UHRF Setup Procedures
This section describes the standard setting up procedures which are necessary
before a night's observing with UHRF. UHRF is normally setup during the
afternoon, either by the afternoon staff or by the support astronomer. Non-expert
visiting astronomers at the telescope are not expected to know how to set
up the instrument.
The following procedures are normally followed in order:
1. Find the existing configuration files
2. Set up using the HeNe laser lamp
3. CCD Rotation using the Laser line
4. Focus the collimator using the laser line
5. Focus the collimator using the THAR lamp
6. Check filters and Lambda Offsets
I also describe how to:
7. Create a configuration file for a new wavelength
UHRF configuration files are created manually for each
wavelength observed. Once a configuration file has been created, it is
kept on disk and can be reused whenever the selected wavelength is observed.
The files are stored in disk$user:[ucles.uhrf_settings]. UHRF files
normally start with an element name, for example the file Na_5980.sdf contains
the settings for the Na 5890 line. Note that UHRF can also be run from the
UCLES account but this currently does not correctly write the header
information.
To look at existing files without running UHRF:
Refer to the UHRF configurations
summary for setup values optimized in previous UHRF runs.
The CCD rotation for UHRF is done such that the laser
line is parallel to the CCD rows. Note that this is different to UCLES where
orders seen from the quartz lamp are aligned parallel to the CCD columns.
Make sure the laser line is in reasonable focus to begin with.
Use an unbinned window such as
ccd: WIN MITLL_CENTRE (central 1044 x 1024 pixels)
Take an exposure. then rotate the frame using the
Figaro rotate task. To run rotation in unix from the obsred account
the following steps should work:
Login as obsred on aatssf then
unix: cd /data/ssf/1/obsred
unix: setenv D /vaxdata/ccd_2/date
unix: ls $D/
unix: cp $D/file file
unix: figaro
unix: rotate file filer
now open up the utilities package
unix: AAT
and select the "CCD rotation correction" routine under
"Setup" - run rotation as usual but set a suitable box around the laser
line, avoiding the overscan region at the top, etc.
Rotate the CCD in Coude room in usual way, then repeat
the procedure to see if the laser line is correctly aligned. Note that output
from rotation gives change needed in mm as for UCLES. For EEV2, a positive
shift requires a decrease in the micrometer reading. One full rotation of
micrometer screw = 0.5 mm. The CCD should be rotated to an accuracy of
0.1 mm or better.
The collimator focus procedure is done with a Hartmann
test by taking two arc frames with the upper, then the lower Hartmann shutter
closed. For example,
ccd: WINDOW MITLL_CENTRE
ccd: SPEED FAST
uhrf: ARC LASER
uhrf: UHU SHUT (shut the upper Hartman shutter)
ccd: DUMMY 15
uhrf: HP OTHER (swaps them round)
uhrf: DUMMY 15
uhrf: UHL OPEN
To determine the focus shift needed, set up a figaro
command file which plots the two spectra on top of each other and gives
the cross-correlation shift in pixels between them. For example the command
file, focuhrf, with the following lines will do this process with a YSTRACT
for the pixel range 50 to 900:
ystract $D/$1 50 900 sp=$1f
ystract $D/$2 50 900 sp=$2f \\
scross $1f $2f xs=min xe=max rec=no
splot sp=$1f wh=y au lab="focus_test" bu=n ha=n col=r ax er \\
splot sp=$2f wh=y au bu=n ha=n col=g noer noax \\
Modify this so that the x-start and x-end values
are within the range of the laser line.
To run the command file:
unix: setenv D /vaxdata/ccd_2/date
unix: chmod a+x focuhrf
unix: figaro
unix: soft /xdisp
unix: focuhrf a b
where a and b are the two Hartmann
exposures taken with the upper and lower shutters closed, respectively.
The cross-correlation routine will give the shift
between the two spectra in pixels. To convert pixels to focus ADU values,
use the following conversion scale:
@1E6, a shift of +1 pixel means the focus value needs to be adjusted by
-28.1 ADU
@6E5, 1 pixel = -45 ADU
@3E5, 1 pixel = -95.5 ADU
The collimator focus for UHRF has two different
drives, giving coarse (UFC) and fine (UFF) control.
