Search the AAO:
General Links
Professional Links

Setting up Nod and Shuffle Observing with AAOmega

There are a number of complexities when observing with AAOmega using Nod and Shuffle mode.  This guide considers those issues that relate to standard Nod-and-Shuffle (N+S) were half of the 2dF fibres are disabled in order to make space on the CCD to shuffle the A (and B) position observations into during B (or A) position observations.  A mini-shuffle option, using all 392 science fibres at once but which can only be expected to give sky subtraction at a level intermediate between full N+S and typical dedicated sky fibre observations, is available but is not yet ready for routine operations.

Observing sequence
Making a Nod-and-Shuffle .fld file
Running Configure
Attaching the fibre mask to 2dF the fibres
Pitfalls

Observing sequence

  1. Define your telescope Nod.  The telescope needs to be nodded onto blank sky every 30-120seconds during Nod-and-Shuffle.  If Cross Beam Switching (CBS) is to be used, then the fibres are configured in pairs so that most (but not all, due to practical limitations) target are observed in both the A and B positions, through different fibres.  CBS  is usually the best mode of operation, since it is less prone to weather mishaps than observing twice as many targets at once, but for twice as long, without using the CBS.  The AAT Nods most quickly in Dec, so North-South offsets of around 30arcseconds are typically best.  The user should check for bright stars in the B position field.  If larger Nods are required (perhaps a number of arcmins) then the telescope settle time dominates the overhead and so RA and/or Dec offsets can be used.  Configure can be used (see below) to generate the correct B positions for CBS observations.
  2. Create a .fld file for Configure.  This should have an appropriate number of targets (this can be a complex definition and depends on the observing mode and requirements, see below).
  3. Create a tdFconstants400.sds file for Configure which has 200 fibre disabled using set_fibre_state.
  4. Configure the field in the normally way, there are a number of options available (Cross Beam Switching, guiding in the A and B positions, allowing some sky fibres for comparative tests, adding calibration stars etc.) more detailed notes on this are given below. 
  5. Configure the field with 2dF as normal.  Note the tdFconstants400.sds file should NOT change for the positioner, JUST for Configure.  Since you have NOT allocated the fibres you wish to mask off, They will remain parked, and so can be masked off even though the robot things they are active.
  6. Slew the telescope to prime focus if it is not there already.
  7. Move the gripper clear, using the tdfeng interface.
  8. Find the metal masking plate(s), this are usually stored against the railing next to the 2dF top end, often cabled tied to the rail.  Some afternoon scouting is a good idea here.  Clip the mask into place.  It is magnetic, but some rivets in the mask have to align with gaps between retractors and the mask is a tight fit.  Take time to make sure it is correctly secured (i.e. does not rock) It must be  a snug fit, it must not fall off onto the mirror.  Take as much time here as you need to be confident.
  9. TUMBLE the field plate into position.
  10. If you have been using two N+S masks, take the first one off the other plate Now.
  11. Slew to the field After you have taken off the other mask if there was one on that field plate.
  12. Take a flat field and an arc frame using the normal CCD observing strategy.  DO NOT N+S these frames.
  13. Acquire your fields with the fiducials and guide as normal.  Note, if you are using CBS they you will probably only have a few of the guide stars in each position.  Don't disable the ones not in the A positions, since you need them in B.  You may wish to have the night assistant perform the Nod by hand at this time, in order to check that the B position guide stars (if any) are acquired. 
  14. Using the CCD control task, as outlined in the AAOmega control task manual, change the CCD mode to Shuffle.  Note DO NOT use the Beam Switch or Cross Beam Switch options, these do something different (they Nod but do not Shuffle).
  15. Set the exposure time, T, to be the time for each sub exposure (e.g. 30sec, 60sec or 120sec).  Shorter exposure will give better data but will have a larger slew overhead.  60sec has been seen to wrok well in good conditions.  
  16. Set the number of Cycles, C, to be the number of A+B pairs you wish to observer.
  17. The total exposure time is given by T x 2 x C since an exposure of length T will be done at each position A and B and a number of cycles, C, of this pattern will be observed.
  18. By default, the Shuffle mode sets the CCD charge shuffling to 10 (or -10 either is fine) pixels.  One should check this is correct (or modify it if required) by selecting the More options option at the bottom of the CCD setup window.
  19. Set the Nod-and-Shuffle telescope offset in the Offset Details subsection.  This region will be grayed out before the Shuffle option is selected.   These number are the arcsecond(s) offset(s) on the sky that will be used to Nod the telescope.  If you are using CBS (switching between two fibres per object, and so observing each target 100% of the time but with a smaller overall target yield) then these numbers should be the numbers that you entered into the Configure Cross Beam Switchingwhen configuring the field (see below).  Note, as discussed above, the telescope offsets quicker in Dec than in RA, so a simple Dec offset (say +/- 30arcseconds) is usually the best strategy.  For larger offsets the time to Nod is dominated by telescope settling time and so RA and Dec offsets can be combined.
  20. Start the cycle, observe your fields and reduce the data in the normal way.  Note that with N+S observations the CCD readout noise is effectively doubled (since you have an A and B position) and the on-target observations is half of the exposure time.  For this reason, even in low resolution Red data, we suggest 40minute exposures (120sec x 10 cycles x 2 positions).  Longer exposures would become dominated by cosmic rays.  3x40mins (with 120sec x 10 cycles) has been seen to work well.
  21. Reset the 2dF observing system (i.e. remove the mask, put the CCD control back to normal) when you have finished with it, this will save a lot of heart ache for others.

