Measuring the sband wide cal  values.

sep,2001

     The sband wide receiver was mounted on the antenna range and its receiver temperature (amp and omt) was measured. It was then mounted on the telescope and the cal values were measured using absorber and sky (see measuring the cals on the telescope using absorber and sky for a description).
    The correlator  configuration used was 25 Mhz by 256 lags by 4 sbc.  1800 to 3100  MHz  was covered in 100 Mhz chunks (13 steps). At each step a 5 second integration with the cal off was followed by a 5 second integration with the cal on. The frequency range  was covered 3 times (each pass separated by about 6 minutes). The high correlated cal was used (diode 1 feeds both polarizations).
    For the computations we:
  1. used 6 K for the cold sky temperature.
  2. measured the absorber temperature with a thermometer : 303 K.
  3. used the receiver temperature measured on the test range.
  4. used 14 K for the scattered radiation into the horn.
    The cal value was computed 3 separate ways:
  1. Using just the cal on/off on the absorber. This needed the measured receiver temperature and it assumed that the absorber noise getting into the amp was 303 K.
  2. Using just the sky. This used the 6 K sky, 14 K scattered radiation, and the measured receiver temperature.
  3. Using the hot (absorber) and cold   (sky) loads to compute the cal. This needs the absorber value, sky value (cold sky + scattered radiation). The receiver temperature is not used since it cancels out.
 The rfi in the spectra was masked out before computing the power in each 25Mhz band. The rfi situation can be seen in the plots of spectra versus frequency for the range 1800 Mhz to 3100 Mhz. These are the spectra taken for the cal off. Black is on absorber and red is on the Sky.  The frequencies with lots of rfi are:
  • 1875 - 2100 Mhz
  • 2300 - 2350 Mhz
  • 2400 - 2600 Mhz

    • The rfi is so strong at 1940 Mhz (cellular phones) that even the absorber (black lines) is full of rfi. After removing the rfi, the rms/mean of the 3 separate cal measurements at each frequency was computed. If it was greater than .015, then that frequency was not used.

    The sband wide receiver also has resonances in the OMT. These are visible in the above plot and are discussed in resonances in the sband wide feed.

        After removing the points with rfi, the cal values versus frequency were  plotted in the figures:

    1. Figure 1 plots the cal value versus frequency. The top plot is PolA and the bottom plot is PolB. The 3 separate measurements at each frequency are overplotted. The colors are:
    2. Figure 2 plots the receiver temperature versus frequency measured on the test range. PolA is black and polB is red. The dashed line is a 2nd  order fit to the measured data. The value at 2580 was not used since it lies on a resonance in the OMT.
        The 3rd order polynomial fit versus frequency only used the absorber and absorber/sky  ratio values. The sky only values were not used since there were few of them and they jumped around.  You can also see a 200 Mhz ripple in the data. This is probably in the directional coupler for the cal value.  It is small enough that I didn't bother to try and fit it.The 3rd order polynomial fit of cal value in Kelvins versus frequency in Ghz is:
    calA[Ghz]= 18.710434    -    9.4532671*f +   3.4901421*f^2 -    .4608357*f^3
    calB[Ghz]=-3.1919432  + 30.9665108*f - 14.7711287*f^2 + 1.8870836*f^3
        The polA cal values are flat (~ 1 K) while the polB values vary by 8 K over the frequency range. The difference is probably the coupler used. PolA has a new coupler that is not painted while polB has an older coupler that is painted (the couplers are in the dewar). You can see that the old cal values both had  large slopes (and they were both painted).

        To check to see how well the absorber was matched to the horn, we moved the absorber 3 inches and then 6 inches below the horn. At each position we measured the cal/TsysAbsorber. It was done 3 separate times at each position. The figures shows the affect of moving the absorber. The variation in the values increase as we move the absorber from touching the horn to 6 inches away. To do this test someone actually held the absorber 3 and 6 inches away (so it may not have been that stable). 

    processing: x101/010920/sbwcal.pro, sbwres.pro, movabs.pro
     home_~phil