VEGAS Baseline Tests for wide-bandwidth Modes-1,2,3

2014-09-15: D. T. Frayer

On 2014 September 04, we took some comparison baseline data using Q-band with VEGAS-mode-2 and the spectrometer. The weather during this test was not ideal (variable sky with high Tsys~140K). Waterfall plots showed that the spectrometer was stable over about 5minutes, while VEGAS baselines were only stable over 15-30sec. The GB-VEGAS team had concerns that this could be related to the Mode-2 BOF. To test the relative baseline performance of the different modes (wide-bandwidth modes-1,2,3), we carried out additional baseline tests on 2014 September 10. The weather was much better during these tests (Tsys~70K), and the weather conditions approximated the expected on-sky conditions for Q-band as scheduled by the DSS (a Q-band project directly proceeded these tests).
  • The baseline stability of Mode-1,2,3 appear fairly similar. Based on the previous tests, there was some speculation that mode-2 and mode-3 may have worse baselines than mode-1 due to the BOF, but there appears to be no justification for modification of mode-2&3 BOF based on baseline performance. The waterfall plots suggest stability of ~2-5 minutes.
  • The Spectrometer data generally have better baseline stability than data collected with VEGAS under similar conditions. The waterfall plots for the Spectrometer shows stability over 20min while VEGAS showed stability over ~2-5min (see attached plots, pages 1-4, below).
  • To test the importance of the effects of possible thermal variations, we opened the VEGAS cabinet during a long observation. We found the stability degraded significantly with the door open (the waterfall plots degraded within 10sec after opening the door, see page 5 of attachment). This highlights the potential importance of maintaining good thermal stability for VEGAS. We should consider investigating this further.
  • Long sequences of NOD observations were carried out with VEGAS-mode-1&2 and with the spectrometer with 1minute scans (1minute ON and 1minute OFF for 30-48 minutes, see plots, pages 6-8). The VEGAS data showed much more baseline structure than the Spectrometer which suggests the possible difficulty of carrying out extragalactic position-switch/NOD observations (with standard 1-2 minute scans). We should test the stability of VEGAS during SubBeamNod observations that position switch on much shorter time scales (e.g., 5-10sec, instead of 1-2minutes). Combinations of current data on ~10sec time scales yield flat baselines with VEGAS, but it is unclear if this will integrate down in good weather conditions as it does for the Spectrometer.
  • baselines_vegas_vs_spec.pdf: baselines_vegas_vs_spec.pdf (plots from session 31)
  • TABLE -- Relative Baseline RMS as a function of time. Below shows a table quantifying the relative RMS change as a function of time in comparison to the RMS measured from the difference between the first two integrations separated by 2sec (we ignored intnum=0 to avoid potential issues with the first VEGAS integration of a scan). For example a tabulated value of 1.10 means the measured noise is 1.10 times higher than the first RMS measured for integrations separated by 2sec. We took multiple scans in each mode to test the consistency of the results with small changes of total power from the sky. The rms ratio plotted has been corrected by the measured Tsys from the noise diodes and the data have been smoothed to a 1.5MHz channel resolution to put all the data on the same scale (e.g., at very high spectral resolution the baseline wiggles are hidden by the channel-to-channel noise). The rms was only measured over the center of the band (to avoid edge effects).

Time

[sec]

SPEC

scan=26

SPEC

scan=30

VEGAS-mode-1

scan=25

VEGAS-mode-1

scan=29

VEGAS-mode-2

scan=27

VEGAS-mode-2

scan=31

VEGAS-mode-3

scan=28

VEGAS-mode-3

scan=32

VEGAS-mode-1

after door open

scan=150

2 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
4 0.95 0.99 1.03 1.02 1.00 1.02 1.00 1.02 0.98
8 1.00 0.97 1.02 1.06 0.98 0.99 1.00 0.97 1.16
16 0.97 1.09 1.05 1.02 1.07 1.00 0.98 1.06 1.14
32 0.98 1.15 1.17 1.11 1.05 0.97 1.02 1.34 2.26
64 0.94 1.00 1.28 1.09 1.05 0.99 1.06 1.06 2.49
128 1.27 1.26 1.49 1.13 1.37 1.01 1.07 1.08 2.19
256 1.24 1.22 1.08 1.47 1.69 1.24 1.07 1.11 3.74
512 2.84 1.22 1.19 1.96 1.84 2.00 1.31 1.23 4.13
1024 2.52 1.04 1.27 2.99 3.57 1.54 1.28 1.27 *
The spectrometer was fairly stable over the entire 20minutes for scan 30, but degraded after 4minutes for scan 26. This indications a variation of the sky during these daytime tests which complicates the comparisons of individual scans. The bold values show when the rms was more than 1.3 times higher than the starting rms. When the cabinet is open, the baseline RMS degrades quickly. All results tabulated are from Bank-A, but we find similar results were obtained by spot-checking other VEGAS Banks. In general, the spectrometer baselines tend to be more stable than VEGAS and when the rms degrades after a period of time, the spectrometer baselines appear smoother (better fitted by a low-order polynomial, while the VEGAS baselines tend to show more structure).

  • Qbandrms_2014.09.10.jpg:
  • Below shows a plot of the measured RMS as a function of "scan" time for simulated ONOFF scans based on the long track scans for VEGAS mode1,2,&3 and the spectrometer.
  • The dotted and the dashed line shows the theoretical ~time^-0.5 relationship for the spectrometer and VEGAS respectively. The dotted (spec) line is above the dashed line (VEGAS) since the Spectormeter is based on 3-level sampling efficiency (K1~1.2 for Spectrometer vs K1~1.0 for VEGAS)
  • For these data, the Spectrometer scans does a better job integrating down with expectations than VEGAS (~60 secs of stable data for spectrometer vs 10-30 sec for VEGAS).
  • For "standard" 2min ONOFF/NODs, the spectrometer rms for these data is about 2.5 times better than that seen for the VEGAS modes
  • Qbandrms 2014.09.10.jpg
  • Currently, VEGAS baselines appear more sensitive than the spectrometer to changes of total power due to sky variations and thermal effects. For extragalactic observations of weak-broad lines, it would be reasonable to continue using the spectrometer until we can demonstrate good baseline performance with VEGAS.

This topic: CICADA > WebHome > GreenBankSpectrometer > CICADAGreenBankSpectrometerCommissioningLogs > CICADAGreenBankSpectrometer2014Sep10
Topic revision: 2015-01-27, RichardPrestage
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback