Search
Text Size:
Smaller Text Normal Text Larger Text

Facebook

Follow us on Twitter

Astro-support issues and how to observe standard stars
FIES astro support: what to do?
FIES tech/service-night afternoon checks: what to do?
Checking the fiber positions on StanCam

How to observe standard stars for Quality Control
Flux-standard stars
Radial-velocity standard stars

Calibration procedures, Quality Control, monitoring
Calibration procedures
Instrument/detector quality control
Stability monitoring
Dying ThAr-lamp phenomenon and QC
FIES change log

Data reduction
Some hints for FIEStool usage
HowTo: Re-run FIEStool when it has failed

Data archive
A list with all FIES data from the FITSheader archive (up to FIyk290133.fits)

Known problems
New top unit very unstable slide (arm) and mask
Top calib unit: unstable arm positions, unstable calibration-fiber alignment
How to change a lamp
Recovering from CryoTiger Failure
More trouble shooting
How to remotely power cycle the FIES CCD controller

Documentation
How to observe with FIES, in the default full sequencer mode (superseded by FIES CookBook)
How to observe with FIES in sequencer mode, with old STANCAM BIAS
How to observe with FIES, non-sequencer engineering mode
Skew readout mode
The FIES solar experiment
Using the FIES fiber shaker
FIBGUIDER INSTRUCTIONS
Fiber-bundle constraints and requirements
FIES acquisition with ALFOSC instead of STANCAM



FIES astro support: what to do?
FIES tech/service night: what to do?

The below lists the setup and checks that need to be done when FIES is to be introduced to a visiting astronomer,
or when using FIES on a tech/service night.


Before a visiting observers' run:

If required, explain the choices (these should really already be known to us at NIRF stage):

  • Resolution low, med, high.
    Low-res has bigger sky-aperture (2.5 arcsec) and can be binned 2x2 without losing resolution. High-res has lowest throughput.

  • SimulThAr mode or sandwiched ThAr spectra.
    If a rad-vel precision of better than 150 m/s is necesary, then use either SimulThAr mode or sandwiched ThAr spectra. In practice, the same rad-vel accuracy can be reached with simulthar as with sandwiched ThAr spectra. But with simulthar the arc lines scatter into the target spectrum of faint targets. Simulthar mode can only be done with med- and high-res fibers. The exposure meter cannot be used with simulthar mode.

    If a rad-vel precision of worse than 150 m/s is OK, then ThAr frames in evening, midnight and morning, are sufficient.

  • Binning 1x1 (default) or 2x2 for low-res. Other binning mode to be considered: 2x1 . For simulthar: 1x1 .
    It is necesary to know the binning mode for the setup of FIEStool.

  • Amplifier B (default) or AB. Latter loses central order for 2x2 binning; it depends on the fiber which order is lost. See table.
    It is necesary to know the binning mode for the setup of FIEStool.


This should all be done in the morning/afternoon before the first night of the run, or before your tech/service night.

  1. Check the fiber positions on StanCam

  2. For all the fibers that will be used (including fiber #5 when simulthar is to be used), check that the lamps are working properly. Determine what expotimes are needed such that the flats have peak counts of about 30000 ADU, and that the brightest ThAr lines in the center of the chip have approximately 15000 ADU peak values.
    Check that the exposure times are not much longer than the ones mentioned here; if they do, then a lamp may be dying!!

    If a lamp shows signs of deterioration, then the lamp has to be changed: change the lamp .
    Call for help when you need it:
        either to change the lamp,
        or to judge whether a lamp needs to be changed.
    Afterwards, write a fault report to indicate a lamp had to be changed.

    An easy way to find out a suitable exposure time for a FIES calibration-lamp setup is the following:

    • for a ThAr exposure
      setup-fibN-thar
      easythar

    • for a halogen (flatfield) exposure
      setup-fibN-halo
      easyhalo
    The scripts easyflat-fies and easythar-fies will take a fast windowed test exposure to determine the countrate, and will set the CCD back to the original window/binning to obtain a halogen or ThAr image with sufficient counts for FIEStool.
    The flats will have approximately 35000 ADU peak counts. For the ThAr frames the 4764.9 line, one of the brightest lines in the center of the image, will have approximately 15000 ADU peak counts.

