[mmaimcal] some antenna notes (as promised, but late...)

[Bryan Butler]

some notes from our last discussion of antenna pointing issues:



the antenna pointing spec is currently:
   0.7 arcseconds offset pointing accuracy, with pointing calibration
   every 15 minutes.  blind pointing accuracy is to be 2 arcsec
   (this is a new spec which peter told me yesterday).


the 4 antenna motion specs are:

 1 - Fast switching
   1.5 deg (to within 3 arcsec) in 2 seconds

 2 - OTFM (On-The-Fly-Mapping)
   0.5 deg/sec - positions read off accurate to 0.7 arcsec, but 
   antennas do not need to be pointed at the same position

 3 - OTFIM (On-The-Fly-Interferometric-Mosaicing)
   0.05 deg/sec - antennas point to same position to within 0.7 arcsec

 4 - slew to new source
   1.5 deg/sec in El, 3 deg/sec in Az



issues we discussed:

 - regarding the pointing spec and offset pointing, min has been looking
   at this in detail lately.  he seemed to think that given 50 12-meter
   antennas (a reasonable number, given our latest estimates), you could
   easily do a very accurate pointing calibration in less than 1 
   minute.  he was considering using the pointing determination scheme 
   currently used at OVRO, which is several repeated triangles at the 
   HP point.  several others of us pointed out that the continuous 
   circles or other schemes may actually be better, and min might look 
   at these other schemes.  min's analysis is similar to, but differs 
   from, robert lucas' analysis in MMA memo 189.

   one thing that might be of use here is to have min actually calculate
   the estimated accuracy in the pointing determination if we are
   willing to allocate 1 minute out of every 15 minutes to pointing
   calibration.

 - regarding the fast switching, we briefly discussed how many FS cycles
   might be undertaken by the array, which concerns the projected
   lifetime of the array.  harvey pointed out that FS cycles don't 
   actually stress the structure very much, since the travel distance is
   so small.

 - again, on fast switching, we briefly discussed whether we need to 
   have a spec on the time required to actually reach the final pointing
   spec (i.e., rather than 2 seconds to get within 3 arcsec, have the 
   spec be X seconds to get within 0.7 arcsec).  

   [peter pointed out to me yesterday that once you get "close" (like 
    to within 3 arcsec), it is only a couple of antenna oscillations 
    beyond that to reach the required pointing, so that if the resonant
    frequency is 10 Hz, then it is only a couple of tenths of a second 
    to get the final bit of pointing accuracy.  see figures 8-10 of 
    torben andersen's MMA memo 231.]

 - regarding OTFM, a brief discussion resulted in the general agreement
   that the spec'ed 0.5 deg/sec rate was sufficient.

 - regarding OTFIM, a longer discussion seemed to reach no conclusions.
   there is still some confusion among us as to how OTFIM measurements
   will actually be undertaken.  this is all tied up in how the total
   power measurement can be done.  if the total power measurement is 
   done separately, in OTFM mode, then the 0.05 deg/sec is probably
   sufficient.  if the total power measurement is done at the same time
   as the interferometric measurement, then the 0.05 deg/sec is not
   fast enough, and you probably need the same 0.5 deg/sec rate as the
   OTFM mode has.  in the case that the total power measurements are 
   done in OTFM mode, you still probably want the measurement to have
   the same SNR as the interferometric measurements, which implies that
   it might take of order twice as long to do a mosaic than if you
   collect the total power at the same time as the interferometric data.

 - regarding slewing to a new source, brief discussion met mostly with
   the telephone equivalent of blank stares, implying to me that the 
   spec'ed rates are probably OK.