[Solved for now] LG 47LN5400 TV LED backlight

[shakalnokturn]

I've saved this from the e-waste. Now trying to decide where the fault lies.

At first I thought it was working, after running for about 35 minutes it lost the backlight. Cycling power would only give a few seconds of light, the TV had to be left 10 minutes after that it ran about 20min. before loosing backlight again.

Once the backlight (BL) stops, if power cycled quickly it will run normally for 10 to 30 minutes. If left running without BL it will need some rest time.

Adjusting the backlight level in the picture menu will give  short flickers on the decrements below "15", no flickering when incrementing, then around "11" it will flicker and loose backlight completely.

Twisting or knocking the TV won't trigger the fault.
When the backlight stops the LED voltages outputs are low.
During "normal" operation I measure 60VDC+23VAC on one LED output, 60VDC+25VAC on the other. Without really knowing what to expect I find that AC component is quite high. (Measured with DMM, haven't scoped yet.)

This looks like a driver problem although I'm having trouble imagining what could have that much hysteresis on the thermal cycle. (5-10 seconds rest will resume at least 10 minutes use.)

Any thoughts or experience with similar faults?

[Miti]

The first culprits in old SMPS that work intermittently are the electrolytic capacitors. Even though they don’t look bad from the top, they can still be dry and have high ESR. If you have an ESR meter you can test them, or better replace them.

[PKTKS]

TRUE may be crapacitors..

But it also may be aged LEDs dropping the booster protection...

That handy tool to check LED strips helps..
But a VISUAL inspection of the strip will also find toasted ones..

Paul

[shakalnokturn]

Thanks for your answers.

After pulling the scope out and making some measurements on the LED supplies the backlight quit completely, only giving a blink on power-up. Measurements were between +LED and ground (not -LED). Not much to be noticed anyway, a little more ripple on the supply to the lower strips than the upper.

Checked the led voltage with the megohmmeter (@1.2mA) both strings read around 60V and lit faintly. All that was telling me was that no string was completely open.
Tied both strings in series got the lab supply out to see how they'd perform with a typical usage current and warm-up.
With my supply at maximum voltage (200V) I couldn't get the current above 100mA. Checking string voltages gave 73V and 127V, the lower side being the suspicious one.

Rechecked both strings supplied separately, upper good, lower bad, flickers at around 90mA. Pulled the thermal cam out out of curiosity, found the hot spot from behind  .

Some panels can be disassembled leaving the LCD face down, this one had to be undone front to back making things a lot more delicate.

Once inside the defect is easily (black-) spotted, no doubt my lab supply testing made things worse, the PCB has a hole charred into it. Went looking for a spare LED strip.but didn't have the right one of course... Checked prices, panel is disassembled on the dining table, can't wait for an order.
Decided to scratch the charcoal away and replace the LED pulling one from a similar spare strip.
Luckily there's enough cathode plane remaining to solder to, the anode has to be wired to the track before the hole in the PCB.

Tested after replacement with lab supply, took a thermal picture just to make sure the replaced LED wasn't running hot due to poor cathode connection. Some of the original LED's run hotter.

Reassembled trying to limit the panel contamination, had to fight the insects too because I was working with windows open...

[wraper]

To avoid repeated failure, it's advisable to reduce LED current in the driver by about 30% by increasing resistance of the current sensing resistor.

[PKTKS]

Rather typical fault.

Replacing the full LED strip is advisable and will not need any further adjust on the sense resistor..
Once A single LED failed others may be compromised.

Cheap repairs can also just workaround the faulty LED but the mileage will vary

Paul

[shakalnokturn]

Could you elaborate on their compromission?

Is it just being used in series? Having one fail in a string where all the LED's are used in the same conditions means the others are at the end of their lives too?

Or is it because one failed that the others may have been abused?
The supply being constant current that shouldn't be the case unless the PSU design looses current regulation when there's an intermittent break on the string.

