search

tart-telescope / tart2ms

Convert TART data to measurement set format
GNU General Public License v3.0
1 stars 1 forks source link

Issue 12 #14

Closed bennahugo closed 2 years ago

bennahugo commented 2 years ago

Closes #12

This is an implementation of the proposed antenna position fix. Things definitely looks a lot better on a plot image

@tmolteno @rubyvanrooyen @o-smirnov please advise? I'm unsure of the exact layout -- I presume the antenna coordinates are the fitted coordinates relative to info dict lat lon alt? However it is anybody's guess which way the handedness were fitted. We need somehow to establish that East and North is the right way here. I'm thinking regenerating UVW coordinates using CASA?

Note I had to put in a kludge to override the observatory name (you can specify e.g. "kat-7" to overrule the default) -- until we can get the TARTs registered with JPL this is what we will need to do (or Tim's group will need to distribute special ephemeris which is a huge maintenance upkeep...)

bennahugo commented 2 years ago

No cigar.... the problem is I don't know what to trust though -- how sure are we about the original uvw coordinates? -- this is not just a difference in hour angle there is a seeming change in the w - ie the declination hand... image

tmolteno commented 2 years ago

These changes look excellent @bennahugo

  1. The antenna layout is visible in the TART telescope web interface (https://tart.elec.ac.nz/viewer/ and switching the API endpoint to Rhodes (in the viewer). This looks correct, but would depend on the coordinate frame.
    image
  2. The TART operates in the local frame with uv coordinates relative to straight up, and it looks like the UV coordinates you have are flattened by projection
  3. The TART array is flat in the tangent plane to the TART location. So the x,y TART antenna positions will result in an x,y,z offset in the geodetic (WGS84) frame used by the ANTENNA table. So the x,y,z positions of each antenna need to be transformed before adding to the array_centroid.
tmolteno commented 2 years ago

Have a look at https://docs.astropy.org/en/stable/api/astropy.coordinates.offset_by.html. This could be used to offset the array_centroid by the antenna positions in meters.

tmolteno commented 2 years ago

I have merged, and we can fix the offsets in another issue!

tmolteno commented 2 years ago

This has been addressed and closed in Issue #15. Antenna positions now match the online view!

bennahugo commented 2 years ago

Hmm ok yes I still don't believe this is correct -- the w direction is supposed to point do declination (which matches the zenethal angle in the equatorial frame. These coordinates clearly do not point to the same point on the sphere probably due to the offsets not being done correctly -- therefore the projected array appears flatter. I will try and redo this until we get a better match -- I fully trust the UVW coordinates coming out of casacore so it is a matter of getting the array configuration correct