EUROPEAN SOUTHERN OBSERVATORY La Silla - Science Operations Department
		HOME INDEX SEARCH HELP NEWS

Practical Hints & Tips

• DRS Initialisation does not find 39 orders
• 'Strange' behaviour of the DRS
• Problem to restart the DRS panel even after rebooting
• Restarting the MIDAS DRS in case of crash
• How to reset the DRS file counter
  
• Focusing
• Guiding on a field-star in the FEROS TCCD
• GRABing the FEROS FibreHead Viewing TCCD image
• When there is more than one image with the same INS name...
• Changing the pipeline's plot regions
• Visitor mode Special settings

• DRS Initialisation does not find 39 orders:
  The DRS initialisation checks that the search for the orders actually finds exactly 39 (no more, no less) orders. If it does not then it is necessary to adjust the search threshold. In the MIDAS window the DRS initialisation lead you through the process:
  
  1. First it will prompt with:
     Do you wish to adjust the threshold interactively? (y/N) :
     Type "y" and press <enter>.
  2. Then it will prompt for the new threshold:
     Press enter to continue or give a new threshold to search again.
     Current Threshold is 0.85 :
     Enter the new threshold and then press <enter>. Normally it is best to adjust the threshold in steps of 0.01. If the search located less than 39 orders then the threshold needs to be reduced, if it located more than 39 orders then you need to increase the threshold. If the search jumps from say 38 orders to 40 orders, then use smaller adjustments (0.005 for example).
  3. The DRS will now perform a new search using the new threshold value. Repeat step 2 continuing to adjust the threshold until the search locates all 39 orders. The just press <enter> to the prompt in set 2.
  The default search threshold of 0.85 results in successfully locating the 39 orders 80-90 percent of the time. When the DRS does not find 39 orders, it is normally because it finds only 38, and then it is normally only necessary to adjust the threshold down to 0.84, or perhaps 0.83.
  It is not known exactly why the threshold required to locate precisely 39 orders varies from day to day, but is probably something to do with aging and/or stability of the flatfield lamp(s).If a threshold of less than 0.80 is required then there is probably a serious problem (or one is about to happen) and the Instrument Scientist should be contacted.

• 'Strange' behaviour of the DRS:
  If reductions do not seem to be running well, for example images are being reduced twice, some images in the the ferosDRS panel file list are turning red (for error) instead of green for (ok), more than one image in the list is blue at once, then the problem could be that more than one instance of the ferosQueueIms and/or ferosReduceQueuedIms scripts are running. In order to make a clean restart to all the processes and to be sure that all of the data has been properly treated, do the following:
  
  1. killall -9 ferosQueueIms
  2. killall -9 ferosReduceQueuedIms
  3. rm -f /data/reduced/.ferosQueueIms/.feros*
  4. rm -f /data/reduced/FEROS/midwork/.ferosDRS.date
     
  5. rm -fr /data/reduced/FEROS/<date>
     Note: above and in all of the below <date> is required in yyyy-mm-dd format.
     
  6. mv /data/reduced/.ferosQueueIms/<date>/Done/* /data/reduced/.ferosQueueIms/<date>/ToDo
  7. mv /data/reduced/.ferosQueueIms/<date>/Error/* /data/reduced/.ferosQueueIms/<date>/ToDo
  8. rm /data/reduced/.ferosQueueIms/<date>/ToDo/.FEROS*
  9. In the ferosDRS panel stop the MIDAS DRS process
  10. In the ferosDRS panel select any other date
  11. In the ferosDRS panel select the correct date
  12. In the ferosDRS panel start the "Queue Images" process
  13. In the ferosDRS panel start the "MIDAS DRS" process (note do NOT press the "Start All" button).
      
  14. In the ferosDRS panel start the "Reduce Queue Images" process
  The above will result in a COMPLETE re-reduction of all the data.
  

• Focusing:
  The focus template is now fully implemented (since 2004-June-06) and can be used for focusing for FEROS. Focus OBs for each of the spectrophotometric standard stars of the Calibration plan have been created and installed in the FEROS Calibrations OT queue. Run the OB closest to the meridian, simply follow the instructions given by the popups from the template. The focus OBs perform in the following manner:
  
  1. Execute a "normal" acquisition to the spectrophotometric standard star. Acquire the spectrophotometric standard star onto the OBJECT fibre and start guiding as usual.
     
