Reading Philips TV logs with an USB-UART adapter

With my Philips TVs I never had the requirement to read the log, as they all had measurable faults or the Service Default Mode revealed everything I needed to know. In case of a two blinks error code, which points to the mainboard, or when the TV won't boot at all, it can be beneficial to peek into the log.

You need

• An USB-UART Adapter. This device maps a serial  (UART) connection across USB to a serial port (COMx on Windows). Device drivers are required.
• A terminal program, which can handle serial ports.

The hardware

There are various types of USB-UART adapters on the market (eBay or AliExpress). First I tried this type:

Those are garbage. They contain an illegal copy of a Prolific PL2303 Revision A chip, which is discontinued since 2012. Read HERE. The problem is that the latest Windows 8 & 10 driver won't work with it anymore. Some articles in the net say that Prolific has changed the device signature in their later revisions to lock out the copies. You need to install an older version of the driver. I wasted way too much time with this rubbish.

I opened mine up and the chip had no marking on it. Sure sign of a copy. Also, the USB plug already started to come off the board.

So I tried another one with the Silicon Laboratories CP2104 chip:

This one's legit. No driver problems, Windows found the driver itself and the device worked.

The software

On the PC you can use PuTTY. A more sophisticated program is RealTerm. It can record sessions, which is quite useful, and has more features than you'll ever need. Both are free.

The Android app Serial USB Terminal by Kai Morich also works fine. You can read the log on your tablet or phone quickly without a bulky laptop. Just put a micro USB adapter in front of the UART device.

The connection

Now this kept me busy for a while due to my own incompetence.

The UART / service socket on the TVs is a stereo 3.5inch type like for headphones. The connections are as follows:

Here is the rub: you need to cross RXD and TXD. Don't connect the RXD on the adapter with the RXD on the TV. Makes total sense once you understand it :-)

A schematic from a ComPair device manual put me on track:

And that's how I built the thing. I attached a 3.5 inch stereo socket to the adapter and used a stereo cable I had lying around:

First tests

I had a working 42PFL9803 sitting in my living room and I tested the device with it. To my surprise I could not get any useful log. The service manual says 38400bps 8N1. I configured everything accordingly and all I got was garbage. The TV sent data but it wasn't readable. I tried many bps setting with no luck. This TV fooled me for quite some time. I thought something was wrong with the UART adapter :-/ I think the older boards use a lower UART voltage level and won't work with this adapter.

Yesterday, I picked up a 32PFL9606 and with this one it worked flawlessly. 

Sweet! I can add one more diagnostic tool to my repertoire. I currently have a 46PFL8007 with the dreaded QFU chassis, which doesn't show any signs of life even though the standby voltage is good. It's not writing a log either. But that's the subject of an upcoming blog post once I have reprogrammed its boot EEPROM, which I suspect.

In the meantime, I glued it into a nice blue box:

RUNTK5351 TCON - defect analysis - ISL98602

The TCONs with the ISL DC/DC chip notoriously go bad. Sometimes they are fixable by swapping the chip, sometimes they are not. I had the chance to play with four broken boards. Two of them got a new ISL and still didn't work. However, I present you a few tips how to avoid fruitless work, because the main video chip may be measurably dead. Also I think I have measured the reference voltages that the ISL should produce.

The next image shows the voltages of a good ISL chip:

If the voltages are all there and there is still no image, the main chip is dead.

Normally though, the TCONs come with an ISL, which produces no voltages at all. Here are the tests you can do to asses whether it is worth changing:

• Test 1: In diode mode, measure the breakthrough voltage of the 1.2V trace. It should be around 0.5V. If it is 0, forget the board.
• Test 2: Attach a lab power supply with 1.2V and current limit around 100mA to the 1.2V supply. The main chip should draw about 0.01A. If not, it is broken.

Alas, I did not yet have a working board in my hand to know how much current the board is supposed to draw from the 12V line. The ones with the fixed ISL and no shorted main chip both pulled 0.39A. The main chip got pretty hot quickly. I guess this is not normal.

Changing the ISL

This is very difficult. I never managed to solder it properly with hot air only. It always took me an extra step with the soldering iron to get the solder to flow at the pins. I failed with a needle tip. It doesn not have enough heat capacity. Spade tips neither worked, even small ones, because they all were too clunky to reach the pins. The only tip that worked was the horse shoe with an excess amount of solder on it. A perfectly rectangular tip would be best. And lots of flux is required, of course.

