SET UP Procedures
SET UP Procedures
First Part
First part
The filters are used (i) to cut out any second order overlap of the blue spectrum, this being important only at >8000 Å (ii) to cut down on red light if you wish to take very long lamp flats or arcs, which are much brighter in the red than the blue.
ICL> wchange
option? [0] 1
to update the filters. It will ask what filters you have put in.
option? [0] 2
to update the grating. It will ask which grating is present.
option? [0] 3
to update the grating offset value.
Current grating angle offset is xxxx mdeg (in Dec 2004 is was
6850)
New value? [xxx] 6850
Once you have updated the information, it will be displayed on the mimic of lpvs1. This display mimics the WYFFOS and INTEGRAL configuration (eg. if you move the filters in and out, the mimic will show that), including the current gratings, central wavelength etc. Check that the displayed set-up is correct (eg. that the filters are fully in/out of the beam). Green writing appears when things are changing, blue when they have completed the change.
The slit which contains the fibres coming from the bundles must be centred on the CCD
in order to get the spectra from all. This step is best carried out with
the SB1 bundle, which is the largest, although in fact it fits very easily on
the CCD. The dispersion axis is X (lines). The names of the bundles for the purposes of
selecting them are:
SB1->STD1, SB2-> STD2, SB3->EXP5, CORONO->STD3
Note the SB3 and CORONO names!.
You should only need to set the slit translation once during your run, nonetheless it is worth checking it each afternoon. The procedure for checking and changing the slit position is:
2.
If the fibres all are on the CCD, then you don't need to change
anything. If they are not, you will need to use the command,
ICL> wslitt xxxx (enter a value between -4500 and 4500).
This will set
the WYFFOS fibre slit to a specified position in microns using the slit
translation drive (it is an absolute movement, not relative).
We suggest movements of about 1000.
A friendly hint for
those of you not so familiar with IRAF. If when displaying your images, you
find that the edges are not well displayed, you can try
using the parameter "fill+" with "display", or
cl> set
stdimage=imt8192.
Rotation
If a flat-field image shows that the
spectra lie along a slope, rather than being horizontal (although only the
central aperture wil be completely horizontal), it will be
necessary to rotate the cryostat unit. Go up into GHRIL, where you will see the
CCD unit sticking out of the front of the WYFFOS bench. To rotate, you have to
loosen (but not undo) the white plastic screws sticking out.
You then rotate the unit to the left, by whatever amount you judge necessary
(ie. guess). Make sure this does not
extend any cables past their length. Then take another
exposure, check, and repeat the whole process as necessary. If you can align the
spectra such that they are horizontally parallel to 10 pixels or so (ie. the
X-coordinate of an arc line
on the top spectrum is within this value of that of an arc line on the bottom
spectrum), you
have done a very good job.
Once the rotation is good, remember to tighten all the screws.
Tilt
To adjust the tilt is a more awkward job. This adjustment will be necessary
if you need to
(i) improve the focus (noticeable if the focus is bad because arc lines
will be non-Gaussian or even doughnut-shaped), and (ii) to even out
the
focus across the CCD as much as possible. The focus can also be changed by
moving the fibres bundles in and out, which is detailed in the next section.
We recommend that you first roughly set the focus using this second method.
You will probably find the focus value somewhere between +1000 and -1000. Then use the method detailed here.
To change the tilt you need to unclip the orange capstans (see
here),
of which there are
three evenly spaced around the CCD unit
(one immediately to view, the others more hidden from view). Hold the
CCD unit and unclip a particular capstan. This will release a copper-coloured
knob, which we usually turn 1/4 or 1/2 at a time (please beware that
we found it very difficult to return to a previous focus condition on the CCD
when returning to a previous knob position). After turning and
reclipping the capstan into place, you can then take an arc exposure and test
the resulting image for the focus (detailed below).
This job can be fiddly because changing the tilt in one
direction (on one of the capstans) will change the focus on particular regions of the CCD,
but then changing the tilt on another capstan may change the focus on overlapping
parts of the CCD, making it necessary to change of the first tilt direction again.
Note added Dec. 2004: capstan A moves the CCD on a V-shaped plane,
capstan C in and out of a hole.
