[vsnet-chat 2625] Re: Tmz85

[Taichi Kato <tkato>]

Re: [vsnet-chat 2621] Re: Tmz85

> Kato-san has raised a good point.  If, for a faint source,
>   (a) all data is collected with the same telescope/camera/filter
>   (b) no intrinsic shape variations change the light curve shape
>         from one cycle to the next
>   (c) no systematic errors are present in your magnitude extractions
> then you can take short exposures, phase your data, and smooth
> the data to improve the signal/noise in the mean light curve.

    The second restriction doesn't always apply to CV eclipses.  CVs
naturally show intrinsic variations (flickering), which play a similar
role in getting an averaged eclipse profile.  Most of quiescent eclipse
observations of dwarf novae are strongly affected by these variations,
so it is usually necessary to average several to tens of eclipse
observations.  In such a situation, high time-resolution is more essential
than the high S/N of individaul data points.

> photoelectric era to obtain the mean light curve of the Crab Pulsar.

    This is another good example.  Even a meter-class telescope can receive
less than a few photons in each sampling bins.

>   However, keep the three restrictions in mind.  I won't go into
> the statistical reasons, but even accepted psf programs like DAOPHOT
> have systematic errors at low signal/noise or bright sky background.

    The same situation is almost always met in infrared photometry, in
which the sky usually overwhelms the faint source, and the exposure time
is severely limited by the noise.   It isn't therefore an unrealistic
requirement.

Regards,
Taichi Kato

VSNET Home Page

Return to Daisaku Nogami

---

vsnet-adm@kusastro.kyoto-u.ac.jp