HETG Orbital Activation & Checkout

back to HETG on Orbit

Contents

• Overview
• HETG OAC Goals
• Notes & To-do
• OAC Timeline for HETG
• Simulations
  • Summary
  • Capella Simulated Observations
  • Capella Spectral Model
• Defocus Analysis
  • Goals and Results
  • Functional Form
• Contacts

Overview

HETG OAC Goals

What are the goals of the data taken during insertion/retraction? Currently they are:

• Verify HETG is inserted: see the characteristic "X" pattern
• Measure HETG parameters and performance:
  • dispersion axes angles on the detector
  • dispersion scale w.r.t. "known" lines
  • zero-order image size vs non-grating size
  • spectral (dispersed orders) focus characteristics (+/- orders, plate focus set) and detector alignment (angle about Z axis).
  • cross-dispersion profile
  • crude efficiency check w.r.t. S3 non-grating observation
  • efficiency vs. energy check with continuum source (Mkn421 or 3C273)
• Verify that the HETG is retracted: don't see "X"
  
• Excersize, debug, verify CXC software including Monitoring and Trends Analysis "gratmon" output.

Notes & To-do

• S/C contacts will be 4 hours ON / 4 hours OFF during this phase. "4 hours" = 14.4 ks.
• Launch-lock release must be performed during ground contact. How long will it take if everything works as expected (including human interfaces and overhead)?
• HETG insertion or retraction takes 3 minutes.
  • Will the first insertion be performed during ground contact? YES. Will there be much human overhead in addition to the 3 minutes? How nervous will people be? Will shouts of "Wait! Do imaging science before putting in the gratings!" be ignored?
• How to verify that the HETG is fully retracted
  • Angle encoders available? YES.
• Any need/desire to measure/verify insertion repeatability? Use the HR1099 observation (and assumptions about Capella vs HR1099 line energies) to check?

OAC Timeline for HETG

See the CXC In-flight Calibration page and the Phase 2 timeline Version 5.0 for specifics of "Phase 2 OAC and Calibration".

The first HETG observations are summarized here:

HETG OAC Observation Simulations

Summary

Based on Herman Marshall's Capella studies (see his colorful In Flight Calibration Observations web page) the HETGS focus test will consist of three 30 ks exposures of Capella taken at 0.0, +0.2 mm, and -0.2 mm with respect to the then-nominal ACIS-S focus position. Simulations and analyses of these specific observations have been carried out to produce flight-realistic results. Three sets of results are created coresponding to the condition that the "then-nominal" focus is actually at a value of 0.0, +0.3 mm, or -0.3 mm with respect to the true focus. The plots in the defocus analysis section , below, show that these cases are clearly identified by the in-hand data.

Simulated data set that would be measured if the then-nominal focus coincides with the real focus location.	Simulated data set that would be measured if, unfortunately, the then-nominal focus actually coincides with +0.3 mm from the real focus.

Capella Simulated Observations

The observations are simulated using MARX 2.20 and include dither. The procedure flight_sim.pro sets up the MARX parameter file, etc. and spawns the simulation. Parameters of the observation are read from the simple "database" file: obsdb.rdb. The observations are then analyzed by the general purpose procedure obs_anal.pro. "Auto-aspect reconstruction" is performed on the simulations by obs_anal.pro using the motion of the zero-order image. This technique produces overall E/dE values in good agreement with expectations and may actually be the analysis method for these first-light images. Due to "apodization" of the time series in the aspect determination algorithm, these 30 ks exposures produce only 20 ks of valid aspect corrected events - this efficiency can likely be improved.

Observation summaries of the simulations including links to line-lists are given here for the 3 simulated sets of three exposures ('3.001 - '3.009 in obsdb.rdb):

• Nominal focus = Real focus
  • Capella-3.001
  • Capella-3.002
  • Capella-3.003
• Nominal focus = Real focus + 0.3 mm
  • Capella-3.004
  • Capella-3.005
  • Capella-3.006
• Nominal focus = Real focus - 0.3 mm
  • Capella-3.007
  • Capella-3.008
  • Capella-3.009

Capella Spectral Model

MARX input spectra representing Capella and HR1099 (a plausible alternative line source) were created by Dave H (nH effects are not included currently.) The desired ("real") model flux is based on ROSAT measurements. The spectra given are not properly normalized; the MARX flux value given yields the desired inband flux, properly normalizing the spectrum. Plots, including flux calculation, were made by plot_marx_spect.pro.

Summary properties of Capella models and BCS data are presented in the table below; this comparison suggests that the simulated fluxes are accurate to within a factor of 1.5 .