We found that the best WYFFOS focus could
only be slightly improved for INTEGRAL, but it could a little. Moving capstan
A 1/4 of a turn anti-clockwise degraded the focus in the top-left
diagonal-half of
the 2-CCDs (ie. the 2 CCDs considered as one)
and improved it on the bottom-right diagonal-half. Moving capstan C
likewise improved the focus on the bottom rows and bottom right of the 2-CCDs,
degraded on the top left diagonal-quarter, and left it unchanged
elsewhere.
Capstan B could not be reached without having
to reconfigure the human body.
The aim when adjusting the focus by moving the CCD unit is to improve the fwhm of the arc lines, and to make them as similar as possible over the whole CCD. You will not be able to achieve this 100%, as the focal plane where the CCD is located is not flat; the focus will be worse at the edges (particular the top-left corner). Once you have made a tilt adjustment, take an arc exposure and measure the fwhm (eg. with "imexam") of the line in 9--12 positions evenly spaced over the (2) CCDs. Recording the correspondence between tilt changes and fwhm improvements will allow you to eventually reach a fairly well focused, fairly flat focus image. We recommend using one of the 1200 gratings for this job, and bundle STD1 or 2. STD3 is very much oversampled, hence not useful for measuring focus changes. In the end, you should have vaules of about 3.5--4.5 pixels over the CCD for bundle SB2, 13--14 for unbinned bundle SB3, and unknown for bundle SB1. In the page of pictures is one showing an example of the variable focus over the field of view (for a not very well focused set-up),
2.
Take an exposure
SYS> glance integral 10
view your arc image and select at least 9 evenly spaced positions on the detector where there are spectral
lines, which are bright and not blended with other lines (using eg. imexam).
Write the (x y) positions of the individual lines into an ascii file,
space separated and with line breaks and no blank line at the end.
3.
Run the IRAF task int_fwhm1 with this input file,
to obtain the fwhm for the selected
positions
(to run the task, you must first
have loaded the "integral" package in IRAF).
NOTE that at present, the entry for "dispaxis" should be
1 for horizontal, not 0. Note also that, to avoid the mean and
standard deviation values returned by this script being strongly affected by
very bad focus values, it is probalby best to avoid the very corners of the
CCD, and later to use a separate inspection to make sure they are not too bad at
your finally chosen focus values.
4. You can then change the focus, if necessary, with wslitf, in jumps of about 500. The pixel focus for SB3 (unbinned), SB2, and SB1 are, respectively, about 13-14, 4.0+/-0.4 , and 2.8 pixels.
For the other method, one does the following:
1. Take two arc exposures (using "arc", not "glance", to save the images)
with the same grating and bundle as recommended above,
one with the right Hartmann shutter in, the other with only the left Hartmann
shutter in, eg.
ICL> whart r
2. Follow step 2 of the first method.
3. Run the task int_focus, which
measures the Hartmann shifts for each selected
position
The parameters to input are obvious, and the aim is to get the Hartmann shift
values (ie.\ the difference between the focus with left or right Hartmann
shutters in) as close to 0 as possible (Note added Dec 2004: at
present this IRAF routine will not run, hence you will have to measure and
record these values by hand. You could measure
the difference in fwhm values between the arc lines resulting
from the left and right Hartmann shutter exposures, and aim to have a
difference close to 0
).
Remember to open the Hartmann shutters after the Hartmann test,
with the command whart OUT, and turn off the arc lamps.
You cannot avoid some of the fibres having bad focus, normally those at the ends of the CCD. To fine tune your adjustments, we suggest the follow iterative procedure. The focus range can be altered from about -3000 to +3000. Use wslitf to set the focus to, say, -3000 and calculate the Hartmann shifts or mean fwhm and dispersion (over the whole CCD). Record the values in a file. Change the focus successively up to the value +3000 (eg. -3000. -2000, -1500, -10000, -800, -600...), measuring and recording the fwhm/shifts each time. A plot of focus vs. eg. the fwhm (which int_fwhm1 produces) will show a minimum, and it is this value of focus you will want. There may be 2 minimum values (separated by a local maximum); that with the smallest dispersion (scatter over the CCD) will be the best.
You should only need to set the focus the day before your first night, nonetheless it is worth checking it each afternoon.
