Observing Procedures
On this page we detail a typical observing sequence, those things the observer usually does
and those things the SA or TO usually do
A few things to remember before you take exposures. It is a good idea to
window the CCD, so that both the CCDs are put into the fits extension 1
SYS> udas_window WHTWFC "[1:4300,1:4200]"
and check that the
CCD is set to the "fast" readout speed for your calibration exposures, and
"slow" for the science exposures (although the difference in readout time and
RON is not large: see the WYFFOS
pages). For observations with SB3, one should strongly
consider binning the CCD by at least 2, as the psf,
even when well focused, is very non-Gaussian. To bin only in the spatial
direction use:
SYS> bin WHTWFC 1 2
although as the fwhm of the lines are so very
wide, no loss of spectral resolution will result from binning also in the
spectral direction.
To acquire your target is it necessary to
move the acquisition bundle into place, using the command, in the
lpvs1 computer:
ICL> isp acq obs
An image of your field of view, which is 22.5 x 30.2 arcsec, will appear (this
is all done by the TO). At
position angle (PA) 0º the orientation on the
TV screen is such that north is up and
east is left (when H down, and V up on the TV switches: if not, the TO will
know what it is). By moving the rotator (ask the TO),
the acquisition (and observing) bundle can be
set to any PA.
The limiting magnitude of the acquisition TV without integration is V~15. Any two of the following filters can be installed in the acquisition system: CLEAR, BG38, GG495, OG530, and RG630.
There are two ways to focus (spatially) the telescope using INTEGRAL. The first is by by visually inspecting the image of a (not too bright) star on the TV acquisition system. A typical telescope focus encoder value is 97.7. This can result in a fairly well focused telescope, which is almost always going to be sufficient for your needs.
However, if you want to be more careful (eg. under good seeing conditions and if you want to have the best spatial resolution possible), there is a second method, which involves reconstructing the image of your focus star. For this you will need to have previously taken and traced a flatfield image for your current setup, so you can then extract (see here) the spectra from the image of your focus star. Note that if you are focused in one bundle, you are in all, therefore chose the bundle which best reflects the seeing conditions at the time (eg. if the seeing is 0.8 arcsec, there is no point using the 2.7 arcsec fibers of SB3), and for which you have a traced flatfield image.
The method for fine tuning the focus requires you to image the data of your focus star and measure its fwhm (ie. essentially the psf of INTEGRAL in that bundle). You take your exposure of the star, extract the spectra (using the flatfield image as the reference image, and the IRAF routine int_apall), reconstruct the image (with imarec or the stand-alone Euro3D visualisation tool), display (eg. on DS9 if not using the visualisation tool), and measure the fwhm (eg. with "imarec" or within the Euro3D visualisation tool), change the focus as necessary, and repeat. See the pages quick look and data reduction for how to use these tasks.
Be aware that the value of focus you get with this method will not be the same as the seeing measurements taken with the monitoring telescope at the ING, as you are measuring the resolution of INTEGRAL, not the telescope. Therefore, it would probably be a good idea to record the atmospheric seeing values over the course of the night.
There is a small off-set between the rotator centre and the position of the central fibres of the science bundles. To determine this offset, and thus centre the science bundle(s), you need to centre a target on the acquisition camera and then find where it is on one of the science bundle's field of view. Do the following:
1. Put the
acquisition bundle into the observing position: ICL> isp acq obs
2. Move
to a bright star (V~8) and place it at the (acquisition)
rotator centre previously determined.
3. Select
your observing bundle (eg. SB2) with the command:
ICL> isp std2 obs
4. Take
a short exposure (~10s): cl> run integral 10
5. Extract the spectra and map the star (int_apall and imarec or
Euro3D visualisation tool).
6. Measure the position of the star on the map with the IRAF commands
imexam or
imcenter.
From the arcsec-to-pixel scale of your map (which is a user input parameter in imarec or the fibre size/interpolation size in the visualisation tool) you know then the offset between what you see in the acquisition camera and where the fibre bundle is pointing (at PA 0º, with the default image display orientation, up should be north and left east on the map).
7. If you wish then to exactly centre the telescope with respect to the science bundles, move the telescope appropriately and re-image, until you have centred the star.
When well centred, mark the position of the star on the TV acquisition screen. This will be the real reference to which to centre your objects during your observing run. Normally, this procedure is not carried out, but if very accurate pointing has to be carried out on many objects during the night, it could would be worth the time.
Centring of a target on the fibre bundle can also be done by visually inspecting the raw images. The pattern of the spectra on the CCD, inspected in conjunction with the fibre distribution pattern, will indicate where your object is (assuming you are not looking at a very complicated field), as the centred object should have spectra in certain fibres only. If you are centred on the star, the central fibre should have the brightest spectrum, and those fibres surrounding the star (which numbers you can get from the fibre-pattern map) should be successively fainter.
Accurate positioning will probably only be necessary for the pseudo-coronographic bundle, where you want to get your bright source on the blocked-trasmission fibres. You should allow yourself some the necessary time to do this. It is also necessary, for this bundle, to have a very good fibre-transmission (twilight sky) flat-field image.
THE MORNING BEFORE
When you first go up to the telescope, the morning before your run, there are various set-up procedures it is wise to carry out then. We will assume that the instrument itself has been correctly installed, the fibre bundles orientated, focused, rotator zero-set, GSS details correct etc. etc. These are all done by the ING staff or the SA. The remaining set up procedures are detailed in the set-up pages, so we do not repeat them here.