Be extremely careful with the microscopic SMD parts around it, especially on the upper right corner (previous image). That one 0 Ohm resistor close to the edge gets pushed away easily.

Philips 46PFL8007 - QFU1.1E LA - no standby LED - Boot EEPROM 7CT3 25P10 reprogrammed

I bought this Philips as defect and didn't ask any questions. The description said it wouldn't switch on anymore. Well, that did not sound so bad. I had fixed another QFU1.1 this year with a dead standby supply.

At pickup the seller told me that the TV had been fixed during warranty with the same symptoms. The mainboard was the culprit then. Uh oh...

When I plugged it in, the standby LED did not light up. Standby voltage was present. The LED is controlled by the main processor (the standby processor section). The only voltage the stdby proc. takes in is the 3.3V standby. Nothing else. The LEDs are fed with the same voltage.

I studied the service manual thoroughly and the only conclusion was that the processor was in trouble. I ran a reflow session in the oven. Did not help.

In the Iwenzo repair forum, I got the hint to reprogram the standby software flash ROM. For that you need two things:

• The software binary for the QFU1.1 platform SPI boot EEPROM. Version 77.02.
• An EEPROM programmer.

My adventures with the UART adapters you can read HERE. This TV did not produce any log whatsoever. The CPU was not running any boot program.

I also purchased an EEPROM programming device called SkyPro USB Programmer. It is made by Coright. The software installed flawlessly on Win10. I had to desolder the Flash ROM 7CT3 and solder it on an adapter board, which then went into the programmer's socket.

The chip sits under the upper right corner of the white heatsink. The sink has to be removed carefully. It is mounted with two spring bolts, which are easy to release, and an adhesive foil. The foil does not survive the process.

I tried a test clip from Aliexpress first directly on the board. This was like lottery. The clip did not attach properly and I got only nonsense results.

The software then identified the Flash as 25P10 (128k) instead of a 25MP05 (64k), which is listed in the service manual.

Now, the hardest problem was to find the right software as there are a number of versions floating around in the net. The QFU1.1 has two variants. One for Fusion 67.0.0 and one for Fusion 77.02.08. This device needs the 77 version.

The 77.04.08 is QFU 1.2 and will not fit. It is used in the xxx8 series, not the xxx7.

To add to the confusion, 77.02.08 is also supposed to work for QFU 2.1. This is only used in the 6007 model, however.

This is the software that worked for me: DOWNLOAD

A peek into the binary files

The first diff shows the good file to the left, which finally revived the set, and the scrambled one to the right, which I read from the EEPROM initially. You see that the first block is wiped out with garbage. There were more garbled blocks further up the address space.

The second diff shows the Fusion 77 to the left and the Fusion 67 to the right. If you have a file at hand and like to investigate which version it is, take a HEX viewer and study the first block.

Notes

I first did not realize that the TV was actually fixed because I didn't insert the flat cable to the TCON properly. It looked totally fine, yet it wasn't sitting right. The sockets on the mainboard have a locking mechanism. You need to push the two black knobs down. I failed to do that and broke off the locking nose on the cable. The cable then does not sit very well anymore.

The TV was stuck in a boot loop because of the cable. The log displayed weird errors about the DVB-S tuner. In hindsight, it must have been trouble with the I2S bus.

I stumbled over the solution while testing another board where I made the same mistake with the cable again! This time I caught it rebooting immediately after I had touched the cable.

I once destroyed a not properly seated cable. A trace went up in smoke. Be very careful with those.

Don't forget to plug all wire harnesses into the mainboard. If the WLAN is missing, you will also get a loop.

So what is going on with those QFU chassis?

A number of devices with similar symptoms are mentioned in repair forums and sold on eBay/classifieds. What's going on here? How can an EEPROM, which is otherwise fully functional, lose blocks of its memory? What I know is that the processor gets really hot. I measured 57° celsius at 21° room temperature with an open back cover. Now extrapolate that to 30° and closed cover. I recon it will be 70° or more. Does the EEPROM get grilled? I don't know.

Panasonic TNPA5330 SN board 7 or 6 blinks - detailed fault analysis and repair guide (TX-P42GT30, VT30, ST30)

I received a TX-P42GT30 with the famous 7 blink disease and it is the fifth Pana plasma with this defect. I am going to present my analysis of the causes of this defect, which appears after four to six years. On eBay I am seeing more of them these days.

