NAOMI PRE-RUN READINESS CHECKS
and DUTY ENGINEER DAILY CHECKS

This procedure includes enough checks to be followed as a "Pre-AO Run" check list, after a long period without observation or when engineering work has been done in GRACE.
On a normal day during an AO observing run it is not expected that  the Duty Engineer performs the whole list of readiness checks (for example the DE usually won't have to flatten the Deformable Mirror). But as the "daily" and "readiness" checklists share many actions in common, it was decided to write only one procedure. Simply, a note is written any time an action is specific to the complete readiness checklist or instead specific to the reduced daily checklist. 

List of abreviations (or names)

• DM: Deformable Mirror

• SG: Strain Gauge (for the control of the DM segments)

• FSM: Fast Steering Mirror

• NCU: Nasmyth Calibration Unit (slide with pinholes, calibration lamps and AG3 camera)

• OAP2: Off-Axis Paraboloid number 2

• OAP1: Off-Axis Paraboloid number 1, it’s in fact the FSM

• WFS: Wave Front Sensor

• OMC: Opto Mechanical Chassis (associated mostly to the WFS)

• USP: Universal Science Port, to which OASIS is attached. By extension, mechanisms on the optical bench.

• FISBA: brand of the interferometer used for mirror flattening.
  

• In GRACE
  • Inspection of the Light Path
  • Temperature and Humidity conditions
  • Inspection of electronics racks
  • Switching from L3 (Tip-Tilt) to CCD39
    

   

• In the WHT Control Room
  • Starting up the AO system on Taurus and the AO console
  • Laser Flattening the DM
  • White Light Flattening
  • Simplex Flattening and image quality checks.
    
  • Detectors sanity checks (OASIS, INGRID)
    

• Temperature and Humidity are seen from the display either below the Met display in the Control Room or in Grace Electronic Room. Humidity should be maintained above 30% (too low it makes the piezo-electric SG unstable). Steep gradients in T or %RH are especially bad for DM stability.
• Check in Grace if the de-humidifier tank needs to be emptied
• Only during an AO run Depending on conditions start-up the humidifier (only a few hours autonomy with full tank)  

• Take out all mirror covers (access from two places: around the DM are the covers for the DM, return Flat, FSM, Fisba small return flat, Fisba; between OASIS and  INGRID is the OAP2 flat). Place the covers in the dedicated plastic rack, on the wooden table next to the flat screen.
  For DE checks during a run: usually if NAOMI was used the night before, the mirror covers are not put back by the observer.
  Note: from February 2007, new optical components for GLAS have been added between the large Flat and the Fisba return flat. Only the dichroic is in the beam and to observe with Naomi only, its covers have to be removed.
  

  

  View of the NAOMI bench for acces to the DM, return flat, FSM, Fisba and Fisba return flat

View of the new components: Laser Dichroic and Fold. Only the dichroic is in the optical path for science.

Where to store the mirror covers

• Readiness checks only  Especially if engineering work has been going on, check if nothing obviously obscures the light path (forgotten tools, mask). More subtle obscuration (e.g. a mask in front of the detector!) will need imaging with WFS to reveal itself. 
• Readiness checks only Check which pinhole is at the Simplex position on the NCU: the 2µm pinhole will be needed for a good simplexing process. If it needs to be changed, the whole series of pinholes is kept in a plastic box on the table next to the cover rack. It is simply screwed into its position.
  
  
  
  close up of the NCU showing pinholes, and the DM
  

• Reboot the Fisba PC (with the mouse, close every window, don't save anything when it's asked, then click on "Shutdown" “Restart”. Ctrl-Alt-Del twice from keyboard will also do it). After the reboot when the program has (automatically) started, adjust the zoom to 3x to better see the window showing fringes (click on upper left corner of that window for menu).

