|
|
International Occultation Timing Association * European Section * |
|
|
International Occultation Timing Association * European Section * |
Abstract
For the use in stellar occultation work a fast, transportable and
relatively cheap camera system has been developed. It is based on the
TC245 frame transfer chip (Texas Instruments). The control program
running in a PC allows real time monitoring of the images as well as
data storage on harddisk. For small image sizes, up to 18 images per
seconds can be recorded. The readout noise of the camera is around 25
e- at a chip temperature of about -30 deg C. The timing of the
images is done by receiving the DCF 77 time signal radio station. In
the report recordings and analysis of lunar occultations, of the
occultation of nu2 SGR and of PPM269153 by Jupiter will be shown to
demonstrate the versatility of the system. First tests on photometric
measurements of planetary satellites will be reported.
Introduction
Most astronomical objects require long exposure times (up to many
hours) to collect enough photons for a sufficient signal to noise
ratio. Their light intensity remains constant over time or their
variation is slow compared on a hourly or daily time scale. However,
there are objects with changing structure and brightness on a much
smaller time scale. Variable stars such as fast eclipsing binaries,
cataclysmic variables or flare stars can change their brightness
sometimes in a few seconds of minutes by a factor of ten or more.
Occultations of stars by the moon, planets and asteroids can only be
observed by CCD systems with a time response in the subsecond range.
For the mutual events of satellites of the big planets (Jupiter,
Saturn etc.), the necessary time resolution is a little lower, but if
possible a minimum resolution of 1 second should be achieved.
Most CCD systems used in astronomy up to now are very specialized to
extremly high sensitivity light detection. They are large area CCD's
(up to 7000 times 9000 pixels!) with readout times of up to many
minutes. Many of them can not be used in a fast recording mode, even
if smaller subareas can be selected. Sometimes the rotation of the
earth can be used, to smear out the image on the chip in the
direction of rectascension and to have a better time resolution. But
the 2-dimensional information is lost, and severe problems may rise,
if a high background (or a lunar of planetary limb) is close to the
object.
On the other side, there exist a large number of systems with normal
video readout, commercially available video cameras or camcorders.
Up to now, these systems deliver a analog read out video signal,
which gives problems for correct intensity calibration and linearity.
The images (50 half frames per second) have to be digitized and
transfered to a computer. With careful linerity checks, they can be
used in some cases, but if the object intensity is small or the
wavelength range to be used is restricted (p.e. the 20nm width
methane band range around 890nm to suppress the jovian light scatter
for mutual event measurment) large telescopes may be needed. The
exposure time per image can not be increased, however by adding more
than one frame to an image after digitization the effective exposure
time can be longer.
There exist a large number of CCD cameras designed for the amateur or
public observatory as well as for small university observatories,
which have smaller sized chips, thermoelectric cooling (instead of
liquid nitrogen) and are connected to a standard PC. However, most
of theese cameras are not designed for image acquisition rates faster
than 1 image per second.
This situation has provoked us to design a small area chip CCD camera
with an image acquisition rate of up to 20 images per second for
small subareas, with thermoelectric cooling, direct computer
interface with a analog/digital converter board, exact timing by a
radio time signal station and a real time software to display the
images immediatly during acquisition. In our report we discuss the
technical data of the system and show some application from
occultation work.
Camera Design, Technical Data and Software
The IOTA Occultation Camera (IOC) has been developed around a
commercially available CCD chip for use in vido cameras in order to
keep the price of the system as low as possible. This allows a broad
distribution of these systems among amateurs, public observatories
and small professional observatories as well. The system is mainly
focussed on the observation of occultations, but is nevertheless
suitable for long time exposures up to many minutes. In order to
have a high quantum efficiency (up to 65%) combined with a good blue
wavelength response without coating or thinning, a chip from the
Texas Instruments series in virtual phase technology with full frame
transfer has been selected. Balancing price and performance, the
TC245 has been chosen. It has an area large enough even for the deep
sky observers under the amateur astronomers as well. It has an
on-chip implemented double correlated sample and hold circuit (DCSH)
and a full frame store section, to allow a fast transfer of the image
for read out and at the same time to give a high quantum efficiency.
Due to the masked image storage area, a mechanical shutter is not
necessary.
The chip is cooled with a single stage thermoelectrical cooler to
give a temperature roughly 50 deg lesser than the outside temperature
of the housing. A glas window in front of the chip closes the housing
to prevent the formation of ice on the chip. The housing can either
be cooled by free convection, with a small fan (enforced air cooling)
or with water or an other coolant.
The read-out noise of the chip is around 25 electrons, a value which
is not too good compared to more modern CCD arrays, but gives a
reasonable compromise. The electronic has to be designed to be of
high linearity and reproducibility. The output signal is amplified
from about 4aV per electron to about 100aV. Together with the A/D
converter card (12 Bits, 7asec conversion time, for the ISA Bus
system) it results in approx. 10 electrons per ADU.
