[evlatests] Referenced Pointing Problem understood

[Rick Perley]

     Maybe.  It could also be that the problem is less frequent when the 
target source is not the same as the referenced pointing source.  Even 
if what you write is true, I'm sure I'm not the only user who uses the 
target source as a referenced pointing calibrator.

   I would feel better if we squash this bug before using the referenced 
pointing mode on science observations...

     Rick


On 09/12/2018 11:07 AM, Bryan Butler wrote:
>
> Please note that not every observation utilizing referenced pointing 
> is affected by this bug.  As I understand it, it only affects 
> observations where a reference pointing scan immediately follows a 
> scan on the same source.  Most observers do not set up their SBs in 
> that way, aside from the possibility of the very first source, so 
> should be immune from this.
>
>     -Bryan
>
>
> Rick Perley wrote on 9/12/18 10:01:
>>      As part of the antenna optics alignment project, my summer 
>> student (Lucas Wilkins) and I discovered an important failure mode in 
>> referenced pointing. The details of this are contained in various 
>> emails posts distributed last May, and will not be repeated here.    
>> Due to the site power issues, it was not possible to diagnose/correct 
>> this problem over the summer.
>>
>>      with power now restored, and the array now functional, Barry 
>> suggested, and Ken implemented, a 'quick fix' to this problem, which 
>> was tested yesterday.  Ken re-ran a Ku-band script used last May, 
>> both in the original mode, and with the 'fix' put in.
>>
>>      The results are very clear:
>>
>>      a) The original script showed the failure mode on every antenna, 
>> for every solution.  Resulting solutions (in azimuth) are erratic, 
>> and in error.
>>
>>      b) The 'fixed' script showed no failures, and the solutions are 
>> consistent throughout the six referenced pointing solutions made.
>>
>>      Barry tells me that the next step is to devise and test a more 
>> elegant solution to the problem, as the current fix is something of a 
>> kludge.  I note that the 'quick fix' solution adds ten seconds to 
>> every referenced pointing operation.  (The added 10 seconds is a 
>> separate scan, with the antennas on source).
>>
>>      Until this is done, and the final solution tested and 
>> implemented, I strongly urge no observing be done utilizing 
>> referenced pointing.
>>
>>      Rick
>>
>> _______________________________________________
>> evlatests mailing list
>> evlatests at listmgr.nrao.edu
>> https://listmgr.nrao.edu/mailman/listinfo/evlatests
>