Next: Remote Observing and Automatic FTP on Kitt Peak
Previous: Displaying Drift Scanning Images from Multiple CCDs in Real-Time During Data Acquisition
Table of Contents --- Search ---
PS reprint
V. F. Shvartsman, I. N. Bernstein, G. M. Beskin, V. N. Komarova,
S. I. Neizvestny, V. L. Plokhotnichenko, M. Yu. Popova, A. V. Zhuravkov
Special Astrophysical Observatory, Nizhniy Arkhyz, Karachai-Circassia,
357147, Russia, E-mail: beskin@sao.stavropol.su
channels (space, colors,
polarization, etc.) simultaneously.
Special software was developed to investigate variability on a
time scale of
sec by statistical analysis of
the time interval distribution between detected photons and ``classical''
light curves. The results of some observations are presented.
Research of fast optical variability of relativistic and variable objects has
being carried out at the Special Astrophysical Observatory of the Russian
Academy
of Science since 1972 in frames of the MANIA (Multichannel Analysis of
Nanosecond Intensity Alterations) experiment.
Its goal is to study the energy
transformation in strong gravitational and/or magnetic fields which is
manifested by very fast optical variability on time scales from
up
to
sec (Shvartsman 1977). Special hard- and software is used
to investigate nonstationary processes on flare stars, pulsars,
X-ray binaries, etc.
To study such a fast variability a special photometrical hard/software complex was created which consists of:
channels simultaneously;
ns, the dead time is 20 ns;
the equipment supports an external synchronization by a precise
standard time service.
sec and, therefore, the dead time of ``MANIA''
complex is
sec as well.
Testing and installation of a coordinate-sensitive detector
are being carried out now. Moreover, any device which can provide a registration
of individual photons can be used as a detector as well.
While studying variability on a time scale smaller than the mean interval
between photocounts the classical
methods for light curve analysis are not effective due to very small
fluxes and, therefore, very large data sizes.
For these purposes a special, so called
-function, method
was worked out (Shvartsman 1977).
The
-function method is intended for variability analyses on the
time scale smaller than the mean interval between photocounts. It is based
on statistical analysis of the time intervals between photons.
The
-function is defined as:

where
is the fraction of intervals of duration
from
to
in the flux from the object;
is the same for the standard flux.
The
-function method is used on times larger that the mean interval
between photons and is an analog of the variances method.
The
-function is defined as:

where
and
are the sample dispersions
of the number of photocounts
and
in the window of duration
of the object and standard flux respectively
and
is the expectation value of the
.
As a standard flux we can use either the flux from a comparison star
or the constant Poisson flux with the mean flux equal to the object's one.
It can be shown that there is a connection between the
variability parameters (amplitude, filling factor, time, etc.),
and the forms of the
-
and
-functions allow us to estimate the parameters. In the
presence of a variability the
-,
-functions show an increase
on times less than the characteristic time of the variability.
If a variability is absent
we can find the upper limits for the relative power of a variable component
with 99.9% probability (
level) on timescales from
up
to
sec. The limits are calculated based on the dispersion of the
-,
-functions of the comparison star.
The algorithms of analysis are described in (Plokhotnichenko 1983, 1992).
The special FLL method for period search was developed (Plokhotnichenko 1992). In this method we choose a ``test'' period and build a set of consecutive light curves with it. We will sum all these light curves afterwards. But due to the difference between the test and the real periods the light curve details will not coincide in phase and they will be broadened in the summarized light curve. To avoid this we summarize the light curves taking all possible values of phase shift and get a set of final summed light curves. Then by means of statistical methods, we find the light curve with the most significant deviation from the noise (poisson) light curve. The period this curve is folded with is the closest to the true one.
These methods are used in the case of a pulsar period known with sufficient accuracy to build a pulsar light curve with a well-shaped profile. Using data obtained in close moments of time we build a set of light curves. The uncertainties in the period cause the phase shift in each light curve relative to the first one. The value of this shift is calculated by the minimization of the squared difference of each couple. We express this phase shift as a function of time, expanding it into Taylor series. Finally we calculate the values of pulsar frequency with its two first derivatives which are used to build the pulsar light curve with a high time resolution. This light curve is used afterwards in search for fine structure and in investigation of pulsar photometric features.
To study variability in a wide time range by classical methods (light curves, FFT and correlation analysis) a special software system was created---Light Curve Analyses (LCA) system for ``Quantochron'' data analysis. It runs under X Windows on Linux (UNIX) systems using the XView library. It allows working with the data in an interactive mode.
to
sec were detected
(Shvartsman et al. 1989a; Beskin et al. 1994).
to
sec in the high optical state was revealed
(Bartolini et al. 1994).
sec were registered
and analyzed (Shvartsman et al. 1988a).
sec was analyzed (Shvartsman et al. 1988).
This work was partially supported by ESO's Support Programme for Central and East Europe (grant A-02-023), by the Scientific and Educational Center ``Cosmion'', by the Russian Ministry of Science and by the Russian Foundation of Fundamental Research (grant 95-02-0368).
Bartolini, C., Guarnieri, A., Piccioni, A., Beskin, G., & Neizvestny, S. 1994, ApJS, 92, 455
Shvartsman, V. F. 1977, Soobsh. SAO, 19, 5
Shvartsman, V. F., Beskin, G. M., & Plokhotnichenko, V. L. 1988, in The Physics of Neutron Stars, Pulsars and Bursters ( Leningrad, 1988), 178
Shvartsman, V. F., Beskin, G. M., Gershberg, R. E., Plokhotnichenko, V. L., & Pustilnik, L. A. 1988a, SvA Lett., 14, 233
Shvartsman, V. F., Beskin, G. M., & Mitronova, S. N. 1989, SvA Lett., 15, 337
Shvartsman, V. F., Beskin, G. M., Mitronova, S. N. Neizvestny, S. I., & Plokhotnichenko, V. L. 1989a, SvA Lett., 15, 590
Vikul'ev, V. V., Zin'kovskij, V. V., Levitan, B. I., Nazarenko, A. F., & Neizvestny, S. I. 1991, Astrofiz.Issled. (Izv.SAO), 33, 158
Plokhotnichenko, V. L. 1983, Bulletin of the Special Astr. Observ., 38, 29
Plokhotnichenko, V. L. 1992, PhD Thesis
Zhuravkov, A. V., Pimonov, A. A., & Plokhotnichenko, V. V. 1992, Astrofiz.Issled., 37, 132