Real-world compact fluorescent bulb life

In the last week I’ve replaced three compact fluorescent (CFL) light bulbs. All were rated to last between 8,000 and 12,000 hours; none had been installed for more than two years. Two of the bulbs were manufactured by Philips. One was installed in a seldom-visited garage 1.5 years ago and had been on for perhaps 1 hour per week. So it lasted 75 hours. I don’t think that it was temperature extremes that killed it because it was stamped “Outdoor”.

How are CFL lifetimes calculated? Are they for turning on the light and walking away for a few years? Does cycling them on and off destroy them much sooner than the published lifetime? My main reason for buying CFL bulbs was that I thought I would be spared the hassle of changing light bulbs, but so far I seem to be running around with replacements about as fast as before (except now it is costing 8X as much).

And who has tried the new LED light bulbs? Are they ready for prime time? I am ready to give up on CFL.

36 thoughts on “Real-world compact fluorescent bulb life

  1. AFAIK the ballast are the ones the fail just as often. Particularly in cold weather. Unfortunately with the CFLs the ballasts are part of the package so you can’t replace them as you can with the big fluorescent lights.

  2. I have CFL lamps that are in nightly service and seldom burn out – the ones that do are many years old. An exterior light that is on all night every night needs to be replaced every ~10,000 hours of operation (and although sheltered isn’t exterior-rated), so the temperature extreme (-20°F – +105°F) doesn’t seem to be much of a factor..

    Edison-socket LEDs are awful so far; poor light output, utterly dismal color – I keep returning the few that I try.

  3. My longest-lasting CFL is in a floor lamp that’s always on; it’s past 12,000 hours. The one in our downstairs bathroom does seem to last only 5,000-6,000 hours; a little low, I guess. The ones outside the front door have been in for three years, getting at least 5 hours of use a night, summer and (Canadian) winter.

    Maybe you got a bad batch, or else have some electrical, vibration, or other environmental issues that cause premature failure?

  4. The story of CFLs is a little more complicated.

    I went through and replaced pretty much all the lights in the house, when CFLs first became cheap. (How many years back was that? 5? 7?) What I used was nothing special – whatever was on sale at Home Depot at the time. Of the dozen or so CFLs, I think one may have failed. (On the other hand, I may have just needed to re-tighten the CFL in the socket.) The light on the front porch has been left on enough nights, it could easily be into thousands of hours. (Though southern California winters are not especially cold compared to snow-country standards.)

    My father went through and replaced with CFLs a few years after I did, and has seen many failures.

    Clearly older CFLs can be reliable, and clearly more-recent CFLs are not. There must be some difference in manufacture. Is there any way to distinguish between reliable and too-cheap CFLs?

  5. Harald: Thanks for the data. I don’t think the failures I’ve seen are a “bad batch”; the three bulbs were made by two different companies and were of two different shapes. One was installed in a location 15 miles away from the other two. I think that I’ve had about 6 or 7 failures out of perhaps 40 CFL bulbs installed in the last couple of years. I’m wondering if the bulb life on the package is really “what you can expect if there are no defects in the product” and there are separate failure mechanisms that have nothing to do with the 8,000 or 12,000 hour “wear out” published lifetime.

    My LG Blu-Ray player was probably designed to last for more than playing 20 discs, but it failed all the same!

  6. Interesting. A couple of years ago I’ve replaced all the bulbs in our place for CFLs. Went for the cheaper ones for most of the house. I was quite annoyed when about 30% started to fail after a little bit less than one year. They had moderate use — I estimate about 1000 hours each.

    They all failed with the same symptoms (dark part starts to develop, increases in size and soon after the bulb no longer works). They were all made in China for the Australian brand Mirabella (stay away!). Sent a long complaint to the manufacturer, but never got any answer. So far I’ve been lucky with a few Philips and G.E. bulbs. But it does seem that some in the market are rated for a LOT less than they advertise.

  7. Year ago I’ve installed 6 CFL. Couple of them failed pretty soon after. One installed in fairly humide environment (kitchen) went out in just few days. The other was from some cheap, low quality source and had very low output from the start.

  8. Phil,
    I’ve had the same experience you’ve had. I’ve stopped using fluorescents for the reasons you’ve stated – I’m not getting any more time and they cost much more. There’s also the disposal issue.

  9. I’ve never had a CFL fail, except for one in an unheated garage which failed very quickly in the winter. My guess is that they’re tested at room temperature and don’t tolerate cold very well.

  10. Not to state the obvious, but make sure the are well screwed in before you replace. I have had two that I thought were blown but then when I really tightened them in they started working again. The way this happened and the amount of force required was unexpected because it was different from my experience with incandescents. Maybe the additional weight caused the contact point to slightly distort…?

  11. Screw-base CFL’s are crap ever since production went to China. Intermittent use, cold temperatures, low hours=long-life incandescents.

