Short summary of the meeting held on 21 May 2015 and suggested names:

This meeting was held to discuss the pointing of the GMRT antennas based on the data that 
has been gathered over the last four years (since October 2010) and logged at 
http://gmrt.ncra.tifr.res.in/~astrosupp/gmrtmon/pnt/pnt.php. The first two years of data 
have been discussed in a NCRA techinical report at 
http://ncralib1.ncra.tifr.res.in:8080/jspui/handle/2301/579.

Collegues from the mechanical, servo, operations, academics were present for the discussion 
meeting.  The meeting had two aims - first was to have an academic discussion to understand 
the trends seen in the pointing data and second was to come up with concrete plans to 
improve the pointing which would finally benefit astronomical observations.

A. There was discussion on the origin of the observed trends in the data namely: (1) sudden 
jumps in the mean pointing offset (2) large and at time varying scatter on the mean offsets 
of an antenna (3) gradual change in the mean offset of an antenna (4) band-dependent mean 
offset.  It was also pointed out that there were a few very good antennas.

(1) Jumps - the discussion concluded that the origin has to be (a) encoder change (b) servo 
software update.  For example C01.  Correlation between the jump and encoder change has 
been noted for a few cases in the technical report but needs to be verified for all times.

(2) Scatter - the origin has to be (a) measurement errors (b) loose mechanical elements.  
For example C02. The measurement errors were not favoured since some antennas which had a 
large scatter subsequently showed lower scatter and had likely been fixed by the mechanical 
group.  However there was a opinion that RFI-corrupted data or low quality factor pointing 
offset might be contributing and should be examined.

(3) Gradual change - unlike the abrupt jump in the offset value, some antennas (e.g. C06) 
show a gradual change in the pointing offset with time and then stabilises.  The discussion 
appeared to conclude a mechanical origin for this.

(4) Band-dependent mean offset - this was believed to be due to the FPS system.  A few 
antennas show this problem in the elevation and azimuth pointing offset.  For example W02, 
C09.  It was suggested that the problems are likely due to misalignment of the feed stool 
which could cause a tilt in the feed and hence these offsets.  

TODO: The mechanical group will look into their log books to ascertain when mechanical work 
was undertaken on the antennas and the servo group will look for the dates on which servo 
software was updated.  This information alongwith the online pointing offsets will help 
conclude further on the first three points mentioned above. The importance of maintaining 
extensive log books was stressed by Prof. Swarup.  Based on their work , it will be very 
useful if the mechanical group could also generate a report on the fixes to the above 
problems.  There was a suggestion that the relevant log book entries can be available on 
the engineering webpage.

Point 4 is more involved and will need separate experiments like the one outlined by BCJ in 
his summary email.

B.  There was also discussion on coordination between the mechanical, operations and servo 
groups so that the problems mentioned in A(1) can be avoided.  According to the current 
procedure, the mechanical group is responsible for changing the encoder after which the 
control room operations calibrates the encoder.  Based on the results, any remaining mean 
offset is communicated to the servo group for an appropriate software update in the servo 
computer so that zero mean offset is seen by the control room user.  It was suggested that 
at least five data points at L band per antenna should be taken after an encoder change so 
that a mean offset could be estimated. It was felt that there is some scope for improvement 
on the feedback mechanism and coordination.  It was suggested that Nimisha will start this 
coordination and pass it on to Subhashis. It would be good to have a member from each group 
who could be involved on a more direct fashion - such as Santaji from operations, Manish 
Patil from mechanical and Thyagarajan from servo (I have mentioned Manish and Thyagarajan 
since they were present in the meeting - I request Mr. Kale and Mr. Sabhapathy to confirm 
this or pl. suggest other names - ngk).  It was also suggested that it would be good to include 
all this information on the engineering webpage for quick and wide access by all groups. 
Santaji can work with the mechanical and servo groups accordingly.

C. The other points which were discussed and possible people who could be involved are: 

(0) A list of reference antennas which are good e.g. C12 (El), S01 (Az), S03 (Az), S06 
(Az), W03 (Az).  Also worst examples of the aforementioned problems so that we have both 
reference antennas to aim for and problem antennas to correct - Santaji, Ishwar

(1) Re-examining the pointing analysis - observing procedure, analysis software, quality factor 
- Subhashis, Santaji, Nimisha, Vasant (if agreeable) 

(2) Cross-checking results of a 2D pointing grid with 1D grids and exploring efficacy of 
smaller grids. - Santaji

(3) Operator awareness on pointing algorithms and importance of pointing - Santaji, Subhashis 

(4) Removing outliers in the 4-year pointing data for estimating std dev- Santaji 

(5) FPS-related offsets - Santaji, Subhashis, Thyagarajan, Ishwar

(6) Errors, if any, introducd by pointing model and coefficients of mechanical pointing 
model - Subhashis, Nimisha