The GBT Flux Calibration Program (GBT15A-486) was a filler project
design to track the flux variability of the sources, to measure the
stability of the GBT noise diodes, and to monitor the performance of
the receivers, pointing model, and surface of the GBT. It is design
to run within short 30 minute blocks time and can be carried out at
any LST range using any one of the available receivers. A primary
goal of the program is to make the GBT calibration consistent with the
VLA calibration scale given by Perley & Butler (2013).
All data taken from this program are publicly available. Contact
David Frayer if you have questions or are interested in the data.
The measurements from calibration program 15A_486 are tabulated
We have concentrated measurements on the VLA standard
calibrators 3C286, 3C295, and 3C123 which are stable calibrator
sources (Perley & Butler 2013). Below shows a figure from the
currently available data that shows the comparison between GBT
measurements and the adopted VLA flux densities
(Figure [ps]) Figure Caption:
GBT/VLA flux density ratios based on the calibrators 3C286, 3C295, and
3C123. The expected flux densities are based on the VLA calibration
(Perley & Butler 2013), while the GBT values are based on the current
adopted noise-diode values, many of which have not been updated in
more than a decade. The plot justifies updating the GBT noise-diodes
for every band, and potentially taking new SCAL observations. Only
X-band is currently in agreement with the VLA calibration.
Median GBT/VLA Flux Density Ratios
The above table shows median results from the GBT observations of
3c286 compared with the expected flux densities from the VLA
calibration program. The results are based on measurements taken from
2015.07 to 2016.03. The frequencies observed within each band are the
standard VLA calibration frequencies. The results for both
polarizations for the GBT have been combined.
Note that the results from Ka and Q-band are derived by on-going
observing projects in good weather since the high-frequency data for
the calibration program was impacted significantly by the weather,
given the calibrators are weak at high-frequency (e.g., small Tsys
differences on the sky between on/off yields large % scatter on
The L-band X-polarization noise diode is too low by ~20%.
The C-band noise diodes are overestimated.
The X-band noise diode values appear reasonable.
The KFPA has shown a wide range of results over the last two
years and highlights the importance of checking that the cal-drain
currents are near 5.0mA before your observations begin.
No data has been taken for Ka-band from this filler program.
Many of the Q-band data points are affected by weather.
The S-band noise diodes are close but slightly understimated.
Below 1 GHz, we should not blindly adopt the VLA calibration
scale since the VLA results are best constrained above 1 GHz.
Since most of the GBT noise diode TCAL values are out dated, we
recommend observing a standard calibrator to calibrate your data.
GBT Calibration vs VLA Calibration
The GBT calibration scale is based on the absolute calibration from
Ott et al. 1994 and Peng et al. 2000 while the VLA calibration is
based on the Perley & Butler 2013 paper. The plot here shows the
ratio between the current GBT calibration scale and the VLA
calibration scale based on different calibrators as a function of
frequency (Figure [ps]). The plot does not
represent any observational data from the GBT or the VLA. It just
shows the relative fluxes assumed for the calibrators as a function of
frequency. The GBT SCAL programs should be updated to match the
source coefficients given by Perley and Butler 2013 to place the GBT
noise-diodes on the VLA calibration scale.
The Ta temperature scale is derived from Ta = Tsys (ON-OFF)/OFF. The
Ta' temperature scale has been corrected for atmospheric attenuation:
Ta' = Ta * exp(tau_o*Airmass). The flux density scale for the GBT is
Snu = Ta'/(2.84*eta_a), where eta_a is the aperture efficiency which
is given by the Ruze equation with a long-wavelength efficiency of
0.71: eta_a= 0.71 exp(-1.*(pi4*esurf/lam)^2.). For default, we have
assumed esurf errors of 250-microns. This will be checked with high
frequency observations. With repeated measurements at low frequency
where the effects of the surface are small, we can directly check the
applicability of the current noise-diode Tcal values by measurements of
the flux calibrators. If we find system discrepancies, we will
suggest detailed SCAL observations across the full receiver band so
that the values could be updated in the GBT calibration database.
The program runs for any LST range and for any available receiver on
the GBT. The ASTRID script chooses the appropriate source based on
the current LST and the system is configured based on the current
receiver in the scan coordinator.
(1) Put Rx of choice in the ScanCoordinator
(2) Run "run_GBTcal" in astrid under project code AGBT15A_486
(3) If time permits repeat run_GBTcal a second time
(4) As time permits, repeat steps (1)-(3) for different Rx's
within the observing session
Each instance of run_GBTcal will take about 10-15 minutes to run
depending on how far the telescope needs to slew.