   

  View of the Fisba PC (to the left of the entrance in GRACE electronics room)

   

  

  How to adjust the size of the interferogram window

• Readiness checks  Switch on the SDSU controllers for all CCDs to be used: WFS (master for CCD39, slave for L3), always AG7, AG3, plus MITLL3 (for OASIS) and/or INGRID.
• Readiness checks  The VME racks should all be alive (green lights…) but this can be checked later on from the AO console. 
• See especially on the WFS Peltier front panel if both green lights are on. This means the flux of cooling liquid is ok.
• Daily checks (essentially)  If it's not the first night of an AO run, and the instrument in use is OASIS, chances are NAOMI has been used in Tip-Tilt only mode. This means that the WFS in use is the Low Light Level (L3) and the signals to the FSM are not transmitted by the C40 processors, but by the new Tip-Tilt PC. But for the checks, you need to switch back to the old system, using CCD39 and the C40's. This implies several actions:
• In the electronics room, switch off the L3 SDSU, switch on the CCD39 (MASTER) SDSU
• In the electronics room, switch off the FSM controller.
• In GRACE, open the door giving access to the back of the electronic racks. Switch from CCD39 to L3 or vice versa (well-labelled switch on small box close to the FSM control rack).
  
  
  
  
  Access door to the back of electronic racks
  
  
  Switch for FSM control by the Tip-Tilt PC or the C40 processors
  
  
  Tip Tilt PC in the electronics room (no action needed on it)
  
  
  
  
• In the electronics room, switch on the FSM controller.
• In GRACE, go to the WFS enclosure (large black box) and from the access on the left side of the enclosure, behind the black cloth, move the metal slide below the front of the WFS camera. It has 4 positions (see label on top of WFS box). The slide is fairly stiff. It seems to click firmly into place, but if the WFS display looks odd, try moving it out and back in again. 
  
  
  
  Access to the WFS behind the black curtain
  
  
  View of the WFS box with the slide for the mirror to send light to CCD39 or L3
  
  
• Another action will be needed in the control room to refocus the camera (as L3 and CCD39 don't share the same focus position)
  

• Back to menu
  
  

  In the WHT Control Room

  ---

At the WHT instrument-control screen (taurus)

• The camera configuration file at first has to be edited on Taurus, but if it’s not the first day of an AO run and you don’t want to interfere with the main instrument, the AO observing system can be brought up on standalone from lpss94.
•  In any case, obssys has to be started on the das machines for the following detectors: AG3 (the acquisition camera), AG7 (the simplex camera), the instrument cameras INGRID and/or MITLL3 (for OASIS), so some access is needed on Taurus at least to edit the config file and startobssys in the orange das windows. For DE checks: see the observing schedule or ask the support astronomer to know which instrument camera will be used at night.
• Readiness checks only To work from lpss94:
  
  • telnet to lpss94 (for example from an xterm in the second working space of the right AO Display scree), login  as whtobs
  • Type set instrument_config_file = /wht/var/instrument.cfg.ao
  • Type obssys, answer 1, 1, then startobssys. In this way, the AO servers (like oasisserver) are started, as well as all observing guis like the AO mimic, NAOMI Observer gui,…
  
  
• DE checks only Start the observing system on Taurus with at least one instrument camera (Ingrid or OASIS) in the configuration file, because in this way all the servers start automatically (like oasisserver).
  
• Start the camera sanity checks by taking bias runs for each camera and examine the noise level with Iraf
  
  Back to menu
  

At the NAOMI control screen: AODISPLAY

AODISPLAY is a pc which allows to emmulate a unix session (via Exceed) either to Navis or lpss42. Use the login name  naomi. The left screen is an 8-bit display and reserved for topgui; the right screen is a 24 bit display and allows to launch more complex (and modern) mimics or gui's. DE checks only  If the AO system has been in use on the previous night, it shouldn't be necessary to log on again.

From AODISPLAY right screen:

• telnet to Taurus for a session as whtobs then type obssys,  answer 1, 1  (but don't start the observing system!)
  
