Pointing corrections jun08

• Page 1: pointing errors all data.Top two plots are za errors vs (za,az). Bottom two plots are az errors vs (za,az).
  
• You can see a 1 azimuth  term in the pointing errors. The az peak is at  az=90,270 while the za peaks at 0,180.
• It has an amplitude of about 9 seconds of arc.
  
• This is a platform tilt in the north/south direction. Tiedown 12 encoder position has probably moved.
  
• There is also a coherent dip in the az errors around az=160. This corresponds to where the azimuth encoder rack gear has a large runout.
  
• Page 2,3: a 1 degree encoder offset was found in the vertex system on 11jun08. It had been there since 21feb08. I don't think it is included in program track mode (since 1 degree*sin(10deg) is > 600 arc seconds...). To check, i split the pointing data up before and after 11jun08..
• Page 2 is prior to 11jun08.
• Page 3 is after 11jun08.
• Page 4: the az, za coverage for the sources tracked. Some sources where tracked on more than one day (to see if things had changed).
  
• Top: all data
• Middle : before 11jun08
• bottom: after 11jun08.
• page 5: Az, Za plot of pointing errors. the length and direction of the arrows represent the pointing errors at each spot on the dish.
• 1 tick mark on the plot is 5 arc seconds
• Each circle is 5 degrees in za.
  
• The  top of the plot is when the feed is in the north (tracking a source width declination < 18 degrees).
  
• The left side of the plot is for sources rising.
• The sources B0038+328 (red) and B0123+329 (green) don't seem to line up with the rest of the observations. These sources were tracked after 11jun08.
  
• Page 6: fit 1,3az terms to the az and za errors.
• Top: the az errors (black)  with a 1az, and 3 az fit (red *).
• 2nd: za errors (black) with a 1 az and 3 az fit (red *).
• 3rd: PntErrs - fit residuals vs azimuth . Black is az residuals , green is za residuals
• 4th: PntErrs - fit residuals vs za . Black is az residuals , green is za residuals
• Some of the az errors above za=15 don't fit very well. These are the 2 32 deg Dec sources.
• The fit values to remove from the current model are in the table below:
  
  

• Values to SUBTRACT from current model

  Model Term	az Err Asecs	za Err Asecs
  constant	.59	-.65
  cos(az)	-.51	-8.13
  sin(az)	9.03	.68
  cos(3az)	-1.65	.27
  sin(3az)	.73	-.46

  
  
• Page 7:Fit residuals (Measured Pointing errors - fit)  plotted by source.  The top 2 plots are za residuals, the bottom two plots are az residuals.
• The rms pointing errors:
  
• before fit: Za: 6.9, az: 8.2 . Adding in quadrature: 10.6 Asecs rms
• After Fit: Za: 3.4, az: 5.6 . Adding in quadrature: 6.5 Asecs rms.
• Page 8:Fit residuals plot vs  Az,Za.
  
• The sources B0038+328 (red) and B0123+329(green) as well as B0940+029 (blue) a stand out. They have large azimuth errors.
• All of these sources transit AO at about 15 degrees za...
• Page 9:Pointing errors after fit installed on sbn model.
• The offsets were installed on sbn on 20jun08 and on the rest of the receivers on 21jun08.
  
• The rms values are close to the fit residuals, but za errors now have a 5asec offset.  We may need to take some more data over other parts of the dish to see if this stays this way.
  

Summary:

• The model corrections were added to sbn model on 20jun08. They were added to the rest of the receivers on 21jun08 (after checking that the sbn correction when the right way).
  
• The 1az term za errors are -8.13*cos(az) + .68*sin(az).
  
• At az=0 degrees we need to add 8.13 asecs to point  at the source. So tiedown 12 is too low.
•  I'm not sure when this happened (since we took no pointing data between feb and may08).  A partial list of tiedown failures can be found here..
  
• The 1 degree encoder correction found in the vertex system doesn't look like it was applied for tracking mode (3600asecs* sin(10) is still about 360 asecs).
  
• The data taken after the mods were installed has a 5asec offset in za. With only 55 points it is probably too soon to say if this is real or not.
  

page up
home ~phil