[wraper]

Replacing the full LED strip is advisable and will not need any further adjust on the sense resistor..
Once A single LED failed others may be compromised.

You miss the fact that TV design is flawed. To achieve high brightness spec for marketing purposes, they stress the LEDs too much when brightness is on high setting. An because of that backlight fails after a few years. By replacing LED strips you don't fix the initial problem why they failed. Not to say replacement LED strips often are less robust than original.

[PKTKS]

Rest asure it will work fine for a decent time period..  you can put warranty

Without trial and error messing unknown design methods

Paul

[wraper]

Rest asure it will work fine for a decent time period..  you can put warranty

Without trial and error messing unknown design methods

Paul

Yeah, it will last a few month warranty repair shops usually provide. But you don't fix the initial issue. Reducing led current is a standard practice experienced TV repairmen do. It barely affects perceptible brightness.

[Labrat101]

I have had the exact problem with the LG 50inch it uses the exact same panel as yours . The output voltage are stamps on the board.
 being the 50 inch has a few more leds . After checking it had 2 shorted leds on almost every strip and 1 open. As already said the LEDs are not fixed 100% good so the solder is not completely under the LEDs causing them to bend when hot and crack in the middle.
This is a fault on most LGs .
And we find a quiet a few in the recycling.
I just ordered a complete new set of back lights for a mere $18 so for a free smart tv I guessed it's worth it. Yes the biggest problem is space when lifting the screen up and not to bend it .
 

[wraper]

https://masterelectronicsrepair.blogspot.com/2020/03/repair-servicing-tv-lg-47ln613v.html

You can reduce the backlight current in TVs with LGP4750-13PL2 power supplies (EAX64905501) as follows. Current sensors in the R811-R818 LEDs - channel 1 and R829-R836 - channel 2. Resistors in the channels in pairs of 4.3 + 3.9 Ohm. You can reduce the backlight current by 25% by removing one pair of resistors in each channel.

[PKTKS]

I have had the exact problem with the LG 50inch it uses the exact same panel as yours . The output voltage are stamps on the board.
 being the 50 inch has a few more leds . After checking it had 2 shorted leds on almost every strip and 1 open. As already said the LEDs are not fixed 100% good so the solder is not completely under the LEDs causing them to bend when hot and crack in the middle.
This is a fault on most LGs .
And we find a quiet a few in the recycling.
I just ordered a complete new set of back lights for a mere $18 so for a free smart tv I guessed it's worth it. Yes the biggest problem is space when lifting the screen up and not to bend it .

Yep I do the same on these cases..

I order a batch of half dozen strips ..
change the most suspicious ones and  good to go.

Unless absolute necessary to alter the sensing network...
it usually works for long periods to the next burn fault..

Cheap reliable and fast repair

Paul

[wraper]

Unless absolute necessary to alter the sensing network...
it usually works for long periods to the next burn fault..

Cheap reliable and fast repair

It's should never fail during lifetime of the TV.

it usually works for long periods to the next burn fault..

This sentence already means you have done a subpar repair. As such fault should never happen again if repair was done properly.

change the most suspicious ones and  good to go.

By the time when some LEDs are starting to fail, every other LED already has cracks in the compound. So not replacing them all is a disgrace of a job.

[PKTKS]

I see the point.

I have stopped doing repair for 3rd part at early to mid 90s.

But for myself I have never stopped.

Usually that kind of non cheap repair I do only for myself.
and only in cases where I have time and will for the job..
more frequently time is lacking the most..

But for the price it should be fine replacing the visible stressed strips.

Cheaper and faster  although if the fault repeats..
Use a hammer as usual..

Pul

[wraper]

But for the price it should be fine replacing the visible stressed strips.

I don't think that doing a repair that costs 50% less but will last a few months to a year instead of basically a lifetime of the device is a good way to spend your money. LED strips are not that expensive to save by not replacing all of them.