The actual cause are loose screws. Why did they get loose? Because of the solder on the contacts. Solder is soft and flows under pressure. Why is this troublesome? Some screws and their contact pads conduct a lot of current into the metal panel chassis. When they become loose, the contact resistance increases and sparking occurs. This in turn leads to very high current peaks, which eventually kills (shorts) a diode. It is also possible that the increased resistance causes the mounting area to heat up and that kills a diode. This diode connects six transistors and another diode and those die immediately.

The next image shows the troubled screw position in close-up. It takes the full current from the DAF30 diode to the panel chassis ground. Other screws, which I'll show you below, pick up the current and from there it flows back to the power supply. Notice how the legs of the diode are discolored through heat! This one got enormously hot:

The mounting hole backside under the microscope. We see blackened, burnt solder:

And this is what a mounting point looks like:

As long as the screws stay under tension, the contact to the chassis will be good.

In the 60 series, Panasonic has learned from the screw disaster. This is an image from a 55STW60. No solder on the holes and screws with spring washers!

They also finally stopped using SMD power transistors and returned to decent heatsinks. Who needs silly super-flat TVs, anyway? I think those scan boards are built for eternity - maybe, provided the capacitors, which get all the heat from below, are holding up well.

In the next image, I marked the other screws, whose holes also had burn marks. The two on the top pick up the current from the chassis, which enters through the screw in the center. You see the already repaired board with my choice of transistors and diodes (see this post). Other screws have a proper bracket on the board or don't carry much current. There, the contacts looked ok.

And here is the section of the circuit where you can see all the affected high-power parts in one glance. The DAF30 diode is marked yellow. The diode and three transistors to the left and the three transistors to the right next to the troubled diode are all shorted when disaster strikes.

From 7 blinks to 6 blinks

6 blinks indicate a problem with the MIR voltage, the energy recovery voltage, which builds up across the blue C631. It must stay in a corridor around 120V.

Dead driver transistor array and control chip (energy recovery H section)
Strangely, a driver transistor plus its control chip die, even though they are not responsible for any of the shorted power transistors. Q441 is still ok, yet those two are dead. I wonder why, but I have observed this twice already, only in 42 inch models though.

Broken IGBT (energy recovery L section)

The boards I have repaired also had a less obvious failure in Q451, a DG302 transistor. It has no dead short, but in diode test mode, it will leak between collector and gate and show a break-through voltage on the multimeter. It may not have this fault at the beginning and develop it once the other defects are repaired and you switch the device on for the first time! But don't worry, it will not destroy any other parts. It is best to routinely replace it.

See also below where I describe how to debug the recovery section.

How to repair this defect properly

• Replace all broken components
• Remove all solder on both sides of all screw mounting holes. Only apply a very thin and flat(!) layer of solder. It helps making a good contact.
• Clean the mounting points on the panel chassis from all black residue.
• For the screw next to the DAF30 diode and the two on the top, replace the original screws with ones with a spring washer. Make sure they are not too long, otherwise you will drill into the panel! The screws will not loosen much once the solder is gone, but for the critical ones I want extra safety.

Debugging the 6 blinks of the energy recovery circuit

Through my own stupidity I damaged an already repaired board and spent hours trying to find the reason for the 6 blinks. The device started, the green LED on the SN board briefly came on and then it immediately shut down. Not enough time to take measurements with a voltmeter. A not 100% working driver transistor was the culprit. Along the way I learned a lot about the circuit.

Get the manual for the TC-42GT30 from elektrotanya.com. This manual is top quality with zoom-able schematics. Page  69, chapter 12.27 SN1 Board Schematic Diagram. I have another for the TX-42 with pixelated graphics, where you can't read the part numbers.

The recovery voltage can be measured across C631. It should be around 120V. The protection activates below 36V and above 157V. The polarity is not important, just focus on the amount.

A multimeter in MIN-MAX mode may not be fast enough to catch the max amount. I used a digital storage oscilloscope in roll mode with 1ms time base. It showed the ramp-up of the voltage beautifully. However, this is not required, because if the protection circuit fires, there are only two possible cases, which a quick multimeter can detect in MIN-MAX mode:

If the voltage is missing, the recovery L section is not working. Most likely, Q451 has a problem as described above. Mine never had a full short. If it has, also replace Q551 and IC522.

