Previous: Acquisition
Up: OPERATING THE TELESCOPE
Next: To check and correct the INT prime focus geometry
Previous Page: Acquisition
Next Page: To check and correct the INT prime focus geometry

Autoguiding

1. Location of a guide star on the PSS or other survey plates

   There is no need for the observer to measure guide star coordinates in advance as selection on the finder field is simple and quick. It is, however, worth checking that there is a suitable guide star in the field accessible to the probe, especially if there are constraints on the exact position angle of the CCD chip on the sky. This is because rotation of the Prime focus assembly is very slow. A scale diagram of the chip and guide-probe fields is shown in Figure 1.10. In order to check that there is a suitable guide star, place a scaled version of the overlay on the PSS with the chip outline centred on the object to be observed. Rotate the overlay to your chosen position angle (180 is standard if you do not care, as it avoids unnecessary loss of liquid nitrogen). Check that there is a star with 6 < < 14 in the probe field. If there is not, try again with the overlay rotated by 180 (if you care about the orientation) or any angle (if you don't, but remember that cardinal points give better displays). It is unheard of for there to be no suitable guide stars at any PA. Make a note of the approximate PA but do not measure the position of the star. If you happen to have the coordinates of a guide star however, then these can be entered on the user interface.

2. Acquiring a guide star using the finder

   (NB Much of what follows - if not all - will be done for you by the INT Telescope Operator.)

   The standard method of acquiring guide stars is to locate them on the wide-field finder display. Having reassured yourself in advance that a suitable star exists (see section 1.6.1) and having zeroed the finder coordinates at the beginning of the night (see section 1.5.2), proceed as follows:

   Select FINDER and DIRECT on the Grinnell.

   Offset the telescope so that the probe field is visible on the finder by pressing the BEAMSWITCH A button on the control desk. This offset is in a coordinate system fixed in the Prime focus unit so its value on the sky depends on the rotator PA. Stars should appear to move along the major axis of the chip as the beamswitch takes effect. The correct offset is set up on startup of the control software and should not be overwritten. To check the value, type ADVISE/TELESCOPE (abbreviable to AD/T) on the user interface console. Aperture A should be set to 18, -580 arcsec. If you have accidentally overwritten it, type:

   UI-> AP/A 18 -580

   The aperture in use is shown on the Information Display with the offsets converted to RA and Dec.

   Now choose GET GUIDE STAR on the Grinnell menu. A display of the probe field will appear on the screen (Fig. 1.10). Move the cursor to your chosen guide star and press ENTER. The coordinates of the star will be transmitted to the instrumentation computer which will then move the guide probe to the correct position. A star will be bright enough for autoguiding if it is visible on the finder with 4 frames of integration at full TV gain. Avoid very bright ( < 6) stars.

   Note. If acquisition of guide stars fails, check whether the TV finder scale has been altered (e.g. by replacement of the TV camera). Nominal values are 3.38 arcsec/pixel in X and 2.75 arcsec/pixel in Y. Use TV_SETSCALE to alter the TV scale values (see subsection 3.1.8).

   

   Now press BEAMSWITCH NOMINAL to return the telescope to its initial position and DIRECT on the Grinnell to unfreeze the display.

   You are now ready to acquire the guide star.

3. Guide stars with known RA and Dec

   Suppose that you want to observe an object called NGC 741 which is already in the user interface catalogue and to enter the coordinates of a guide star directly (e.g. if the finder is broken). Type the following on the user interface console:

   

   N.B. Object and guide star positions must be in the same coordinate system (e.g. mean 1950).

   

   At this point LIST on the user interface console will display your catalogue with positions and associated guide stars. Now type:

   

   and press the GUIDE STAR button in the row labelled GUIDE PROBE on the A&G panel. The probe will move to the correct position and you will be in the same state as at the end of 1.6.2.

4. Guide stars with known probe coordinates

   If you know the probe coordinates (e.g. from an earlier observation), type:

   Adam:> GDXY < x y>

   on the instrumentation computer console, where x and y are the coordinates printed out by
   PFSTATUS, shown on the Mimic display and written in the observing log. These are in encoder units (arcsec) despite PFSTATUS's allegation that they are in mm.

   In this case, you may well wish to acquire the guide star by moving the telescope rather than the probe, as this ensures that your field is in exactly the same position on the chip as it was during your earlier run (see 5, below).

5. Acquisition modes Move probe or telescope?

   There are two ways of moving the guide star into the centre of the autoguider field so that the guiding loop may be locked: by offsetting the telescope, leaving the probe stationary or by moving the probe whilst the telescope tracks undisturbed. Their advantages and disadvantages are:

   • Move probes. The field viewed by the detector will not shift by more than an arcsecond or two. Location of a guide star is normally much less accurate than the pointing of the main telescope (because of finder flop), so this is the default mode. It will not work if the interprocessor link is broken.

   • Move telescope. If the guide star position is known precisely, either in RA and Dec by accurate measurement or in probe x,y from a previous observation, then this mode can be used to make a blind offset. This is not very important with a CCD, but acquisition is also significantly faster than if the probes are moved and the process is independent of the interprocessor link (IPL). If you do not have an accurate position for the guide star, then you should centre it in the autoguider field first by moving the probe (from the instrumentation computer using GDXY if the IPL is broken).

   In summary, normally, move the probe. If you have a probe position for the guide star or if the IPL is broken, move the telescope.