  3. Make sure to switch off all the lamps.

  4. Make a new output directory for FIEStool:
    • ssh -X fies-pipe@elena
    • mkdir /raid1/data/reduced/fies/44-210
    • cd datared
    • ./FIEStool.py
    • in FIEStool: change the "Output directory" to the one just created
    • in FIEStool: "Settings" --> "Save current settings to default"
    • in FIEStool: "File" --> "Quit"

  5. Create the Sequencer script directory for the observer, e.g.
          mkdir ~/scripts/44-210
    and copy in some example scripts from the directory ~/scripts/staff/FIEStemplates/ .

  6. Explain the following to the visiting astronomer:
    • Software startup
    • FIESTA mimic, and calibration arm !!
    • How to start and abort exposures
    • How to start and abort scripts
    • setup-fibN-star, setup-fibN-thar, setup-fibN-halo
    • fies-calibs
    • simulthar and extra calibrations for simulthar-mode (arcs and flats for fib #5)
    • lamp 1,4,6,7 on/off command
    • Telescope
    • camera probe !!
    • setup-tel-fies
    • exposure meter
    • Target acquisition
    • How to use FIEStool
    • That there may be ALFOSC overrides....
    • What more??

  7. Help the visiting astronomer make some SEQUENCER scripts,
    using the templates in directory ~/scripts/staff/FIEStemplates/ .

  8. The visiting astronomers should be advised to obtain a full set of calibrations every afternoon, and possibly additionally every morning.

  9. At night, during the first acquistions, see if there is a large offset when stancam.fiboff (or stancam.acquisition) tries to bring the star close to the fiber. If the offset is more than a few pixels, then consider running stancam.fibposcalib in order to make the offset more accurate.


Checking the fiber positions on StanCam

At the start of each run check the fiber positions for all fibers, using the StanCam filter that will be used. Note that for different StanCam filters the fiber positions are slightly different. The filters B, V, R, SDSS g', and Clear#126 can be used for FIES; typically the B filter is used.

In the afternoon, please follow the software-startup section of the general observing instructions.

In the afternoon, check the fiber positions in fiber-viewing mode (camera-probe-fies).
The instructions below are for fiber #4 with the blue StanCam filter.

  • Sequencer window: camera-probe-fies
  • Sequencer window: ccd-filter 2 (blue = default)
  • TCS: o-m-c y
  • switch on some dome lamps

    1. Sequencer window: setup-fib4-star
    2. Sequencer window: fib4window
    3. StanCam Sequencer: exp 1 (check required expo time!)

      • ds9: zscale
      • ds9: zoom to fit

      • StanCam Sequencer: fibmark no
      • StanCam Sequencer: fibmark yes

      • Use
            fibpos-man 4
        to redetermine the green-cross position.

      • StanCam Sequencer: fibmark no
      • StanCam Sequencer: fibmark yes

      • Iterate step 3 until the mark is centered on the hole

    4. Repeat the above (step 1-3) for the other fibers.

  • Switch off the lamp!
  • TCS: close-mirror-cover
To check the fiber positions in direct StanCam mode:
At night, once a star has been well centered onto a fiber, flip the camera probe back to the StanCam position (cam-probe-ccd). Then take a full-frame unbinned StanCam image and use stancam.fibpos-man to record the X,Y positions of the star=fiber. Or use the command stancam.fibposcalib to do all that from within a script.
Note that there is a flip in Y when moving the camera probe from StanCam direct imaging to the FIES fiber viewing position.


Quality Control: How to observe flux standard stars for FIES

In the afternoon, please follow the software-startup section of the general observing instructions. At night, secure a spectrum of one of the following stars on fibers #1, #3 and #4 (low-res, med-res, high-res). No other calibration images/spectra are necessary.

