Back to Astronomical Checkout summary.

Checkout of Spectrometer Mode 1N2-0A-50-9: 50MHz, 9 level, 2 samplers, 1 quadrant

Dave Hogg, Richard Prestage, 12th September 2002

As part of TSPECTEST_02 and TSPECTEST_03, eighteen pairs of 300 sec OffOn scans were performed with the L-band receiver, observing two galaxies which had previously been observed as part of the Fisher GBT HI galaxy survey. To the extent tested by only ~3 hours hour of observing, we consider this mode to have passed its astronomical checkout, with the caveat that 30% of our 300sec OffOn pairs showed bad baselines. Excluding these scans, the rms noise values integrate down as expected, and the HI galaxy line profiles are a good match to that obtained by Fisher. Further details are given in the sections below.

Contents

TSPECTEST_02/03 - 18/23 August 2002

Observation Summary

Data for TSPECTEST_02 were acquired on 18th August 2002 (EDT). Data were acquired using M&C version 3.5.0, but with a test version of the LO1A synthesizer manager, which had upgrades to improve the frequency switching behaviour. This was expected to have no impact on position-switched observations. The version of the Spectrometer Manager used was the build of Tue Aug 13 13:55:10 GMT 2002. The "daily" version of aips++ was used for data reduction. This corresponds to Version 1.8, and builds between approximately 165 and 180.

These observations were made on U9179o, a galaxy chosen from the Fisher GBT HI Galaxy Survey. The data are available in raw form at /home/gbtdata/TSPECTEST_02. An aips++ measurement set containing the scans discussed below is available at /home/thales/scratch/gbtmsdata/T50_MHZ_9LEVEL_1N2-0A-12-9 in MS TSPECTEST_02_SPECTROMETER_A/ .

For the purposes of Spectrometer checkout, six pairs of 300 second OffOn scans are relevant, as follows: 

   25-26
   27-28
   29-30
   31-32
   33-34
   35-36
   37-38
Three of these spectra, 30,34 and 38, show bad baselines. Apart from discussion of the comparison with the Fisher data, no further results from this dataset are presented here.

Data for TSPECTEST_03 were acquired on 23rd August 2002 (EDT). Data were acquired using M&C version "test" a pre-release of 3.6. This behaved well, and this M&C version was subsequently released with no changes. The version of the Spectrometer Manager used was the build of Thu Aug 22 19:53:44 GMT 2002. The "daily" version of aips++ was used for data reduction. This corresponds to Version 1.8, and builds between approximately 165 and 180.

These observations were made on SL1720+6236, an extremely faint galaxy chosen from the Fisher GBT HI Galaxy Survey. The data are available in raw form at /home/gbtdata/TSPECTEST_03. An aips++ measurement set containing the scans discussed below is available at /home/thales/scratch/gbtmsdata/T50_MHZ_9LEVEL_1N2-0A-12-9 in MS TSPECTEST_03_SPECTROMETER_A/ .

For the purposes of Spectrometer checkout, twelve pairs of 300 second OffOn scans are relevant, as follows: 

   32-33
   34-35
   36-37
   38-39
   40-41
   42-43
   44-45
   45-47
   48-49
   50-51
   52-53
   54-55
Three of these spectra, 33, 35 and 55 show bad baselines.

Example Spectra

Passband Edges

Figure 1 (postscript) shows a typical total-power passband for this spectrometer mode. It can be seen that the IF filter band edges are not quite aligned with the spectrometer passband. For this (and other) reasons, the first ~1.5MHz (~250 pixels at this resolution) of the reduced spectrum is significantly noiser than the remainder. Results for a typical scan (46) are shown in Figure 2. (postscript).

Baselines

In many cases,baselines obtained with this spectrometer mode with the L-band receiver were good. However, three scans out of six on the first night, and three out of twelve on the second were markedly worse than the remainder. We inspected these scans integration by integration, and saw no signs of individual integrations which were obviously contaminated by RFI; rather the baseline shapes varied smoothly throughout the scan. Figure 3 (postscript) shows an example of one of the bad scans (TSPECTEST_02 35) as created using DISH d.calib(),d.getc(). No baseline has been removed. Figure 4 (postscript) shows the spectrum obtained by averaging the nine good scans from TSPECTEST_03, then averaging the two polarizations, with a linear baseline removed. The feature at ~1403.5MHz is the HI galaxy.

Comparison to 12.5MHz Results

Figure 5a (postscript) shows a comparison of the 50MHz scans obtained on U9179o during TSPECTEST_02 with those obtained in 12.5MHz mode on the same object. Each line corresponds to the average of the good scans, with a linear baseline removed. As can be seen, the agreement is reasonable. By inference then, these data also agree with the Galaxy Survey results for the same object.

Comparison to GBT Galaxy Survey Results - SL1720+6236

Figure 5 (postscript) shows a region around the spectrum of SL1720+6236, chosen to be suitable for comparison with Fisher's spectrum. Note that due to a temporary quirk in dish, this has been plotted against radio velocity; Fisher's spectrum is plotted against optical velocity. Although the S/N is very poor in both spectra, the general details agree (including the line velocity, once the radio/optical conversion is taken into account).

Our result:

Fisher result:

Noise Performance

Table 1 contains the RMS values in mK for a single spectral channel for the individual scans from TSPECTEST_03, and for various averages. 
	  		   Table 1

	                 RMS Values (mK)

               (a) Individual Scans of 300 Seconds

	  Scan Number    Chan X      Chan Y      Both
               33          15.5        41.0       21.5
	      35          17.5        40.4       24.4
	      37          15.3        15.1        9.79
	      39          16.2        15.8       12.1
	      41          13.5        13.1        9.27
	      43          13.5        13.0        9.37
	      45          13.4        13.1        9.37
	      47          13.7        13.5        9.30
	      49          13.0        12.9        9.35
	      51          13.0        13.3        9.51
	      53          13.3        13.6        9.47
	      55          13.0        14.8       10.2

          Mean of 12      14.2        18.3       12.0
          Dispersion       1.5        10.5        5.2
          Mean of 8       14.0        13.7        9.76
          Dispersion       1.2         1.1        1.0


                       (b) Averaging Scans

	  Scan Number    Chan X      Chan Y      Both
            33&35         13.1        14.2        9.43
            37&39          9.95       11.0        7.64
            41&43          9.76        9.41       6.90
	    45&47          9.85        9.66       6.80
	    49&51          9.42        9.34       6.73
	    53&55          9.57       10.3        7.15
         37&39&41&43       7.32        6.91       5.09
         45&47&49&51       7.14        6.70       4.91
	 37 to 51 incl     5.55        4.82       3.77
 	 33 to 55 incl     5.37        4.61       3.65
Assuming that the Tsys is 19 K, and that the channel width is 6104 Hz, the rms expected in an observation of 300 seconds is 14.1 mK. The observed values of rms for the eight best scans pairs, 37 through 51 are plotted in Figure 6 (postscript). The solid line shows the expected variation, and the observed values are in excellent agreement with the expectations.

Observing Logs

TSPECTEST_02

TSPECTEST_03

Glish reduction procedures

After the data were inspected interactively, the following glish procedures were used to automate the production of the above figures. Note some interaction with dish is required to set baseline regions.

Back to Astronomical Checkout summary.