Polarization angle of 'xcal' calibrator 3C286 after polcal dose not match to setjy parameter

[Rsikezudta]

Use '3C286' for xcal while excute worker polcal, the polarization angle(PA) is not correct.
'setjy' from the log file had set the polarization information of 3C286 to the model seems no problems, but the CORRECTED_DATA plot by worker inspect show the PA of 3C286 are 0 not 33 deg (0.575959 rad).
I also check the MODEL_DATA plot using inspect, the PA plot are 0 deg not 33 deg.

Meanwhile, I imaged the calibrator 3C286 after polcal, but the PA was 28 deg, not 0, Did not match with 'setjy' 33 deg.
If there are some problems, how can I fix it?

Here to show a part of logs, plots and image.

INFO	setjy::::+	##########################################
INFO	setjy::::+	##### Begin Task: setjy              #####
INFO	setjy::::	setjy(vis="/stimela_mount/msdir/1529521526_sdp_l0-cal.ms",field="3C286",spw="",selectdata=True,timerange="",
INFO	setjy::::+	        scan="",intent="",observation="",scalebychan=True,standard="manual",
INFO	setjy::::+	        model="",modimage="",listmodels=False,fluxdensity=[14.7172],spix=[-0.4507, -0.1798, 0.0357],
INFO	setjy::::+	        reffreq="1.47GHz",polindex=[0.098],polangle=[0.575959],rotmeas=0.0,fluxdict={},
INFO	setjy::::+	        useephemdir=False,interpolation="linear",usescratch=True,ismms=False)
INFO	setjy::::	{'field': '3C286'}
INFO	Imager::open() 	Opening MeasurementSet /stimela_mount/msdir/1529521526_sdp_l0-cal.ms
INFO		Clearing all model records in MS header.
INFO	setjy::::	CASA Version 5.8.0-109
INFO	setjy::::
INFO	imager::setjy()	Using channel dependent flux densities
INFO		Clearing all model records in MS header.
INFO	imager::data selection	Selected 135420 out of 2620560 rows.
INFO		Clearing all model records in MS header.
INFO	imager::data selection	Selected 135420 out of 2620560 rows.
INFO	imager::ft() 	Fourier transforming: replacing MODEL_DATA column
INFO	imager::createSkyEquation() 	Processing after subtracting componentlist /stimela_mount/msdir/1529521526_sdp_l0-cal.ms_setjy_spw0_3C286_0.856GHz58289.8d.cl
INFO	imager::createFTMachine() 	Performing interferometric gridding...
INFO	imager::setjy()	Flux density as a function of frequency (channel 0 of each spw):
INFO	imager::setjy()+	  Frequency (GHz)    Flux Density (Jy, Stokes I)
INFO	imager::setjy()	     0.856         18.3354
INFO	setjy::::	##### End Task: setjy                #####
INFO	setjy::::+	##########################################

CORRECTED_DATA phase-freq

MODEL_DATA

Image of 3C286 after polcal by CARTA

[Athanaseus]

@Rsikezudta thanks for reporting this.

There have been recent updates to caracal. I'm curious, which version are you currently running?
Please also provide the full log as this may help us narrow down the problem.

Kind regards

[Rsikezudta]

Thanks for your reply.

Caracal was executed on ilifu serverion and which container I used is STIMELA_IMAGES_1.7.7.
Also following the log, the caracal version is 1.1.1.
Here is the full log file.
Should I change to recent version?
slurm-9906533.zip

Thanks in advance

[Athanaseus]

Hi @Rsikezudta,

It may be worthwhile to run with the latest version.
However, upon reviewing the corrected data and the log, it appears that some spectral channels were not flagged properly.
The phase scatter is likely caused by unflagged RFI in those channels.

You can address this by enabling flag_spw for the calibrator and
specifying the problematic frequencies, including band roll-offs (note that the end channels are not inclusive):

flag:
  flag_spw:
    enable: true
    chans: '*:856~880MHz , *:1658~1800MHz, *:1419.8~1421.3MHz'
    ensure_valid: false

Out of curiosity, could you also share the bandpass gain solutions that were applied (e.g., .B1)?

Hope this helps!

Best regards

[Rsikezudta]

The RFI channels may very useful for me, thank you.

Here I show the gain solutions B1 which were applied.

Best regards

[AnnalisaB]

Hi all, my colleagues and I are having the same issue on several datasets where 3C286 is observed. We almost never recover the model we specify. After some reasearh, we have found this memo, where there's a lot of useful info:
https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://archive-gw-1.kat.ac.za/public/repository/10.48479/bqk7-aw53/data/Absolute_linear_polarization_angle_calibration_using_planetary_bodies_for_MeerKAT_and_JVLA-REV-C.pdf&ved=2ahUKEwjW3MKo9MyJAxVMlP0HHQWHAx4QFnoECBkQAQ&usg=AOvVaw1sjkLIB0z8gfCH2pfAIBYg

Using this model, a PhD student, Alessandro Benati, manages to obtain a good match in UHF. He has 3C286 observed during night-time.
I don't have the same luck with my L band observations, where 3C286 is observed during day-time.

