ISIS Setup
Overview
Roughly one month before the observations take place you should have
read the observer proposal which is located at /home/ingpatt. The
proposal should contain all the details of the required configuration
(CCDs, dichroic, gratings, order sorting filters and central
wavelengths). If this is not the case you should request a detailed instrument setup in the email that you send to the observer. You should also check that the
required configuration is met by the engineering schedules. If this is
not the case contact the WHT manager.
Starting the observing system
- Log on the main system terminal TAURUS. Check the username and
password on computer screen.
- Type startobssys on the orange window and select any
option (check the white board for the right one).
- Now, type obssys and startobssys
on the pink window
selecting the right option again. This will bring up some extra
windows.
- A detailed procedure can be found here.
Physical setup: Dichroic, dekker and gratings
At the control room check that no one is moving the telescope. For
security reasons turn the telescope power off, lock it with one of the
keys (don't forget to unlock it at the end!) and write it down at the WHT
log book. You can now proceed to change the dichroic and dekker (note
that even if the dichroic or dekker unit you need is the one on the mimic
you should check that the dichroic and dekker units are physically in.)
To make the change pull the dichroic out of the beam for easy access:
TO> bfold 0
Then protect the slit unit with the dekker:
TO> dekker 1
Then unlock the slit door:
TO> slit_door open
You can now open the slit door located on top of the red cryostat. Inside
you will have access to the dichroic, filters and dekker. Remove the filter slide (RFILTA) to protect filters from being accidentally damaged. Pull the pin
that holds the dichroic and carefully slide the dichroic out. Put the
dichroic back in its sealed box at the bottom of the WHT cabinet.
Take the new dichroic out of its box and carefully slide it in
and lock its pin holder.
Proceed in a similar way with the dekker unit. Just pull it out, put
the dekker back in its box and introduce the other one in the unit. Be
careful to arrive to the end and that the dekker is locked.
Close the slit door securely by snapping the locks. Finally lock the slit door
in the control window:
TO> slit_door close
and put the dekker at long slit position:
TO> dekker 8
and update the mimic:
TO> setdichroic <dichroic name>
TO> setdekkerset <dekker name>
The dekker name can be either observing, standard or polarisation.
Here is an image of available dekkers.
For the standard ISIS runs the observing dekker should be used. It
has the same clear part as the dekker standard but also eliminates ghosts in
the blue arm caused by a dichroic.
Now click on the "Update filters" button in both the "ISIS Observer" and "ISIS Eng." tabs of the Instrument Control Console so that the dekker drop-down menus are updated.
Now you can proceed to change the gratings, e.g.
TO> setgrating red R1200R
to change the red grating to the R1200R.
Then answer yes to access the grating door. The doors are then unlocked and
can be opened manually. To release a grating hold it by its handles, push the "open" switch and gently pull it out from its holder. Place it back in its box and insert the new one.
You must ensure that when you load a grating in
the grating cell that
the blaze arrow on the back of the grating points towards the left
for the red arm, and towards the right for the blue arm, i.e. gratings should be inserted
blaze-to-collimator. If the grating is inserted in the opposite sense, i.e.
blaze-to-camera, then throughput is reduced by up to an order-of-magnitude.
Never touch the surface of a grating.
Once a grating is in place hold it by its handles and push the "close" button
(place your fingers away from the three grating clamps!). Finally, securely close the grating doors by snapping the door locks.
You are now ready to align
the dispersion axis with the CCD columns, ensure that the CCD is
not tilted with respect to the incident beam, focus the
spectrograph, and finally check that the collimator position for the
spectral focus also produces a good spatial focus, e.g. that there
is no astigmatism. It is good practice
to firstly check the focus and CCD tilt, and then the rotation
(so that lines used by the setup scripts are reasonably and uniformly sharp
across the CCD). If you are going to use a dichroic, check
the rotation with the dichroic deployed since the dichroics
can introduce some image rotation, especially in the blue arm.
Be aware that configuring
the setup is an iterative process, and that there is "cross-talk"
between e.g. rotating the cryostat and and adjusting the cryostat tilt, and
vice-versa. Therefore if you need to modify the tilt after
setting the rotation, re-check the rotation again. After the final
mechanical intervention for rotation and tilt, tighten bolts and capstan
clamps, and make a final check that rotation and tilt remain acceptable.
Finally, with the spectrograph
anastigmatic (see below) fine-tune the focus by adjusting
the collimator (this process does not involve mechanical intervention
with the cryostat).
Set the
slit width to ~0.5-0.7 arcsec; don't close the slit more than this since
the intrinsic line-width will not be fully sampled on the CCD.
TO> slitarc 0.7
and set the red
and blue central wavelengths to the values requested by
the observers, e.g.