    LEDs are improving rapidly, but are cost-effective only in high-hour applications. I have used Cree LR6 can replacement modules, and they are excellent. For replacement of regular bulbs, I would try Philips, but I don’t have any personal experience with these.

  12. I purchased two 6 Watt Pharox60 LEDs in November. Overall, it has been a good early adopter experience but there’s room for improvement.

    – Emits an excellent white light comparable (to me) to incandescent
    – Very low power consumption = no guilt lighting
    – The bulb should fit in almost any fixture

    Not So Good:
    – Not as bright as the advertising would have you believe. I would subjectively equate the Pharox60 to a 40 Watt incandescent.
    – It’s picky about dimmer types. I would avoid applications where a dimmer is used since it could damage this very expensive bulb.

    I have used a few other LEDs. They have all had issues with light color (blue or yellow, not white).

  13. I replaced out bathroom halogen (50watt, small pot lights) with early LED lights about 4 years ago. They have been trouble free, except the early models have a distinct blue colour cast. I recently replace all of the kitchen and basement lights (same small 50watt halogen pot lights) with LED; early days, but no problem yet. Newer LED’s are available in a warmer colour. The LED’s are all rated at 7watt, and are somewhat more directional than the halogen floods they replaced. In real terms they seem to be equivalent to about a 35watt Halogen. Still I am very pleased with the result.
    I have had similar luck with CFC’s, they do not seem as robust as other bulbs.

  14. I think that I’ve had about 6 or 7 failures out of perhaps 40 CFL bulbs installed in the last couple of years…

    ..My LG Blu-Ray player was probably designed to last for more than playing 20 discs, but it failed all the same!

    Have you had problems with other devices failing prematurely? This sounds like it might be a voltage spike/floating neutral or other wiring problem.

  15. Scott: Voltage spike? Could it be from switching on the 480V 200 amp aluminum mill that I’ve got in the living room? I think there is could be some inductance there. Oh wait, that was in my old house.

    As noted above, the failures occurred in two different locations separated by 15 miles. I always thought that the power company did a good job at delivering 117V AC, so I don’t own any surge suppressors (except a handful that are bundled into power strips that I wanted to buy for other reasons).

    Equipment plugged in without a surge suppressor and that operated failure-free for 10-20 years in these locations: Sony and Panasonic televisions, incandescent light bulbs (until their service life in hours was exceeded), aquarium pumps, lights, and filters, Dell, IBM, HP, and Toshiba personal computers, Panasonic telephones, microwave ovens, refrigerators, washing machines, dishwashers (except for a Bosch dishwasher that required 8 service visits; the one excuse that Bosch did not offer was NSTAR delivering non-standard voltage), digital photo frames, network-attached storage, Ethernet switches, wireless routers, a Sonos whole-house music system (6 separate boxes, each plugged into what NSTAR recklessly claims is 117V), a couple tube amplifiers and preamplifiers, an LP turntable, multiple CD and DVD players, digital controls and timers on a GE gas kitchen range, clocks, coffee machines with and without digital clocks and controls, coffee grinders, range hoods, furnace igniters, water heater igniters, air conditioners, VHS VCRs (who remembers those?) and treadmill.

    Maybe we should start a new thread for folks who’ve actually experienced a voltage surge at home other than from a nearby lightning strike (not an issue here in Massachusetts though perhaps with another 20 years of global warming we will be getting Texas-style weather).

  16. The “old bulb” hypothesis is possible; all of mine were purchased in bulk several years back. I haven’t had to restock yet, so I have no experience with newer manufacturers, but I’ve seen that trend with almost every other electronic device in the last few decades, so it doesn’t surprise me…

  17. Strictly speaking, they are tested in accordance with IESNA LM-65-01. But that standard costs $15 and I didn’t feel like spending it just to read the procedures. See for the latest US DOE specs.

    Consumer Reports does testing from time to time also. Check them for product ratings. One of their quick bits of guidance is to limit yourself to CFLs that pass Energy Star ratings. The poorly made units often don’t spend the money getting the Energy Star rating because cheaply made units have more problems passing those tests.

    My experience has been that locations with good air flow (cooling) are lasting 4-10,000 hours. The major killers seem to be heat, humidity, and below freezing temperatures. Those three are rapid killers of CFLs.

  18. I’ve had about the same luck with CFCs. a few under daily or constant use seem to be fine, but the infrequently used ones are fairly short lived. Assuming it is thermal expansion wear issues and such with the ballast as stated.

    I have six or so low-watt LED desklamps in my shop, and they seem to be ticking along fine. A tad on the dim and color-variable side, but the voltage use is negligible, and multiples do well for putting light immediately on my projects.

    On a side note, they do seem to be running at some noticeable multiple of 60hz, not as bad as the LED christmas lights going at 30hz, I think maybe exactly at 60hz. My digital cameras get very weird trying to meter and expose with them.

  19. rjh: Hmmm.. heat, humidity, and below freezing temps. I would say that the “outdoor” Philips bulb in the garage was subject to all three of those at various times during its 15-month/75-hour life.

