[Aavso-photometry] Re: Lightbox usable for Ic filter flats - Summary of the list responses

[Wolfgang Renz]

Hello

First a summary of the responses I got.

Bruce Gary:
I ran into the same problem of needing very long exposures for B-filter when using
an incandescent bulb in a light box. I had better results with a halogen flood lamp.
My plan was to upgrade to one of those "sunlight lamps" that are touted in magazine
ads. I abandoned light box flats when I rotated the light box and got different flats. I
concluded that there are many ways to get incorrect light box flats. So I switched to
illuminating a white screen with another white screen and a halogen flood lamp. This
was a lot of work, and sometimes the screen was not uniformly illuminated. So I
switched again, this time to sunset zenith flats, sometimes with a T-shirt cover (to
prevent stars from showing up).
My advice for anyone relying upon light box flats is to do a reality check using sunset
zenith flats. 
Doing good flats is difficult! One of my life's goal is to learn how to make better flats.
Good luck with your light boxes.

Gary Walker:
I have done some experiments with white LED's, and the ones in particular, are
made by Linnrose Co.  They seem to have no emission lines, but a continuous
spectrum when viewed thru a simple spectrascope--I use the one sold my Learning
Technologies Inc.
I use 12 of them, facing away from the white milky plastic. Exposures of 8 seconds
give about 16,000 counts in V and R.  In B the same exposure gives about 4000
counts, and in I, it gives 1200 counts.  
I too have thought about adding a blue and IR LED, but have only looked quickly at
a blue one. No lines visible in spectrascope. Use of mixed colors probably means
adding an adjustment circuit or potentiometer, which is easy.
I would reccommend that you look at any white LED's that one uses for flat fielding
with a spectrascope.  Some are made up of 3 emission lines in RGB, which would
not be good for flatfielding.  Not all White LED's are made the same.

Arne Henden:
I continue to voice my opposition to using white LEDs for photometric flatfield light
sources.  There are two generic types of such LEDs:
   (1) the blue LED with phosphor coating.  This gives a sharp peak in the 400nm
       range, followed by a dip and then a broader peak towards 550nm.
   (2) the RGB LED, formed by having three LEDs (red, green, blue) in the same
       package and illuminated simultaneously. This gives three peaks with valleys.
Gary, if you have a spectrum of the Linnrose LEDs that covers 300nm to 1000nm,
I'd like to see it.
Remember, basic LEDs are narrow-band illumination sources. They do not mimic
blackbody emitters, whereas stars are much closer to blackbodies. Either type of
white LED has zero response at U, a widely varying response across B, and zero
response at Ic.
They are not good sources, especially at the blue and red wavelength tails of the
CCD response, where small changes in the lamp illumination have quite large
effects in interpixel response factors.
It is *far* better to use an incadescent source.  For VRI, almost any small grain-of-
wheat lamp will give sufficient, uniform illumination. for U and B, you often need
more output to give a decent flat in a short exposure; you can accomplish this with
a second bank of lamps that have more output. You can further filter those brighter
lamps to push the mean output into the blue with just blue colored plastic. Much
easier than worrying about nonuniform sources like LEDs.
The color imaging folks can get by with LED lightboxes because the RGB filter
responses are covered by the LEDs; they don't worry about photometric accuracy;
they don't worry about very blue or very red wavelengths. Don't fall into that trap
when attempting photometry, except perhaps if you are only using V and R filters
(where the white LED is relatively smooth response and the CCD pixels have
almost uniform response to light as well).

Lew Cook:
Has anyone thought of combining both white light LED's and small incandescent
bulbs in a light box? Seems like this would give good results in the blue end of the
spectrum and partially fill in gaps in the spectrum, where they exist.

Siegfried Bergthal on wide field sky flats:
Why not make four flats, rotating the camera by 90° in between and averaging the
four flats.

More to come.

Cear skies
  Wolfgang

-- 
Wolfgang Renz, Karlsruhe, Germany
Rz.BAV = WRe.vsnet = RWG.AAVSO