Low S/N for bright FLUXSTD

Some of FLUXSTD have low S/N for their PSF flux. Here is a plot for visit=92557.

If I separate the upper/lower half of this broad correlation and plot the (RA,Dec) distribution of FLUXSTD, I got this figure. The red points are the lower half (i.e., FLUXSTDs with low S/N for their brightness). The cobra positioning accuracy was not great at the field edge and I am likely looking at that effect. But, I will look at the raster data to make sure if things are consistent.