Now can check the wavelength centring and desired wavelength range, for each
grating/central wavelength you want for that night. There is no particular
need to do it also for each fibre bundle, although it does no harm to check.
Currently the left
CCD is the red, the right CCD the blue.
An example sequence of commands would be:
The spectral focus and slit translation only need to be done once per run, but it is worth checking them each afternoon.
The rest of this page details the set-up procedures that are carried out mainly by the ING technical staff. To summarise the set-up procedures that are carried out by the SA and/or the observer:
In GHRIL by the SA
Ask if all the technical set-up has been carried out, including the rotation
of the CCD.
Check that WYFFOS is in the correct position (reflection/echelle).
Install the grating and filters, and place those needed during the run
somewhere handy.
Check that the flat and arc lamp switched are on but that there are no bright
light sources when the main lights are swtiched off.
In the control room by the SA and observer
Enter the correct gratings etc. into lpvs1 with
wchange.
Centre the fibre bundles (slit translation).
Check the camera tilt and focus, probably having to fine tune (which can
take a long time).
Focus INTEGRAL via arc exposures.
Finally, check the wavelength centring is good and also that your datafiles (not
from "glance" however) have all their headers (incl. those from WYFFOS).
Take some biases and dome flats, trace the latter for use later than night in
image reconstruction (see
data reduction).
Second part
The end of the acquisition bundle that is
in front of the TV camera can be focused
using the focus knob (on the black box in the lowest of the mechano-like shelves).
(see there pictures)
In principle there are two procedures,
both requiring the presence of two people: one in the GHRIL room
(moving the focus mechanism), the other in the control room, looking at the TV
screen to check the results.
(A) Put the acquisition bundle into the calibration position (ICL> isp acq cal). Turn on the white flat lamp (ICL> ila w on): it may be necessary to reduce the brightness compared to its use for taking flat-field exposures. The knob for this is located below where the arc lamps are (see pictures). Take images with the acquisition camera GUI in the control room ("start"). When focused, you will be able to make out the individual fibre cells and see many dust balls. If the focus remains unsatisfactory the following procedure should be used (as a last resort).
(B) The bundle is dismounted from the SP (see INT-IAC-PRC001), and connected to a dedicated focusing system. This system allows a test image (eg. a number) to be seen through the bundle.
In order to aid night-time acquisition, the outline and centre of the bundle should
be marked on the TV screen. The TO should know the N-E directions. The
the field of view is about 20x30 arcsec.
It is also possible to adjust the tilt of the fields of view. For this, you need to unscrew whichever fibre bundle (again, see pictures). Take off the round cover, and undo a small screw on the top with a shortened, small Allen key. You can then turn the fibre bundle, and check the results.
In the standard configuration the inner (outer) field is to the night
(left) of the TV screen.
The GHRIL rotator has hardware switches limiting its motion to slightly
less than one revolution. The TCS software limits are set about 1 degree
inside the hardware limits, constraining the rotator mount position angle
to a range of 3 to 358 degrees. If one were to over-rotate, the fibre bundles
would be stressed and pulled.
Owing to the fact that the absolute encoder "wraps" at a certain
point, it is very important to zero-set the incremental encoder against
the absolute at a known physical position. That position is mount PA 180,
and it is the "park" position for INTEGRAL. When the rotator is
at this position, the "180" label marked on the rotator in indelible ink
is at the 12 o'clock position when viewed from the GHRIL bench. Before
observing with INTEGRAL it is important to ensure that the rotator
is in this position (to within twenty degrees or so is adequate).
The local controls can be used to move it if necessary. The incremental
encoder can then be zero-set against the absolute in the usual way. It must
not be zero-set without a visual check that the rotator is in the required
position.
The GSS with configurations for WHT INTEGRALINNER and WHT INTEGRALOUTER
is selected when you log on to the GSS account and select WHT.
To start the
WYFFOS mimic, log in to lpss3 (from an xterm),
with username: wyffos; password: wht_wyffos,
and
type
start_eng
To start the INTEGRAL mimic, log in to orion.roque (from an xterm) with username: integral; password: wht_integral, and type
start_eng
The low level Software Users Manual has the reference INT-IAC-MNL-001
