Hacker News Reader

Comment by throwaway290 | Hacker News Reader

And only one of them is directly pertinent. None of the "decades of data" takes into account correct exploitation. But all talks about leds are about perfect spherical cow in vacuum that doesn't exist for average consumer

I am sure leds technically do live longer than bulbs. But the difference is not significant enough in real life.

View on HN

Look up Dubai Philips LEDs. The problem with most consumer LEDs is that they are overdriven, so their life is short. In Dubai, the Sheikh basically mandated that the bulbs need to be underdriven, so the lifetime is about 25x that of incandescents. They only deliver about 1/4 the power to the individual “filament” that most similar Philips lamps would at the same light output.

https://hackaday.com/2021/01/17/leds-from-dubai-the-royal-li...

That is interesting.

Since only apparently anecdotes count as pertinent for you, here’s one. I haven’t bought a bulb in over 12 years. My LEDs have simply not gone out. I had changed bulbs nearly every 6 months before that.

Maybe spring the extra couple of dollars and get high quality LEDs.

Not anecdotes but practical data. Yes if you power cycle tungsten all the time at maximum brightness they will not live long (when I was a kid we used them this way and changed them often). Read my comment about correct exploitation

Leds themselves are often fine for long time. It is the circuits they are powered by that are very often crap, poorly designed, specified, too cheap. So heat can kill it. Or like my last cheap powerful bulb in kitchen that flickers when I have certain controlled resistive loads on.

I wonder how many tests are run in actual enclosures for example. Which for example might not dissipate enough heat.

One of the dead lights that died was above the stove so could be heat related. Other 2 maybe just bad circuits...

I think it is too expensive to run tests in real life changing conditions

Reply on news.ycombinator.com