Test of Spectrometer Configuration 4W2-6XY-200

D. Hogg and F. Ghigo

November 18, 2002

Observations made at CBAND October 29,2002

Summary

In this configuration the noise on a wide-band spectrum is dominated by the systematic effects of baseline ripple. In one test in which six OFF/ON pairs were observed, the noise for a narrow signal reaches the value expected, if one scan which is clearly poor is omitted. In another group of six scans four were poor, and one was bad. The noise in the average was higher than theoretical, by 20%, reflecting the effect of the baseline ripple.

The inferred sensitivity of the GBT at 5 GHz is 1.93 K/Jy and 1.83 K/Jy for receiver channels one and two, respectively.

The velocity observed for the formaldehyde absorption line, 4829.569 MHz, is consistent with the expected radial velocity (+5.8 km/s) for molecular lines in the source TMC-1.

Channel Two shows a large anomalous narrow feature in the central channels. The feature shifts with a shift in Local Oscillator.

This configuration of the Spectrometer is ready for general use at CBAND in studies of spectral features which are narrower than 10-20 MHz.

Observations and Analysis

The observations were made at CBAND, and included measurements of a continuum calibrator, of TMC-1 and of a comparison between two "OFF" regions.

I. The continuum calibrator 0502+2516. 

Scans #74and #75 are an OFF/ON pair of duration 30 seconds each, made at
a center frequency of 4829 MHz. Because a large anomalous feature is
present in the central pixels of the second polarization, a second
pair of scans (#76,#77),  was taken, centered at 4840 MHz . The
integration time was increased to 5 minutes for each of the OFF and ON.

The spectra from the two receivers in the two pairs of scans are shown
in Figure tday11#75+77.ps. The central feature in the spectrum of CH2:YY
moved with the change of LO, but the baseline structure remained the same,
in each channel, to a remarkable degree. This implies that the amplitude
of the baseline ripple is not reduced by a more rapid switching cycle.

In the spectral region between 4818 and 4828 MHz the amplitude of the
signal is nearly constant. Table 1 summarizes the parameters of the
signals in this region. In developing this table we adopt 2.37 JY as the
flux at 5 GHz of 0502+2516, and we take the spectrometer parameters
appropriate for the 3-level configuration. The system temperature
is 22 K, and Hanning smoothing was applied.

                                Table 1.

Scan       Amplitude         Gain             RMS in mK
              Ta*            K/Jy        Observed    Expected
          RX1     RX2     RX1    RX2    RX1    RX2
75        4.58    4.43   1.93   1.87    35.7   43.5     44.9

77        4.66    4.44   1.97   1.87    21.1   26.7     20.1

It is seen that the gain of the antenna is close to the expected
value of 2, and that the rms is close to the expected value over
a narrow range in the spectrum.

The variation of the system gain is given in Table 2.

                                Table 2.

Frequency Range                Gain       K/Jy

      MHz              Rx Channel 1            Rx Channel 2

                    Scan 75    Scan 77      Scan 75    Scan 77

  4745 - 4790         1.89       1.92         1.80       1.80

  4780 - 4825         1.96       1.99         1.88       1.88

  4845 - 4890         1.89       1.91         1.81       1.82

  Rx Average            1.93 +/- 0.04           1.83 +/- 0.04



II. The Molecular Line Source TMC-1



This source was observed in a set of six OFF/ON pairs (scans #78-
#89), the OFF and the ON requiring five minutes. each. Inspection
showed that the first scan, #79, had more baseline structure and
significantly higher rms than the other scans, and it was not included
in the subsequent analysis.


Figure tday11#81pla.ps shows the average of the scans #81 - #89. The
feature in the center is an anomalous spike which appears in the second
polarization (CH2:YY) only. The formaldehyde absorption line appears
at approximately 4830 MHz. Table 3 summarizes the rms values observed
for the scans, and for the averages.

                                Table 3.

Integration                   RMS, in mK
Time (secs)     Expected     RX1     RX2     Both
    300            20.1      19.3    21.7
    600            14.2      13.0    14.6    15.1
   1200            10.0       9.76   10.0
   1500             8.98      8.47    9.70
   3000             6.35                      6.56

The rms values are close to the expected values, with the first
polarization being slightly better, and the second polarization being
slightly poorer. Averaging the two polarizations on scan #87 results
in a significantly higher rms than for the other scans, and leads to the
high value in the table for "Both" at 600 seconds.


Figure tday11#81avc.ps shows the detail of the absorption spectrum
of formaldehyde. The minimum in the profile occurs at pixel 18093,
or a frequency of 4829.56907 MHz. This differs from the laboratory
rest frequency 4829.6639 MHz (Jewell memo) by  -94.8 kHz, which corresponds
to a velocity difference of +5.9 km/s, in satisfactory agreement with
the expected value +5.8 km/s (Jewell memo).



II. Two "OFF" Positions Near the Molecular Line Source TMC-1



Because of the possibility that the amplitude of the ripple in spectral
baselines was a function of the total power offset between the OFF and
the ON positions, we made a sequence of six pairs of OFF/ON scans using
for the 'on" position the reference 'off' position used in the previous
observations of TMC-1. All scans from #91 to #101 except scan #93
were usable. Scan #93 had a large offset between the OFF and ON positions,
and showed a large baseline ripple which increased the effective rms by
a factor of 2.5 over that seen in the other scans.


Figure tday11#91avc.ps shows the data for five scans averaged, for
each polarization. A linear baseline has been removed, and the residuals
show an rms of 11.7 for CH1 and 10.2 for CH2. The spectra are flat,
with structure of amplitude approximately 20 mK. The following
table shows that the rms is higher in CH1, by about 20% in a five
minute observation. Both are higher than the expected rms, by
37% and 13% respectively. The reduction in the rms achieved by using
five scans is slightly greater than the expected root(five).

                               Table 4.

Integration                   RMS, in mK
Time (secs)     Expected     RX1     RX2     Both
    300            21.0      28.9    23.6
    600            14.8      18.8    16.4    11.0
   1200            10.5      13.4    12.0
   1500             9.39     11.7    10.2
   3000             6.64                      7.86

For reasons that are not clear the spectra in this set of observations
show more baseline structure than was found in the observations of
TMC-1 itself. The only scan which reached the theoretical sensitivity of
21 mK in each of the receivers was #101, though the second channel was
also very good in scan #97. Receiver One showed an rms in excess of 28 mK
in all scans except #101. Thus this test produced the somewhat surprising
result that the data for the observations of TMC-1 were better than those
for the observations of blank sky. However, TMC-1 is not a strong continuum
source, so that undoubtedly the system was is reasonable balance both
OFF and ON during the observations of TMC-1.