To be clear, I'm not looking for LED bulb recommendations based on CRI etc., there are plenty of those on Hacker News and elsewhere. I'd prefer data that is publicly available (like the Backblaze stats) but I'm fine with something behind a paywall like Consumer Reports if the data is accurate and up-to-date. Again, I'm looking for failure statistics based on a large number of bulbs in actual use (i.e. hundreds or thousands), not anecdotes or recommendations based on a small sample size.
Here's some background and explanation:
I'm not a professional electrician or lighting specialist, just a regular person who lives in an apartment, and I'm tired of replacing LED bulbs that start flickering after roughly a year or service or even just a few months. Sure, many of the LED bulbs have lasted since I moved in six years ago, but many have not. I'm aware of the general reason why LED bulbs fail: the manufacturers want to minimize component count so they use fewer LEDs and drive them at higher power, causing them to run hotter and burn out faster, especially if the passive cooling from the heatsink is inadequate. (There are other reasons, too, like poor quality control on capacitors or other components besides the LEDs, bad soldering, etc.) I also know that enclosed fixtures and dimmers are a risk factor, but those are not a factor in the replacements I've done in my apartment. I also know that power fluctuations or poor sockets can be an issue, but many of the bulbs here are still fine after more than six years, and the ones that have burnt out are not isolated to a particular socket, suggesting this is an issue with the bulbs, not the power quality or sockets. I'd also like to emphasize this is not restricted to a brand: I've seen premature failures from Cree, Maxxima, GE, Intertek, Maxlite, and Phillips brand bulbs. Nor is it a particular brightness or form factor, I've had failures from lumens ranging from 480, 800, 830, and 880 lumen bulbs. I've seen failures from bulbs using LEDs mounted on a flat PCB and also from filament-style LEDs, so bulb design is not necessarily reflective of failure rate.
I have a subscription to Consumer Reports and normally I would use their light bulb buying guide, but it hasn't been updated since 2017 and their website says "Consumer Reports is not currently testing Lightbulbs". A lot of their recommendations are no longer manufactured and some of them are wildly overpriced on Amazon, probably because they're selling old stock and know people are buying based on the Consumer Reports guide. The whole LED bulb situation right now seems like an Akerlof-style market for lemons, where buyers have no way to know bulb failure rates and so manufacturers have no incentive to improve them. I'd like to change that if possible for my own purchasing decisions, but until I have actual data I don't see a way to do that.
bigyabai•2h ago