Calibration Project Charter

Project Manager: ToneyMinter
Project Scientist: JimBraatz
Project Tracking Number: This should be obtained by GBT program manager


Project background

Currently, GBTIDL and sdfits use a scalar value for $T_{cal}$ and subsequently for $T_{sys}$. This introduces baseline structures into the data. It also causes systematic errors in the calibration. At band edges this can keep multiple spectral windows placed side-by-side from agreeing in their flux level. These problems can be greatly reduced by the use a "vector" $T_{cal}(f)$, $T_{sys}(f)$, etc.

Observations looking for weak and/or broad spectral lines are affected by the residual baselines left by the current standard, $T_{cal}$ calibration. These observations include high-redshifted CO observations, H 2p-2s fine structure line (both Galactic and red-shifted Epoch of Re-ionization), and the search for positronium. When several spectral windows are setup to be contiguous in freuqency, the resulting data do not line up correctly in their fluxes using scalar calibrations and they should line up once vector calibrations are used. Also affected are the cases where the source has a significant amount of continuum emission.

Additionally, GBTIDL calibration procedures currently use default opacities representative of good weather, but fixed at a constant for any given frequency. These defaults can introduce an error in the flux density scale. The problem could be alleviated by having the GBTIDL calibration functions call on the cleo weather database to retrieve an opacity more representative of the actual weather conditions at the time of the osbervation. A procedure to reduce GBT tipping scans would also help the observer calibrate flux density more accurately.

Project Justification

It is known that $T_{cal}$ values have significant frequency structure. The current calibration scheme uses scalar values for $T_{cal}$ and $T_{sys}$:

$T(f) = \left(S_{sig}(f)-S_{ref}(f) \over S_{ref}(f)\right) * T_{sys}$


$T_{sys} = \left(S_{calon}(f)+S_{caloff}(f) \over S_{calon}(f)-S_{caloff}(f)\right)*T_{cal} - \left(1\over 2\right)*T_{cal}$.

As one can easily see, even if the observed $S_{sig}(f)$ signal spectrum as a function of frequency is offset from the observed $S_{ref}(f)$ reference spectrum as a function of freuqency by a constant value, the structure of $T_{cal}(f)$ will not be taken into account and baseline structure will be artificially introduced into the resulting $T(f)$ spectrum.

By switching to a frequency dependent $T_{cal}(f)$ and $T_{sys}(f)$ calibration scheme, the artificially introduced baseline structure can be greatly reduced. Also, observational data, combined with $S_{cal}$ observations and vector $T_{Cal}(f)$, will lead to significant improvements in the calibration of GBT data. If this is not done then the scientific capabilities of the GBT will remain hindered to observations of weak spectral lines, broad spectral lines and spectroscopic observations of sources with strong continuum emission.

Several cases where vector $T_{Cal}(f)$ will improve the calibration are:

  • For observations where the source has strong continuum emission, the induced baseline structure from using a scalar $T_{Cal}$ can inhibit the detection of broad spectral line. The use of a vector $T_{Cal}(f)$ will improve our ability to detect weak, broad lines.
  • At high frequencies and in marginal weather when there can be a $T_{sys}$ change between an "on" and an "off", scalar $T_{Cal}$ will introduce baseline structure. The use of a vector $T_{Cal}(f)$ will improve high frequency observing in marginal weather, which will open up more time for high frequency observing.
  • $T_{Cal}(f)$ easily can vary by 50% in some receivers. Any spectral line that is not centered exactly in the middle of the bandpass could have its calibration be in error by up to 50% as a result of using a scalar $T_{Cal}$. This is an unacceptably large amount of error in the calibration, and the observer may not even be aware of its existence.
  • When using a spectral line backend for continuum observations you get the wrong flux and spectral index if you use a scalar $T_{Cal}$. The use of a vector $T_{Cal}(f)$ will give more reasonable results.

Overview of Deliverables

  • A design document and MRs for the calibration database.
  • A calibration database file that accompanies the sdfits file for the project.
  • New GBTIDL get(ps,nod,fs, etc.) routines that use vector Tcal, Tsys, etc. within the calibration.
  • A new GBTIDL routine to reduce Scal observations and to place the derived Tcal values into the calibration file.
  • GBTIDL routines to get/put data from the calibration database file.
  • GBTIDL documentation updated.
  • An interface with the Cleo Weather Predictions to allow "default" opacity corrections.
  • A GBTIDL routine to reduce tipping scan data to provide a "measured" opacity.