The chip is controled directly by the computer -any PC from 386SX
onwards can be used- without any microcontroler. Only the voltage
levels have to be converted, from TTL level (0V and +5V) to the more
complex CCD chip levels. This is done by analog switches (CD4053),
which provide high speed switching in about 100nsec from -9V to +2V.
By controling the chip entirely by the PC, the software can be
changed easily. Even complex operations, such as single line
spectroscopy with a rate of up to 1000 samples per second can be
programmed with ease. The system time of the computer is set by an
external radio time signal receiver of the DCF77 at a radio frequency
of 77.5 kHz connected to the serial port of the PC. Every minute an
internal routine compares the received signal with the system time of
the PC and corrects it if necessary. This works in an radius of about
1000km to 1500km around Frankfurt/Main, Germany. For other countries
in the world, special arrangements may be with GPS reception have to
be done.
The software written in Fortran 77 and Assembler (time critical parts
of the software) allows the control of the camera during image
recording. The size and position of a subarea on the CCD chip can be
freely selected. The smaller the actual readout area, the faster the
image can be recorded. For an array of about 50 times 50 pixels (far
more than necessary in lunar occultations) up to 18 pictures per
second can be recorded. During program execution, simple measurements
can be done, giving the mean value in a selected area and its
standard deviation in order to control the chip performance. A single
frame mode can be selected to store images in FITS mode.
The following table summarizes the main technical data of the IOC
system:
Sensor: TC245 made by Texas Instruments
Pixelsize: 8.5um x 19.75um
Pixels: 755 x 242 pixels = 6.68mm x 4.82mm
Pixelsize 2x: 19.0um x 19.75um
by binning: 3x: 27.5um x 19.75um
Technology: Frame transfer in virtual phase technology#
Sensitivity: 390nm - 1050nm
Q.E.: max 65%
Readout: Double correlated sample and hold on chip
Output ampl: 4uV per electron
Readout noise: 20 to 30 electrons
Cooling: Single stage thermoelectric cooler
Typical temp: Chip temperature -30deg C
Cooling of housing by free convection, forced
convection or liquid coolant
Preamplifier: Operational ampl., DC coupled, 0.1mV per electron
A/D board: 12-Bit AD774 (Analog Devices) 7asec conv. time
Separate Sample and Hold, ISA-Board
Conversion: 10 electrons per ADU
SOFTWARE:
Max. acquisition rate: 25asec per pixel
Size of subarea free selectable
Position of subarea on chip free selectable
Internal or external binning selectable
Timing: DCF-77 receiver via RS232 interface, update of system time
every full minute
Image rate: 50 x 50 pixels at 18 images per second
Examples and Applications of the IOC
The occultation of nu2 Sagittarii by Jupiter on the 6th of March 1996
was one of the most spectacular occultation events in the last years.
The IOC camera system has been used to record this event from the US
west coast. In order to suppress the brightness of Jupiter compared
to the 4m4 K0 star a methane band filter has been used. It is
centered around 890nm with a HWB of 20nm and a maxcimum transmission
of 87% (Dr. Hugo Anders, Gesellschaft f. Dunne Schichten, Nabburg,
Germany). It improves the contrast of the star relativ to Jupiter by
approx. a factor of 20. The occultation was recorded at the Anza
observatory of the Orange County Astronomers (about 10 miles inland
from Mount Palomar Observatory) with their 55cm Kuhn telescope. In
order to extract the star intensity at the limb of Jupiter the
following approach was done: After dark image subtraction and flat
fielding an unsharp masking filter was applied to all images in the
recorded data file. The parameter of the mask was choosen not to
change the intensity of the stellar image but to sufficiently reject
most of the remaining Jovian image. The position of the star was
marked by visual inspection of all the images and an automatic
process was used to determine the background around the stellar image
and subtract it from the star. The intensity of the star was
calculated for each image and transfered to a data analysis program
to analyze time series data (Beisker, 1994). The data were normalized
for full star intensity set to 1 and zero star intensity set to 0.
Fig.1 shows the reappearence of the star behind the Jovian limb.
A similar approach was done for the occultation of the 8m7 star
PPM269153 observed from the Cuno Hoffmeister Observatory in Namibia
(a C14 on a fixed site) and at the Specola Vaticana (60cm Cassegrain
telescope) in Castel Gandolfo (Italy). The camera clearly recorded
the occultation of this very faint star, the disappearence as
observed in Namibia is shown in Fig. 2.
Besides theese two events, numerous lunar occultations have been
recorded with up to 18 images per second to give a timing accuracy of
up to 0.055 seconds.
After the PHEMU97 meeting in Catania, an expedition was sent in July
1997 to Australia to successfully record at three independent
stations the occultation of a 12m7 star (event TR176) by Triton. As
first attempts to observe mutual events for the PHEMU97 campaign, 11
events have been recorded so far.