What are these black lines on the pcb? Do they divide the circuits?

Looks like separate low and high voltage areas, the white chips in between are probably optocouplers of some kind. My bet would be maybe bad caps, could be a mosfet too.

These things piss me off because the power boards are generally shit compared to the rest of the set, and all in all it's probably cheaper just to buy a new board, but if you can fix that one and do it well, maybe it will last.

>>1079003
That's what I'm hoping. The set itself is pre led thin line bullshit. It looks great and sounds good enough. I'm looking for insight on what might cause this though. It definitely seems like a thermal issue. If I read correctly you are telling me that those big black lines are isolations between circuits, if so I should only have to test to the left bit of it, seeing how once the back lighting cuts out I end up with everything else working. Does that sound correct?

>Anybody have any ideas where i should look?
it's almost always the electrolytic caps.
checklist for bad caps:
>smaller than a D cell battery?
>rated over 200 volts?
>right next to a heatsink?
score 2 out of 3 and there's 99% chance that cap will test low value/leaky/high ESR.

>>1079021
Shouldn't it be pregnant looking?

>>1079028
visibly burst, leaking or burnt seemed too obvious to mention

>>1078959

It's an isolation barrier. Mostly for safety reasons, no direct electrical connections are allowed across it. Everything is coupled by optoelectronics or electromagnetics only.

Sounds like a thermal issue, as you've ascertained. No good advice to take care of it, since that board doesn't have any obvious temperature sensors on it. I assume it's going into thermal shutdown since it still works after shutting off, rather than going up in smoke.

>>1078959
it shuts off because the backlight tube is worn out and needs too high a voltage to maintain arc. The shutoff is a safety feature.

The board isn't bad, replace the backlight.

>>1079155
What I had originally planned was to change out the ccfl's for led. Anyone have a decent source for something like that? I haven't pulled the cover off the inverter board yet but I bet there are at least 6 tubes. I'm looking for something cheap enough with a driver board included.

>>1079155
I agree with this guy.

I have a couple videos up on youtube showing how to convert ccfl to led.

>>1079288
linx

>>1079304
You'll find plenty of videos with a search on youtube.

>>1079215

if you're gonna try the CCFL to LED conversion, you'd better practice on an old monitor first. the tricky bit is removing the lamps without destroying the metal and plastic bits that hold them in place.

>>1079325
Naw, the hard part is removing and installing the lcd panel with out breaking it.
Most those clips, sockets and holders aren't needed once you go led.
Just need the plastic tips to help keep the glass from getting pushed in.

>>1079021
Yeah, caps are cheap enough you could recap the board and maybe even relocate any caps that might have heat issues. I've repaired several TVs and LCD monitors and gotten much longer life outta them. I do keep a wide range of various caps in my home lab parts cabinets though, so ymmv when it comes to cost/convenience. I'm sure my repairs were less than $10 a board, but I also have hundreds of $ of parts already.

>>1079453
Ah, also fuck I just realized it's a ccfl set, so yeah just do an LED conversion.

>>1079457
>>1079325
If you're going to do a CCFL to LED conversion, kiss your color saturation goodbye.
Sets designed for CCFL backlights can be changed to LEDs, but need special 'component phosphor' LEDs or the color rendering goes to shit. And no, none of those cheap kits on eBay use component phosphor LEDs.

>>1079495
Well what's the best plan here? I have not attempted a conversion yet. I have pulled quite a few bad lcds though. Is there a test to se which ccfl is pulling to much? Like I said it works just fine for a bit. I only paid 15 dollars for it as well. If I match lumens almost to the t would it really make that much difference?

>>1079536
A cheap LED conversion will work, but the colors will suck. They might suck just a little or they may suck a whole lot, but they will be inferior to what the CCFL gave.
Unless you're willing to make a big project of it, just replace the old CCFLs with new CCFLs.
>Is there a test to se which ccfl is pulling to much?
Do you have a scope with a kV probe? If not, no. It's not pulling too much (implying current), it's requiring too high a voltage to push the needed current. CCFLs have a finite life. They wear out. Yours are worn out.
>Like I said it works just fine for a bit.
Yeah, and then shuts off because the arc voltage climbs too high. Strike voltage is higher than the running arc voltage. Eventually it won't even strike.
>If I match lumens almost to the t would it really make that much difference?
It depends. At a minumum, reds will shift to dull orange, greens will become yellowish. If it's an IPS panel, everythign will also take on a sort of cyanish-green hue. That sound ok to you?

>>1079541
>ln-s4695D
just looked that up, I think it's a SPVA. You might get away with white LEDs not sucking too hard then. Your saturation will still suffer, but at least white will probably stay white.

>>1079547
By suffer do you mean get worse or become less saturated?

>>1079796
both, really. The colors will be less accurate and generally less saturated.
The RGB filters in CCFL-based LCDs are tuned for the phosphors and spectral lines but out by the backlight CCFL. Generic white leds put out only two colors: a strong blue peak and a broad yellow peak.

The blue is much stronger and narrower than the CCFL's blue, and it's usaully shifted toward cyan (470-480nm, instead of the 440-450 the CCFL produces).

There's no strong red or green output from a white LED, just a huge broad peak at yellow that leaks through the relatively weak/overlapping red and green filters in the panel. Normally the panels don't have to filter yellow well because the original CCFL output has no yellow.

So even when white is exactly correct, all the colors will be kinda screwed up. They'll still be red green and blue, but not very pure and not very saturated.

>>1079028
Not always. You won't always notice leakage, either, unless you look closely. Sometimes, only a little leaks out, then it dries to the bottom of the cap, which is enough to render the cap useless.

>>1079547
Would RGB LED strips be useful for this sort of thing?

interesting

>>1081279
Sure, but they're less stable than white LEDs designed for this, and you have to monitor and regulate the three colors separately...
Some panels use RGB LEDs, mostly to get highest possible saturation. Chomatic aberration can be really startling with a pure RGB backlight.

>>1079028
No, the caps have just given birth, and are now experiencing postpartum depression.

>>1081777
You are killing me :-)