[Phys-L] Re: light bulb transient

[Brian Whatcott <betwys1@SBCGLOBAL.NET>]

At 09:41 PM 11/14/2005, Mike, you wrote:

> John Sohl made some measurements and posted them
>  on a web page.  I have
> done the same.  My results can be found at
>
> www.bluffton.edu/~edmistonm/light.bulb.pdf
>
> My results are similar to John's but different.
>
> (1) I don't understand the oscillations in John's plots.  If the time
> scales are correct, the oscillation period is about 200 ms giving a
> frequency of about 5 Hz.  This might seem to be a vibrating filament
> except John's fluorescent bulb also showed the 5-Hz behavior.  That
> makes me think his time scale is incorrect.
>
> My filament clealry showed 120-Hz oscillations in light output caused by
> the 60-Hz AC power.  The magnitude of these oscillations is considerably
> more than I would have guessed.  I wasn't sure they would be visible at
> all, but they amounted to about 18% of the average light output.
>
> (2)  John saw his bulbs heat faster than they cooled during turn-on
> versus turn-off.  I saw the opposite in terms of overall behavior.
> However, in the 120-Hz oscillations I observed the cool-down was
> noticeably slower than the warm-up.
>
>
> Michael D. Edmiston, Ph.D.

I salute the experimental data adduced by John and Mike.
I question the use of AC if the impetus is to investigate localized
overheating in bright spots at the turn on transient though.
  DC would show variations a little better?  In the experimental
  subjects,  it is clear that turn on flash is not seen.

Page 18 of this tech report from GE R&D touches on the topic
of filament hot spots.
    <http://crd.ge.com/cooltechnologies/pdf/1998crd027.pdf>

It carries an interesting reductio ad absurdam, the gist of
which follows:
    if filament evaporation were uniform along its length,
the expected life would be infinite (!)



Brian Whatcott    Altus OK    Eureka!
_______________________________________________
Phys-L mailing list
Phys-L@electron.physics.buffalo.edu
https://www.physics.buffalo.edu/mailman/listinfo/phys-l