[daip] dbcon question

[Eric Greisen]

bob at physics.ucdavis.edu wrote:
>> bob at physics.ucdavis.edu wrote:
>>> hello, i ran reway and the weights varied alot from baseline to baseline
>>> which does not seem reasonable. is setting weights equal to integration
>>> time really 'bad'?
>> What algorithm and parameters did you use with REWAY?  It is not
>> unreasonable for weights to vary by factors of 2 or even more and in
>> some bands the antennas really do vary a lot and so weights will be more
>> wide spread (remember they are 1 / sigma^2).
>>
>> Eric
>>


> AIPS 3: APARM      *all 0                  (1) > 0 rolling T buffer to
> AIPS 3:                                       find baseline weights
> AIPS 3:                                       A(1) times
> AIPS 3: BPARM      *all 0                  (1) Set weight to zero if <
> AIPS 3:                                        B(1) samples in window

This means that you have computed the weights over no time smoothing at 
all.  Better results usually arise from APARM(1) > 0 with or without 
additional smoothing.  You can plot data weights with VPLOT or ANBPL to 
see what they are doing.  Note that residual RFI will make weights look 
much lower for affected samples.

ERic