BARS Observation Report
INT Prime Focus observations February 9th-14th 1996

Observers : J.E. Beckman, J. Cepa, M. Prieto, S. Gottesman, S. Lourenso

PERFORMANCE


Telescope

From 9th to 11th the encoder of the focus was broken, and although a
replacement was set, the problem could not be fixed properly till the 12th.
The result was that the Instrument Control System (ICS) could not control
the telescope focus, and then, the focus had to be set manually by
hardware, the automatic temperature tracking did not work, neither the
{\bf FOCUS} command to take several exposures of the same star in different
positions of the CCD using different focus. The lack of temperatur tracking
was not a severe problem since we were using the second half of the nights,
where the temperature is more stable. Anyway, to focus the telescope was
somewhat slow, since we had to focus again the telescope because we were
using different filters than that used on the first half of the night.
Also on the 10th, the display of the telescope focus was moving by itself
from time to time, up to 1mm, but not the focus itself which did not move.
Some time was lost on this night when trying to recover the lost focus and
then defocussing instead of focussing.

The old autoguider system was not able, from time to time, to detect a
guide star. The cure seems to shutdown the Telescope Control System (TCS)
and to begin again. This caused some more additional time lost.


Instrument

Sometimes, the filter wheel refused to change the filter properly,
and had to be reset or driven to another filter, and then back to
the filter needed.

The ``friend of the astronomer'' i.e. a black cloth wrapping the
Prime Focus Cone Unit (PFCU) and fixed using black sticky tape, was not set.
Before the PFCU rebuilding, it was used to avoid light leaks and their use
was compulsory. After rebuilding, some test seem to indicate that the black
cloth was not necessary any more. We checked that by taking a BIAS frame
with the mirror covers open, the fluorescent lights on, and the R filter in
position, and we could not see any evident structure on the bias, which was 
flat up to 0.4%, with a Read Out Noise (RON) of 4.5 ADU. So, although we
suggest to future observers to check it again, it is not necessary anymore
to put the ``friend of the astronomer''.


Filters

	Broad band set

>From the filter set available at the Isaac Newton Group of Telescopes, we
select the largest (50mm) U filter to avoid vignetting if smaller sizes are
used. The B, V and R Harris were preferred to the Kitt Peak ones even taking 
into account that the efficiency of the B Harris is lower than that of the 
B Kitt Peak, because the interference Kitt Peak set is degrading. 

	Narrow band set

To choose the narrow band filters we gave preference to the following
criteria:

1- To keep the number of filters to a minimum to ease to obtain
flatfields (see next section).
2- To use filters of the same characteristics for on--line and continuum
to ease the continuum subtraction, even if the exposure times had to be
longer.

With these criteria, we took into account:


1- The non--converging beam of the PFCU of the INT causes a shift of the
central wavelength of interference filters of ~ -4 A.
2- The interference filters are designed to work at 10 C. It 
exists a change of central wavelength as a function of the temperature of 
0.3 A per C.
3- The effects of the non--converging beam are not limited to a shift of
the central wavelength of the filters, but the spectral response is also
affected by broadening the filter, which it does no longer resembles a
top--hat, and diminishing the peak transmission. This implies that, if a
redshifted interference Halpha filter is used as a continuum, its
central wavelength has to be well beyond that of the on--line filters to
avoid line emission entering the blue side of the continuum filter.

Then, we select the filters 6595/45 A and 6625/45 A as the on--line
filters, and 6685/45 A as the continuum filter.


Moon Phase

The first nights were bright, and the latter ones gray. Since the first
nights we used the narrow band set, Moon phase do not seem to affect the
frames.


AUXILIARY FRAMES

Flatfields

Since the ICS runs on a Perkin Elmer 3220 instead of a Sparcstation, the
CCD readout time is more than 2 minutes for the Tektronics 3 chip (1124x1124)
in quick readout mode. ``This is not a bug, but a feature''. This implies
that, till a Sparc20 substitutes the elderly Perkin Elmer, then achieving a
readout time of 40 seconds instead of more than 2 minutes, it is very
difficult to obtain good sky twilight flatfields for all the six filters of 
the filter wheel in one observing night. Given that, we chose to keep the
change of filters to a minimum to be able to use the flatfields of previous
nights. Then from 9th to 12th we used three Halpha filters (two for the
line and another for the continuum) together with R (Harris), I (Kitt Peak)
and Z (Gunn). The 13th and 14th the Halpha set was substituted by U
(CuSO4, 50mm), B (Harris) and V (Harris). The ITP team of the first part of 
the night used always R, I and Z, and did the flatfields of this filters during
sunset twilight, and we did the flatfields of the other filters during dawn
twilights, so that a suitable set of flatfields could be obtained, and we
can use the R and I flatfields of the first ITP team. 

Until a modern computer replaces the Perkin Elmer, we might suggest the
following procedure for taking dawn (sunset) twilight flatfields: 


* Point the telescope to a ``blank'' field (their coordinates are on 
the INT Prime Manual) near (far of) the Sun shortly (say less than 30 min.) 
after (before) the beggining (ending) of the astronomical morning (evening) 
twilight. 
* Window the chip to get a, for example, 50x50 pixels window near the 
origin. 
* Change the display to get a full colour range from 0 to 65000 ADU.
This allows to estimate the count level of the flatfield at once.
* Take {\bf WINK} exposures of one minute or more (1 second to few 
seconds) till the counts begin to rise above the bias (to go below 
saturation) level.
* Unwindow the chip.
* Take a {\bf RUN} exposure of the same exposure time than the trial.
* Further exposing times may be guessed by an experienced
observer, or determined using the paper by Tyson & Gal (1993, AJ 105, 1206)
* When the required exposure time goes below or around one second (one
minute), the time to take flatfields may be extended by slewing the telescope 
to another blank field further away (nearer) the Sun.