   In order to switch modes, type

   UI> AC/TEL (= ACQUISITION-MODE/MOVE TELESCOPE) to move the telescope and

   UI> AC/PR (= ACQUISITION-MODE/MOVE PROBE) to move the probe

   on the user interface console.

   

6. Setting up the autoguider and acquiring a guide star

   This procedure is the same whether the telescope or the probe is to be moved. Select A/G on one of the TV switching channels. If the autoguider has not been turned on (signified by a blank screen), press the button below the fourth mini-monitor at the top right of the control desk (red = on/white = off). This turns on the HT supply. It is interlocked with the main dome lights. The autoguider display should come up with a frame and an ``idle'' message (Fig. 1.12(a)). If it does not, press SHUTDOWN on the autoguider panel. If there are still no signs of life and the autoguider panel appears to be active, call the Duty Technician.

   Now press FIELD. This causes the autoguider to scan a field about 2 arcmin square (the outer parts of the field are heavily vignetted at Prime). If all is well, an image should appear on the 32x32, 3-bit autoguider display (Fig. 1.12(b)). The program selects the brightest pixel and marks it with a cross - this will normally be the guide star. If the ``image'' occupies a single pixel, repeat the FIELD after pressing WAIT to stop integration, as random events look very similar to faint stars. If an image is found consistently in the same place, then it is probably real. The ``magnitude'' of the star is said to be roughly . Magnitude 14 is acceptable for guiding.

   If you are happy with the guide star selection, press ACQUIRE, if not, press WAIT and pause for thought (see ``Problems'', below). ACQUIRE passes the coordinates of the star to the telescope control computer and requests that the guide probe be moved so as to position this star at the centre of the autoguider field. This is done either by sending a message over the interprocessor link to the instrumentation computer (for probe movements) or directly (by moving the telescope). When the repositioning is complete, the autoguider does another field integration. It then switches from raster to cross scan and starts to measure the position of the star with greater accuracy. This is the ``pulling-in'' phase as is indicated on the screen. Messages are sent to the control computer, which attempts to move probe or telescope to minimise the error. When the error drops below a preset threshold, scanning ceases and x and y profiles of the star are displayed with a message ``OK to guide?'' (Fig. 1.12(d)).

   At this point, press GUIDE unless the profile is double-peaked (double star or out of focus) or very weak. Wait until the guiding display (Fig. 1.12(e)) is complete and then press the LOOP ENABLE button on the A&G panel, which should turn green. Note that the loop is locked by the action of switching from loop disable to enable rather than by being in the loop enabled state. Wait until the cross (star position) is centred in the circle (autoguide field). You are now ready to observe.

   The autoguider display should be monitored during observing in case the signal becomes weak (signified by the message ``cloudy''). On the right of the guiding display (Fig. 1.12(e)) are histograms of the last 128 guiding errors in x and y (incorrectly labelled RA and Dec or possibly vice cersa). On the left are: current state (GUIDING), warnings (e.g. CLOUDY), an approximate visual magnitude, a measure of the relative transmission (RT, on an arbitrary scale) and seeing estimates from the profiles. (Warning: these are not linearly related to the true seeing).

   The cross should always remain within the circle. If it strays, you have probably lost the guide star (is there any signal?)

   If you need light in the dome (even tungstens, unless they are very dim), turn off the HT.

7. Problems
   1. There is no star in the field. Is the beam obscured? Can you see anything through the telescope? Are mirror covers, dome, and shutters open? If you typed in RA, Dec or x, y, were they correct and within the limits (otherwise, the probe will not be moved)? Is the star visible on the finder with 4 frames of integration? If not, it will be too faint for the autoguider. Was the star close to the boundary of the box on the finder? If so it might actually be over the limits. Have you gone back to beamswitch nominal? It sometimes happens that, because of hysteresis in the finder, the guide star is out of position in declination. Try moving the probe by 50 arcsec in an appropriate direction (see Fig. 1.11).

   2. The autoguider selects the wrong star (e.g. a close double). The automatic selection of guide stars may be overridden using the light pen. Do this in WAIT state.

   3. Double-peaked or other non-Gaussian profiles. Is the star double? (look on the finder!). If so, you probably cannot guide on it. If both profiles are double-peaked, the autoguider is probably out of focus. Refocus using GDFOCPOS and SETDGFOCUS on the instrumentation computer. Asymmetries in the profiles often occur when the autoguider does a scan before the acquistion movement has quite finished. This is nothing to worry about - it should guide correctly.

   4. The autoguider seizes up whilst guiding. Press SHUTDOWN to re-initialise the autoguider and repeat the acquisition sequence.

   5. The autoguider refuses to acquire by moving the probes. Have you asked the probes to go somewhere illegal? Is the interprocessor link active on both computers? If not, try reloading it.

   6. The autoguider or A&G panels will not respond. Do a D M on the telescope control computer's system console. Are the tasks AUTOGUID and AGBX (respectively, for the two errors) present and active (A)? If not, cancel and reload the relevant one:

      RESTART AGBX

      (or the same, with AUTOGUID instead of AGBX)

   7. The Grinnell is re-initialised during normal operation (e.g. by pressing the RESET button by accident). In this case, the Prime-Focus overlay may not appear because the Grinnell may think that the Cassegrain system is required. To fix this, type:

      Adam:> RESETTV