Misc

Splatalog interfacing like ALMA?

GBT

1. Mosaic
   1. ( Ref 1; Section 7.7)
      1. A RALongMap can take up to 25 seconds to start; scans <1 min have significant overhead
      2. Fields of Views larger than 10 - 20 beam FWHM should use RALongMap or DecLatMap; smaller FOV should Daisy Scan
      3. Scan lengths should be under 15 minutes
      4. There should be four switching periods per integration
   2. Raster scan lengths or petal lengths should be at least 30s so reduce stress on telescope (ref 1; Section 6.4.3)
2. Front-end/BackEnd
   1. DCR has minimum integration time of 100ms

   2. If high frequency and if doing position switched observations and if the source is not extended, then use the Astrid Nod() procedure to observe (Ref 1; Section 6.4.2)

   3. VEGAS
      
      1. Table 8.2 (ref 1) has minimum recommended switching periods with VEGAS observations that use a noise diode; Table 8.3 has the switching periods that do not use the noise diode.
      2. Switching period, frequency switch, integration time must be the same for all banks; the banks can have different spectral resolutions, bandwidth and number of subbands (spectral windows). The subbands in a single VEGAS bank must be identical except for the central frequency.
3. Calibration
   1. Table 7.1 of (Ref 1) gives recommended Point/Focus Spacing

   2. (Ref 1; Section 7.4) Flux density calibration : Peak/Focus on calibrator, followed by observation in same spectral line setup as program sources.

   3. Observing Strong Continuum Sources (ref 1; section 7.9)
      1. Might want BalanceOnOff() here instead of Balance()

   4. When there is a large difference in power received by the GBT between two positions on the sky, it is advantageous to balance the IF system power levels to be at the mid-point of the two power levels. Typically this is needed when the source position is a strong continuum source. One should consider using it when the system temperature on and off source differ by a factor of two or more. (ref 1; section 6.4.1.10)

   5. Nod should only be used with mutli-beam recievers (ref 1; section 6.4.2.5)
   6. AutoOOF Strategy (ref 1; section 7.2)
      1. recommended for f > 40 GHz
      2. only available for Rcvf26_40 (Ka-band), rRcvr40_52 (Q-band), Rcvr68_92 (4mm Reciever or Wband) RcvrArray75_115 (Argus), and Rcvr_PAR (MUSTANG)
      3. Choose a bright calibrator: preferably at least 7 K in the observed band (4 Jy at Q-band, for example)
      4. Allow ~25 minutes for AutoOOf
      5. AutoOOf is not necessary for extended source – if the science is not impacted by the primary beam variations
      6. Solutions are valid for:
         1. If at least 2 hours after sunset, solutions are good for 4 hours (nighttime)
         2. daytime is less predicable and depends on Sun/cloud cover: 30 minutes to 4 hours
         3. periodically examine peak scans for problems (not predicable)
   7. Balancing Strategies (ref 1; section 7.3)
      1. "There are not any set-in-stone" rules" – great.
      2. Balance IF system after performing a configuration
      3. minimize the number of times you balance when observing (oh my)
      4. you may balance when a chance in source equals a change in system_temp + t+source > 3 dB
      5. avoid balancing while making maps
      6. don't balance between signal and reference observations (on/off)
      7. don't balance between observations of targets and calibrators

...