[1] Vedder and Canizares, ApJ 270:666-670, 1983 July 15
[2]  XSPEC model vmekal+vmekal with parameters from Nancy Brickhouse 

Defocus Analysis

Goals and Results

The goals of the defocus analysis are:

• to determine (and/or confirm) the overall best focus for the HETG, i.e., narrowest LRF.
• to measure the ACIS-S tilt - that is an offset in HETG best focus that is linear with a line's position away from the zero-order, most dramatically a difference between plus and minus order line widths.
• to confirm the Observatory focus stability since last plate focus (using zero-order data)
• to incidentally further verify the HRMA model by separating the mirror shell pairs (using focus data for the HEG and MEG diffracted orders)

The Capella observations and routine analysis ( simulations, above) produce detected line lists of FWHM values for Gaussian fits to the spectral lines. There is a wealth of data here specified by: the line energy, HEG or MEG, plus or minus 1st order, and the defocus condition (0.0, +0.2, -0.2). The zero-order FWHM vs defocus is also measured. The results here are shown for these specific measured widths:

• Zero-order: one measurement (all energies, MEG+HEG)
• MEG: 0.653 keV and 0.826 keV lines, m= +/- 1
• HEG: 0.82 keV and 1.022 keV lines, m= +/- 1

The plots below where produced by oac_anal.pro and carry out a simple quadratic fit of FWHM^2 vs defocus, see Functional Form below. The FWHM values in a defocus set were normalized by the minimum measured FWHM in order to keep the plots on the same scale; other presentation methods will be investigated. The three plots correspond to the case where the actual nominal focus is 0.0, +0.3 mm and -0.3 mm with respect to the real HRMA best focus. The data allow these cases to be distinquished.

• Capella-3.001-3.003, Nominal focus = Real focus :

  

  
  
• Capella-3.001-3.003, Nominal focus = Real focus + 0.3 mm :

  

  
  
• Capella-3.001-3.003, Nominal focus = Real focus - 0.3 mm :

  

  
  

Plots covering a wide range of defocus are presented below to show the expected global variation of width vs focus. Previous MARX 2.11+ and 2.04 simulations are included to show effects of variation in the mirror model. Note that it is unlikely that a full range set like these will actually be obtained in flight!

• Capella-3.001, etc., MARX 2.20

  The MARX 2.20 distribution was used to carry out these flight simulations as described in simulations, above. Aspect and ACIS-S pixel blur were accurately included through "auto-aspect solution" carried out by obs_anal.pro. This is a composite of the three cases presented above.

• Capella-3.001, etc., MARX 2.11+

  The MARX 2.11+ development version has corrected HRMA tilts and I've used Mike Wise's new scattering files - together they give close agreement to XRCF data and SAOSAC-flight. The Gaussian sigma is set to 0.262 arc sec 2D rms.

• Capella-3.001, etc., MARX 2.04

  The MARX 2.04 distribution parameters were used to carry out the flight simulations. This distribution uses HRMA scattering files that produce too much scatter and slightly incorrect mirror shell tilts. Aspect and ACIS-S pixel blur were approximated in MARX with a Gaussian source with sigma = 0.220 arc sec 2D rms.

Functional Form for FWHM vs Defocus

The width of a feature, FWHM, as a function of defocus, dX, can be approximated as:

      FWHM   =   SQRT{ (w_const)^2  +  ( beta(dX-Xo) )^2 }
with
           w_const = FWHM at best focus
              beta = FWHM per defocus distance
                Xo = axial location of best focus

      FWHM^2 =  w_const^2  + beta^2 (dX-Xo)^2

             =  (w_const^2 + beta^2 Xo^2)    1
                            - 2 beta^2 Xo   dX  
                                 + beta^2   dX^2

         FWHM^2  =  a + b dX + c dX^2
and so
         Xo = -b/(2c)

Error estimate and counts needed

The measurement of the FWHM has a fractional error of order 1./SQRT(number-of-counts-in-line-order); the error on FWHM^2 is twice this. Looking at the MARX simulation plots, the zero-order or MEG curves have a slope (at dX ~ 0.4mm) of order :

  dFWHM^2 / FWHM^2   |
  -----------------  |                          ~  200%/mm   
     defocus(mm)     | at dX ~ 0.4 mm,
                         where (FWHM/FWHM(min))^2 ~ 2

OAC Contacts

This web page is: http://space.mit.edu/HETG/flight/oac.html.

Please send any comments and updates to Dan Dewey at dd@space.mit.edu.