The past simulation result cannot give us the accurate source position. One of the reason is that the mask in the Geant4 geometry was at neither the nominal position nor the as-built position. Therefore, we firstly shifted the mask aperture to the nominal position, that is, the top of mask tile is at BAT_Z = 100.3 [cm].
We know that the mask offset
MASKOFFX= -0.109 / [cm] Offset of mask in BAT_X (BAT Cal Memo 2003-06-03) MASKOFFY= +0.078 / [cm] Offset of mask in BAT_Y MASKOFFZ= +0.105 / [cm] Offset of mask in BAT_Z
were derived from actual calibration data by Craig. But, even if we take this offset into account, we could not obtain the accurate source position from Geant4 simulation.
From the conversation with Craig and Derek, there are at least two reasons which make us adjust the maskoff[xyz] keywords. One is from the actual arrignment of the mask and the detector plane. The other is from the possible differences of position definition in the data analysis code.
Therefore, we concluded that many effects are included in these values, and they do not necessarily represent the actual offsets of the mask position. Then, we need to determine Geant4 simulation's own mask offset values by simulating test runs.
Craig suggested to make three runs,
With these three runs, we have enough information to estimate all three mask offset parameters, including a cross-check. The two 45 degree runs make sure that the focal length is right.
|
IMXcent IMYcent IMXwid IMYwid Peak Cts Bkg Cts Bkg Var -0.00104 0.00308 0.00279 0.00279 280270.5 -207.6 1876.4
  |
IMXcent IMYcent IMXwid IMYwid Peak Cts Bkg Cts Bkg Var 0.99811 0.00311 0.00282 0.00282 118646.3 -76.3 1128.1
 |
From these results, we succeeded to derive the mask offset values as follows:
maskoffx = +0.104 [cm] maskoffy = -0.308 [cm] maskoffz = -0.085 [cm]