Make sure, once you are happy with the set up, that you take a reference arc exposure(s) and lamp flat-field exposure. Calibrate the arc exposure (see additional instructions here and here). The arc is so you can easily later check your wavelength setting during the night (which may not always set to the same wavelength for the same input command). For the flatfield, which you need to extract and inspect any night-time observations, you should run int_apall (see here ) identify and trace the fibres (there is no particular need to also extract the spectra of the flat-field). Retain these images. Decide whether you want to bin the CCD if you are using bundle STD3, as if you all, all calibration images will have to be taken with the CCD similarly binned. Make sure you have a traced flatfield and an arc for each grating/wavelength range you will want during the night.
DURING TWILIGHT
Twilight is the time for the sky flats. These are necessary because the
sky light follows the same path as your astronomical light, and thus
flat-field/throughput
corrections require this image. In this case, of course, the wavelength range
is important; you should be within 10% of your range while
taking these flats. Again, the blue end of the spectrum, where INTEGRALs
response drops off, will be faint compared to the red end. Aim for the usual
30,000 counts or so over as much of the CCD as possible: usually 3 flats are
sufficient. Blue wavelengths first in the evening, red wavelengths
first in the morning. To take sky flats you can use the command
SYS> sky
integral 10
"name"
You need to focus the WHT, either through the acquisition camera or by mapping the focus star. This was detailed here. The pointing checks are done by the TO. You can determine the centring for the science bundles following the procedures which were outlined here. Take a few bias images as well. Darks are not necessary.
OBSERVING DURING THE NIGHT
It is a good idea to take lamp flats every time you change grating during the night, and an arc, as physical changes to WYFFOS or INTEGRAL can change the aperture/spectral traces/focus on the CCD. There are also a few things to consider while observing:
Position Angle
The sky PA of INTEGRAL can be set to any value, by the telescope operator. One
should note, however, that as the telescope follows your object across the
sky, at some point
it will need to stop and rotate 360º, before carrying on observing.
The time this will occur depends on the PA you have set to.
Therefore, ask the TO when it will happen for the current PA, and necessary
and possible, change the PA accordingly.
Meridian Crossing
As your object goes
from rising to setting the atmospheric refraction will be changing rapidly.
It is best to avoid taking very long exposures at this time.
Sky Spectra
The sky fibres are located a distance of 45 arcsec from the centre of the
field of view. If you are observing an extended object, you may not get any
true sky spectrum. For this reason, it may be worthwhile taking a separate sky
exposure (long enough to get reasonable s/n in your sky lines; you can use
these sky lines, eg at 5577Å, to adjust the total sky fluxes to those in
your science exposures). This would be a good idea, also, because of the
variable focus across the CCD, which
leads to spectra of variable
fwhm in the spectral (astronomical or sky) lines. Varying with spectrum (aperture)
*and* with wavelength, this makes sky subtraction in the presence of sky lines
very awkward, as large residuals can remain. If your object has an emission
line at the same wavelength as a sky
line, as you get for PNe for example, this can be a problem. A separate sky
exposure would solve the problem.
Dithering
We also recommend you consider dithering between exposures. This will not
always be necessary, however note that (1) the fibres are not contigous, hence
there are small gaps between them through which flux from your objects will be
lost, and (2) this will be a problem if you are undersampling the seeing,
eg. as occurs when you use the 0.9 arcsec-sized fibres and the seeing is about
or less than this.
As the psf is not properly sampled, the star will not look circular, and will not
have the same shape with wavelength and between exposures. You may also lose a
different percentage of the flux between your flux calibration and
astronomical exposures, as the seeing changes and because the positions of the star(s) on the
fibres will not be the same between astronomical and flux calibration
exposures. The importance of this effect has been roughly calibrated by
the INTEGRAL team, and you can contact us for
the most recent information -- it is not, in any case, large, and will only be
a problem for the most fussy of observing programmes.
Dithering would solve this problem beccause you will move a small amount (less
that 1 fibre), and combining this data afterwards will effectivelly increase
you spatial sampling (although not your resolution). However, it will add
to the data reduction and analysis you will have to do, possibly in a way you
may not wish to, so again, we recommend
you contact us (or others) for advice.
Adjustment of the Rotator Axis and Orientation
This is work carried out by the TO, who knows the system best. To determine the X and Y coordinates of a star on the TV acquisition screen, change the Sky Position Angle (SPA) by 180 degrees. The star describes a semi-circle reaching a final location on the TV (i.e., [Xf,Yf]). The centre of this circle [(X+Xf)/2,(Y+Yf)/2] should be the rotator centre. Repeating this process will lead to a more accurate determination of the rotator centre.
To minimize tracking uncertainties, select a star near the pole.
After the rotator centre has been determined a CALIBRATE should be done.
If you need to check that the guiding bundles are properly mounted, the following check can be done:
i) Locate the slide unit at (65237,
0.). For that use the commands iixs
and iiys.
ii) Align a star at the rotator centre
iii) Applying an off-set of 212 (386)
arcsec (telescope to the south), the star should appear at the centre of the
inner (outer) guiding bundle.
With the handset in XY mode, at SPA=0 degrees, the up button will cause an object to move down on the TV screen while the handset left button will cause an object to move left on the TV screen.
Use the WHT INTEGRALINNER and WHT INTEGRALOUTER configurations of the GSS for finding guide stars within the inner and outer fields.
The ICL commands iixslideabs and iiyslideabs are used to move the XY slide to the GSS indicated position.
INTEGRAL uses its own arc and flat lamps. These are located up on the mechano-like shelves in GHRIL which are part of the INTEGRAL construction. It will not normally be necessary to fiddle about here, but it could be that you need to adjust the brightness levels. The indicators for the arc and white light lamps are provided on the page of pictures. The arc lamps should be set in the 5-10 mA, if all is well (although currently it makes no difference what the dials are set to). The two arc lamps provided are Cu-Ar and Cu-Ne. The on-off switch is located next to the dials. The adjustment knob for the white lamp is extremely sensitive.