LG 37LH7000 - LVDS and panel connector contact problems and TCON board defect

Every TV repair is like a journey into the unknown. It's just never boring.

A friend asked me to find her a not too big TV and I got a LG 37LH7000 for 45€ on eBay. The seller said it had green stripes across the screen. The photo on eBay wasn't very clear but I was confident that it could not be a defect in the panel, and everything else is fixable.

When I switched it on at home, I got this:

Damned! This very much looked  like a panel failure, because it had horizontal lines in it, and a rule of thumb of mine says that for horizontal structures the panel is very likely to blame. However, the vertical unregular blocks indicated a TCON problem.

I also heard a dzzzzzzt, dzzzzzt buzzing sound from the back in regular intervals. The small horizontal white noise band at the top of the screen reacted to activity on the menu button. A good sign. So there was a signal arriving at the TCON and from there on it wasn't processed properly.

First rule of TV repair: don't give up too quick.

I first wiggled the panel and applied pressure to the frame at various positions. Sometimes, the thermally bonded connections to the panel go bad and external pressure can make them work again. Not here, though.

I then removed the panel connectors and LVDS connectors, applied contact spray to them, re-seated everything, and the image was back - almost:

The buzzing was still there and corresponded perfectly with a regularly appearing noise pattern, which you can see at the top. I then fiddled with one of the TCON connectors some more and the TV seemed to work flawlessly that evening and I celebrated my cheap repair. The next day I switched it on and the noise pattern was back.

It was time to dig into the TCON board. I noticed that the buzzing sound won't disappear once the panel was disconnected. Good news again, because the panel could finally be ruled out as the problem maker.

I poked around on the board with my scope to find something which correlates with the noise pattern. I found that the positive charge pump of the TPS65161 DC/DC chip, which provides all voltages for the panel, was dropping out exactly in the same interval as the noise appeared. A clicking sound on the board accompanied each drop-out.

This is what a good charge pump output looks like:

My repair journey became a little strenuous and frustrating then. I first replaced the TPS chip, because it's dirt cheap (3€). That didn't improve things. Then I unsoldered the ceramic caps, which smooth the charge pulses. One measured a little strange, but testing the board with an electrolytic cap instead neither improved the situation. All diodes and transistors measured ok.

Close up of the TPS65161:

I then unsoldered a zero ohm resistor, which was conveniently placed - presumably as a fuse - in series with the charge pump's output. This made an easy current measurement possible. Measuring currents on a tightly packed SMD board is a very difficult task, which can not be performed without destructively cutting microscopic traces. I like to avoid that as long as possible.
There were 60mA flowing until the pump dropped out. According to the data sheet this was already too much, and the panel had not even been connected, yet! Something was drawing excess current on the TCON. At that point I gave up, because this discovery pointed towards the big processor chip.

I ordered a new board from flattvparts.co.uk (excellent shop, by the way) for 50€ including shipping.

The new board measured ok without drop-outs, but the regular noise pattern was still there! What the...!! In a mild rush of desperation I used a glass fiber pen to polish the LVDS connector once again, this time also at the mainboard, where they were literally baked to the socket and went loose with a snap. Well, some of the gold plating is gone now, but what the heck. I applied silicon grease to help it against corrosion, a trick I saw in a youtube video by @norcal.

A glass fiber pen is actually a little too abrasive for delicate contacts:

In his book LCD Repair Tips 4.0, Kent Liew mentioned using an eraser to polish the LVDS connector. I did not have any suitable eraser at hand (who still uses pencils these days...). I think I'll stay with the glass fiber pen. Used gently, it is able to polish contacts without removing too much material.

The TV is stable now. It feels a little unfinished, because I'd rather be able to say "this part was broken" instead of  "this contact should work now".

I've learned a lot from this journey. A bad LVDS contact, which is not detectable even under a 10x magnifier glass, and won't go away even with the best contact spray, can produce the wildest phenomenon on a screen. And it most likely killed the TCON board with the intermittent pulses that resulted from it.

About the device and the service manual

The device's build quality is mixed. The boards contain Japanese quality caps. I measured the ESR of all caps on the power supply and they were perfect. The ones on the main board, which I could reach, were also in good condition. The frame is solid and the HDMI ports are screwed to the back cover.

On the other hand, the wiring is sloppy like in many brands, besides Sony and Panasonic. After 5+ years, the tape glue fails and everything starts falling off. It looked like a mess in there.

The service manual has scalable schematics of the main board. The power board, inverter board and TCON (as always) are missing.

Update

Oh boy, this TV kept me busy! After two weeks the funny lines on top appeared again and my friend even reported temporary "split screen". As it turned out, the LVDS cable was to blame. One of the contact traces had lost a little piece. I fiddled in a spare cable and since then it is working ok again. Let's see for how long. I also took the opportunity to completely tear apart the panel. I cleaned the diffusor plate, which had accumulated dust. This had shown as dark bands and blotches in the image.

Update 2

This thing is driving me nuts!! After it had worked flawlessly for a week, I moved it aside for a while and didn't use it. When I switched it back on, the noise patterns appeared. This is a total mystery. It seems as if something is going bad when it is not used. Normally, that would be capacitors. I measured all of them without findings. All voltages are good. Neither could I find anything in the LVDS signal, which correlates with the noise. I think I'll give up on it.

Update 3

I bought another defect set of the same type in hope to swap boards. Alas, the second TV had the same fault :-( It turns out that the only factor was the running time. All the things I did had no correlation to the actual fault. The more I worked on the thing, the better it would run. I swapped the main and TCON with no success. It must have been a scan driver error in the panel.

Screw that, I removed all the boards from both sets and the rest went into the trash.