Verification of the flux calibration in the visual grism: nominal mode

Overview

The new v-grism flux calibration is valid over the whole detector, and makes also a correction for coincidence-loss. The first integrated software version is UVOTPY-0.9.7.3.

With the choice of data, we have an independent set from the spectra used to construct the effective area. Of course, spectra from the same targets are used.

Method

Calibration spectra that were not used in the flux calibration are reprocessed using the new flux calibration and then compared to the known stellar flux. The spectra for verification are located on the detector between the areas with the spectra used to determine the effective areas for the flux calibration.

../_images/anchor_positions_1000_cal_spectra1.png

The anchor positions of the calibration spectra marked by target. The boxes are enclosing sources used to derive an effective area for the location of the box.

Data used for verification

The approximate location of the anchor of the spectra, their obsid, and number of the fits extension is listed. First the cool stars are listed, followed by the hot white dwarfs.

WD1121+145 spectra

anchor obsid ext % err plot aspect correction
845,1304 00055250072 2 20 WD1121+145 Figure 1 None
912,1256 00055250072 3 20 WD1121+145 Figure 2 None
1015,1280 00055250072 4 15 WD1121+145 Figure 3 None
1080,1230 00055250072 1 15 WD1121+145 Figure 4 None

GD108 spectra

anchor obsid ext % err plot
aspect

correction
631, 989 00055050016 1 20 GD108 Figure 1 yes
924, 924 00055051003 1 20 GD108 Figure 2 yes
949, 920 00055050012 1 20 GD108 Figure 3 yes

GD153 spectra

anchor obsid ext % err plot aspect correction
615,1050 00055501004 1 15 GD153 Figure 1 yes
952,1550 00055502008 1 10 GD153 Figure 2 yes
1019,1563 00055502004 1 20 GD153 Figure 3 yes
1456, 889 00055503020 1 10 GD153 Figure 4 yes

Result of the verification

Errors were read off the figures as being typical. The errors are typically 20% or less. For a large part these are due to inaccuracy in the coincidence loss correction and the Poisson error in the counts.