[Aavso-photometry] Why bother to calibrate?

[Michael Koppelman]

We do it because we are compelled to eliminate any systematic source  
of uncertainty within our control to increase our confidence in our  
measurements and to make our measurements as similar to those of other  
observers as possible.

In a nutshell, although your random errors may be similar, the  
systematic errors are now uncontrolled. Now your results are affected  
by temperature, illumination issues and characteristics of the chip. I  
think over time you'd start to recognize the influence of these  
systematics during meridian flips, night-to-night zeropoint  
differences and position on the chip.

Like you say, depending on what you are doing, it's possible that none  
of these effect your ToM or ephemeris.

I admit I'm guilty of using old flats sometimes, but I always do bias/ 
dark/flat. Always. :)

M.


On Jan 3, 2008, at 8:07 PM, Tom Richards wrote:

> It's cat-amongst-the-pigeons time. I think I know some answers to the
> following, but I'd like to get a discussion going.
>
> I just did an unfiltered time series on T Pyx, V~15.5, 216 images. I
> measured the images in MaxIm, first without calibrating them, then
> calibrating them (the full bottle: bias, bias-subtracted dark, and
> bias-and-dark-subtracted flat). And it's a horrible-looking flat  
> thanks to
> numerous chip blemishes.
>
> Uncalibrated, the sigma on the check star (V~15.40) was 0.0128 mag.  
> For the
> calibrated images this improved all of 1.5 % to 0.0126 mag, a  
> negligible
> change. And the standard error on the absolute difference in measured
> magnitudes for T Pyx frame by frame, was 0.0011 mag, again  
> negligible. The
> maximum difference was 0.010 mag.
>
> So, for simple time-series work, just finding light curves, why  
> bother to
> calibrate? And a relevant second question: who does not bother to  
> calibrate
> for simple time-series work? Be honest now!