Specific Project Objectives & Success Criteria

  • Success will result in:
    • The ability to use vector calibration with position switched, nodding, or frequency switching data.
    • The ability to reduce Scal observations and to add the results to the calibration database.
    • sdfits writing a calibration database file with the current fits file.
    • Ability to use opacities determined from CLEO Weather Prediction program in calibration.
    • Ability to reduce tipping scan data.
    • Updated GBTIDL documentation
    • Release of the new calibration routines within GBTIDL

Primary Stakeholders & Roles

  • Observers: the customer
  • Toney Minter: Project Manager and programing and testing
  • Jim Braatz: Project Scientist and programing and testing
  • Bob Garwood and/or Paul Marganian: sdfits and calibration database file programing
  • Ron Maddalena: algorithm development, CLEO interface and guidance
  • Karen O'Neil

Key Assumptions

  • Creation of the MR for the format and access to the calibration database will be done by T.M. and J.B.
  • No more than 3 FTE weeks will be needed to program the calibration database and sdfits
  • All work in GBTIDL will be done "at a high level" and by T.M. and J.B.
  • If the calibration database and sdfits changes cannot be made in three weeks, then a work-around of copying the receiver fits file as the calibration database file will be used (minor sdfits change) along with keeping the calibration data in a global variable within GBTIDL. The cost will be that you will have to redo any Scal, etc. determinations each time you (re)start GBTIDL. The GBTIDL code needed will not depend on this other than the new get/put functions for the calibration database.
  • T.M. will have to spend some time learning the GBTIDL code.

  • Out Of Scope
    • Calibration corrections due to wind moving the telescope.
    • Calibration corrections from the servo system errors.
    • Correction for other weather factors besides the opacity.

Any additional information needed

  • Staffing Requirements:
    • Toney Minter: 25%
    • Jim Braatz: 25%
    • Bob Garwood and/or Paul Marganian: <= 3 FTE weeks
    • Ron Maddalena: <10%
  • Communications
    • meet with Scientific Staff at beginning of project to present plan and ask for comments and ideas
    • Bi-weekly meeting of all involved staff
    • Email and phone conversation between T.M. and J.B. as needed.
  • Preliminary Schedule:
    • PHASE I
      • mid-November 20007: WBS created
      • end November 2007: meet with Scientific Staff, update plan
      • end of December 2007: MRs created
    • PHASE II
      • January 2008 : programing begins
      • end of January 2008: calibration database programing changes done in both GBTIDL and SDFITS
      • end of February 2008: GBTIDL programing done, testing begins
      • end of March 2008 : work complete
  • Prior Documentation That Is Relavent


The following people agree that the above information is accurate:
Preliminary project team members:
DONE ToneyMinter
DONE JimBraatz
DONE BobGarwood
DONE RonMaddalena
DONE PaulMarganian
GB Program manager:
DONE KarenONeil

Use DONE to sign.

-- ToneyMinter - 07 Nov 2007

Error during latex2img:
ERROR: can't find latex at /usr/bin/latex
\begin{math}\displaystyle $S_{sig}(f)$\end{math}
\begin{math}\displaystyle $T(f) = \left(S_{sig}(f)-S_{ref}(f) \over S_{ref}(f)\right) * T_{sys}$\end{math}
\begin{math}\displaystyle $T_{Cal}$\end{math}
\begin{math}\displaystyle $T_{sys}$\end{math}
\begin{math}\displaystyle $T_{Cal}(f)$\end{math}
\begin{math}\displaystyle $T_{cal}$\end{math}
\begin{math}\displaystyle $T_{sys}(f)$\end{math}
\begin{math}\displaystyle $T(f)$\end{math}
\begin{math}\displaystyle $T_{sys} = \left(S_{calon}(f)+S_{caloff}(f) \over S_{calon}(f)-S_{caloff}(f)\right)*T_{cal} - \left(1\over 2\right)*T_{cal}$\end{math}
\begin{math}\displaystyle $T_{cal}(f)$\end{math}
\begin{math}\displaystyle $S_{ref}(f)$\end{math}
\begin{math}\displaystyle $S_{cal}$\end{math}

This topic: Main > TWikiUsers > ToneyMinter > CalibrationProjectCharter
Topic revision: 2009-10-15, ChrisClark
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback