from:	Hauyu Baobab Liu 
to:	melvyn wright 
cc:	Jun-Hui Zhao 
date:	Thu, Jan 5, 2017 at 12:47 PM
subject:	Re: SMA SWARM full polarization observations towards Sgr A*

Dear Dr. Wright, Jun-Hui,


Yes the data from ASIC for ants 678 are good.
But there were some more flagging due to phase jumps and other issues.
The observations rotated the waveplates for ants 1 and 2 to calibrate cross rx phase offset.
But it turned out that ants 1 and 2 have more phase jumps,
which caused all ants to be flagged since the cross rx phase offset became un-calibratable.
The flagged data is about ~40% of our on-source time.
I am not sure whether the phase jumps and the waveplate rotations were just coincidence.

With the Aug. 20 observations only,
the achieved Stokes Q and U RMS at the center of our mosaic field is around 7 mJy/beam.
Miriad-invert reported that the theoretical RMS is ~10 mJy if assuming pointings are not overlapped.
Given that our observations were mosaic, the achieved RMS seems reasonably consistent with the theoretical values.

I am trying if we can use at least the short baseline part of the Aug. 21 observations,
which however may not help much with the sensitivity.
We are allocated one subcompact array track and one extended array track for the upcoming semester.
I will reduce the number of mosaic fields.
Hopefully we will also do some more tests before our tracks, to really enable using SWARM for full polarization observations.
It seems like the cross rx gain phase offset is not varying with time (or at least not varying in a few hours timescale).
Once this is confirmed, one scan on a bright source cycling through sufficient waveplate rotations may allow calibrating this effect out.

There is another issue, since the Sgr A* polarization detected from SMA and JCMT are not consistent.
The JCMT polarization seems to be subject to a ~45 deg rotation.
This is significant enough, such that the JCMT image and the SMA data do not have the same sign with Stokes Q and U.
So one of those calibrated observations were wrong, but is more likely to be JCMT.

Since most of the JCMT polarization observations were for magnetic field strength in star-forming regions,
they don't care that much about absolute polarization position angle.
When I visited JCMT last August, the staffs told me that they only compare the polarization segments derived from SCUBA-2 observations towards OMC-1,
with those derived from SCUBA-1 observations.
They felt the orientations of those polarization segments are "more or less" consistent,
but there was no quantitative analysis, and was not documented with any report or memo.
If we cannot resolve the problem of 45 deg rotation of JCMT, I wonder whether it is fair to use the SMA results of Sgr A*
to calibrate the position angle offset of JCMT.
It might not be the best idea since Sgr A* is a variable.
But this may be better than not doing anything with it...

Best wishes,
Baobab