[a-benati]

Hi all, my colleagues and I are having the same issue on several datasets where 3C286 is observed. We almost never recover the model we specify. After some reasearh, we have found this memo, where there's a lot of useful info:

https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://archive-gw-1.kat.ac.za/public/repository/10.48479/bqk7-aw53/data/Absolute_linear_polarization_angle_calibration_using_planetary_bodies_for_MeerKAT_and_JVLA-REV-C.pdf&ved=2ahUKEwjW3MKo9MyJAxVMlP0HHQWHAx4QFnoECBkQAQ&usg=AOvVaw1sjkLIB0z8gfCH2pfAIBYg

Using this model, a PhD student, Alessandro Benati, manages to obtain a good match in UHF. He has 3C286 observed during night-time.

I don't have the same luck with my L band observations, where 3C286 is observed during day-time.

To expand on this, in UHF band (@816 MHz) the model of 3C286 of that memo predicts:

• pol angle: 23.8 deg
• pol fraction: 6.4%
• RM: 0.12 rad/m^2

And I obtain:

• pol angle: 24.2 deg
• pol fraction: 6.6%
• RM: (0.00 +/- 0.02) rad/m^2

While using other models, such as the one implemented in CARACal, I'm not able to recover the values of the model that are specified.

[bennahugo]

Have you done ionospheric corrections? There can be up to -6 rad / m2 offsets

…

On Sat, Nov 9, 2024, 11:24 Alessandro Benati ***@***.***> wrote: Hi all, my colleagues and I are having the same issue on several datasets where 3C286 is observed. We almost never recover the model we specify. After some reasearh, we have found this memo, where there's a lot of useful info: https://www.google.com/url?sa=t&source=web&rct=j&opi=89978449&url=https://archive-gw-1.kat.ac.za/public/repository/10.48479/bqk7-aw53/data/Absolute_linear_polarization_angle_calibration_using_planetary_bodies_for_MeerKAT_and_JVLA-REV-C.pdf&ved=2ahUKEwjW3MKo9MyJAxVMlP0HHQWHAx4QFnoECBkQAQ&usg=AOvVaw1sjkLIB0z8gfCH2pfAIBYg Using this model, a PhD student, Alessandro Benati, manages to obtain a good match in UHF. He has 3C286 observed during night-time. I don't have the same luck with my L band observations, where 3C286 is observed during day-time. To expand on this, in UHF band the model of 3C286 of that memo predicts: - pol angle: 23.8 deg - pol fraction: 6.4% And I obtain: - pol angle: 24.2 deg - pol fraction: 6.6% While using other models, such as the one implemented in CARACal, I'm not able to recover the values of the model that are specified. — Reply to this email directly, view it on GitHub <#1604 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4RE6XCB462HL2TNARG7JLZ7XIFHAVCNFSM6AAAAABQDJHMGGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINRWGE2DGMBXGQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[a-benati]

@bennahugo no, I didn't apply any ionospheric corrections to my data, while in the model of the memo, if I understood correctly, they are applied before performing the fit, right?

[AnnalisaB]

@bennahugo no, I didn't apply that yet. That would be the next thing to try. Do I uderstand correctly that you've used TECOR in AIPS?

[bennahugo]

You need to at least do a per scan correction of both your calibrators and targets to account for diurnal effects. You can get the highest cadence measurements using ALBUS

…

On Sat, Nov 9, 2024, 18:30 AnnalisaB ***@***.***> wrote: @bennahugo <https://github.com/bennahugo> no, I didn't apply that yet. That would be the next thing to try. Do I uderstand correctly that you've used TECOR in AIPS? — Reply to this email directly, view it on GitHub <#1604 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4RE6QPQP3I2WMPTHHZUUDZ7YTBTAVCNFSM6AAAAABQDJHMGGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINRWGI2TQMZTGY> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[AstroRipples]

Hi everyone,

Just to throw my 0.02€ into the discussion here on 3C286, I thought I'd share my results for 3C286 // J1331+3030 from an observing block I've been working on.

Below you can see the values measured on the images given by running WSclean in full polarisation on the calibrated data:

The left panel shows the Stokes I flux density (left-hand axis) and linearly-polarised flux density (right-hand axis), centre panel shows the polarisation fraction, and right panel shows the polarisation angle, measured in 48 channels across the MeerKAT L-band. The orange datapoints show the values reported in Taylor & Legodi (2024), whereas the shaded region and horizontal line show the values reported in Table 1 of the MeerKAT polarisation calibration wiki page.

For this observation (CBID 1639795398) in 2021, we had two 5-minute scans on 3C286, separated by some parallactic angle. I've processed the data using CARACal following the standard calibration procedure. Assuming I recall the CARACal workflow correctly, as we have two scans, the polcal worker does not attempt to set a model, but derives the polarisation properties using the observed gain tables.

The fractional polarisation we get is nicely consistent with what's expected for 3C286, although there's clearly an offset in the polarisation angle. I haven't looked into this much further yet, as I've been primarily concerned with other datasets where I have 3C138 (I'll open a new issue to discuss that source later this afternoon) but as this observing run took place during early daylight hours (sunrise at 03:30 UTC, pol cal scans circa 06:00 and 08:36 UTC) I suspect that ionospheric RM is the culprit. A quick back-of-the-envelope calculation suggests an ionospheric RM of ~2 to ~3 rad/m2 would give rise to the observed offset with respect to the values reported by Taylor & Legodi.

Will report back with more once I've had the chance to run RMextract ...