If the amount is too high, the recovery H section is not working. Strangely, Q441 does not die, but its companions Q531 and IC502. This is a total mystery to me.

Spare parts

The driver IC I got from  HERE and the transistors from HERE. So far, they seem legit and work ok.

In THIS BLOG POST I talk about possible replacement for discontinued parts. I am trying my luck with the FGD4536 for all the power transistors, including the DG302.

Marantz SR7007 - not starting - rapidly blinking LED - defect transistor MMBT5551 in ASO protection circuit

I always wanted to peek into an AV receiver even though I know that these devices are notoriously difficult to service due to their tight packing of boards. But sometimes, we just need challenges, right?

The SR7007 caught my eye on eBay. It is still a quite young model (2012) and is technically mostly identical to the 08 Series of 2013. I need preamp outputs and this one had all channels. Nice.

When I switched it on, I heard one click, the display went on and then after about three seconds it switched into protection and the LED blinked rapidly about twice per second.

I suspected a problem with the power amplifier because after the delay normally the speaker relais should switch. Knowing nothing about this device I first tried my luck by isolating (aka plugging off) parts of the power amp board. I got lucky by pulling the high voltage supply. The receiver started and switched the relais after 3 sec. I found nothing suspicious on the power transistors. They all measured identical. Neither did I see any burned parts.

I studied the thing a little more and identified a cable with five lines as the protection signal cable, which goes to the main digital board.

When I pulled that, the receiver would again go into protection, albeit with a different blink code (1x per second).

Now it was time to get the service manual, which is only available for money. 12$ is not a bad price, so I got one and it proved to be worth every penny.

With the high voltage unplugged from the power amp board, the device could be booted in a diagnostic mode where it displayed the last protection error. After the steps in 3.2. shown below, you have to press the Status button on the device. The description is somewhat confusing.

I did that and I got this:

For ASO (no idea what this means), the manual states the following:

Now that did not make any sense. The receiver had never switched the speakers on and neither was there any speaker connected. However, it pointed me to the direction of the ASO protection circuit. Something had to be wrong with it.

Q7001 does a logical OR on all ASO signals supplied by the 7 power channels ASO circuits. The collector signal goes straight to the digital board:

The red line is high voltage plus. It made perfect sense that when this voltage is missing, the ASO protect signal would not go to the digital board. I confirmed through measurement that the transistor switched on during boot. So at least one of the channels must have had a problem.

Now about the actual ASO protection circuit (here: surround back right channel). The marked transistor senses excessive current and when switched on, R7744 passes a plus signal on to Q7001 on the previous image. This circuit is identical in all 7 channels.

The inconvenience was that all the parts for the protection circuit were on the underside of the board. The unit is easy to remove, that was not bad at all. I decided to bend the legs of the power transistors to get access to the underside. Better than unscrewing all of them (horror!)

Now I checked all those 22k resistors, which go from the sensing transistor to the ASO switching transistor's base. And lo and behold, one was off by 3kOhms. The surround back right channel had a problem.

The only possible failure could have been Q7739, the sensing transistor. It measured ok with the diode tester, but it showed a lowered resistance between C and E. I unsoldered it and presto, the resistor R7744 measured normal again. My transistor tester did not recognize the MMBT5551 transistor at all. It showed me two resistors instead.

It is the little guy in the center of the image:

I replaced that sucker and the device went back to normal. A cheap fix, time-intensive and very rewarding. I learned a lot. I always like to improve things after I have fixed them, but in this case I didn't come up with anything. It is hard to tell what happened here as the power stage was still in good shape.

While I was at it I took the chance to adjust the idle current of the power stage as well. It is nicely documented in the manual. The SBR channel was off the most. Was that a coincidence or some further symptom of  the incident? I don't know.

About the Marantz SR7007 in general

I never had an AV receiver and I was curious how it worked. Well, the sound is considerably worse than my highly modified Benchmark DAC2, which feeds Focal SM9 active speakers. That was no surprise at all. I have seen its guts and I knew.

What did disappoint me was the image quality via HDMI. One might think that digital signals will not get compromised in digital processing. Not true. The passed-through image definitely lost sharpness to a degree I don't accept it. So that was no option. I am watching movies via my PC and its optical output worked well with the receiver.

