HandsOn Calibration

1
Hands-On Calibration

• Ron Maddalena
• July 13, 2007

2
Preliminaries

• Change directory cd /home/scratch/sdscal
• Start GBTIDL
• Access data
• filein,T_TCAL14MAR07.acs.raw.fits
• summary
• Note For 8 GHz receiver, used observing
  technique of OffOn with, for example, scan 6 an
  observation toward blank sky and scan 7 toward
  3C286. Used 4 windows
• Note For 12 GHz, dual feed receiver, used
  observing technique of NOD. Source in feed 1 for
  scan 31, in feed 2 for scan 32. Used 2 windows.
• getps,6,ifnum0
• Try different windows to see whats different
• header
• Record elevation, UT date and time
• getnod,42,ifnum0
• For the adventurous
• .compile getscalquad.pro

3
Typical Position-Switched Calibration Equation
4
Putting it all together
Remove AveragingSolve for Tcal
5
What Do We Need?

• ? from graph, assume gain is elevation
  independent
• Ap from dish diameter
• Calculate Air Mass from elevation of observation
• S from a catalog (e.g., Ott et al 1994, AA, 284,
  331)
• Table pp 333-334
• Functional fit p. 335
• Note that S will vary significantly across wide
  bandwidths
• t from weather models
• At Linux prompt, type cleo forecasts

6
What Do We Need?

• t from weather models
• At Linux prompt, type cleo forecasts
• Select Curves tab
• Enter date and UT of the observations
• Enter frequency range for the receiver (e.g.,
  7-11 GHz, 11-16 GHz)
• May want to select Write Out Results to create
  an ASCII file of results
• Click on Process
• Read opacities off of graph

7
What again are we calculating?
8
How to Calculate (RefOn-RefOff)/(Sig-Ref)

• Use the commands emptystack, select
  avgstack, copy, subtract, divide, and
  scale
• Emptystack
• Clears anything that peviously has been done with
  the stack
• Select,scan6,calF,ifnum0,plnum0,fdnum0
• Finds all data that meet this selection criteria
• Avgstack
• Averages together the data found by Select and
  places into Data Container (DC) zero
• Copy,0,9
• Moves the results to another DC for later use
• DC 9 will now contain RefOff
• Repeat for scan6, calT, place into DC 8 to
  create RefOn
• Repeat for scan7, calF, place into DC 7 to
  create SigOff
• Repeat for scan7, calT, place into DC 6 to
  create SigOn

9
How to Calculate (RefOn-RefOff)/(Sig-Ref)

• Summary
• DC 9 contains RefOff
• DC 8 contains RefOn
• DC 7 contains SigOff
• DC 6 contains SigOn
• Create Sig and place into DC 10
• Add,7,6,10
• Scale,0.5,10
• Similarly create Ref and place into DC 11
• Create Sig-Ref and place into DC 12
• Subtract,10,11,12
• Similarly create RefOn-RefOff and place into DC
  13
• Create (RefOn-RefOff)/(Sig-Ref) and place into DC
  14
• Divide,13,12,14
• Show,14

10
What again are we calculating?

• Finally, scale DC 14 by ?, S, to determine
  Tcal. For example, using fictitious values
• scale, 0.51234/222, 14
• scale, 1/1.38e-16exp(-0.12/sin(33180/!pi)), 14
• etc.
• show,14

11
Check for non-linearity

• If system is linear, than
• (SigOn-SigOff) (RefOn-RefOff ) 0
• Model the response curve to 2nd order
• Pout B Pin C Pin2
• Our observations provide
• Pout Refoff when PinTsys
• Pout Refon when PinTsysTcal
• Pout Sigoff when PinTsysTA
• Pout Sigon when PinTsysTATcal
• Its easy to show that
• C (Sigon- Sigoff )-(Refon- Refoff)/(2TATcal)
• Try to estimate a value for C using subtract,
  divide and scale
• Things are really a bit more complicated since we
  really measure Pout and want to determine Pin .
  Must invert 4 simultaneous linear equations.

12
Now for the real easy way

• Getscal,7,6,ifnum0,plnum0,fdnum0,tau0.05,
  ap_eff0.55,smth1
• Show,13 (Tcal, assuming linearity)
• Show,3 (Tcal, assuming non-linearity)
• Show,15 (Tsys, assuming linearity)
• Show,5 (Tsys, assuming non-linearity)
• Show,11 (Source flux)