TPTCSPNT_080408 K-band gain calibrations, April 8, 2008
The plan is to do peak/focus on 3C147 and Mars until they set,
Then switch to 3C286 and 3C295 until they set about 8am tomorrow morning.
Weather conditions were mostly clear, but becoming foggy or light overcast.
First, Galen and I went up to the receiver room to flip a switch that enabled MCB control
of the KA-band receiver cals.
Observing from about 00:45 to 12:00 UT.
Observing Log
Start observing about 8:50 EDT; scan 1 on 3C147 -- antenna not in scan coordinator!
Start again with scan number 2.
Using scheduling block "Kgaincal_2"
--config: 22.236 GHz center with 320 MHz bandwidth, dual beam; tracking feed 1.
-- Tsys about 46 K; winds calm < 5 MPH at start; clear skies.; Tambient=8.5C
-- active surface, Zernickes on, FEM on; dynamic corrections on.
-- quadrant detectors etc running, but Inclinometer temperature files not updating?
-- Cleo predicted opacity about 0.08 at start, a bump up to 0.12 sometime during the run.
-- MARS diameter April 8 is 6.66x6.62". Tant about 3.5K
10:15 pm -- leave it with Dave
Scans 2 - 301 peak/focuses on Mars and 3C147
02:01 AM -- switch to scheduling block "Kgaincal_1" (3C286 and 3C295)
Scans 301 to 628.
End of observing at 08:02 EDT (12:02 UT)
Altogether, 125 good peak/focus sequences.
- AZ/EL coverage for this session:
The system temperature Tsys, measured at each peak scan, was fit to a simple
function of elevation:
Tsys = Trcvr + Tatm{ 1 - exp(-τA)}
where the airmass A = 1/sin(elevation)
The Tsys values were calculated by the program "prepoint" which uses the GFM utilities;
the noise cal injection is used to convert units to Kelvin, using previously measured
calibrations of the values of Tcal.
- The results of the fitting, and the assumed Tcals, are listed in Table 1.
| Pol | Trcvr | Tatm | τ | rms | Tcal |
| LCP | 19.53 | 259.9 | 0.1031 | 1.37 | 5.213 |
| RCP | 20.60 | 279.0 | 0.1024 | 1.46 | 4.333 |
All values, except for τ, are in Kelvin. "rms" is the rms deviation of the data from the fitted curve.
Figure 2a shows the graph of Tsys vs elevation, and the fitted curves; Figure 2b shows Tsys plotted vs airmass.
We can compare the opacity, τ, derived from this dataset, which is sort of an average over the 11 hour observing session, with
the Cleo predictions, which are shown in Figure 3. The predictions show the average τ varying from about 0.08 to
0.11 over the observing period. So the τ of 0.103 derived from the Tsys data is in reasonably good agreement.
We can also compare with Cleo's prediction of the atmospheric temperature, Tatm. Figure 4 shows the Cleo predictions,
in which Tatm ranges between 270 and 274K, again in not too bad agreement with the fits to the Tsys.
- Fig 2A: Tsys vs Elevation and fitted curves.:
- Fig 2B: Tsys vs Airmass and fitted curves.:
- Fig 3: Predicted Opacity, April 7-8.: The horizontal bar shows the duration of the experiment.
- Fig 4: Predicted atmospheric temperatures.: The horizontal bar shows the duration of the experiment.
Tatm is not expected to be different at different polarizations. We can use the Cleo prediction for Tatm instead
of fitting for it. The Cleo prediction for atmospheric temperature, averaged over the time period of the
observing, is 272.4 K. If we use that, then we get two different τs : 0.0964 and 0.1064.
Of course there should be only one τ that applies equally to both polarizations.
We can bring the LCP and RCP data into alignment by increasing the LCP Tcal by 3% and correspondingly decreasing
the RCP Tcal. We thus correct our Tcals. The results of the fitting of an opacity curve are listed here in Table 2:
| Pol | Trcvr (K) | Tatm (K) | τ | rms (K) | Tcal (K) |
| LCP | 20.01 | 272.4 | 0.1013 | 1.42 | 5.376 |
| RCP | 20.01 | 272.4 | 0.1013 | 1.41 | 4.201 |
Figure 5a shows the corrected data and the fitted curves versus elevation, and Figure 5B versus airmass.
Using the calibrator flux information from
http://www.vla.nrao.edu/astro/calib/manual/baars.html
we obtain the following for our observed calibrators:
| Source | Flux Density at 22.236 GHz |
| 3C147 | 1.815 Jy |
| 3C286 | 2.539 |
| 3C295 | 0.941 |
Gaussian profiles were fitted to each cross scan. After each two azimuth scans, the pointing LPC is updated so that
the elevation scans should go through the peak. Thus we used the peaks of the fits to the elevation scans to determine
the antenna temperatures (Tant). The Tant is corrected for atmospheric absorption and divided by the source flux
density to obtain the gain, G, in Kelvin/Jansky.
- G = Tant * exp(τA) / S(Jy)
The Gains are plotted in Figure 5 versus elevation, and also shown are fits of a quadratic function of the form:
- G(z) = Gmax * (A0 + A1*z + A2*z^2 )
where z = the zenith distance = 90 - elevation.
The resulting gains are 1.87 and 2.02 K/Jy at LCP and RCP, respectively. But these values seem unreasonably high.
- Figure 5: Gains(K/Jy) and fitted curves:
We can predict the gain from previous measurements and estimates of the RMS surface errors.
The gain, G, relates to aperture efficiency, η as follows:
- G (K/Jy) = Ta'/S = Ag η / 2k = 2.845 η
- Ag is the geometric area of the telescope,
- Ta' = Tant * exp(τA) ; (where A is the airmass).
The main factors contributing to the aperture efficiency are the pattern illumination efficiency, ηpat,
which is about 70% for the GBT feeds, and the surface efficiency, ηsurf, due to surface irregularities.
The surface efficiency is given by the Ruze formula:
- ηsurf = exp{ -(4*π s / λ)^2 }
- s is the rms surface irregularity.
Earlier calibrations have yielded s=390 microns, which would imply that
- ηsurf = 0.8762 at 22.236 GHz.
Thus the combined aperture efficiency would be
- η = 0.61
- or G = 1.74 K/Jy.
Note that G=1.74 is about what was measured in the previous gain measurements of December 2005.
Since that time the surface model has improved due to the addition of the Zernicke polynomial terms.
This would result in flattening out the gain curve, but is not expected to increase the peak in
the gain curve by a significant amount. A measured G=2.0 implies that there are no surface errors,
i.e., s=0.
A likely explanation of this paradox is that the TCals have changed since December 2005.
Another contributor to the discrepancy is a possible error in the opacity. The
-- FrankGhigo - 08 Apr 2008
coverage_080408.gif 