To set the focus, first use the coarse control.
This sets the focus position to within about 30 ADU. Then use the fine control
to adjust it to the final value. The fine control has a maximum range of
around +/- 100 ADU and can be driven out of range. If this happens, the response
will be something like:
¨obey" UHRF_FFOCUS - action complete - text: 49623adu
!!
instead of the requested UFF position. In this
case, you will need to reset the focus value using the coarse control - drive
it up by 100 units or so (if that was the direction you were trying to go),
then re-issue the UFF command. It is possible to set the focus position to
an accuracy of 5 ADU.
> R UF
It may be necessary to repeat the focus procedure
if the initial focus was poor. After focusing on the laser line, open both
Hartmann shutters and take another frame to measure the intrinsic instrumental
resolution. Use the Figaro emlt routine to measure the FWHM of the
line in pixels the convert this to Angstroms. Note that the instrumental
resolution for UHRF can only be measured on the laser line as the ThAr lines
are resolved. A best FWHM of 4 pixels or less should be achievable at 1E6,
corresponding to a resolution >900,000. The ThAr lines will give 5-6 pixels
at best for 1E6.
If the configuration file for the selected wavelength
already exists then configure UHRF in the usual way:
uhrf: ARC THAR (it may be necessary to type
this twice, as the lamp may not always strike)
uhrf: FETCH filename
uhrf: CONFIG
If necessary make adjustments to the echelle angles
UT and UG to recenter the order on CCD. Make sure that the selected central
wavelength allows for enough arc lines for wavelength calibration. Also
check that the velocity shifts of the sources are within the observed range.
Record the final grating angles. These can be changed
in the configuration file using the UHRF commands LIST and ISET
(AAT staff only).
To focus the collimator using the THAR lamp follow
the procedure described above for the laser line. The collimator must be
refocussed each time the central wavelength is changed.
Make certain that the visible wavelength range is
correct by checking the arc-spectrum against an arc atlas. To orientate
an EEV frame to the standard orientation (wavelength increasing left to right
and bottom to top) use the Figaro rotate and irevx routines.
Previous hardcopies of UHRF arc spectra are kept
in a folder at site. It is helpful to add any new plots to this file. A
grey scale plot of the XMEM display can be made using the Figaro irevy
and imageps routines. Alternatively write the file to FITS format
using wdfits. Then grab the image using xv_fits and save it
as a postscript file.
Here are some example parameters used to set up on
Na_5890:
Date: 24/3/97 @ 1E6
UXD = V
UF = 47473
SECOLL = 22.624 (32551ADU in listing)
UT = 38939
UT_offset = 0
UG = 20310
UG_offset = 0
SA = -0.205 degrees
SA_offset = 0
pixels = 24 x 96
SLIT SLICER (except for FLATS)
BR OUT
FM UHRF
UHRF uses two filter wheels which are aligned are on
the same optical axis. Each filter wheel can hold up to eight filters. To
select a filter, set one filter wheel to the clear position and the other
to the position with the correct filter. For example, the command:
uhrf: SF1 7 # sets the filter in position 7
of slit filter wheel 1
For blue wavelengths it is advisable to use a bandpass
filter to cut out scattered red light. Examples are UG11 centred at 3300
and BG12 centred at 4000. Note that the blue filters have some transmission
losses and may also have strong red leaks.
For red wavelengths, it is advisable to use a 'cut-on'
filter such as RG630. These have a sharp cut-off profile at low wavelength
edge.
The response of different filters can be plotted using
Steve Lee's programme curves.
The position seen by the TV is offset from the true
position due to the different optical paths. This wavelength-dependant position
offset is minimized using a LAMBDA offset. This is an essential procedure
for UHRF! First tape or place a filter or known wavelength in front of the
TV lens. This should be chosen to be close to the observing wavelength so
that the lambda offset is small. Standard filters include the B(4400), V(5500),
R(7000) and I(9000). Most filters are kept in wooden box in the dark room
on 6th floor. The night assistant should then be requested to do a lambda
offset from the filter wavelength to the wavelength of the observations.
The lamp filters LF1 and LF2 should both be in position
1 (clear) except when the quartz lamp is on. This is mounted in the LF2=8
position. To select the correct quartz lamp it is necessary to type :
uhrf: arc quartz
uhrf: lo quartz
The filter wheel is returned to the correct position
when the lamp is turned off.
