In writing a km3 scale simulation for the antares detector I've tried to enter some properties of water for optical photons other than the ones used in ExN06 (I did temporalily use the ones in ExN06 and they worked just fine). Changing absorbtion length and energy of the photons in the tables PPCKOV[NUMENTRIES] and ABSORPTION1[NUMENTRIES] didn't give any problems (the changes were rather small by the way). However when I changed the RINDEX1[NUMENTRIES] from G4double RINDEX1[NUMENTRIES] = { 1.33, 1.33, 1.33, 1.33, 1.33, 1.33, 1.33, 1.33, 1.33, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.34, 1.35, 1.35, 1.35, 1.35, 1.35, 1.35, 1.35, 1.35, 1.35 }; to G4double RINDEX1[NUMENTRIES] = { 1.3435, 1.344, 1.3445, 1.345, 1.3455, 1.346, 1.3465, 1.347, 1.3475, 1.348, 1.3485, 1.3492, 1.35, 1.3505, 1.351, 1.3518, 1.3522, 1.3530, 1.3535, 1.354, 1.3545, 1.355, 1.3555, 1.356, 1.3568, 1.3572, 1.358, 1.3585, 1.359, 1.3595, 1.36, 1.3608}; and ran a high energy muon (10 GeV) through my km3 scale detector consisting out of pure H2O and within that the Optical Modules (placed according to the 1999 proposal) which also consisted out of pure H2O I got the following error: > !!! the global time goes back !!! : 48.6022 -> -1.79432e+07[ns] > !!! the poper time goes back !!! : 0.00460323 -> -1.21661e+07[ns] > G4ParticleChange::CheckIt > pointer : 0x8667550 > ----------------------------------------------- > G4ParticleChange Information > ----------------------------------------------- > # of 2ndaries : 0 > ----------------------------------------------- > Energy Deposit (MeV): 0 > Track Status : Alive > True Path Length (mm) : -35.3 > Stepping Control : 0 > Event Biasing : not used > Position - x (mm) : -4.9e+08 > Position - y (mm) : -3.61e+09 > Position - z (mm) : 1.54e+09 > Time (ns) : -1.79e+07 > Proper Time (ns) : -1.22e+07 > Momentum Direct - x : 0.124 > Momentum Direct - y : 0.912 > Momentum Direct - z : -0.39 > Kinetic Energy (MeV): 0.243 > Polarization - x : 0 > Polarization - y : 0 > Polarization - z : 0 > Touchable (pointer) : 0x86689e8 > G4ParticleChange::CheckIt > > *** G4Exception: Aborting execution *** > Please correct this problem. I can't get my simulation working properly without the actual values for the refraction index. Thanks, Robbert Mirani intern for the Antares Project for the NIKHEF
*** Bug 14 has been marked as a duplicate of this bug. ***
I changed the index of refraction in the ExampleN06 (ExN06DetectorConstruction.cc) to those used in the Antares Project for the NIKHEF. I am running OS: Linux with release: geant4.0.1; and tags 'geant4-00-01-ref-03' for everything but /global, where I use tag 'global-V00-01-01'. I have also compiled source/track and /example/novice/N06 with G4DEBUG and G4VERBOSE on, and G4OPTIMISE off. ExampleN06 'Itworksforme' even with these changes. Seems to me, there is a memory leak / overwrite happening somewhere (see also problem report #15). To find it, the very application which produces this bug needs to be debugged.
*** Bug 15 has been marked as a duplicate of this bug. ***
Turns out, when in G4Cerenkov.cc the method, GetContinuousStepLimit calls GetAverageNumberOfPhotons the result can be (very small but) negative. This is possible because I use a finite constant in GetAverageNumberOfPhotons: const G4double Rfact = 369.81/(eV * cm); The next line in GetContinuousStepLimit then checks for if(MeanNumPhotons == 0.0) return DBL_MAX; G4double StepLimit = fMaxPhotons / MeanNumPhotons; Now, when MeanNumPhotons is negative StepLimit becomes negative, and then probably something in the transport goes haywire. This also explains why a change in index of refraction caused the problem. Itwould have happend anyway, sooner or later, even with the old values. Of course, it's a very rare occasion, when a step is very near to the Cerenkov threshold, that this bug happens. The numerical value for 'MeanNumPhotons' was -0.00004! All that is needed is the change to: if(MeanNumPhotons <= 0.0) return DBL_MAX; The corrected version of G4Cerenkov.cc has been committed to the head of the repository (in /source/processes/electromagnetic/xrays/src This change was tested with Robbert Mirani's ANTARES application. Regards, Peter Gumplinger
