*****
What we know so far:
There is an oscillation in PSUM in a few antennas (sometimes in pdif). It changes the count levels more than two times. If we set rq with a psum in a level and it changes, that could produce the bandpass scattering we see on the calibration table.
The oscillation happens on the same subbands and antennas where we see the scattering. Below example of a psum oscillation (ea17), a pdif oscillation (ea18) and a good antenna (ea01) - project 21B-111
If the rq calibration table is not applied, the oscillation is gone.
What we do not understand so far:
What causes such oscillation? An RFI is probably triggering the issue. But which one needs to be clarified. The RFI source at 6.2 GHz presented in the report is seen in a few antennas, close to AZ 150 and 320, but it does not match with the same antennas where the psum oscillation is found. There are also projects where an evident RFI is not seen on ea17, but the oscillation is there.
I could see the oscillation in ea17 during the tests, with different frequency tuning, and even in the compact configuration, but not in any other antennas closed. Why we did not see it in the other antennas is still unclear. The RFI at 6.2 where clear seen.
What could be:
It looks like ea17 has an issue with the attenuator. And that issue is more evident close to an RFI. Tuning change did not decrease the oscillation.
What I would like to ask:
We are at a more extended array once more. Please let me know if someone finds another C-band project with such scattering in the bandpass calibrator table due to any northern antennas.
More tests:
Yesterday I pointed to azimuth 150, 170, and 190. and I have changed tunning. I am trying to see if I can see the psum oscillation on the north arm.
*****
Paul/Vivek: The most likely explanation for the scatter seen in A2C2 calibration solutions is strong RFI near 6200 MHz that is compressing the 3-bit samplers. This signal is strongest in antennas near the ends of west (ea24) and north (ea17, ea18, ea22) arms. The likely source is one of the nearby hilltop transmitters, and the pattern of antennas affected has to do with direction to the source combined with antenna pointing.
Dave: There is a standard Point-Point 30MHz channel allocation centered at 6197.24 MHz which fits.
Vivek via vla-test-maint: https://science.nrao.edu/facilities/vla/observing/RFI/Oct-2021-B-configuration/C-Band_spectra_202110B
Dave update: I have not had a chance to make any measurements at the site yet but the interference appears to be consistent with what was identified in https://bugs.nrao.edu/browse/WO-7315 first opened in July, 2020. The interference is from a microwave link transmitting from Gray Hill to Smith-Bagley west of Datil. This link traverses the North Arm and almost directly across BN7. The interference detected on the West Arm is likely a sidelobe of the microwave antenna.
*****
2021-22 B config:
always - ea18 [N28], ea22 [N24], ea24 [W36]
sometimes - ea06 [N32], ea17 [N36], ea26 [W32]
2022 A config:
ea05 (W40), ea07 (E72), ea13 [W48], ea19 [E48], ea21 (E56), ea22 [N24], ea26 [W32], ea27 (E40)
*****
Easily seen in finalcals stage of pipeline, "final amp time cal" and "final amp freq cal" tabs.
example pipeline products here:
/lustre/aoc/projects/srdp/problem_data/C_band/
previews below:
---
---