:03pm
As a new owner of an edge-lit LED (still in the return window haha), I am very interested in seeing more data on this.

And, as a home theater enthusiast, I am very interested in this in general! This is a very important thread that needs some support. The quicker data can be compiled, the quicker we can see if there is a serious issue here.

ChadB, I'm in Ohio and have been seriously considering contacting you to calibrate my new set. So that would give us another set of data to work with. With the holidays coming, though, money is tight....
Reply
Reply
post #24 of 43
12/3/10 at 6:02am
Here's my thoughts on my 55LHX results.
I agree they look bad. But I wouldn't call this a scientific comparison. The 1st calibration, at first I couldn't even remember for sure if I profiled the C5. I did too many calibrations in the 10+ months to remember. After talking with the owner and checking my old meter profiles, I determined that I did. I don't think I mounted the i1 on a tripod the first time; that's something I just started doing recently. So, I probably used the LCD counterweight mount the first time, and the second time I tripod mounted farther from the screen. That combined with using 2 different i1 spectros (even though both were in calibration, there can be a small difference) and not checking to be absolutely sure the 10 pt settings didn't get changed, I'm not going to condemn the 55LHX.
Reply
Reply
post #25 of 43
12/3/10 at 12:38pm
Thread Starter 
I just had an interesting idea...

Could someone buy a display model LED set that has been on display for at least 8 months in torch mode and a new, identical TV and compare them side by side (with a light meter to record the numbers) This would probably be the easiest way to see if this issue is worth pursuing or not. 8 months may not be that long, but running in torch mode they will be pushed much harder than a home display so it should give us an idea of the longterm results that we would see.

Things to measure.

Peak light output.
Calibrated light output. (same backlight setting on both units)
If there is a noticeable difference, can it be compensated for with proper calibration?

Anything else that might be worth checking?
Reply
Reply
post #26 of 43
12/6/10 at 3:20pm
I was at a box store the other day and was thinking the exact same thing. I no longer have the Sony EX710, I decided to go with the LG LD550 for several reasons - one of them being this possible issue. I am still very interested in this topic, though.
Reply
Reply
post #27 of 43
12/9/10 at 2:14pm
Reply
Reply
post #28 of 43
12/27/10 at 10:46pm
Thread Starter 
I did a little more digging into LED tech and DLP's using LEDs and I found a lead, the main factor in their aging seems to be heat. If the temp is outside of their range it causes stress on something in the LED, and causes it to dim at a quicker pace. A water cooled LED lamp in a DLP has remained consistently bright far longer than any other LED based DLP lamp. (saw it while digging through UMR's posts, don't have the link handy)

With how thin many HDTVs are getting thanks to these things, there isn't much room for the heat they generate to go, mainly due to the lack of airflow. It may only be slightly higher than their intended range in a normal calibration, but even that would be enough to cause problems.

Thicker Backlit models would have fewer issues in theory due to more room to disperse heat, but I can't confirm that. The local dimming array will still cause a slow but steady uneven wear regardless.


I'll continue to try and wrap my head around this issue until definitive proof is found, it's all I can do at this point, try and explain how this might happen.
Reply
Reply
post #29 of 43
2/17/11 at 9:52am
I'm guessing this problem might be brand specific or even model specific.
A properly driven LED should not loose intensity and should not produce
enough heat to cause degradation unless they are being over driven.
I own a Sony 55XBR8 with over 8000 hours and there has been no loss in
light output of the LED backlighting and the panel produces almost no heat.
Reply
Reply
post #30 of 43
2/19/11 at 8:37pm
.......
Reply
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration

AVS Top Picks

AVS › AVS Forum › Display Devices › Display Calibration › LED longterm performance?