Problem 374

Summary: incorrect nucleon counting, momentum checking and angular distribution in TwoCluster
Product: Geant4 Reporter: dennis.herbert.wright
Component: processes/hadronicAssignee: dennis.herbert.wright
Status: RESOLVED LATER    
Severity: normal    
Priority: P3    
Version: 4.0   
Hardware: All   
OS: All   

Description dennis.herbert.wright 2002-05-17 19:17:53 CEST 
In G4ReactionDynamics::TwoCluster the neutrons and protons need to be accounted
for separately in order to get charge and energy conservation.  The same is true
in the AddBlackTrackParticles method which only accounts for average nucleon
mass.

The MomentumCheck routine should be removed since it is possible for a particle
to have more momentum after a reaction than it did before the reaction.  It
looks like this routine was put in to correct an earlier error in the momentum
sampling.

The angular distribution in TwoCluster has a large backward peak which is not
right.  Currently cos(theta) is sampled from t = (1/b)log(1 - random).  It
should be sampled from t = (1/b)log(1 - factor*random) where factor is
1 - exp(b*(tacmin-dumnve)).  This guarantees a proper behavior around 180
degrees.
Comment 1 Hans-Peter.Wellisch 2002-06-04 04:17:59 CEST 
ok. Part 2 will be fast, part 1 will come once
the oszillation stuff is ready.
Comment 2 Hans-Peter.Wellisch 2002-06-17 04:08:59 CEST 
it turns out that twocluster has rather more to fix than you mentionned.
This will become an active thread for some time, and work has started.

Tank you for getting me onto this one.

Many greetings,

Hans-Peter.
Comment 3 dennis.herbert.wright 2007-01-01 15:21:59 CET 
Parts one and two have been fixed.   Part three (angular distribution) will not
be fixed because it would upset the parameterization of the model.   When a new
parameterized model is developed, this will be done correctly.