Some general comments:

1- Do not take flatfields of less than one second exposure since due to the
shutter speed, one side of the chip will be more exposed than the opposite 
side.
2- It is convenient to take several flatfields in the same filter to
make a master flatfield cosmic ray--free by means of a scaled median. 
Also, it is advisable to take several flatfields in the same filter of the 
same blank zone, shifted few tens of arcsecond each on different directions, to
easily remove the stars that might be present by taking a scaled median.
The flatfields may be taken on differents nights as long as the filter
wheel has not been touched.
3- Good flatfields has from 20000 through 40000 counts, although
flatfields of lower or higher count may be used with caution.
4- The order at which the flatfields for the different filters should be
taken are: R, V, I, B, U and narrow band (narrow band, U, B, I, V, R for
sunset twilight), for the Tektronics CCD using the Broad band Harris or
Kitt Peak set. Although given the long readout times, it is difficult to
obtain good flatfields for B, V and I filters on a single twilight.

Although we were told by our Support Astronomer (in spirit, at least), that
dome flatfields could be used for narrow band filters, we prefered not to
take the risk, and get as many as possible sky flatfields. In general, sky
flatfields are recommended, because dome flatfields (at the INT and 
JKT) have structures, at least using broad band filters. If necessary, dome
flatfields for narrow band filters may be taken with mirror covers closed,
telescope pointing to the zenith or opposite to the dome shutter, dome lights 
closed (all and always), and with the dome slightly open to get some Sun
indirect light inside.


Standards

To calibrate broad band observations, a field (SA104) which includes several 
faint Landolt photometric UBVRI standards (Landolt 1992, AJ 104, 340)
of different colours (to derive the colour correction for the U (CuSO4) 
filter) was observed in all broad band filters twice each night in which broad 
band observations were taken.

The spectrophotometric standards Feige34 and Feige66 (Massey et al.
1988, ApJ 328, 315) were observed in all the Halpha filters every night
in which Halpha observations were taken to calibrate the narrow band 
filters. 

For extinction corrections we will take the V extinction measured by the 
Carlsberg Meridian Circle (CAMC).


Exposure times

We decided to take 1800 seconds exposures through the narrow band filters
and the U broad band (initially we planned to do two 1800 seconds exposures
for the U band, but time constrains due to weather conditions forced us to
limit to a single half an hour exposure), 1200 to 1800, 900 and 600 seconds 
for the B, V and (R, I) filters, respectively, repeating exposures using
shorter exposure times in case of saturation of the nuclear zone.


NIGHTLY COMMENTS 


February 9th

The first part of the night was clear, but the second part was misty, with
humidity rising periodically above the security level (90\%). Although the
dome was opened during the intervals at which the humidity dropped, no
useful observations could be taken because these intervals were too short, 
and the night was lost (well, not completely because we could determine the 
focus shift from broad band to narrow band filters with the new provisional 
encoder reading, and to estimate good exposure times for standards). 


February 10th

Some time lost at the beggining of the second half of the night due to
focus reading shifting by itself although it was not really moving (only
the reading was moving). The first objects could not be observed due to 
the unability of the autoguider to find guide stars or to guide on them.

We observed NGC 4151, NGC 4596 and NGC 5383 in Halpha and their
corresponding continuum. NGC 4151 is impressive, to avoid saturation due to
the nucleus we took, apart from the 1800 seconds exposure, several shorter
exposures, and we had to take an exposure as short as 15 seconds to avoid
saturation!.  

The seeing was around 2.2 arcseconds at the beggining of the second half of
the night, going down to 1.3 arcseconds after a couple of hours. The whole 
night was photometric. Sky clear at dawn.


February 11th

No instrumental problems.

We observed NGC 3359, NGC 4123, NGC 4340 and NGC 5850 in Halpha and
their corresponding continuum. NGC 3359 and NGC 4340 looked very beautiful
(sorry for the subjective comment). The nucleus of NGC 4123 was saturated
so that we took an additional 300 second exposure.

The seeing started at 1.7 arcseconds at the beggining of the second half of
the night, going down to 1.5 arcseconds after $\sim$four hours, and
worsening at the end of the night. The whole night was photometric. Sky 
clear at dawn.


February 12th

No new instrumental problems.

We observed NGC4314 in H$\alpha$ and their corresponding continuous. NGC3359, NGC4123, NGC4596 and NGC4151 were observed in R and I filters. NGC4314 showed very interesting ring around the nucleus in Halpha filter. All of them were saturated in the nucleus and we took several short additional exposures. We observed the field of photometric standards around Landot104334 in R and I filters. We don't need to do a color correction in these filters and so don't need to observe the calibration stars in all photometric bands. We did flatfield for the Halpha filters in the morning.

The seeing was 1.2 arcseconds the first part  and 1.9 arseconds from TU=5.


February 13th

We observed NGC3359, NGC4123, NGC5384, NGC5850, NGC4596, NGC4340, NGC4145 in U filter. We observed the Landot104 field in UBVR and I filters at the begining of the observations to do the color correction in U filters. We taken just two flatfield in U filter.

The seeing was about 2.8 arcsecond during the night.


February 14th


The seeing was very bad at the begining of the observations, about 3-4 arcseconds.
We observed the field of calibration stars in UBVRI filters and NGC4314 in U filter. When observing the galaxy and the exposition time was 1749 seconds we had to stop the exposition for clouds. This image can be not very reliable. The clouds kept the rest of the nigth and the temperature going down untill -6 degree. We only could do flatfield of dome iluminated for a lamp in UBVRI filters.   


Lost due to weather conditions: 2 (1/2 nigths). 

HOME