  2. Once the acquisiton template is complete the focus template will offset the telescope by the amont specified in the OB (note, all focus OBs have been created with default offsets of +120 secs in both RA & Dec, but not all OBs have been checked to make sure this is adequate). The objective is to offset to a nearby field where faint field stars (compared to the spectrophotometric standard star) can be used with ~3sec integrations of the TCCD to measure image quality.
  3. The template will change the TCCD exposure time. Again, all templates have been created with a default TCCD expsoure time of 3secs for field stars (and 0.001 sec for the spectrophotometric standard star).
  4. The template will then step the M2 focus from a Focus Offset (with respect to the theoretical equation) of -200 to 0 in steps of 20 (again these are the default values in the OBs created). At each step it will "GRAB" 3 TCCD images, display them, measure the image quality and plot in the MIDAS PCO session.
     Note: Before starting the scan, the template will preset the focus to an offset of -100 where it will grab one image. You will need to select one or more stars in the MIDAS PCO 09 display window (which will be created the first time you use the template during the night). Simply click on one or more stars with the left mouse button, when finshed click anywhere in the display window with the right mouse button. This selection of stars is probably the MOST critical phase in the execution of the template. Keep the following in mind when making the selection:
     
     • One star seems to be the good number...
     • The focus offset of -100 has been chosen based on the experience that the measured focus ofset tends to be somewhere between -200 and -50. Therefore this is a reasonable guess. More importantly the image quality in this image should be close the best thereby making it easier to choose faint stars that even at the best focus will not be saturated.
     • Select stars in the central part of the TCCD, as intrinsic image in the TCCD degrades significantly near the edges of the TCCD field of view (due to the lenses used between Fibrehead and TCCD). In fact it is best to use stars in the half of the area bewteen the two fibres.
     Other useful tips...
     
     • A good range of scan values seems to be -400 to +100 in steps of 100. These limits have been implemented in the CalPlan SpecPhot Standard star focus templates and the "Focus Anywhere" OB in the FEROS CalPlan OT queue, but are NOT the current default values when creating a new focus template, therefore when executing a focus OB created by a VA check what FOCUS.OFFSET.START, FOCUS.OFFSET.END and FOCUS.STEP parameters they have used. Needless to say, the IP will be updated at the next opportunity so that the aboce are the default values...
       
     • In Bob Engineering mode, it is possible to adjust the end limit of the scan "on the fly", i.e. during the execution of the OB. e.g. if you are scanning from -400 to 0, and by 50 there is no clear infication that the image quality is getting worse again, you can change the FOCUS.OFFSET.END parameter on the fly to extend the scan (in this case to say +150...)
       
  5. After completing the focus scan, the template will fit parabolas to the measured image quality major and minor axes and then calculate the focus offset corresponding to the minima, which will be reported in a template popup. Follow the instructions in the popup to set FiltOff and AAFT in the TCS Control Panel.
  6. Typically in this mode the template takes less than 10mins to complete. It can also be used in a mode where a SCIENCE spectrum of the spectrophotometric standard star is taken at each focus offset value, but since this mode of the OB typically takes 40mins to complete it is used only during technical nights to verify the correspondance between optimum imge quality and optimum S/N in the science spectra.
     
  Incase of problem with the focus OBs, manual focusing may be necessary:
  
1. For the first FEROS observation of the night, focus the telescope manually using the "normal" TCCD RTD (not the TCCD AG RTD) and the judging the image quality using the x & y FWHM parameters in the "Pick Object" window.
2. Once you have established optimum image quality, in the TCS-Control panel set the FiltOff by clicking the "Set>" button in the M2 Focus subpanel.
3. Set the AAFT to 15.
4. Click the "Apply" button in the M2 Focus subpanel.
5. Remeber to set the FiltOff to 0 (ZERO) and click the "Apply" button in the M2 Focus subpanel whenever you switch back to WFI.
6. When you switch back to FEROS, either repeat steps 1 to 4 if you feel the FEROS focus FiltOff should be redetermined (e.g. because you are a pointing a long way from where you previously determined the focus, or because it was a long time ago, or because the Temperature of the Long Surririer has changed by 5 degrees or more), otherwise simply reset the AAFT to 15 and then click the "Apply" button in the M2 Focus subpanel.

• Guiding on a field-star in the FEROS TCCD:
  
  1. In the E2P2 Real Time Display panel turn OFF Dynamic Centering.
  2. Accurately acquire target onto fibre as usual.
     
  3. In the E2P2 Real Time Display panel click the "Set Reference" button and then in the Pick Refernce Star panel select the "Box and Refer.", "Pick Object" and "SKYFIB" radio buttons and adjust the box size if necessary, and then click on a "suitable" field star for guiding in the TCCD image. If the Pick Reference Star panel was already open you will need to click the "Pick Object" button in the bottom left corner of the panel (not to be confused with the "Pick Object" radio button) before clicking on the "suitable" field star.
  4. In the Autoguiding panel click "Start Guiding".
     Note: Steps 3 & 4 above must be done quickly to minimise decentering of the object on the fibre.
     