From the (!!) STANCAM (!!) sequencer window you can execute the following:
    fies.fies-FS
or
    fies.fies-FS -tutorial
This script sets up and slews the telescope, will lead you through the process of target acquisition, and will do the actual FIES exposure. The script will allow you to put the same star subsequently on different fibers. When an exposure is finished the script will initiate reduction of the obtained spectrum (as qc-user@elena), and a few minutes later the results will be available in an ASCII database. The results contain the average system efficiency [in %] in the BVR bands.
In case you want to manually start the reduction, open a terminal and type
    ssh qc-user@elena "cd fies/fluxstandard; ./fluxstandard.py FIqb230123.fits"
with the filename of your choice.


Detailed instructions on how to observe the standard stars without the use of the script are listed below.

name TCS catalog entryspectral typeexp time
G191-B2B SP0501+527 05:05:30.6 52:49:56 Mean 2000 +0.030 -0.090 11.9DA0 (blue) 300  
HD 93521 SP1045+378 10:48:23.5 37:34:13 Mean 2000 +0.000 +0.010 7.0O9Vp (blue) 60  
BD+33 2642 SP1550+330 15:51:59.9 32:56:53 Mean 2000 10.7B2IVp (blue)300  
BD+28 4211 SP2148+286 21:51:11.1 28:51:53 Mean 2000 10.7sdOp (blue) 300  
Feige 110 SP2317-054 23:19:58.4 -05:09:57 Mean 2000 11.9DOp (blue) 300  

NB the above exposure times are for FIES, for target acquisition with StanCam use someting on the order of 0.1 to 5 seconds. Click on the objects, or see the ING standard-star page for finding charts and more info on these stars.
The visibility plot of today!


Observing steps (the general observing instructions explain how to do these steps in detail).
First fiber #1:

  • start the software in the afternoon as indicated in the general observing instructions

  • To load the flux-standard catalog into the TCS type:
        read /catalogue/JHTFS.cat

  • sequencer: type mask 5 to open Hole #1, arm 5 to move the calibration arm away from Hole #1. Switch all FIES lamps off (lamp N off).

  • StanCam Sequencer: put the star close to Hole #1 , and check the telescope focus for CAM-probe-SPLIT (default focus 22720, with foc-del 0)

  • StanCam Sequencer: accurately center the star on Hole #1. See 'acquisition hints' in the general observing instructions

  • sequencer: type fies.expose nnn to expose (see table for expo time)
Now fiber #3:
  • Sequencer: type mask 2 to open Hole #3
  • StanCam Sequencer: accurately center the star on Hole #3. To do this, leave the cam-probe at c-p-split, and do fiberoffset fib=3. Then fib3window and continue like normal with the acquisition onto the fiber.
  • sequencer: type fies.expose nnn to expose (see table for expo time)
Now fiber #4:
  • sequencer: type mask 3 to open Hole #4
  • StanCam Sequencer: accurately center the star on Hole #4. To do this, leave the cam-probe at c-p-split, and do fiberoffset fib=4. Then fib4window and continue like normal with the acquisition onto the fiber.
  • sequencer: type fies.expose nnn to expose (see table for expo time)


Quality Control: How to observe radial-velocity standard stars for FIES

Currently we have a short- and a long-term monitoring program.
For monitoring the sky should be clear, and the seeing better than 1.3 arcsec.

In the afternoon, please follow the software-startup section of the general observing instructions.
Execute fies-calibs -fib N for all fibers you want to get RV data for. For monitoring these are fibers 3 and 4 (med-res, high-res).

