[daip] BP TABLE

[Eric Greisen]

Andrey Chuprikov wrote:
>> The weight field in the BP table is not used a great deal.  
>> It signals failed BP solutions and 
>> is used in time-interpolations of BP solutions. 
> 
> Dear Dr. Greisen, 
> 
> Thank you for your explanation.
> I still have a question.
> 
> As I understand, 'SPLIT' uses the BP solutions.
> Then, how are used weights from BP TABLE for time-interpolations ?
> Are real/imaginary parts multiplied by corresponding weights before 
> Interpolation or anything else ?

The weights may or may not be used depending on the DOBAND setting. 
With DOBAND=1 a weighted average of all BP solutions for each antenna 
separately is generated.  With DOBAND 2 and 3 the solution for a given 
time is a weighted interpolation (or selection with "nearest") while 
DOBAND 4 and 5 are DOBAND=2 and 3 but ignoring the solution weights.

These are interpolations so that the result is a multiplication by 
weights and then division by sum of weights.

This is a VERY minor point.

Eric Greisen