  
• Run the oasisgui: startoasisgui ; if everything is in error it’s probably a server problem, which shouldn’t happen if the observing system has started properly on Taurus.  The command startoasisserver will give an error if the server is already running.

• From oasisgui USP page float the bench (takes 5min to stabilise); move the NCU to simplex position. Disregard the ADC (not commissioned)

• If Ingrid or OASIS is going to be used try moving mechanisms from Ingrid, OASIS pages. Some (enlarger or imager wheels on OASIS) might require more than one attempt. It’s very unusual when none of the mechanisms on the same tab (USP, Ingrid or OASIS) can be be initialised. In that case the VMS rack in GRACE controlling these mechanisms needs to be rebooted, then all mechanisms initialised from oasisgui.

  

  oasisgui INGRID page

   

  

  oasisgui OASIS Engineering page

   

   From AODISPLAY left screen:

• At the prompt in a console or xterm (lpss42 or navis) type restart or NaomiRestart then topgui & . Note for DE checks: If TopGui was still running from the night before, restart will kill the process.

• The left half of Topgui is static, the right page gives access to SDSU control, C40, Mirror Control, Light Path, etc either from the icons or from the Tools menu.

  

  

  TopGui icons and Tools menu

• From the C40 page  check if all 16 processors are in green. Left ring is for WFS, right for Strain Gauges (ie DM control). The SG icon is pink which means the loop is open, that’s normal.  NaomiRestart usually clears most processor problems. If not: when a second restart doesn’t help see the “special case” recipe in the green Naomi Software folder, at the tab Trouble (hand written in the margin)
  
  

  

  TopGui C40 Control page

• In C40 Control under WFS click on Start to get WFS into framing mode and in the Display Control click on Continuous

• Check noise level on the WFS, either visually (with Max level in the Display Controlset at 100, no diamond structure is visible on the image) or more precisely from the WFS Display page, with a Grab, Save as Fits and then display of the file on iraf, using imexam. An rms value of 7 to 10 is normal. To run Iraf, you will need to start an xgterm, in it from the home directory type ds9& then cl. The grabbed  files for the WFS are kept in the directory /opt/Electra-save-dir/Grabs/WFS
  
  

  

  WFS Display page, from which to Grab an image of the WFS

• Sanity check for the WFS SDSU: try changing to Mode 10 (while the WFS is framing, but not in continuous display because this would give a Timeout error) then switch back to Mode 1. If Status turns to Error, power cycle the SDSU power supply and VME crate (centre electronic rack, bottom).

• After an SDSU power cycle or if it gives repeated Timout errors the WFS needs a Re-Init from TopGui WFS control (SDSU) page

  

  SDSU Camera Control page, from which to Re-Init the WFS SDSU

   

  Switch on the Fisba Monitor, on channel A
  

  Note: a display of the Fisba screen can also be started from a web browser: http://gracecamserver1 and select "DM" as the source.
  

  Back on AODISPLAY

• Readiness checks only  Initialise the pick-off probe and other mechanisms for the WFS:
  from TopGui’s EngGuis – Mechanisms DM bring up 3 cascaded windows: NAOMI Mechanism control mechs.dl, right-click on NAOMI systems to start Wavefront Sensor wfs.dl, right-click on WFS Components to start Pickoff Mechanisms pickoff.dl. Click Init, Start, wait for the Idle status, and click Index, Start. It takes a few minutes to initialise, and this will leave the pick-off probe at (X,Y) = (5.32,3.31). From TopGui send it back to 0, 0.
  Note: right-clicking on the second brown window allows access to other mechanisms: WFS Filter wheel and WFS ADC; this is the way to initialise them.
  

