One dead LED killed a working TV

A television repair technician repeatedly revives apparently dead, expensive TVs by bypassing a fault caused by one failed backlight LED. The panel itself still works, but the firmware detects an LED current outside a narrow tolerance and shuts down the entire set. LED aging is an ordinary, expected failure; treating it as a fatal fault turns a minor repair into the loss of a working television. The piece cautions readers not to open a TV because dangerous voltage can remain after unplugging it, and focuses instead on the firmware design. It argues that the shutdown need not imply deliberate planned obsolescence: whatever the manufacturer’s motive, the engineering mistake is classifying a recoverable condition as fatal.

The broader lesson is to classify failures before choosing an error-handling mechanism. Bugs, such as violated preconditions, should be exposed with assertions or contracts and generally stop execution; expected problems, such as invalid input or an unavailable file, belong in ordinary control flow, potentially represented with C++’s std::expected; genuinely rare failures that a distant caller can handle are appropriate for exceptions. The television’s LED fault belongs in the second category: firmware could log it, continue operating, adjust brightness, and signal that service is needed. A hard shutdown should be reserved for genuinely dangerous conditions such as overheating or overvoltage. Choosing exceptions, error codes, or other mechanisms is secondary to deciding what kind of failure occurred.