# Bugs: srcflux

There is an error in the computation of the model independent NET_FLUX_APER and the NET_PHOTFLUX_APER values reported in ${root}.flux output file when using psfmethod=file or arfcorr or marx. The fraction of the source flux in the background region is getting incorrectly scaled by the ratio of the area of source-to-background. Consider the default case where the background area to source area is approximately r=24 (background annulus outer radius is 5 times the source radius) and has typical PSF fractions of a=90% in the source region, and b=9% in the background region. The error in the fluxes will be the ratio of $$\frac{(r*a-b)}{(r*(a-b))} = 1.106$$, so a a 10.6% over-estimate of the photon flux. The same error is propagated to the LO and HI confidence interval limits for both NET_FLUX_APER and NET_PHOTFLUX_APER values. The actual values for r,a,b can be obtained from the columns in the${root}.flux file: r = BG_AREA/AREA, a = PSFFRAC, and b = BG_PSFFRAC.