  

  

  

  

  How to intialise the Pickoff and related mechanisms from the engineering gui's
  
  

• DE checks essentially  If during the previous night the L3 WFS has been used (with OASIS in Tip-Tilt mode), then the camera stage position has to be adjusted so that the Shack-Hartmann spots are  focussed on the CCD39 WFS.  Usually the position to focus on L3 is around 172mm and for CCD39 it's around 210mm. You need to open the  Pickoff engineering gui as described above, then type focussed in the Camera stage, press Enter then Move and Go.
• Check on TopGui if the SG (Strain Gauge) gain is 0.8, otherwise type SG SetGain 0.8

• From the Light Path page load the latest DM flat (usually from last night). On the Fisba monitor with the interferogram, the DM segments will still show lots of fringes.
  
  
  

  TopGui Light Path page

   

• From the C40 page close the SG loop.
  DE checks only If temperature and humidity conditions have not changed dramatically since the end of the night, the flat should still be good enough (on the Fisba display all segments visible and uniformly lit, no fringes or only a couple of fringes visible). If not, a laser (Fisba) flat is needed; see the procedure below.
  
  Readiness checks only  Probably fringes will be lost (completely blurred) on several segments. After a long period of inactivity (> week) try out several previous flats if needed. For Fisba laser flattening process to work fine, an acceptable starting point is when interference fringes are seen on each DM segment.   

• From oasisgui USP page switch on the simplex light, Iris at -150

• From TopGui Light Path page open (if it's not already ticked as open) the WFS Shutter.

• Send the pick-off probe to 0,0. 

• Spots should be visible on the WFS Display. Otherwise (and if the pickoff is at the right position, NCU light is on, shutter open) something has been left on the light path! You can also suspect the CCD60 / CCD39 switch below the WFS to be in the wrong position. 
• Centre approximately the spots on the DM grid (a segment is 2arcsec wide) by entering values in the Pickoff X and Y boxes and clicking Apply, then click on Autocentre from the General Scripts gui to recentre precisely. 

  

  

  General Scripts gui, for AutoCentre, and result of the AutoCentre. To close that pop-up window, just locate the mouse cursor in it and press Enter
  

Laser (Fisba) Flattening of the DM (Readiness checks only)

• Go to the air conditionning console (upper left corner above the TO console, next to the met station) and set Fan speed to 20% with the “knobs” menu 

• From TopGui Engineering Menu launch FISBA. Two pop-up windows appear; one is for controlling the interferometer and launch the flattening process, the other is a display of the interferogram like the one on the monitor above. At first the display is black; click on One Shot to populate the display. The normal Setup is with X gain=0.5, Y gain=0.5 and Z gain=0 which means the laser flattening algorithm does not move the segments in piston.

Fisba interferometer control window

Interferometer display

• If not all DM elements are not visible (i.e. Fringes are completely blurred on some segments) when the SG loop has been closed on one of the latest flats in the list then it is necessary to tweak these segments manually from TopGui Mirror Control page.
  For this, select X -Y-Piston then click on any segment (which turns in blue) and play with the arrows (only in X and Y; it's not the purpose of the laser flattening to act on piston, this will be done later by the White Light flatenning). If the segment turns to red when you select it, a limit in X ,Y or piston has been reached.Set all 3 values to 0 and try moving the segment again.
  It often helps to look for the spot on the WFS Display window in TopGui when you move a segment; then when the spot is roughly in its red box corresponding to the DM segment, look at the Fisba display and see how to get fewer fringes. A good starting point for the automatic algorithm to work is when there are no more than 3 to 4 fringes in each segment.

Mirror control page to move the segments one by one in tip, tilt or piston

• To start the laser flattening automatic process, click on Iterate in the interferometer control window.
• Two counters: Cycle and Phase are decrementing down to 0 and the algorithm stops. A pop-up appears to remind us to turn up the air conditioning fan speed. It's not critical to do that straight away as other laser flats can be needed even after a White Light Flat or a Simplex, but don't forget that temperature shouldn't be allowed to change too much around the DM during the whole check-up procedure. 
• Save this flat from TopGui Light Path page by clicking Save button at the DM line (under Mirrors).
• All DM elements should be uniformly lit (all black or white, fringes disappear). It can happen (especially if the previous flat dates back to weeks ago) that one or several segments are lost during the Fisba process. Then tweak them again by hand, save as a flat and do another laser flat (and save it).
  When the laser flat is good, all segments on the DM are parallel but maybe not in the same plane. This will be corrected by the following process, White Light Flattening.
  

