Frequently asked QuestionsΒΆ

  • `I have a spectrum which shows a rise in the flux at the shortest wavelengths.

How do I get rid of that? <upturn_in_blue>`

  • `UVOT photometry has discovered some patches in the image with reduced throughput.

What does that mean for the grism spectra? <small_scale_sensitivity>`

Question:

I have a spectrum which shows a rise in the flux at the shortest wavelengths. How do I get rid of that?

Answer:

The sensitivity becomes low at the shortest wavelengths. This tends to increase the noise in the spectrum which can be mistaken for a rise in flux.

Another possibility is that a weak spectrum from another source falls over the spectrum but not the background. The extra counts from the interloper are also treated as coming from the target. A small constant level of extra counts over the spectrum will become more apparent where the sensitivity is low, since there the factor converting counts to flux is larger.

Sometimes the reason is a bad choice for the background region. Make sure that the contamination of the background region is limited to small patches. The background is derived using sigma clipping and smoothing. The implicit assumption is that the background varies in an approximately linear way.

Question:

UVOT photometry has discovered some patches in the image with reduced throughput. What does that mean for the grism spectra?

Answer:

Although most of the UVOT images show only large scale sensitivity changes, and the in-flight calibration LED illumination of the detector shows no small scale patches of lower sensitivity (within 1%), some patches with lower throughput were found in stray light and source photometry.

As this is a recent development (Autumn 2014), this matter is still under investigation. The cause is not known at this time, but the throughput for a sky source and the on-board LED differ, leading to the conclusion that the reduced throughput patches are not due to the detector.

In the calibration spectra, some White Dwarf spectra show broad dips. This was observed in only a handful of spectra, and will be reported here in the future.

Bottom line is that the patches cover only a very small part of the image area, and most data are not affected.