Abstract -- Hits in the SMD between AgSTAR and AgML are slightly inconsistent, showing slightly more (or less) energy deposited on an event-by-event basis.  We have asserted in the past that the effect is consistent with a < 1mm shift in the SMD planes between the geometry decriptions.  In the attached PDF we show a comparison of two identical AgSTAR geometries, with the SMD plane shifted by a known amount, which is varied between 0.03mm and 1.0mm.  We find that the discrepancy between AgSTAR and AgML is consistent with a shift in the SMD plane of less than 0.1mm.

The problem appears to be a bit more serious than originally thought.  There was a bug in the function which converts pseudorapidity to angle, resulting in a systematic (and large) error in the placement of SMD strips.  This function was used to calculate two geometry constants: the min and max angles of the EEMC fiducial volume, so it was easy to eliminate the function and replace with the correct angles.  Once this was done we get the expected correlation of hits between the EEMC and ESMD in AgML and AgSTAR.