SC mcsmear resolutions

The measured Start Counter timing resolutions that are obtained from ST_Tresolution (for example) are currently stored in the CCDB table START_COUNTER/time_resol_paddle, and are used as parameters in the reconstruction software and mcsmear. However, these resolutions are actually due to a convolution of the start counter and tracking resolutions. The "true" start counter resolution can in principle be obtained from MC simulations.

The following plots show the resulting output resolutions as a function of "true" input resolution.

HDGeant	HDGeant4

The results are fit with a linear function, output = p0 + p1 x input

HDGeant:   p0 = 49.85 +- 3.40   p1 = 0.906 +- 0.014
HDGeant4:  p0 = 63.65 +- 3.75   p1 = 0.882 +- 0.016

The results from both versions of HDGeant are fairly similar, however HDGeant4 is still in alpha/beta mode, so I propose to correct the start counter resolutions obtained from data using the inverse relation true = (measured - p0) / p1 using the HDGeant parameter values.

--- July 5, 2017, sdobbs