notes_on_aipy_to_miriad.txt

The antenna positions as read in through aipy 

antpos = aa.antpos_ideal.T.flatten()

appear to be meters, in an ENU coordinate system: (x,y) looks like the hex, z is nearly 0

but that's not what official miriad does, and is not what pyuvdata is expecting

per Bryna:

miriad positions should be in an ECEF frame that’s rotated to put the x axis through the local meridian. 

(uvfits uses that convention as well, except that uvfits wants them to be relative to the telescope location (which is in ECEF, not the rotated ECEF))

pyuvdata has functions to do ECEF_from_ENU + rotECEF_from_ECEF

the pyuvdata object has positions in ECEF (not the rotated frame) and they’re relative to the telescope location

Proposal: import pyuvdata into add_uvws_and_antpos.py and convert antpos to rotECEF frame and write out.

Hopefully, reading back into pyuvdata.UVData.get_enu_antpos should give what's expected