Making a Nod-and-Shuffle .fld file

All of the usual considerations for SAconfigure and Complex configurations apply equally to N+S observing, with some additional considerations.

Running Configure

A temporary version of the tdFconstants400.sds fibre availability file is needed for Configure.  This temporary file has half of the fibre disabled in software (although they will remain active for the 2dF robot) so that they are not allocated when setting a field up.  This leaves space on the CCD for fibres to be shuffled into during observations.  To make this, follow the instructions on how to use the set_fibre_state utility.  This should be done on a user account on aatlxa.  The current tdFconstants400.sds file which needs to be modifed can be found on the ftp site or from:

cp /instsoft/2dF/positioner/tdFconstants400.sds .

The default directory from which Configure will source it's setup files must then be changed in order for Configure to pick up the new tdFconstants400.sds:

setenv CONFIG_FILES /path/to/the/setup/files/

When you start Configure check it picks up this new file (disabling half of the fibres on the field plate) AND NOT the defaults off the /instsoft account.
You may want to copy over the astrometric files as well (note that Configure looks in the default locations if it cannot find any of the required files, so you should not need to copy these files when running on aatlxa).  These are found either on the AAOmega ftp site, or from :

cp /instsoft/2dF/config/tdFdistortion0.sds .
cp /instsoft/2dF/config/tdFdistortion1.sds .
cp /instsoft/2dF/config/tdFlinear0.sds .
cp /instsoft/2dF/config/tdFlinear1.sds .

When you start Configure, CHECK that the intended modified tdFconstants400.sds file was used.  If the xterm is restarted, then one must reset the environment variables each time.

Note that at present (May 2007) the software disables every other 2dF pivot.  This is not quite what is wanted, since it is the Slit position and not the 2dF Pivot position that is important. This will be fixed as soon as possible.
  1. Load your .fld file in the usual way
  2. If cross beamswitching, select Commands -> generate CBS pairs. Enter the CBS offset in the pop up, and select the Include fiducials flag (default).  Your target list will now double in size (although no second position is generated for any sky positions).
  3. Select Allocate to bring up the target allocation menu.
  4. If you are Cross Beam Switching, then select the CrossBeamSwitching flag.  This gives extra weight to objects that can have a fibre allocated to each part of the pair of fibres.  Some targets will still only get a science fibres in the A or B positions, but many more will get both.  Note, this is NOT the same as the Weight Close Pairs option.
  5. If you are using CBS, then you might select the Weight fiducial target pairs flag.  Typically you will have to inspect the fiducial allocations here anyway and check that a sensible distribution of Fiducials has been allocated.  There is an argument for allocating perhaps 6 Fid to the a position and 2 to the B position.  This way one can acquire well in A and then trust a small Nod-and-guide in B will be fine.  Alternatively, allocate the fibres 4 to A and 4 to B.  Note, you CAN use different stars in A and B where required, just make sure you have enough A or B positions, and not all 8 B positions.
  6. Save the .sds file and configure it in the normal manor with the positioner.

Attaching the fibre mask to 2dF the fibres

Half of the 2dF fibres at the top end need to be disabled and masked off.  This is achieved in software at the Configure level (by simply not allocating half of the fibres), and is described above.  At the hardware level, half of the fibres are left park while half are configured.  The parked half are then masked off using a magnetically clamped mask which covers the parked positions.  The masking system includes an interlock with deactivates the gripper Z axis, preventing the griper being lowered to the field plate when the mask is in place.   To apply the mask follow the steps outlined above.  Take care here, it would be easy to damage fibres, or not to fit the mask correctly and have it fall off.  Take all the time you need to secure the mask correctly and carefully. The mask fits securely when mounted.  The three magnets tightly grip the base of the outer ring of the mask.  However, there is a lip on the inside of the ring which can catch on the top of the retractor over one of the magnets.  The ring will feel secure but will in fact only be attached to two of the magnets.  Check the outer ring is flush with all three magnets or it may fall off.
  1. Slew the telescope to prime focus if it is not there already.
  2. Move the gripper clear, using the tdfeng interface.
  3. Find the metal masking plate(s), this are usually stored against the railing next to the 2dF top end, often cabled tied to the rail.  Some afternoon scouting is a good idea here.  Clip the mask into place.  It is magnetic, but some rivets in the mask have to align with gaps between retractors and the mask is a tight fit.  Take time to make sure it is correctly secured (i.e. does not rock) It must be  a snug fit, it must not fall off onto the mirror.  Take as much time here as you need to be confident.
  4. TUMBLE the field plate into position.
  5. If you have been using two N+S masks, take the first one off the other plate Now.
  6. Slew to the field After you have taken off the other mask if there was one on that field plate.

Pitfalls

All of these seem quite common mistakes.  They are all quite serious, will cost time to fix, and can be avoided by a careful observer who takes proper notes and follows the instructions above. Please mail comments and suggestions to : Rob Sharp (rgs@aao.gov.au)