White light flattening of the DM (Readiness checks only)

• Switch on Simplex light, Iris = -150
• Check that the dichroic used is the 50/50 (important to have a broad spectral range)
  
• Set WFS Integration Time to 25ms or 50ms
• Spots have to be bright enough for the algorithm to converge
• From the General Scripts window do an Autocentre
• Check focus of the WFS; the spots separation has to be close to 8 pixels.To do this, go to TopGui WFS Control page (3rd button from the left, or from the Tools menu) and click on the Alignment tool. A new window appears, check separations: Sep X and Sep Y which should be very close to 8 (or symetric around 8).
  
  
  
  If necessary, adjust the WFS fore-optics focus by a few tenths of millimeter from the EPICS guis. In the exemple below, the position was shifted to -0.6mm (the nominal "focussed" position would be 0).
  
  
  
  To adjust the Fore-optics stage position, place the cursor in the Demand Position window, type the value, press Enter so the value is in white, then successively press the buttons Move and Start. I won't move if you don't click Start. Note that Demand Position and Current Position are not necessarily in the same units (it's a mystery).
  
  
• Type AutoWhite from Navis prompt
• Answer 0 (leave alone) to the question about Integration Time 
• The process takes more than 5 min; the X-shifted and Y-shifted lenslet are successively used
• Note the result: we can consider the algorithm has given a satisfactory result if:
  
  • There are more than 120 measurements
  • X and Y gradients are less than 200 in absolute value
    
    
    
  
  
  In the screen capture above, the number of measurements is 163, and the X , Y gradients are a bit high, but acceptable. Note that a new DM flat is automatically saved and its name is the same as the previous laser flat, with a suffix .white appended. This is why it's extremely important to save a laser flat just before doing a white light flat; otherwise the coefficients found by the white light flat algorithm will be applied to the latest laser flat in the directory (which is maybe a week old) and the mirror flatness will be ruined! If it happens you need then to start again by a laser flat.
• As shown above, the profile of the DM can be examined with the command plotflathist (followed by the absolute name of the flat file).
  

• For specialists: If Gradients are high, take out a global tip (X) or tilt (Y) with AddZernike tip –(gradient/100) then the Fisba return flat will need adjusting in Grace, followed by a new sequence Laser/White light flats.
• After Laser Flat and White light Flat, don’t forget to restore air flow to >80% in Grace, otherwise there will be a temperature drift in Grace in the optics room.
  

Perform a Simplex to optimise image quality (Readiness checks only)

• Fix the bench and wait a few minutes before air completely evacuates.
• If the simplex has to be done on INGRID, do the detector sanity checks first and leave the instrument configuration ready for simplex (J, H or K bands).
  
• Start the simplex control window using either AG7, for visible light (if OASIS will be used), or Ingrid for infrared. This is done either by typing in a Xterm window mirroralign --camera AG7 (or -- camera INGRID) or from TopGui on the left side under the black WFS display window, in the Flattening line, at Simplex click to select the camera. Two new windows appear: one for the display and one for the control.
  
  
  
  Simplex control window (here with AG7)
• From the control window you can turn on/off the image grabbing mode (check/uncheck Continuous Display), modify the size and position of the display window (X and Y are the pixel coordinated of the upper left pixel), change the camera integration time (Exposure).
  

  
  
  
  Simplex image display
  

  
  
  
  
• To simplex with Ingrid, take out OASIS and Simplex mirrors (from USP tab on oasisgui),otherwise (simplexing with AG7) you need both mirrors in.
• Use the 50/50 dichroic to simplex with AG7, and the IR dichroic to simplex with Ingrid.
• Turn on the Continuous Display. The window position is the one that was entered by the previous user, so normally the image of the light source should be visible. If it's not there, you can display the whole CCD and look for the spot coordinates, by unchecking Clip Images. Then place the cursor on the spot, its coordinates are shown below the display.
  
