What we've done and settings used:

170809: manually tracked sources.

•  turret = 160 was felix's estimate of turret position for center of horn.
  
• offsets needed to point:
•  azOff=-.025 deg = -90Asecs (gc)
• 45asecs sky /turret deg.
  
• So increase turret by 2 degrees (increasing turret  moves in ccw direction).
  
• zaOff=-.0134 deg = -48.2 Asecs
• temporarily corrected  with model:
• -93.52 - 48.2 = -141.72 (should have been)
  
• ** I accidentally used the cos(az) term=-97.13 so o got -145.37
  

170811- calibration scans, first try. (top)

    Used 15:00 to 16:30 to run some calibration scans on cbw.
 The conditions were far from ideal for pointing,calibration tests (rain, no tiedowns).

The horn was sitting at the initial position (when first installed).

• used same freq as cband: 4500,4860,5000,5385 MHz
• the cal values are probably not too accurate (high cal, low cal give different Tsys values). sefd is ok.
  
• 15:00 - 15:28 B1117+146 (about 1Jy) at 5ghz
• lots of rain, temp 73F
  
• tiedowns not working. avg jack height about 10"
• platform probably 2 inches high.
  
• 15:37 to 16:18 B1328+254 (3.2 Jy at 5ghz)
• rain stopped, still lots of clouds, tiedowns still not enabled.

• Page 1: gain,tsys,sefd, beam width
• sefd: 6 to 10 Jy (5385 probably has rfi).
• beamwidth: 55 to 68 asecs...
  
• at 5385 cband gets down to about 52 asecs..
  
• this could be coma, or just out of focus because of the rain.
• Page 2: coma, sidelobes, beam efficiencies.
• coma parameter gets to .2.. cband is less than .1.
• sidelobe -13 db average
• 
  
• The table below compares the B1328+254 source at za=10 deg using cband(mar17) and cbw (today)
  

Beam maps:

    Beam maps were reconstructed from each calibration scan  for B1328+254.

•  8 maps were made for each of the 4 frequencies
  
• for each frequency, 4 maps/page, 2 pages.. each at a different az,,za.
• The contour levels are spaced by 3db (the outer edge of the red circle is 3db down.
• the axis are +/- 3amin great circle in az,za offsets.
  

• the coma lobe
• It is in a negative za offset. It gets up to about -9db.
• a coma lobe in the -za offset direction means the horn is in the +za direction (relative to the optical axis (more info)
  
• The horn needs to come down hill
• This is also supported by the large pointing offset
• cbw current model offset:-141.7 asecs
  
• cbh  model offsets -93.5 (most other horns have offsets around -85 to -100 asecs ..except lbw).
  
• -141.7 - (-93.5) = -48.2 asecs
• If the error was only because of the horn positioning then it would have to be moved:
• -48.2 asecs / (46.79 asecs/inch) = 1.03 inches downhill.
  

processing: x101/170811/x102cbw.pro

170814: move horn .391 inches in za direction. (top)

    Felix moved the horn .391 inches in the za direction (toward the secondary..uphill). Manual pointing checks were then done to find the new pointing offsets. This move was in the wrong direction (i had gotten my signs backwards..). But it let us check the plate scale in the za direction... except that it was raining and the platform was not in focus...

• horn moved .391 inches in -x direction
  
• plate scale: 46.79 asecs/inch (more info)
• Measured offsets: (az,za): (-.003,-.006)deg  or (-10.8,-21.6) asecs
• pointing error from move: -.391 * 46.79= -18.3 asecs
• The model offset used was -145.37 instead of -141.72..
  
• So the horn started an extra 3.6  asecs  downhill.
• so the correction needed should have been -18.3 +3.6 = -14.7 asecs.
• The measured error -14.7 was 7 asecs (.002 deg) greater than expected
• this is probably within the accuracy of the measurements:
• i was slewing the  telescope at .001 deg/sec  while looking at the oscilloscope for the peak..
• the measurements on the first day had the platform out of focus..so the original offset -141.72 may have not been the same offset needed for 14aug17.
  

170815: move horn downhill, then calib scans (top)

• Felix and company moved the horn downhill in za .628 inches from the original horn position on 170809.
• The measured (correct)  za model correction for original horn position: -141.72 asecs
• the measured za model correction after the move: -112.06 asecs
  
• The plate scale for za motion (when za above the paraxial ray): 46.79 asecs/inch
• the computed move distance: (141.72 - 112.06)/46.79 = .634 inches.. so this is pretty close to the actual move of .628 inches.
  
• This means that the 170811 model offset of -141.72 was probably pretty close.. the error in the 170814 measurement was probably caused by the rain on 170814.
  
• We needed to increase the az position by 16 asecs (GC) (which means the az position was 16 asecs low).. Instead of updating the turret position, i updated the az model offset.
• a turret correction would have moved the turret position to:
• positive az motion  (CW) is  to decreasing turret angle.. so
  
• 162.0 - 16asecs/(45asec/turret deg) = 161.44 deg turret position.
• if the horn  was positioned -16 asecs az offset from the paraxial ray,  a coma sidelobe would appear on the +az offset of a beam map.

calibs scans on B1328+254 (3C287).