TO> cenwave red 7500
CCD rotation
For a well-aligned cryostat you should aim to have the pixel shifts
between top-bottom as reported by the rotation script <0.5 pixel, preferably
<0.25 pixel. Good alignment is not only necessary to observe extended
sources; for faint point sources it helps the spectrum to be traced and
extracted without significant degradation of signal-to-noise.
Start by closing any IRAF+DS9 sessions (but not the RTD ones) and open a new session by typing
iraf
in a terminal window. Then load the
isis package within IRAF,
ecl> isis
When aligning the rotation of the cryostat you should also set the dichroic to the
correct position, since the dichroics especially in the blue
arm can introduce a rotation of the image [use: in (bfold
2), out (bfold 0), or mirror (bfold 1)].
TO> agcomp
TO> complamps cune+cuar
and take a test exposure:
TO> glance red 10
This exposure will allow you to compute the exposure time to obtain
reasonably intense arc lines and to define the CCD windows such that no
dekker vignetting is seen.
This is critical to obtain good results when
using the calibration scripts (the rotation measurement is done by
analysing an arc image. The CCD must therefore be correctly windowed in the spatial direction so that the arc lines cover the full spatial extent of the window).
TO> window red 1 "[585:1550,1:4200]"
(should be OK for red arm+REDPLUS)
TO> window blue 1 "[585:1550,1:4200]"
(should be OK for blue arm+EEV12)
TO> window red 1 "[1:1072,440:739]"
(should be OK for red arm+QUCAM)
TO> window blue 1 "[1:1072,440:739]"
(should be OK for blue arm+QUCAM)
TO> window red 1 "[1:1124,190:840]"
(should be OK for red arm+Tek4)
TO> window red 1 "[650:1750,1:4700]"
(should be OK for red arm+MARCONI)
Take an exposure, making sure no line is saturated.
Do not bin the CCD.
TO> arc red 2 "Rotation test"
Within an IRAF window (opening a terminal and typing 'iraf' will do the trick), run the isis_rotation task with usemode=1.
isis> isis_rotation r234567
The
isis_rotation.cl IRAF script extracts the spectra of an arc image in two windows separated spatially and cross-correlates them to find any pixel shift. The shift is then converted into a detector rotation offset in micrometer units.
A common output of the rotation script is:
pixel shifts between top and bottom : 1.08
corresponding angle in degrees : 0.0935
corresponding micrometer offset in mm : -0.282
The pixel shifts are the important numbers; the other two numbers are estimates from a factor conversion and may be ocassionally wrong (e.g. if the cryostat is rotated with respect to the nominal position).
Loosen the cryostat rotation bolts, which attach the cryostat mounting
ring to the spectrograph mounting ring, with an Allen key. Never loosen
the three recessed bolts, which fix the spectrograph mounting ring to
the spectrograph. Apply the
recommended micrometer offset to the cryostat, clockwise or anti-clockwise as
required, and re-tighten ~3 of these bolts.
Iterate this procedure until the rotation is within limit.
When both rotation and tilt are acceptable retighten all rotation bolts
with the Allen key. In order to
not overtighten them, engage the long end of the Allen key
in each bolt, and rotate the short end. Finally, back-off the
rotation micrometer so it is not in contact with the stop to
ensure pick-up noise is not transmitted through it to the cryostat.
Bear in mind that the spectral and spatial directions are NOT perpendicular in ISIS. Therefore, rotation of the cryostat to set the arc lines horizontal (which is what most observers want to simplify sky subtraction), makes the spectrum of a star not precisely aligned with the
CCD columns. The effect is best noticeable with the low dispersion gratings.
If
alignment of the dispersion with the columns is paramount, this can be
checked by taking a tungsten lamp spectrum with a narrow dekker, e.g.
0.3", deployed and running the rotation script in Option 2 (tungsten
lamp).
Note also that for the 1200 gratings, arc lines are slightly curved
by approximately 2 pixels from centre-to-edge of the standard window.
CCD tilt
Take two exposures, one with the left Hartmann shutter closed and one
with the right Hartmann shutter closed. Do not move the Hartmann shutter
until the frame has been read out. The order in which the exposures are
taken is not important. Make sure that spectral lines are not
saturated and that the collimator is in the anastigmatic range (see CCD
focus below). Do not bin
the CCD.
Using the setup done for the rotation test, you can use one of the
following scripts:
1-
this script to take the
necessary images in both arms
2-
this script to take the
necessary images with the red arm only
3-
this script to take the
necessary images with the blue arm only
This is basically what the scripts do:
TO> rhart l
(left Hartmann shutter closed)
TO> arc red 4 "test hart l"
TO> rhart r
(right Hartmann shutter closed)
TO> arc red 4 "test hart r"
TO> rhart 0
(Both Hartmann shutters open)
Now analyze the data with IRAF; first display one of
the spectra taken with the red arm and note the
y-coordinates of three, well-spaced spectral lines, i.e. near the bottom, centre
and top of the detector. The positions need be accurate only
to within 10 pixels. Try to choose lines that are strong, that are
not doublets or blends, and that are not too close to other lines. Run
the isis_tilt task, e.g.
isis> isis_tilt r234568 r234569 y1 y2 y3
where y1,y2,y3 are the y-coordinates of the three fiducial lines, e.g.