At night, use script fies.fies-RV, to secure a spectrum of one of the following stars.
!! The highest-priority target is HD182488 !!

name TCS catalog entryspectral typeradial velocityexp time
HD182488 19:23:34.0 33:13:19 Mean 2000.0 0.0 0.0 6.4 G8V -21.55 120

HD3765 00:40:49.3 40:11:14 Mean 2000.0 0.0 0.0 7.4 K2V -63.30 300
HD38230 05:46:01.9 37:17:05 Mean 2000.0 0.0 0.0 7.4 K0V -29.25 300
HD79210 09:14:22.8 52:41:12 Mean 2000.0 0.0 0.0 7.6 M0V 10.65 300
HD115404 13:16:51.1 17:01:02 Mean 2000.0 0.0 0.0 6.6 K2V 7.60 120
HD151541 16:42:38.6 68:06:08 Mean 2000.0 0.0 0.0 7.6 K1V 9.40 300
HD197076 20:40:45.1 19:56:08 Mean 2000.0 0.0 0.0 6.5 G5V -35.40 120

NB the above exposure times are for FIES; for target acquisitioning with StanCam use someting on the order of 0.5 seconds.
RV standards visibility plot of today.

Please use fies.fies-RV to do the whole observation. The script will let you cycle through all desired fibers, i.e. fibers 3 and 4 for monitoring.

For long-term monitoring (on each full Nordic service night), please execute the script twice (2x fibers 3 and 4) consecutively on the same star.
For short-term monitoring (on tech nights indicated on the schedule), please execute the script 8 times (8x fibers 3 and 4) consecutively on the same star.


FIES calibration procedures and instrument/detector quality control

These calibrations have to be initiated and executed by staff or observers.
  • Calibrations for data reduction
    For the data-reduction tool FIEStool, a full set of calibrations should be taken at the start of each run. Use the script fies-calibs : instructions are given above.
    The visiting astronomers should be advised to obtain a full set of calibrations every afternoon.

  • System-throughput monitoring
    At least once every other month standard-star observations should be obtained on technical and/or Nordic-service nights.

  • Radial-velocity stability monitoring
    Long-term monitoring: an RV-standard taken every full Nordic service night.
    Short-term monitoring: a 2-hour series of RV-standard data taken once per month.
    See the description above.

  • Light-pollution testing
    It has repeatedly happened that LEDs and/or the CCD-controller display inside the white box were not taped off; this gives rise to low-level (about 10 ADU in 300 sec) diffuse scattered light on the red side of the CCD, when exposed with the shutter open.

    To test for light pollution make a 300 sec exposure with the dome- and calibration-lights switched off. Do this every time after electronics components have been changed (controllers, pressure sensor, etc.), or in general when the CCD controller has been touched.


The calibrations described below are done automatically and do not require any assistance from staff, unless an error or anomaly is reported through email warnings.

  • Monitoring of gain, RON and bias levels, for all amplifier modes.
    This is being done automatically every Sunday afternoon 13:30, through a cronjob by obs@verushka:
    > crontab -l
      30 13 * * Sun /usr/local/bin/Xvfb-run /home/softremote/obs/fiesmonitoring/fiesGR.main >> /home/softremote/obs/fiesmonitoring/crontab.log 2>&1

    The job starts the complete FIES data acquisition set of programs, but only if no other FIES observing system is running. Then it takes the required data, which is saved in /data/fies/FIESgainron/ , and fires off in parallel the Pyraf reduction script fiesgainron.py as qc-user@elena.

    The results are kept in an ASCII database elena:~qc-user/fies/gainron/fiesgainron.database , and are graphically visualised on the Detector Quality Control pages.

  • FIES stability monitoring
    This to be able to monitor echellogram shifts as a function of atmospheric parameters, or before/after changes of the instrumentation or staff activities.
    Images using the top calib unit (halogen, ThAr) are taken automatically, once every other hour, through a cronjob by obs@verushka:
    > crontab -l
      20 1,3,5,7,9,11,13,15,17,19,21,23 * * * /usr/local/bin/Xvfb-run /home/softremote/inst-scripts/FIES/monitoring/fiesmon.main >> /home/softremote/inst-scripts/FIES/monitoring/crontab.log 2>&1
    The job starts the complete FIES data acquisition set of programs, but only if no other FIES observing system is running.
    Then it takes the required data, which is saved in /data/fies/FIESstability/ .

