I've looked at two images of the IO beam dump steering mirror from Oct 2012 and Nov 2016.  Comparing the two images suggests the beam may have moved enough to be clipping on the edge of the steering mirror, which would cause significant scattering.  The image from Nov 2016 does not show the entire steering mirror, so there is some non-trivial potential for error in identifying the location of the beam on the mirror.  My estimate for that error is around 25%, so if the beam is 25% closer to location of Oct 2012 the risk of clipping is small, but if the beam is 25% farther away from the location of Oct 2012, the risk of clipping is high.  Attached: image of beam dump and sterring mirror from Nov 2016, analysis of images (2.4MB)