[ArXiv] Swift and XMM measurement errors, Sep. 8, 2007

From arxiv/astro-ph:0708.1208v1:
The measurement errors in the Swift-UVOT and XMM-OM by N.P.M. Kuin and S.R. Rosen

The probability distribution of photon counts from the Optical Monitor on XMM Newton satellite (XMM-OM) and the UVOT on the Swift satellite follows a binomial distribution due to detector characteristics. Incident count rate was derived as a function of the measured count rate, which was shown to follow a binomial distribution.

Discrepancy between the mapping the 1σ Poisson incident count error onto measured count error and the mapping the 1σ binomial measurement count error onto incident counts was illustrated with an example of large counts per frame. Although this discrepancy for small counts per frame was commented to be negligible, the authors urge to use their binomial distribution based formalism to derive the errors in measurements.

One Comment
  1. vlk:

    That is an interesting approach to pileup! (Or so I gather from a quick reading.) Rather than try to model the effect of pileup (which is what happens when more than one event is detected within a single frame, thus collapsing the multiple events into a single one), they have taken the opposite route, of simply setting the error bars to be larger.

    Note also that this just modifies the error bar on the counts, and not on the energies, which is what Yaming has wrestled with in the context of ACIS data.

    09-11-2007, 9:26 am
Leave a comment