The serviceability is not so bad, really. The service manual is very good. The boards, which have the highest expected failure rate (digital board and power amp board), are easy to remove. I think none of the receivers have a bottom lid anymore for an easy checking of the power transistors. They replace whole boards these days and don't mess around.

Replacement of Panasonic Plasma parts (DAF30, 30F131, RFUH25, DG302, RJP30H2A, RF1501N)

The original Panasonic parts for the NeoPlasma series 30 and 50 are slowly going extinct. As I am a big fan of those devices, I was spending some time to find alternatives. I was able to source all parts either from AliExpress or DigiKey.

The parts listed here are susceptible to failure on SC or SN boards, respectively, and SS boards.

Use this information at your own risk! This is my result of doing research and experiments. So far, all of the Plasmas I have fixed are still running fine after 1-2 years. If the parts were fake, they would have blown up after a short time.

You can never be sure whether a chinese seller has the same quality parts on stock after some years. Take my tipps as a starting point. I wrote this article in early 2017.

New Kid in Town? TGD40N40P

I found yet another potential alternative to all the IGBT transistors: TGD30N40P.

This type is manufactured in Korea and is around since 2012 according to the datasheet. I will test them soon. Stay tuned.

I think that the small number of sources on Aliexpress is a good sign. The market does not yet seem to be swamped with crap copies.

IGBT Transistor GT31F131

Can be replaced with FGD4536. That I know for sure, because I fixed a TX-P50GT30 and a TX-P55VT30 with those successfully. Alas, this device is also discontinued.
My source has dried out. Stephen Foxall found a legit source HERE.

IGBT Transistor DG302

Datasheet
This transistor seems to be the strongest in the bunch. 250A peak and 40A continuous current. It's good to have a bunch of them on stock. I don't know any exact replacement for this one. It's the gold nugget of the circuit.  The FDG4536 might be a viable candidate. It is difficult to tell because the interval for the 250A peak current is not specified in the minimal data, which is available for the DG302. The FGD goes up to 220A for half-sine, pulse-width 1µsec and its switching times are even lower than the DG's.

In the meantime I have tested a 42VT30 with the FGD instead of DG and it runs perfectly fine. The FGD run at around 55°C with the original heat sink glued on top and that is totally normal.

I am happy to finally share a source of legit DG302 on Aliexpress! They measure exactly like the originals and one did work flawlessly recently in a 42VT30. 

IGBT Transistor RJP30H2A

Datasheet
Again, I fail to see the distinctive difference to the DG302 und 31F131 besides the 5A less collector current. I recon they could have built the whole thing with DGs exclusively. This type isn't used in the 50 series anymore. I have replaced them successfully with FGD4536. Even Panasonic uses an 31F131 instead in the scan board of the TX-P55VT30.

Generally, the problem with the data given in the sheets for the maximum  pulse current is difficult to compare as the manufactures use different pulse lengths. The F131 is rated for 3µsec, the DG302 datasheet doesn't tell anything about it, and the FGD4536 is rated for <1µsec. This leaves the hobbyist with trial and error as the only option.

Diode DAF30 (DA3DF30A)

Datasheet
Can be replaced with STTH20R04G. The datasheets are a perfect match. If you want more juice, the STTH30R04G will deliver it, it's a monster diode. I recently used STTH20 twice as substitute and it runs perfectly fine with normal temperature.

A STTH30 recently also worked fine in a 42VT30.

Diode RFUH25

Datasheet
Why Panasonic is using this diode alongside the DAF30 is a mystery to me. The specs read the same. Maybe there is some subtle detail I don't understand. I think the STTH will fit here, too.

Diode RF1501N

Datasheet

And yet another diode, which looks the same as the others. From the specs I cannot see any significant difference to the DAF30 and RFUH25. The DAF is a few nanoseconds quicker at recovering. They are dirt cheap and available from DigiKey or Mouser or even cheaper HERE on Aliexpress. The chinese source is legit. I have tested and used the diodes successfully.

Control Chip BD8639FVM (SMD Marking D86)

Bought some from THIS source without problems. Update: has been reported to have become a bad source.

Driver Transistors CPH5524 (SMD Marking 3Y)

Bought some from THIS source without problems.

Dell 2709W Ultra Sharp - blue power LED constantly on - unresponsive - main board reflowed

I have a DELL 2410 and I like it a lot. When I saw the defect 2709 on eBay for a reasonable price, I thought I could try a monitor for a change.

