ALFA Performance at 1665 MHz
C. Salter and T. Ghosh

On 26 July 2004, with the "OH filters" installed in the ALFA receiver, 17 cross-scans were made on the unpolarized point source, B2120+168 (3C432). The spectrum of this source has been model fitted for the standard Arecibo catalog of calibration sources, the rms differences between all the flux densities used for the fit and the model itself being about 8%. The four values at 1.4 GHz used in the fit give S(1.4) = 1.622 Jy, with a mutual standard deviation of about 5%; the 1.4 GHz flux density from the model fit is 1.577 Jy, a deviation from the mean measured value of about 3%.

The 100-MHz passband recorded was centered at 1.665 GHz. Each of the arms of a cross scan had a great-circle length of 31.5 arcmin. Successive cross scans cycled through successive beams in the order Beams 0-6. A single cross scan (on Beam 2) was rejected as then the source passed too close to the Arecibo zenith to track. We note that all the cross scans used in the present analysis were situated in the range, 2 deg < ZA < 9 deg.

The total-power time series of each arm of a cross scan was fitted by a baselevel plus three Gaussians at the positions of the central beam and the first side/coma lobes. A single value for the noise diode temperature, 20 K was adopted after looking at Phil Perillat's measurements, was entered for both polarizations and all beams. Phil cautions against using the values he derived, as detailed in his memo, and thus no confidence should be placed on the values of Tsys and K/Jy derived. Below we will present only SEFD values, which are derived independently of the noise diode measurements, and depend only on the flux density adopted for the source. However, the internal spread of the derived values of Tsys for the different beams and polarizations does produce an interesting result (see below). The peak source deflections measured were corrected for the pointing offsets. However, the positions available with the data were (at that date) recorded asynchronously with the data taking, and hence are less accurate than should be.

In the plots displayed below;

• Beam 0 is marked by a plus sign
• Beam 1 is marked by a star
• Beam 2 is marked by a filled square
• Beam 3 is marked by a diamond
• Beam 4 is marked by a triangle
• Beam 5 is marked by an open square
• Beam 6 is marked by a multiplication sign

Pointing (Figs. 1 a-d)

As expected from the then-asynchronous nature of the position information, a spread of measured pointing errors of about +/-30 arsec is found in each coordinate. The mean offsets averaged over all the beams are -11.5 arcsec in azimuth, and -6.0 arcsec in zenith angle.

Half Power Beam Widths (HPBWs) (Figs. 2 a-d)

The HPBWs averaged over all beams and for the two polarizations are;

It is seen that the Poln A HPBW of 3.04 x 3.32 arcmin is significantly smaller than the Poln B values of 3.33 x 3.92 arcmin, especially in the zenith angle direction.

System Equivalent Flux Density (SEFD) (Figs. 3 a-b)

The mean SEFDs computed separately for Beam 0 and Beams 1-6 for the two polarizations are;

The higher rms's for Beam 1-6 is attributable in part to differences between the individual beams. For example, for Poln B, Beams 3 and 4 always had SEFDs higher than the mean, with Beams 5 and 6 always having SEFDs lower than the mean.

Interestingly, the ratio of SEFD(B0)/SEFD(B1-6) has the same value of 0.76 for both polarizations. This lower value for Beam 0 is expected from the optics of the system.

Apparent System Temperature (Figs 4 a-b)

For Poln A, the total spread of apparent system temperatures (assuming the same noise diode intensity for all beams) over all beams and measurements is 16% of the mean value. This includes statistical noise fluctuations, RFI, baseline fitting and changes with the (albeit limited) ZA range.

In contrast, the spread for Poln B is 68% of the mean value. Clearly the noise diode coupling to polarization B is less uniform than for Polarization A.

Coma Lobe Level

For Beams 2 and 5, given the ALFA array orientation at that epoch, the cross-scan azimuth cuts passed right through the maxima of their coma lobes. The ratio of the Gaussians fitted to the main beam and the coma lobe yield the following values;

There is some indication that the coma lobe levels may be a little higher for Poln B.

Concluding Remarks

It is of interest to compare the above values for SEFD (~4.3 and 5.7 Jy, for the orthogonal polarizations of Beam 0) with those from ALFA at 1.4 GHz (~2.75 Jy for Beam 0), and the L-Band Wide receiver (LBW; ~2.2 Jy). The quoted values are for the same range of zenith angle.

The measurements made here also show that Poln B has degraded much more by 1.665 GHz than Poln A. The SEFDs for Poln B are 30% higher than Poln A, while (consistent with this) the beam area of Poln B is 30% the larger. Also indicating that all is not well with Poln B, there are indications that the noise diode coupling for polarization B is less uniform than is the case for Polarization A.

The values derived here for SEFD and HPBW agree well with those measured recently by Phil Perillat from an independent data set of "spider scans".

Figures