Y'know, Mythbusters actually did a segment on that. They looked at how long lights of different types need to be on to "amortize" the startup power surge. The specific target myth was the one that says "if you're leaving a room with fluorescent lights for longer than N minutes, it costs more energy to switch the lights off and turn them back on when you come back in than it does to just leave them on" (for various cited values of N ranging from about 3 to 15).
CFLs were, I think, the longest of the types they tested ... but it still only took something on the order of one second.
no subject
CFLs were, I think, the longest of the types they tested ... but it still only took something on the order of one second.