Qualification of quartical transfer calibration

[bennahugo]

Neither a bug nor a feature, but I have agreed with @landmanbester to make raw flagged data available from the Lunar project for this purpose. I will give you short flagged L-band data to start with (30 minutes total off transit for the moon).

There are a number of things to test, namely:

• bandpass and amplitude transference to secondaries
• phase transference from secondaries to target
• parallactic angle corrections
• crosshand phase corrections when |X-Y| phases are less than 90 and more than 90, which has the potential to reverse signs
• transference to the moon of the crosshand phases

The datasets should show the handedness correctness of the corrected data.

Rick is also making similar VLA databases available to me which I will process with CASA and give you similar sets to test the circular feed path.

[landmanbester]

I have made a start on this here. This seems to give slightly worse initial images than @bennahugo's CASA reference. Possible reason for this is transferring G phases from primary to the secondary (reference calibration only transfers amplitudes). Discarding the phases is currently a bit of a kludge because some of the G phases end up in K. Would be nice to have the pure delay term as discussed in https://github.com/ratt-ru/QuartiCal/issues/192. It would also be very convenient if one could discard the phases by say, loading in a complex term as a pure amplitude term

[JSKenyon]

Ok, I think that both of these requests will be in main once the tests pass - I will then get them into stimelation ASAP. In principle, using K.type=pure_delay and G.interp_mode=amp should allow you to confirm your suspicions @landmanbester.

[landmanbester]

Awesome, thanks for the rapid response @JSKenyon

[bennahugo]

Ok I've made the preflagged linear lunar databases available that we took with MeerKAT in 2021 for all three bands: UHF, L and S(0). This will give good testing grounds because the H-V phases are spread between the systems, with S markedly wired backward and UHF varying on a receiver by receiver basis - I think this is more to do with the signal chain than the actual dipole wiring and should be treated as calibration error as such.

As discussed this has the potential to flip signage in the -arctan(V/U).

The databases contain a bandpass and leakage calibrator PKS 1934 (just use the Reynolds-Stevens 2016 model in setjy to set the scale -- it should be good enough for this test). 3C286 as the HV phase calibrator and a nearby gain calibrator that the gain phases and delays can be computed from @JSKenyon

The Moon ephemeris must be treated as a special case because it is not stationary in ra/dec, so you will need to compute the ra / dec for each few timestamps.

As discussed this should be a good test to check the handedness of the corrected data complies with IAU convention

[bennahugo]

Not quite there yet, see https://github.com/ratt-ru/QuartiCal/issues/251#issuecomment-1639717258

[landmanbester]

@bennahugo I want to have a go at this. Can you point me at the lunar data again please?

[bennahugo]

just to understand what you want to do end to end? The test case I reported in #251 was just KGB. Do you want to retest that?

[landmanbester]

I want to test all the points you mention at the top. @JSKenyon is working on a crosshand phase solver, I think we will hear about it in the meeting tomorrow. But want to get started so long. Step 1 - locate data

[bennahugo]

ok #251 data is here: bhugo@clapton:~/LunarPolarimetry/absleakage_study/s0/s0_offsets.flagged.avg.uncorrected.ms S band data for you here: bhugo@clapton:~/LunarPolarimetry/S0Band L band data: bhugo@clapton:~/LunarPolarimetry/Lband UHF data: bhugo@bruce:~/UHF_MOON_MK/UHF

[bennahugo]

If you want a real test I can give you P band data for the VLA. The only HV phase calibrator is DA 240 which has a resolved northern lobe which will put an arctan solver to the test