• 6 calibration scans were done on B1328+254.
• There were  clouds (but none were black) and the tiedowns were working
• The telescope was in focus and we had tension in all of the tiedowns.
• scans were done from 15:50 to 16:26.
• The za coverage was 9.7 to 6.6 (before transit).
• We used the standard cband frequencies: 4500,4860,5000, and 5285
• The cal values may be a little off..

Plots made from the calibration scans:

• Page 1: gain,Tsys,sefd, beamwidth
• Page 2: coma, 1st sidelobe, beam efficiencies
  
• Page 3: pointing errors
• Comparing the cbw with cband for this source at 4500MHz and za= 8deg:
  
• 
  

• measure	cbw	cband
  gain	7 K/Jy	7.5 K/Jy
  Tsys	34 K	29 K
  SEFD	5 Jy	3.9 Jy
  beam Width	63 Asecs	61 Asecs
  coma parameter	.1	.07
  1stSidelobe avg Hght	-16 db  down	-15 db down

   
• The sefd is a bit higher for cbw
• the coma is also higher fo cbw.
• The first sidelobe average height is about the same..
  

•  az strips +/- 3 beams
• za strip +/- 3 beams
• -az,+za to +az ,-za  (6beams long)
• -az,-za to +az, +za (6 beams long)

• the 1st 60 points is the az strip, the 2nd 60 is the za strip, 120-179 and 180-239 are the two 45 degree strips.
• Each pattern has been normalized to the peak value in the pattern.
• The vertical is db below the peak.
• The red dashed lines separate the 4 strips.
• The top frame is polA, the bottom frame is polB
• Each page is a separate frequency.
• The solid green lines on page 1 show the positions that were used to compare the sidelobe levels 11aug and 15aug.
• You can see that there is coma. It is at negative ZA (it is easier to see the the beam maps below).

The 3rd set of plots  compare the  sidelobe height at different horn positions (.ps) (.pdf)

• Calibration scans were taken on 11aug17 on B1328+254.
• The feed was moved down hill .628 inches on 16aug17 and then calibration scans were redone on B1328+254.
• The plots shows the sidelobe height for +/- 3 samples about the green line in the previous plot: strips sample 168 and 193.
• Black is data taken on 15aug17, red is the data from 11aug17.
• PolA and polB are over plotted.
• the solid line is sample 168 or 193. Since the sidelobe position may not line up on each pattern, i included +/- 3 samples about this.
  
• Top frame: sidelobe height about sample 168
• Bottom frame: sidelobe height about sample 193
• You can see that moving the feed downhill decreased the size of the peak coma lobe by about 2db
  

Beam maps:

   Beam maps were reconstructed from each of the 6 calibration scans  for B1328+254.

•  6 maps were made for each of the 4 frequencies
  
• for each frequency, 4 maps/page, 2 pages.. each at a different az,,za.
• The contour levels are spaced by 3 db (the outer edge of the red circle is 3db down.
• the axis are +/- 3amin great circle in az,za offsets.

Summary:

• A coma lobe was seen in the negative za direction in all patterns.
• A coma lobe in the negative za  offset is caused by the horn being to high in za.
  
• the feed was moved .628 inches downhill from its original position.
• this improved the peak coma lobe by about 2 db.
  
• After these measurements, it was found that the plate holding the cbw wedge/receiver was mounted rotated 180 degrees. Rotating the plate to the correct orientation should  bring the horn downhill even further.
  
• We will  repeat these measurements after the move.

processing: x101/170815/x102cbw.pro

170829: beam maps after horn moved 2 inches.

•  A new pointing model offset was computed from the distance moved (or at least what they reported it moved).
• manual offset while tracking a source were  used to refine the az,za offsets.
• The new offsets were inserted in the model
• calibration scans (spider scans) were done using B1328+254 (3.2Jy) and B1607+268(1.8Jy).
• td 8 lost tension for the first part of B1328+254).
  
• Beam maps were made for each calibration pattern,.

Plots made from the calibration scans:

Beam maps:

   Beam maps were reconstructed from each of the 8 calibration scans  for B1328+254.

•  8maps were made for each of the 4 frequencies
  
• for each frequency, 4 maps/page, 2 pages.. each at a different az,,za.
• The contour levels are spaced by 3 db (the outer edge of the red circle is 3db down.
• the axis are +/- 3amin great circle in az,za offsets.

•   After moving the horn downhill 2 inches from the 170815 position
• the largest sidelobe has now moved from -za to -az.
• The peak sidelobe is now greater than -9db. The standard calibration plots show the average sidelobe height).
  
• Using the alfa multibeam sidelobes as a guide, the cbw horn needs to be moved closer to the stairwell (turret angle should increase).
  
• the za sidelobe are still offset a bit to the -za direction