If you're an avid fan of EEVBlog, and have been following Mr. Dave's Youtube videos, then you would be aware of his latest defective 4K Dumpster TV find. The main GUI was functioning hunky dory (hope I spelled that correctly), but the video stream was spewing a senseless array of colors.  

After downloading a technical manual on the product and re-heating suspect BGAs, he managed to pinpoint the chip causing all the trouble. It was the LG1614-A1, a pivot where all interconnections to other components on the board converge.

Below is a download link to the service manual (I think LG already unclassified the document since the product has been in the market for quite some time now):

My personal opinion is that these pins (below) in the LG1614 were the ones that had a cold solder or bad connection to the board.

The other portions of the LG1614 were routed to supplies of various voltage regulators (AVDD and DVDD) and DDR ports (with the DDR's reference voltages) so they're most likely not the culprits. It's most definitely not the SCL/SDA (I2C) or CS/DI/DO/SCLK (SPI) ports because they're (most probably open-drain ports) used to communicate with the IC with a serial interface (like an FTDI). The 4 TMODE pins are for setting (or entering?) test mode. The UART and crystal oscillator (most likely a Clapp) are out of the question. That leaves us with the HDMI pins. (After all, a famous detective once said - "Once you've factored out the impossible, whatever remains, however improbable, must be the truth.")

After Mr. Dave re-heated the chip with a handy hot air re-work station and hot plate, the result of an image check on the TV changed dramatically. With high hopes that the problem would be solved after a few more iterations of heating, the TV, in an unfortunate turn of events, no longer entered the main GUI after boot-up.

But the troubleshooting session performed by Mr. Dave was commendable and it is noteworthy for future reference and debug (Watch the video below!!).