VEGAS Bank Discrepancy Problem - 10 December 2014
The same receiver/feed/polarization data acquired through different Banks of VEGAS show different line intensities and system temperatures. The affect appears more pronounced for Modes 21 and 20, less so for Mode 5. The same effect is not visible in the GBT Spectrometer.
When routing signals from the same receiver beam/polarization to different VEGAS Banks, the resulting system temperatures and line intensities vary, although they should be essentially identical. This result was reported by Dana Balser in July, but at the time we feared it might be conflated with other VEGAS problems. After most of these have been solved, we retested this configuration on 10th December 2014, and the problem remains.
Commissioning Tests - 10th December 2014
Commissioning tests were performed Wednesday 10th December 2013, 2:30 - 5:00pm ET. Weather was very overcast, with some light snow. Data was taken into AGBT14B_431_05. Observations were made at X-band towards the Radio Recombination Line source W3 (which also has strong continuum). Dana's log file is attached here.
Data were taken using the configuration file
Configuration "!lineH91BankA_H" configures VEGAS to Mode 21, eight banks, eight subbands, and with the same line, at the same rest frequency, placed in each subband. VEGAS was balanced on source, and then an OffOn observation with 6 minutes scan duration was performed.
A typical "getps" GBTIDL spectrum is shown here:
A typical spectrum with a baseline removed, and a Gaussian fitted to the line is shown here:
Dana Balser wrote a short script to debaseline and fit a Gaussian to each of the 64 spectra. The system temperature Tsys, and line intensity Ta was recorded for each subband.
For the eight spectra within a single bank, Tsys and Ta were essentially identical. However, Tsys and Ta varied between Banks (see table below).
The same experiment was repeated with Modes 20, 5, and the GBT Spectrometer
21 20 5 ACS
- Line intensity as a function of bank for different modes:
- System temperature as a function of bank for different modes:
- Line intensity as a function of system temperature for different modes:
Potential Diagnostic Tests
- Run a similar test with mode 3 (will rule out the GPU code) - needs a broader line, or a continuum source.
- Run a similar check on a source with no continuum (but Dana sees that the problem is there on a weak continuum source)
- Put an N-way splitter after the converter rack, and bypass some (all?) of the VEGAS analog section.
- Reroute cables, and see whether the effect (e.g. highest Tsys) moves with the analog system or the digital system