Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.naic.edu/~phil/pnt/turretscans.html
Äàòà èçìåíåíèÿ: Mon Apr 22 20:40:49 2002
Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 04:12:54 2012
Êîäèðîâêà:

Turret scans.

apr02

Why (or why not) use turret scans?

To measure the telescope response of a source, you can take two or more strips through the source by moving the azimuth, moving the dome, and possibly moving both together. Moving the azimuth through small motions and then reversing directions causes oscillations in the azimuth arm especially when the dome is at high zenith angle. At low zenith angle a constant angle on the sky requires a large move of the azimuth arm because of the 1/sin(za) dependence.
An alternative is to move the horn in the image plane. The motion is great circle (constant angle) on the sky so there is no 1/sin(za) dependence. It also limits any oscillations to the feed tower rather than the entire azimuth arm (this may or may not be a good thing). This motion can also be done much faster than moving the azimuth arm. An az strip will take on the order of 1 minute while the turret/feed can perform an 8 second sine wave for the same distance on the sky. The drawback with the turret motion is that you are moving away from the paraxial ray to do the motion. The results that you get have been convolved with the field of view response of the optics. This can be a small problem if you are interested in pointing offsets. It will be harder to deconvolve if you are looking at the beam pattern generated.

How turret scans are done.

    A typical turret scan will drive the dome from -5 arc minutes to +5 arc minutes in two minutes at constant velocity. This will cut through the source in the za direction. While the dome is moving, the turret will perform a sine wave of 8 turret degrees in 15 seconds (1 turret degree=45 arc seconds on the sky). This motion gives a 2-d sampling of the source/beam. If the beam is 3 arc minutes (lband) and the dome has been offset by 5 arc minutes, then the turret will perform 120sec/15sec=8 cycles of the sine wave during the two minutes. There are 10/3=3.3 beams in the 10 arc minutes so the turret will cut through the source (0 turret offset) 8*2/3.3 = 5 times (*2 since a complete cycle cuts through the 0 position twice). You can fit a 2-d gaussian to the sampled data to get the pointing offsets, amplitude, and beam widths of the measured beam. This is like making a beam map every 2 minutes (as long as you remember that is has been convolved with the field of view response in the azimuth direction).

    You need to sample the data fast enough so that the rapid motion of the turret through the source does not smear out the source peak. Each polarization is detected with a 20 millisecond time constant and sampled at 100 hz with the 12 bit A/D's in the radar interface sampling system. The dome, turret, and sampling system are all started on a 1 second tick to synchronize things. The 2d-fit includes the phase of the turret sine wave as one of the fit parameters to catch any delays in the turret system. The pointing errors need to include the time constant delay when figuring out which sample should be the source transit. You can solve this by tracking a source rise to set using turret scans that have the dome going up and then down. Look at the zenith angle errors for the up and down going scans. Systematic offsets occur when the time constant is incorrect (since the time constant gives a positive za error going uphill and a negative za error going downhill).

    The distance the dome moves and the amplitude of the turret scan need to be tailored to the receiver that is begin used. The 10 arc minutes barely gets to the first sidelobe of the 3 arcminute lband beam. If you used the same parameters for the 30 arc second xband system, the turret would only cut through the source once or twice. Below is a table of parameters that have been used for the various receivers:

receiver	beam width	dome motion Amin #beams	turret Amplitude	turret period secs
lband (1400)	3.2 Amin	+/-5 = 10Amin , 3 beams tot	8 turdeg, +/-6Amin=3.75 beams tot	15
sband (2380)	2 Amin	+/-5 = 10 Amin,5 beams tot	8 turdeg, +/-6 Amin=6 beams tot	15
sbhi (3500)	85 asecs	+/-5 = 10 Amin,7 beams tot	8 turdeg,+/-6 Amin=8.5 beams tot 3 turdeg,+/-2 Amin=3.2 beams tot	15 8
cband(5000)	60 asecs	+/-5= 10 Amin,10 beams tot +/-5= 10 Amin,10 beams tot +/-3= 6 Amin ,6 beams tot	2 turdeg,+/- 1.5Amin =3 beams tot 3 turdeg,+/-2.25Amin=4.5 beams tot 3 turdeg,+/-2.25 Amin=4.5 beams tot	8 8 8
xband(8800)	35 asecs	+/-1.5=3 Amin,5 beams tot	2 turdeg,+/-1.5 Amin=5 beams tot	8

 home_~phil