Dear Nimisha et al.,

Following on yesterday's meeting, I have the following suggestions for follow up action item

1. As I understand, there are two immediate, easy to identify problems, which can be followed up
   (a) Jumps in azimuth offset
   (b) frequency dependent offset in some antennas

2. I believe 1(a) is not a problem per se, but a better coordination between mechanical, servo and operations will allow understanding all these as these can be (a) due to mechanical adjustment (b) encoder change.
I suggest that
   (a) Santaji produce a table for each antenna listing the start and end
       dates between which this change was seen
   (b) Mechanical/servo then check their respective logs to see if any
       encoder realted changes took place in that interval. If yes, this
       be entered in the table along with the reason
   (c) In future, any changes in servomechanical system related to
       encoders be intimated to a mutually agreed coordinating astronomer
       (I suggest Subhashis), who can coordinate between operations,
       mechanical and servo actions to bring azimuth offset to zero. The
       relevant information be aslo emailed to all addressee of this mail
       and put up on a web site.
   (d) SOP for C
        i) mechanical/servo changes (either due to a problem noticed/
           response to call sheet/ response to feedback from operations)
           communicated to nodal astronomer plus coordination meeting
        ii) Multiple (at least 5 preferably more and L Band pointing) to
            be carried out by operations as soon as possible after the
            change and results communicated to nodal astronomer
        iii) Feedback to servo / mechanical from nodal astronomer on
           new measured mean offset
        iv) zeroing of offset by servo
        v)  Multiple (at least 5 preferably more and L Band pointing) to
            be carried out by operations to verify zero offset
    (e) Servo zeros offset twice an year after feedback of full cycle from
        operations
1(b) will allow understanding the jumps and 1(d),(e) will keep offset zero as suggested by Prof Swarup.

3. One possible experiment to tackle 1(b) is as follows
   (a) Santaji chooses the worst antenna from the data displayed yesterday
   (b) Santaji estimates the mean offsets of each frequency from L Band
   (c) Feed counts for this antenna are adjusted to take care of the
       above for each feed position for this antenna
   (d) Repeated multi-frequency pointing (at least 5 points in each
       frequency, maybe more) are carried out by operations immediately
       and results communicated to nodal astronomer
   (e) If the offsets vanish/diminish, the problem is solved, otherwise
       experiments to test other more complicated explanations such as
       out of plane tilt as suggested by Subhashis be taken up
   (f) Pointing model incorportaes a term for feed tilt


4. In addition, better analysis of pointing data presented yesterday is also warranted to anser the following questions
  (a) What is the typical spread of measurement as a function of frequency
      ? A histogram for each antenna is needed with estimated rms at each
      frequency and this will tell us contribution of RFI to the scatter
  (b) How do the measurement depend on quality factor ? This is more
      involved and required analysis of all pointing table files and
      choose only measurements with Q factor above 3 or more. Does this
      reduce scatter and removes outliers ?
  (c) Repeat of 4(a) to see how this modifies the scatter.
  (d) What are the typical error bars on measurements ? A table would
      be better.
I believe Santaji with help from 2 other operators can carry out the required analysis with supervision from Nimisha or Subhashis


On servo side, we will take 1(b) and 1(c) in SGM next week and can come
back with information as soon as Santaji produces 1(a).

These are my suggestions for whatever it is worth for a better understanding the obvious problems. There are other issues not so
easy to tackle, which I believe should be taken up once these are
satisfactorily resolved.

Bhal Chandra