This section gives an example of how to solve for
the grating angles to set up a UHRF configuration file for a new central wavelength.
This process is very tedious and should normally be done well in advance
of an observing run.
In general it is best to solve for new configurations
using the 3E5 resolution with the EEV CCD as this gives the most wavelength
coverage. The grating angles are the same for each resolution so, once created,
configuration files can be used for any resolution. In the following example
we set up a configuration file at 3859 A using 1E6.
The aim of this excercise is to find initial values
for the parameters:
UXD = cross dispersing element
UT = echelle theta angle
UG = echelle gamma angle
UFC = collimator focus position
USC = Secondary collimator focus position (in pseudo-mm)
The steps taken to fix these parameters are as follows:
- Determine the order number and UXD using Appendix
C of the UCLES manual. This shows that at 3859 Angstroms UXD = B and m
= 149.
- Run the ray trace programme /epping/sdr/UHRF/ras/routines/22decuhrf.e.
This gives model solutions corresponding to the blaze peak of the selected
order. Note that this is only an initial estimate of the solution. The output
values from 22decuhrf.e are incorrect due to errors in the model and to
non-linearities in the adu to mm conversions.
Enter the target lambda and order number. Record
the UFC and USC readings in both ADU and mm and the wavelength and order
of the blaze peak.
UFC = coarse focus
USC = secondary collimator focus
For this example
UFC = 27803.649, ADU = -42.082 mm
USC = 21221.706, ADU = -9.012 mm
BLAZE PEAK = 3846.974 in order 149
(XBD grating has blaze peak at 3959.42 A)
- Now run two further programs in /epping/sdr/UHRF/ras/routines/,
ufc.e and usc.e. Enter values in mm from previous
output and record the new values in ADU. These correct the previous values
for non-linearities in the system.
UFC ip = -42.082 mm, output = 27769 ADU
USC ip = -9.012 mm, output = 21154 ADU
STATUS = 0 (all OK)
These ADU values are now 'real'.
- Set the initial values for UT and UG. For all
orders the UT of the blaze peak is 33000 ADU. UG values are given in appendix
G of the UCLES manual. We now have:
UT = 33000
UG = 15825
UFC = 27769
UXD = B (corresponds to 9 on the Syntel computer)
Enter these values manually into UHRF
- Convert the USC value to 'pseudo-mm' as follows:
uhrf: RAW
uhrf: USC 21154
uhrf: REAL
and record the corrected value of USC in mm.
This value is now in pseudo-mm. Here we have USC = -8.983 mm.
- Take a quartz frame. If necessary adjust UG to
centre the order(s) on the XMEM display.
- Take an arc frame to identify the blaze peak
for the order:
At this stage we have determined values which
correspond to the blaze peak of the selected order (149). To check this,
identify the arc lines using the arc atlas. If necessary focus the collimator.
If OK carry on, if not rerun the programmes and check your values.
- Step across the order to the desired wavelength
by taking repeated arc frames until the the selected wavelength is reached.
Do this bit by bit so that the arc lines can always be identified. We did
this in about 15 goes!
When line is correctly centred on CCD - redo
focus for last time and check it.
- Record the settings in a configuration file:
The solution here for 3859 was:
UXD = B
UT = 27950
UG = 15803
UFC = 26796
USC = -8.983 mm (Note USC has to be in pseudo-mm)
To record these settings. In UHRF type
uhrf: CENTRE
and answer the questions and set MODEL = EMPIRICAL.
Enter the filename, UT, UG and UFC from above. To see the settings type:
uhrf: LIST
Finally we have to fudge the USC setting to give
pseudo-mm.
uhrf: ISET UHRF_SECOLL -8.983
The configuration file, name.sdf, is now available
on the disk and can be used at other times with:
uhrf: FETCH filename
uhrf: CONFIG
Enter the new solution into the Epping html file
/local/www/UHRF/cookbook/uhrf_settings.html. Keep a copy of a grey
scale image of the XMEM display and any spectral plots with identified
arc lines for reference.
Page created: 7th December 1998 by Jessica
Chapman
Last Updated by Stuart Ryder on 3 March 2005