+ 1 for waiting for it to fail. Keep in mind that while lamps have an estimated lifespan, there can be a really wide margin. I believe the way they are tested is to burn the lamps in relatively large batches, and when 50% have burned out,
call that the average life (or round to the nearest whatever). With manufacturing being what it is, I'm sure they will all be reasonably close. But in addition, your particular
venue's electrical service (or
dimmer's
trim settings) could have an
effect on this. My
venue gets 127v. Yours could get 119v. This will have some
effect. To confuse matters more, avg lamp life does not necessarily account for how long your fixtures may spend at ~90% rather than FL, which measurably increases lifespan. On the contrary, if you want a little more light reading, study up on the
Halogen Cycle. Running continuously below a certain %
threshold can have an adverse
effect on usable lamp life, in that the
envelope will become clouded with deposited
tungsten and really kill your
intensity. Obviously, there is a
bit of a balance in getting the most out of your lamps, and knowing what is going on in the back-end of things.
Overall, it's just too hard to pinpoint when a lamp will fail, and replacing proactively without potentially/probably wasting a bunch of money. Then we get in to
power fluctuations,
HVAC vibrations, thermal
shock,
etc...
Now, is there a way to tell when they're about to go? If the glass is bubbled (as mentioned above) or if the
filament is sagging close to the glass (I've heard of some people flipping the lamps if this is the case - YMMV), it is a good idea to just
drop a new lamp in. I haven't ever noticed any difference in
intensity between a new lamp and an old one. Especially not to the
point where it would require a replacement. Wiping the dust off the
gel would likely have a larger
effect.
** Now if you see that the lamp pins are scorching/pitting, replace the lamp
and socket right away.