Pointing run tactics

The items below are instructions or logical constraints or timing constraints that I believe govern the choice of pointing sources during a pointing run. These may help other JCMT staff executing a pointing run as well as providing a basis for an algorithm that could completely specify the sequence without further human interaction.

      0. I shall assume the use of SCUBA and the restriction of sources to blazars and small-diameter planets.

1. At start, note the LST, tau_cso, seeing and location of planets. Decent pointing runs may be made with tau_cso < 0.15 provided the submillimetre seeing is < 1".

If a planet is available, use it for the first pointing and for focussing the telescope in all (Z,X,Y) axes. n_integrations (hereafter ni) = 1 will suffice for all sensible tau_cso. Point again - this is perhaps your first useable datum.
2. If the 2 or 3 hours duration of the pointing run occurs around dusk or dawn, the focusses will need redoing every hour - on something suitably bright.

3. A useful pointing run could be comprised of 20 data, although 30 or more is ideal.

4. These data should be uniformly fill the relevant part of (az,el) space : 5 < az < 445, 25 < el < 85. Ideally this gives ~6 points per quadrant, or 1 point per 15 degrees of azimuth, and, within each quadrant, 1 point per 10 degrees of elevation.

   [ Data acquisition occurs at a rate of ~10 per hour. ]

5. Additionally, the schedule for some particularly useful sources with 10<declination<30 must be anticipated in order to observe them when they attain el>80.

6. These same sources may be usefully observed when rising, and, 20-30 minutes later when setting, and/or when transitting, all with el>80.

7. One such source transitting in the south and one in the north during the course of the run would be ideal. There are <20 such (blazars) available so their passages through el>80 are important events to catch. (Note that only 4 of them are brighter than 1 Jy at 850um - marked * below at 2005 Jan).

   These sources include

      0149+218  0235+164* 0528+134* 
      0735+178  0745+241  0748+126 0754+100* 
      OJ287     1012+232  
      1147+245  1156+295  VirgoA 
      1354+195  1413+135  1538+149
      1606+106  1923+210
      2230+114* 2234+282  2251+158*
   

8. Check current brightnesses in the pointing catalog, but those fainter than 0.5Jy may require ni=6. But they are fairly critical to the model generation, so it's worth spending this time on them.

9. Observing a couple of sources with 15<el<25 *may* help to pin down the flexure term (TF in TPOINT), but it is best to concentrate on the sources brighter than 1Jy; the large airmass makes observing fainter sources difficult, time-consuming and ultimately frustrating.

   Observing at el<15 is unusual and it may be argued that the model need not be accurate there. The main problem with obtaining pointing data at el<15 is that atmospherics may dominate the elevation pointing errors and these data will skew the determination of the pointing model.

10. At this point, it is worth remembering that the idea is to measure pointing errors at particular positions of (az,el) : if the duration of each POINTING is so long that az or el cannot be precisely defined (to within, say a degree or two) then the value of the datum to the dataset is reduced. For this reason observing (quickly) a bright source is preferable to spending longer on a faint source, unless that source fills a particular gap in the (az,el) plane.

11. Otherwise, the order in which to observe the sources should ideally be randomized. However, practical matter such as the need to minimize the time elapsing implies a need to minimize slew times. At the same time there should be some randomness in the sequence so that hysteresis effects do not accumulate : ie. successive targets should be randomly left and right, up and down : successive slews should be randomly CW and CCW, up and down . . . (it's non-trivial to meet all these constraints . . . - just do your best !)

12. Introduce also the occasional v. long slew, and the occasional 360deg slew - to the same target in the 1st or 5th quadrant.

    The randomness demands of 12 & 13 and the efficiency demands of 12 are essentially incompatible and a compromise is desired.

13. In poor or variable conditions return regularly to bright sources or planets to reestablish good pointing.

Return to POINTING home page .