I've been using compact fluorescent lamps almost exclusively for ten years or more. I pay the extra price in the name of the energy efficiency that everyone claims for them. Even so, I've always been skeptical about claims of energy efficiency, especially when I start to think about the energy embedded in CFL lamps. Just looking at the two types of lamps (CFL and incandescent) I have to conclude that the energy it takes to assemble a CFL is many times greater than what it takes to make a standard incandescent bulb.
One of the reasons I've been skeptical about all this is because CFL's, in my experience, never seem to deliver the hours of service claimed by the manufacturers. I don't care how efficient a lamp may be, energy return on energy investment is compromised by poor durability (and greatly so, I think). I rather suspect that, all things considered, in the aggregate I've been using more, not less, energy by using CFLs. Who knows, though? In the summertime my air conditioner would work even harder if it had to deal with extra heat from incandescents. Maybe it's all a wash.
In the past couple of years I've taken to marking each CFL with the date it entered service in order to get a better sense of this. It's stupid, but when a CFL I've marked has failed, I generally just glanced at the date, bolstering my suspicion, and screwed in a replacement CFL.
This morning, though, I took the time to think about this a little bit.
A 26 watt (100 watt equivalent (yeah yeah, I know)) CFL that I had installed on November 9 2005 (480 days ago) failed in the bathroom. The bathroom has to be one of the more severe environments for a light bulb because of the number of on/off cycles as well as total running time. I thought about the usage pattern in that bathroom and figured that the lamp probably experienced about 15 on/off cycles per day, averaging 20 minutes or so of on time per cycle, which came to about 6 hours on per day. Over its cumulative service life of about 2880 hours, this bulb would have gone through about 75 kilowatt hours of energy, costing about $6 at 8 cents per kilowatt hour. It would have gone through about 7200 on/off cycles in the process.
Now, 2880 hours in nowhere near the lifetime claimed by CFL manufacturers. According to the EPA's Energy Star website, these things ought to last 6,000 hours or more.
Well, after a bit of looking around I found a document about durability testing of CFLs. The part that caught my eye was about stress testing, where they cycled CFLs on and off for the lamp's lifetime (five minutes on, five minutes off). What they found was that two manufacturers' lamps failed "very prematurely". Here's the relevant figure (click it for a bigger view):
I generally buy whatever CFLs Costco, Home Depot or Walmart are selling, which I assume are the cheapest things they can find. It wouldn't surprise me to find out that the lamps I buy are from the two sub-par manufacturers. I know the test cycle pattern doesn't match the usage of CFLs installed in my bathroom, but from the results of the test and my own experience, it seems reasonable to look for higher quality lamps. I'll see what the people at Lights of America have to say, and otherwise try to identify higher quality CFLs to buy. If I'm still here in a few years maybe I'll remember to post how the better lamps perform. On the other hand, maybe I'll revert to lower-wattage halogen incandescent bulbs for the bathroom. Apparently moisture can affect CFL's, so maybe I'll go with a higher-quality weatherproof CFL for the bathroom.
Should government ban incandescent bulbs, as is being done in Australia and as California is considering? Well, it seems to me highly likely that if we ban incandescents without, at the same time, banning sub-par CFLs like those in the graph above (presumably what I've been buying), it seems that such a ban would be worse, from a global perspective and considering embedded energy, than doing nothing.