The defect showed as follows:

When plugged in, the blue power LED came on immediately and none of the other buttons reacted to touching. The device was totally unresponsive.

In this thread on Badcaps I learned that reflowing the main board revived some of the sets. Before that I did my voltage check routine and found nothing suspicious. The 23.5V were there and all the secondary voltages looked good as well.

The right chip under the heat sink was warm, the left was stone cold and its clock wasn't oscillating. It seemed as if the right main BGA chip would activate the left after booting.

So, my brand new reflow controller got something to prove its value. It is a clever device, which controls a normal el cheapo pizza oven precisely via a feedback sensor, which I taped directly on the BGA chip.

Before I ran the standard profile, I poured some liquid flux under the chip. I don't know if this had a part in the success, but I thought it wouldn't hurt.

And voilà! The monitor ran flawlessly again. Hallo Frau Johansson!

During my research I came across an interesting software, which is able to reset the factory defaults in a monitor. It's called softMCCS and is available here: http://www.entechtaiwan.com/lib/softmccs.shtm

It identified my two monitors. First, the defect 2709 would be found, but DDC/CI was marked as "not supported". After the repair, it read "supported". So this is a nice tool to check whether the processor is working ok or something else is broken.

Rest in Peace

14 months later the error came back. A second reflow did not fix it. I also learned that the larger electrolytic capacitors did not like the oven heat. They were bloated afterwards. I'll remember that for my next reflow task.

LG 50LN5406 - flashes shortly - LED backlight defect

I finally got my hands on one of the LG trouble children. The TV flashed the LG logo very briefly and then switched off again. The seller had a repair shop diagnose it and they said that the LED backlight was broken. Their repair offer was 320€. Wow. That's more than the used price of this device.

The LG of this age all die sooner or later due to one or more dead LEDs. The backlight in standard setup is too aggressively set up and they overheat.

I did not bother measuring anything and went straight on to a backlight repair.

Back cover off, nothing spectacular to see:

Step 1: Remove the front bezel:

Step 2: Unscrew the metal panel bezel, unplug the flat cables from the TCON board and carefully slide the buffer boards out of their seat:

Step 3: Remove the panel. This is the most delicate part of the operation. A 50 inch panel bends a lot when not handled correctly. I always wear gloves to avoid finger prints. Do not force a blade or similar objects between the panel and the frame to get a grip on the panel! A simple trick is to first unclip the black frame, raise it a little and then reach under the panel. That way it is easy to lift the panel with one arm diagonally supporting it from below and the other hand holding one edge to keep bending under control.

Be careful with stuck buffer boards. Make sure they slip out without force.

Step 4: Remove the panel frame and the foils. This is simple. Just wrap a tape around somewhere to keep them together.

Step 5: Remove the reflector sheet to get access to the strips. The four plastic standoffs come of easily when you squeeze them with pliers from the backside.

Okay. Measuring time! A good strip settles at 27.4 Volts. The bad strip yielded 36.1 Volts and it was obvious why. One LED was dead.

I removed the right half of the strip with a spatula, taking care of the double-sided adhesive tape, which will be reused.

Now, my new preheater got something to work on. The plan I had was to heat the strip up until the LED's solder would melt. Nope, not working. The material is too thick and a large copper trace shielded the heat from the top. Anyway, the lens came off very nicely under heat. The LED was so heavily damaged that it broke into crumbles. I removed what was left of it with the iron.

This looked terrible. The cathode pad was burned black all over. It almost seemed as if the soldering went wrong on this one and it heated up too much.

I cleaned it with a fiber glass pen and some Isopropanol:

This looked much better. Next time I will remove the sticker, too, before I work on the LED :-P

The SEOUL 2835 3V 1W fit perfectly. I put some leaded solder on the pads and the strip back on the heater. With just a short heat shot from the top with my heat gun at 260°C the LED snapped into place by itself on the molten solder.

Click here to buy the original LEDs. The LG Innotek LEDs perform identically to the SEOUL, but some might feel more comfortable with the exact match.

I glued the lens with plastic glue. Super glue did not quite do the job.

And here we are, all shiny and new! The new LED was indistinguishable from the others:

Evangeline Lilly is looking happy!

What to do to prolong the life of this TV

Turn the backlight intensity down to 50%! The image is still good enough and the LEDs are substantially less under stress.