  20. My understanding is that the rating is for continuous usage. Apparently switching on/off drastically reduces the life. I just had three GE CFLs die and I am annoyed enough to write GE and demand replacements. Why pay more at purchase time and *disposal* time if the bulbs don’t actually last longer?

  21. The $7 ones from 2000 lasted 9 years. The $5 ones from 2010 are real flaky. These are very bright 100W replacements, not the standard 60W replacements. As the Chinese optimized parameters over the years, they figured out how to get bulbs which die as fast as incandescent bulbs.

    Unlike CFL lights, you can build LED lights for less than what China is asking if you have the education.

  22. From :

    The ideal use for a CFL is in lighting fixtures, such as table lamps, in which the screw-in end is below the unconfined bulb. Sure enough, all but one of my dead CFLs came from enclosed downward-pointing lights with the screw-in end above the heat-generating bulb. In fact, in the whole house, I counted just three upward-pointing, unenclosed lights that would be appropriate for CFLs. Energy Star’s Web site, states that “CFLs perform best in open fixtures that allow airflow, such as table and floor lamps, wall sconces, pendants, and outdoor fixtures” (see Compact Fluorescent Light Bulbs). The Web site also recommends installing CFLs in fixtures that you use at least 15 minutes at a time or several hours per day, which excludes such areas as closets and laundry rooms.

    Not sure if any of this coincides with your failures; it doesn’t with mine, since the one that fails most often is mounted horizontally :).

  23. I tried some of the off brand cheap CFls at Lowes and their life was worse than a normal light bulb. I switched to GE brand CFLs and have yet to have one burn out. They cost twice what the ones the stores pushes cost but since they last a lot lot longer it is worth it. I can;t find the reference but one note I read said that a good indicator of quality and longevity is the size of the ballast part of the light bulb. The cheapies had a very small ballast section whereas the GE ones are quite large. (Of course not the cheapies will probably just make the plastic part around the ballast bigger without actually improving anything so I’m not sure how long that indicator will work.

  24. Harald: Thanks for the EDN link. I hadn’t thought about it until I saw that article, but the most recent three failures were all downward-pointing bulbs. I have had a handful of failures in upward-pointing bulbs, but they were after a fair number of hours of usage. The fixtures for the failed downward-pointing bulbs were not especially tight and the “Outdoor” bulb that failed was in an entirely bare fixture on the garage ceiling.

  25. I’ve noticed the same problem with downward facing bulbs. Our bathroom fixtures where notorious for failing early since the heat rises from the bulb into the ballast. The design also traps the heat around the base of the bulb, and I could hear the ballast making a lot more noise before each bulb failed. Also, the white plastic had turned a noticeable brown color around the point where the glass exited the base. I’m thinking of reducing the individual wattage size since the aggregate wattage is pretty high.

  26. I have them all over the place. Changed nearly all of my ceiling lights right after we moved in three years ago, at least 8 of them, and they’re all still there–all ceiling mounted with the screw in end facing upwards, above the heat generating lightbulb part. No failures at all. On the flip side, the guys in the hardware department at the local wal mart have fifty plus years of experience in the hardware department, and they all think cfls’ suck, and that the 8 to 12 thousand hour claims are bs. The ones in my ceiling are GE. Luck

  27. most people should eventually realize that CFLs are not the holy grail, that’s when the consumers will be introduced to the next energy silver bullet – the high efficiency incandescent. There appears to be a potential in heat recovery and conversion into light in bulbs using a traditional filament. GE announced they will be just as efficient as CFLs

  28. I have a cheap Ikea CFL in a cheap Ikea hanging paper lantern, it hangs downwards. I have left it on continuously for 3 or more years at least and it is still going. It is however one of the lower power 11-15W ones.

  29. We’ve given up on CFLs. Our experience with about two dozen bulbs is that they don’t last nearly as long as incandescents (typically 3 months between installation and failure in home fixtures) and they are fire hazards.

    One bulb actually caught fire to the point of a continuous flame thereby causing more than a little alarm. The next two we caught smoldering because of the smell.

    The bulbs came from a variety of manufacturers and incandescents have no trouble in the same sockets.

    Our conclusion is that CFLs are yet another feel-good ‘green’ swindle (and no one ever had to dispose of an incandescent by following some protocol).

  30. I switched over around 2002 and about 2 years later about 1/3 died – all from a capacitor failure in the ballast. Home Depot, house brand.

  31. Unless you run these lamps for a minimum of 15 or 20 minutes at each use the life is appreciably shortened. Incandescent’s or LED ‘s are the choice for short run time installations

  32. I’ve had great experience with FEIT CFLs, and correspondingly horrible luck with Lights of America CFLs.

    They have electronics, and just like with any other sort of electronics, quality varies widely.

    Light quality also varies, but overall i like most CFL colors better than incadescent.

Comments are closed.