[Halld-offline] changes to hit generation in FDC and CDC in HDGeant

[Simon Taylor]

Hi.
This is a heads-up for those of you who stay near the bleeding edge of 
software repository activity.

I have made some changes in hitFDC.c and hitCDC.c  to make the 
simulation of the hits in these two detectors more realistic.  Instead 
of generating
a single hit at the distance-of-closest approach to a particular wire as 
a charged particle passes through a given layer, I generate clusters of 
ionized electrons  randomly all along the path of the particle within 
the cell and compute separate drift times for each.  Note that this 
means that the cluster
with the smallest drift time does not necessarily correspond to the 
distance-of-closest approach of the incoming particle's trajectory to 
the wire and that clusters can be sufficiently separated in time to 
cause a second hit (or several more hits) on the wire.  The current 
induced on the wire for each cluster is then convoluted with the
ASIC response and eventually the cable/flash-ADC response will also be 
modeled.   I have parameterized the drift times based on Garfield 
simulations
using the nominal gas mixtures for both the FDC and the CDC.   This 
means that the drift times for the CDC can be as large as ~700 ns if the 
ionization occurs close to the straw wall.  Long drift times of a few 
hundred nanoseconds are also possible for FDC hits.  This impacts the 
minimum
look-back window for the readout to on the order of a microsecond. 

I have modified the Kalman filter to use the new time-to-distance 
relationships.

Simon