500,2000,3500 respectively.
Apply the recommended capstan turns, if any, in order A,B,C.
The capstan micrometers should be backed-off when turning the capstans.
You should aim to have the top-bottom and left-right tilts
restricted to <0.5 pixel, preferably <0.25 pixel.
To
adjust capstan A, loosen its clamp, and loosen
slightly the clamps of capstans
B and C (to prevent any buildup of mechanical tension when applying the adjustments
to capstan A), while supporting the cryostat with your other hand. Then proceed similarly with capstans B and C.
(see also
the sa notes). When adjustments
are finished, ensure all three capstan clamps are tight, but don't
overtighten to avoid introducing further tilt in the system. It's
good practice to take and analyse a further set of spectra to ensure
the tilt remains within acceptable limits after final tightening.
Never completely loosen all three capstan clamps simultaneously.
Proceed similarly for the blue arm.
When you're happy with the capstan settings, back-off their associated
micrometers again if you've used them to measure the capstan positions, to avoid transmission of pick-up noise to the cryostat.
CCD focus
The nominal collimator values with no extra refractive components
(dichroics, filters, polarisation module) between the slit and the
collimators are 5700μ for the blue arm, and 9100μ for the
red arm. The spectrograph should be focused with the collimators
within 3000μ of these nominal values, otherwise
it will be astigmatic due to de-collimation of the beam
incident on the grating. If the spectrograph were astigmatic the best
focus on a spectral line
would result in a degradation of the spatial resolution along the slit,
and hence of signal-to-noise in the extracted spectrum.
If the optimum spectral focus set by moving the collimator
would move the collimator beyond
this nominal range, then instead the spectrograph should be focused by moving the detector
using the capstans
so that the collimator remains within the anastigmatic range. So, conceptually the focus
procedure involves setting the red and blue collimators
to 9100 and 5700 respectively with no additional optics in the beam,
reducing the Hartmann shifts as reported by the focus script to <1 pixel
by adjusting the respective cryostat capstans, and
then fine-tuning the focus to its optimum value by adjusting the
respective collimator values.
If you subsequently deploy a dichroic or filter then the appropriate
focus offset must be applied manually by
adjusting the collimator.
Alternatively,
if you focus the red arm with a dichroic or filter deployed, then
the anastigmatic range of the collimator
is incremented from that quoted above by the corresponding focus offset.
Note that a dichroic does not introduce a focus offset in the blue
arm because the blue light is merely reflected off the dichroic surface.
Therefore, the focus procedure is:
TO> rcoll 9100
TO> bcoll 5700
Take two exposures, one with the left Hartmann shutter closed and one
with the right Hartmann shutter closed. Don't bin the CCD, don't move
the Hartmann shutter
until the frame has been read out, and ensure that spectral lines are
not
saturated. For example:
TO> rhart l
(left Hartmann shutter closed)
TO> arc red 4 "test hart l"
TO> rhart r
(right Hartmann shutter closed)
TO> arc red 4 "test hart r"
TO> rhart 0
(Both Hartmann shutters open)
Now run the isis_focus task
isis> isis_focus r234568 r234569
If the Hartmann shift reported is >1 pixel turn each cryostat
capstan by the
recommended amount (see
sa_notes). Alternatively, if the Hartmann shift reported
is <1 pixel
apply the recommended collimator shift, i.e.
TO> rcoll "value"
TO> bcoll "value"
where value=original_value+shift.
Repeat this sequence until the recommended collimator shift is
200&mu or less and the collimator value is in the anastigmatic
range. Finally, you can confirm that the spectrograph is indeed
anastigmatic by taking a tungsten spectrum with a narrow
dekker, e.g. 0.3", deployed and checking that the spatial extent of
this spectrum is commensurate with the dekker, but experience shows
that if the collimator settings
are set as described the spectrograph will be anastigmatic.
Please, DON'T FORGET
TO NOTE the A,B,C, & D setting DOWN in the CCD capstan settings
folder
Re-focusing when changing dichroics
When the spectrograph is focused
there are fixed offsets for the red collimator in case the dichroic is
changed
(see
useful information)
Setup for spectropolarimetry
For spectropolarimetry focus the spectrograph as described above without the
polarimetric module (half-wave plate, quarter-wave plate,
calcite) in
the beam, i.e. set
TO> rcoll 9100
TO> bcoll 5700
and proceed with the set-up in the usual way.
Once the set-up is ready introduce the order sorter filter and
polarimetric modules
TO> fcp calcite
TO> hwin
(or qwin depending if linear or circular
polarimetry will be done)
and offset the red collimator by 9260 and the blue collimator by 9580.
Using the hartmann test check the final focus.
