George J Bendo and the SPIREICC Basic Equations Derivative of flux calibration curves from empirical analysis Flux calibration curve integral of above equation V 0 is a zeropoint voltage that is selected to match the background in dark sky ID: 604017
Download Presentation The PPT/PDF document "SPIRE Flux Calibration: Implementation" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
SPIRE Flux Calibration: Implementation
George J. Bendo
and the SPIRE-ICCSlide2
Basic Equations
Derivative of flux calibration curves (from empirical analysis):
Flux calibration curve (integral of above equation):
V
0
is a zero-point voltage that is selected to match the background in dark sky.
K1, K
2
, and K
3
are unique parameters for each bolometer that are derived using the techniques discussed here.
An additional K
4
term is used to scale the RSRF-weighted flux density to produce monochromatic flux densities (assuming υF
υ
is constant).Slide3
Calibration Strategy
PCal
flash observations taken against backgrounds with different surface
brightnesses
give the relation between V and 1/ΔV. These data can then be used with the derivative of the flux calibration curve to give unscaled versions of K1 and K
2 and a scaled value of K3.Fine scan observations in which each bolometer scans over Neptune are used to scale the K1
and K2 parameters.Slide4
PCal Flash Observations
Instrument was pointed at a series of locations near
Sgr
A* during
PCal flash observations to get 1/ΔV vs V measurements. Each bolometer falls within a region that should be ~75% of the peak surface brightness.
Additional PCal flash measurements against dark sky were used to constrain the part of the curve near V0.Slide5
Results: Unscaled Derivatives of Calibration Curves
Dotted lines:
Voltages
for V
0
and V0-VPeak (Neptune)Slide6
Results: Unscaled Derivatives of Calibration Curves
Dotted lines:
Voltages
for V
0
and V0-VPeak (Neptune)Slide7
Results: Unscaled Derivatives of Calibration Curves
Dotted lines:
Voltages
for V
0
and V0-VPeak (Neptune)Slide8
Neptune Fine Scan Observations
In these observations, each bolometer in SPIRE passed across Neptune in a very fine pattern. This gave measurements of the peak and background voltages that could be used to scale the calibration curves.
To derive the peak and background voltages, we fit the timeline data (not the map data) with two dimensional Gaussian functions. An example is shown below.Slide9
Results: Scaled Calibration Curves
Solid line:
V3-0
flux calibration
Dotted line: V2-3 flux calibrationSlide10
Results: Scaled Calibration Curves
Solid line:
V3-0
flux calibration
Dotted line: V2-3 flux calibrationSlide11
Results: Scaled Calibration Curves
Solid line:
V3-0
flux calibration
Dotted line: V2-3 flux calibrationSlide12
Problematic Bolometers
Some of the dead, noisy, and slow bolometers could not be calibrated because of difficulties with performing analysis on the sample. The following number of bolometers were not calibrated:
PSW 7
PMW 1
PLW 1 16 bolometers in PSW and 7 bolometers in PMW saturated on Neptune. To calibrated these bolometers, we used fine scan observations of 3C 273. We calibrated the signal from 3C 273 for the “good” bolometers, and then used the signal from 3C 273 to scale the bolometers that saturated on Neptune.Slide13
Sources of Uncertainty(Individual Bolometers)
Uncertainty from fits to
PCal
flash data.
Uncertainty in scaling terms.Uncertainty in model flux densities of calibration source.Slide14
Uncertainty from PCal Fits
Because of degeneracy problems with the nonlinear equation fit to the
PCal
flash data, we cannot derive simple uncertainties in the K parameters from the fit.
Instead, we used a Monte Carlo approach to create plots that show the uncertainty in the calibration curves as a function of voltage.
The typical fractional uncertainties from the PCal flash fits derived this way are ~0.0002. The maximum uncertainties do not supercede 0.005. Slide15
Uncertainty from PCal Fits
Dotted lines:
Voltages
for V
0
and V0-VPeak (Neptune)Slide16
Uncertainty from PCal Fits
Dotted lines:
Voltages
for V
0 and V0-VPeak (Neptune)Slide17
Uncertainty from PCal Fits
Dotted lines:
Voltages
for V
0 and V0-VPeak (Neptune)Slide18
Uncertainty from Scaling Terms
The table below lists the uncertainties for the scaling of the calibration curves for individual bolometers based on the uncertainty in the peak voltage measurements of Neptune or 3C 273.
The maximum fractional uncertainties in the PSW and PMW arrays are for bolometers that saturated on Neptune. Most other bolometers have uncertainties much closer to the mean.
Array
Mean
Fractional Uncertainty
Maximum Fractional
Uncertainty
PSW
0.0098
0.11
PMW
0.010
0.12
PLW
0.0016
0.040Slide19
Tests of Flux Calibration
The new Flux Calibration Product was used to process standard large and small scan map data for two sources:
Neptune (primary calibrator)
Gamma
Dra (secondary calibrator)The results can be used to gauge the random uncertainty in the flux calibration as well as variations in the flux calibration over time.
Measurements of the flux densities were performed on the timeline data. Mapping the data increases the dispersion in the flux density measurements, as binning the data into map pixels will effectively blur the data. The source extraction tools currently included in HIPE (DAOPHOT and Sussextractor
) both systematically undermeasure flux densities, and Sussextractor also functions very poorly for >100
mJy
sources.Slide20
Neptune Measured/Model Flux Densities
Triangles: Small scan map
Squares: Large scan mapSlide21
Neptune Measured/Model Flux Densities
Triangles: Small scan map
Squares: Large scan mapSlide22
Neptune Measured/Model Flux Densities
Triangles: Small scan map
Squares: Large scan mapSlide23
Neptune Measured/Model Flux Densities
Array
Measured/Model
Flux Density Ratios
All
Large
Map
Small Map
PSW
0.995 +/- 0.007
0.996
+/- 0.009
0.994 +/- 0.003
PMW
0.993
+/-
0.010
0.997
+/-
0.013
0.992
+/- 0.004
PLW
1.001
+/-
0.003
1.000
+/-
0.003
1.002
+/- 0.003Slide24
Gamma Dra Measured Flux Densities
Triangles: Small scan map
Squares: Large scan mapSlide25
Gamma Dra Measured Flux Densities
Triangles: Small scan map
Squares: Large scan mapSlide26
Gamma Dra Measured Flux Densities
Triangles: Small scan map
Squares: Large scan mapSlide27
Gamma Dra Measured Flux Densities
Array
Model Flux Density (
Jy
)
Measured Flux Densities (
mJy)
All
Large
Map
Small Map
PSW
251
258 +/- 3
260 +/- 2
258 +/- 4
PMW
127
139 +/- 4
140 +/- 5
138 +/- 4
PLW
61
74 +/- 5
75 +/- 5
73 +/- 5
These are monochromatic flux densities without color corrections (and without a
u
in color).Slide28
Conclusions
Although a few individual bolometers have calibration uncertainties of >5%, each array as a whole can measure flux densities with the following accuracies:
PSW 1.5%
PMW 1.7%
PLW 0.5%The accuracy of the flux calibration for >100 mJy sources will primarily be limited by the Neptune models.