[shakalnokturn]

So far only Labrat101's photo gives me some form of credible proof on what actually destroys the LED's.
Sure running lower current is likely to improve their lifetime too. Modding the current sense at least is user-proof, when you know about the problem may as well turn the backlight down in the picture adjustments. (There may even be a service mode setting?)

[PKTKS]

Well this issue is not "new" at all.

There are a half dozen considerations that can be raised.. like:
- these LEDs are high power  variation thus temperature always play a role
- these LEDs are frequently operated in excessive "HIGH" brightness settings..
- the panels are frequently placed into a bad ventilation setup leading premature failure.
- the boosters are 100% times related to the secondary winding of a PFC SMPS
- the dependency creates a nasty relation between the LED booster and PFC booster.
- in short term they compete when the panel is in high demand of power and heat..

the list goes on.

Usually I turn the brightness to minimum "acceptable" to avoid premature failures..

Mileage vary
Paul

[Labrat101]

Firstly LG especial all fail after about 5 to 7 years .
The manufacturers like this way . They are designed to work for so many Hours.
if the warranty says 5 year warranty don't expect to get more  .
The LEDs rely on the double sided tape to cool them onto the back plate . The tape
looses heat transfer efficacy . If you check closely about 30% of all the LEDs are NOT soldered onto the stripe correctly ie only the ends are bonded These will warp and crack in the middle and will short opposed to going open circuit . This is part of the design every LED that shorts  puts more stress on the next . and the snowball effect starts . The over current wont kick in to about 5 or more LEDs have failed . Maybe.
Removing one resistor from the current sensor is fruitless turning the brightness to 55 % and increase the contrast to 90% has the same effect .
 Always replace the entire set other wise you will just have to do the whole job again in maybe a month and this time you will have a second chance to Crack the Screen.
and a totally wasted your time and have ... Still a non working TV 
 Every LG that comes into repair maybe 90% have a back light problems . some are better..  LEDs all got shorted and burn out the driver chip . 
 Just love LG good money in repairs.
 

[wraper]

The LEDs rely on the double sided tape to cool them onto the back plate . The tape looses heat transfer efficacy

It's not like that. Those circuit boards are made from single sided CEM-1 with poor thermal conductivity. So something like 50 microns of adhesive tape does not matter whatsoever. LEDs fail because of excessive heat, time they spend overheated and thermal cycling. It's not like cooling worsens over time. And it's not like some LEDs overheat because adhesive comes off, it holds well. I inspected LED strips under microscope, 100% of the LEDs had cracks in the compound, even though most of them were still functional.

If you check closely about 30% of all the LEDs are NOT soldered onto the stripe correctly ie only the ends are bonded These will warp and crack in the middle and will short opposed to going open circuit .

That's some wild claim  . Why "correct" soldering would save them from warping? Also why 100% of the LEDs become cracked even though only 30% of them are soldered wrong?

Removing one resistor from the current sensor is fruitless turning the brightness to 55 % and increase the contrast to 90% has the same effect .

1. You calculated the number completely wrong, it reduces current by 25%. And max brightness loss is even lower since LED efficiency worsens at high current. Perceptible brightness loss is barely noticeable since eyes are not linear. 2. It's not fruitless, it limits maximum current/brightness and thus maximum heat.

turning the brightness to 55 % and increase the contrast to 90% has the same effect .

And then you should explain this to everyone who uses the TV to never set the maximum brightness, even though it's often a default setting. And hope they will never forget over the years. Reducing maximum current limits maximum brightness/LED current to safe level. Increasing contrast is lame since it compromises image quality.

Always replace the entire set other wise you will just have to do the whole job again in maybe a month and this time you will have a second chance to Crack the Screen.

True, disassembling LCD panel bears a serious risk. And skimping on new LED strips may turn into cracked glass and TV going to the dumpster.

[Labrat101]

What every makes u happy.