🚀 go-pugleaf

In article <Bx93z4.n2B@cs.dal.ca> legrady@ug.cs.dal.ca (Tom Legrady) writes:
>
>I don't know about the exact relationship of heat and lifetime for
>electronics comnponents, but there is an inverse exponential effect on
>the life of high-intensity lamps, such as the ones used for stage
>lighting. I have to fudge the exact numbers, since I've been out of
>theatre for a number of years, but dropping a lamp to 90% full
>intensity generally doubles the expected life of a lamp. Dropping to
>80% results in an unlimited expected life. Since our reaction to light
>is non-linear, dropping to 90% has little effect on preceived
>brightness. Since the lamps used cost $30 - $50  or more apiece and
>have an expected life ranging from 300 - 2000 hours, doubling the life
>is a desirable things to do. Of course, if you drop it, it still dies. 

What about photographic floodlights?  These things have expected lifetimes
of just 2-8 hours.  I thought if I ever use the lamps and bulbs that I have
I would build a control box that allows the lamps to be dimmed until the
moment they are needed (for a still photo), then increased to full intensity,
then dimmed again.  That would save on heat, lamp lifetime, power
consumption, etc.  Several ways to dim a lamp come to mind:

  1. Two identical ones in series
  2. Diode in series (halfwave rectification)
  3. Electronic dimmer.

But there are a couple of things I am not sure about.  Lamp dimmers cause
light bulbs to buzz, surely not a good thing for a fragile, overheated
filament you are trying to extend the life of.  The diode approach also
messes with the AC waveform.  Two in series sounds pretty good, but when
I switch over, the lamps will be briefly off, then fired up to full
intensity.  Would this switching transient and associated rapid heating
or cooling of the filament be just as bad for the bulb as leaving it on
high intensity in the first place?  I'm guessing this because most bulbs
I've seen burn out did so when turned on.

Anyone have any experience in this?

Markus Wandel
markus@pinetree.org <-- NOT the source of this posting