Unity Gain Test

[ronniyjoseph]

This is basically repeating Sievers, 2017.

Is corrcal able to find unity gains and under which conditions?

Perfectly Redundant Array

• pure covariance based
• covariance + (an increasing number of) source models

Final goal: Take MWA redundancy metrics and analyse statistical behaviour of corrcal

Future work Realistic Arrays:

• increase position offsets to see behaviour

• increase positions offsets + increase number of source models included

• Introduce beam variations (change more and more antennas in the array)

• increase beam variations + increase numbers of source models (Note: for corrcal to succeed this requires re-calculation of the covariance matrices, which is fairly straightforward for phased arrays, less so for dishes)

To really push it:

• increase beam variations + increase position errors
• increase beam variations + increase position errors + increase number of included source models

[piyanatk]

@ronniyjoseph

• pure covariance based
• covariance + (an increasing number of) source models

Can you clarify what you mean by the above?

And is this overlapped with #27?

Also, I think it is probably better for planning and documenting purpose to split each bullet point into it own issue. We can consolidate them in a unity gain test project or something like that if needed too.