  5. If in the next acquisition the OBJFIB is set to TARGET FIBRE, the reference position should automatically revert back to the OBJFIB allowing easy acquisition of the next object using the "Centering" button in the E2P2 Real Time Display panel.
     Note: sometimes this feature of the templates does NOT work (probably a problem with the environment or something...), therefore to reset the reference fibre back to the OBJECT fibre for the next acquisition, in the Pick Refernce Star panel it is necessary to select the "Box and Refer.", "Pick Cursor" and "OBJFIB" radio buttons, adjust the box size if necessary, then click on the "Pick Object" button in the bottom left corner of the panel, then accurately click on the TCCD image at the exact coordinates of the OBJECT fibre (as already displayed in the E2P2 RTD Data panel).
     Don't forget to turn Dynamic Centering back ON before starting the autoguiding.
     

• GRABing the FEROS FibreHead Viewing TCCD image:
  Remember that after you have acquired the target onto the fibre and started the FEROS AutoGuider, the acquisition template will then...
  
  1. Stop the AutoGuider
  2. GRAB the image
  3. Restart the AutoGuider
  So do not be surprised when immediately after starting the guiding (and the clicking OK in the Bob popup that told you to start the AG) the guiding switches off for a few seconds then switches back on
  A related issue here is that when running the same OB more the once consequetively you don't want to waste time acquring when the target is already on the fibre and guiding, so you might be tempted to skip the acquisition template, however the better policy is to instead set the "New Preset" flag in the acquisition template to FALSE so that the FFHV TCCD image is stil GRABed.

• Restarting the MIDAS DRS in case of crash:
  If for whatever reason the MIDAS DRS session crashes or if someone simply exits from the MIDAS DRS session by accident and it is necessary to restart it...
  
  1. In the ferosDRS panel start the "MIDAS DRS" process
     
  2. You must then reset the initalisation.
     If a standard initialisation (without FFsol search method) was being used then in the blue MIDAS window:
     
     • Type: $ ls -altr ThAr*INIT.tbl
     • Type: init/feros ThAr<NNNN>
       where <NNNN> is the number of the most recent initialisation (i.e. the last initialisation listed by the above ls command).
     If an FFsol search method initialisation was being used then in the blue MIDAS window:
     
     • Type: $ ls -altr FFsol*INIT.tbl
     • Type: init/feros FFsol<NNNN>
       where <NNNN> is the number of the most recent initialisation (i.e. the last initialisation listed by the above ls command).

• When there is more than one image with the same INS name:
  This happens when images are acquired after UT 12:00, and then all images in wferos:/data/INSROOT/FEROS/SYSTEM/DETDATA are moved into a date directory so that DETDATA is then clean and ready for the coming night, and then during that night another image is acquired with the same template and the same index number. For example, at 11:00 you acquire one OBJSKY test observation with, for example, INS file name FEROS_ech_obs_objsky_0001.fits. You then move it into the date directory for the previous night. During the night the observations acquire another OBJSKY observation. Its INS file name will be FEROS_ech_obs_objsky_0001.fits also. But on the w2p2off machine the DFS will be clever and name the first image FEROS_ech_obs_objsky_0001.fits and the 2nd one FEROS_ech_obs_objsky_0001_0001.fits! In order to properly process the second image do the following:
  
  1. Open an xterm on wferos as user feros.
  2. Type: ferosDRS4Bob FEROS_ech_obs_objsky_0001_0001.fits
  (obviously if this actually happens replace FEROS_ech_obs_objsky_0001_0001.fits above with the correct file name). The image should then be properly processed by the FEROS MIDAS DRS and properly archived on ferosdrl and copied to drl2p2 for the VAs copy.
  Note: Even if you fail to notice this nothing is lost since the image will ALWAYS be properly archived within the NGAS system, and hence eventually retrievable from Garching.

• Changing the plotting regions:
  You have to copy the file autoreduce_plot.prg which is in /home/e2p2ops/FEROS/FEROS-DRS/midwork to the directory /home/astro/FEROS-DRS/midwork giving it another name (autoreduce_plot_user.prg for example) . After, you have to edit the following part of the file
  
  

! H-alpha   

  assi/grap G,1  

  plot {p3}{p1}{p2} ? 6520,6600  

  label/grap "H\_\alpha" 35,17,mm ? 2 1  

  copy/grap postscript  

  $ mv postscript.ps {OFLABEL}-Ha.ps  

! H-beta  

  assi/grap G,2  

  plot {p3}{p1}{p2} ? 4820,4900  

  label/grap "H\{\_\beta \}+He" 35,17,mm ? 2 1  

  copy/grap postscript  

  $ mv postscript.ps {OFLABEL}-Hb.ps  

ln -s autoreduce_plot_user.prg autoreduce_plot.prg

Send comments to <ls-feros@eso.org> Last update: Sep 12, 2007