    ToDO: automated analysis and reporting is pending.

  • CCD temperature monitoring.
    This is now part of the general TMS. See graphs .
    The sensors are sampled with 1-minute interval.

  • Dying ThAr-lamp phenomenon
    From experience we know that from one day to another a well-functioning ThAr lamp can go into a state in which the ionisation balance is gradually changing, and subsequently the strength of lines will change accordingly. Over the course of several days-weeks, the lamp will gradually die.
    Lamp life time for the top-calib unit is typically between 0.5 and 1 year.

    The phenomeneon is picked up by the QC routine triggered by fies-calibs. The routine logs the number of ThAr lines found in the ThAr frames that are exposure-level scaled to have 15000ADU in a particular blue line (4765 Angstrom, see above). The log is in fies-pipe@elena:
    If an anomalous number of lines is found (50-90% of nominal), then a warning is sent to the talker, and an email to 'duty'. If less than 50% of the lines are found, then ERROR-level warnings are sent.

    If the warnings persist, or an ERROR is given, then it is likely that the lamp needs to be changed. Some more info, in particular on how to diagnose the phenomenon, is given in the trouble shooting page.


FIES top calib unit: unstable arm positions, unstable calibration-fiber alignment

Due to a design flaw of the calibration fiber that connects the lamps in the top calibration unit with the adapter, the positions of the arm that shines the calibration light onto the science fibers 1-4 have to be recalibrated every now and then.

Symptoms: the count rates of the halogen or ThAr lamps at one or more of the 8 arm positions drops with respect to the normal values.
See this table for normal exposure times aiming for 20000 ADU in the flats and ThAr spectra.

Software solution: start the engeneering software, see how to observe with FIES, non-sequencer mode, and start the fib8server with the -e flag.
This allows to position the arm with the command
  arm 30000
Use easythar/easyhalo on your laptop at different values of the arm position (close to the programmed position, with steps of 5000 units), find the optimal arm position (i.e. highest count rate), and let Jacob put the new value in the fib8server program.
Do this for all 8 arm positions (4 fibers).

Mechanical solution: rotation: If the above does not work, then the calibration fiber has to be rotated slightly in the socket of the calibration arm on the adapter. There is no rotational alignment precision nor stability, which is exactly the cause of the problem. Currently the fiber is taped to the socket, in order to keep it fixed in rotational sense.

Mechanical solution: XY slide: As of 2009 we have an XY slide mounted, with which one can accurately position the calibration fibers. When properly clamped, and with the rotation fixed by tape, no re-alignment was necessary until June 2015, when a sudden jump occured possibly during an instrument change. Procedure (can take a few hours):

  • cut some cable ties and take down the fiberhead-assembly plate. Put it on the side table with wheels, with the ARM down. Be careful to put some big spacers under it to make sure the ARM does not touch the table, and can move freely through computer control
  • loosen the main securing bolt and the two black bolts on the XY slide. One of the latter can only be reached with pinky finger, and only when the plate has been put on the table (on spacers to protect the ARM!).
  • use setup-fibN-thar/halo to position the arm/mask in any of the 8 positions
  • use easythar/easyhalo on your laptop (the obssystem should be up and running, eg in the CR) until it show the predicted exposure times and then press CNTR-C. Record the predicted expotime as a function of the XY slide micrometer positions.
  • optimize F3 ThAr first. I got X=3.2mm, Y=2.8mm (with XY defined as normal when the plate is mounted)
  • check the expo times in the other 7 positions. If OK then secure first the main bolt, and then the two XY bolts.
  • if NOT OK then find a compromise to get flux through all 8 positions. I used X=3.1mm, Y=2.9mm.
  • after securing the bolts check again the expo time in all 8 positions
  • bundle it all up nicely again
  • now continue with the Software Solution (see above) to do fine tuning


Back to top Last modified: 28-Jul-2017