• Before starting the simplex algorithm, reduce the window size to 30x30 pixels and recenter the image. If the window is too large the simplex algorithm has too much information to process and takes longer.
• When simplexing with AG7, adjust the Filter Focus position from oasisgui USP page (there is a simplex position at 2.20mm but lately it seems better focused at 3.0mm). If the image is defocussed the simplex algorihm will try to correct this by adding a focus term on the DM.
  
• Adjust the Simplex lamp Iris from oasisgui USP page and the camera integration time so that the maximum intensity (in the Max: box under Clip Images) is between 10000 and 20000. It has to be bright enough at the beginning, but not too much to avoid saturation as the simplex converges. Typically set the simplex iris at -250 to -290, and 0.015s to 0.030s for the camera exposure time.
• Click on Simplex
• Monitor the Merit Function (MF): image is considered good when MF < -0.015 (with AG7) or MF < -0.03 (with Ingrid)
  
  
  
  
  
• Usually image starts improving after 800 to 1000 cycles.
• If after 2000 cycles MF doesn’t improve, stop simplex then start again: the algorithm may be stuck in a local minimum
• When the MF is satisfactory, stop the simplex algorithm by clicking the Stop adjust button.
• Check the image quality: disable Continuous Display in the simplex control window,  then from Taurus take an image: run AG7 0.1
• Display the .fits file with Iraf, use imexam j and k commands, or r for the radial profile. If the simplex was good the image should look symetric, with FWHM of the order of 2 pixels. The example below  shows a good, but not perfect image: there is still energy in the wings around the central peak and the gaussian fit is not perfect.
  
  
  
  

  Example of a good image profile

• From TopGui save again the current shape of the DM as a flat. Now the mirror flattening process is complete and NAOMI is ready for observing. Before leaving the instrument to the astronomers, it's just needed to check that the detectors themselves (OASIS or INGRID) are healthy and receive light.  
  

Checking the image quality on the Simplex Camera AG7 (Readiness checks only)

• From oasisgui USP page move OASIS and Simplex flats to be IN the beam. Edit the value of Filter focus, set to 3.0mm. With the Simplex light ON, set the iris to -250.
• From Taurus ICS window take an image: run 0.1 AG7
• Display the .fits file with Iraf, use imexam j and k commands, or r for the radial profile. At this stage if the image looks good enough, fwhm of the order of 3 pixels, no further flattening of the DM is needed. Otherwise refer to the complete procedure for mirror flattening.
  

Checking light reaches OASIS 

• From oasisgui USP page fix the bench (it takes a few minutes for the air to flow away). Simplex light ON, set the iris to -150, use 50/50 dichroic. Then take the Simplex flat OUT, OASIS flat IN; Filter focus to imaging.

• From oasisgui OBSTOOL page, in Set Demand select  Observing mode Imaging, with any filter as Spectral config, for example MR516. Then in Configure tick Oasis and click on Configure Selected Items.

• The instrument mechanisms can also be configured one by one from OASIS (Eng) page, it only takes longer and there's a risk of forgetting something. You should check that you end up with something like this:
  
  
  
  
• From Taurus set the standard CCD window: window oasis 1 "[1:2059, 1035:3140]" then take an image: glance oasis 1 and display the file s1.fit with iraf, the spot should be visible.
  

Checking light reaches INGRID (only if it will be used on the following night)

• From oasisgui USP page fix the bench, move OASIS Flat OUT of the beam. Use the IR dichroic. Switch on the Simplex light, with Iris to -150.
• From oasisgui INGRID page, configure Ingrid, for example choose Simplex H and take the pupil imager OUT.
• From Taurus tke an image: glance ingrid 1 and display s1.fit with Iraf (use the package ingridql instead of the basic ing package and display the file using the command idisp. The spot should be visible.