Geant4 Cross Reference

Cross-Referencing   Geant4
Geant4/processes/electromagnetic/adjoint/src/G4ContinuousGainOfEnergy.cc

Version: [ ReleaseNotes ] [ 1.0 ] [ 1.1 ] [ 2.0 ] [ 3.0 ] [ 3.1 ] [ 3.2 ] [ 4.0 ] [ 4.0.p1 ] [ 4.0.p2 ] [ 4.1 ] [ 4.1.p1 ] [ 5.0 ] [ 5.0.p1 ] [ 5.1 ] [ 5.1.p1 ] [ 5.2 ] [ 5.2.p1 ] [ 5.2.p2 ] [ 6.0 ] [ 6.0.p1 ] [ 6.1 ] [ 6.2 ] [ 6.2.p1 ] [ 6.2.p2 ] [ 7.0 ] [ 7.0.p1 ] [ 7.1 ] [ 7.1.p1 ] [ 8.0 ] [ 8.0.p1 ] [ 8.1 ] [ 8.1.p1 ] [ 8.1.p2 ] [ 8.2 ] [ 8.2.p1 ] [ 8.3 ] [ 8.3.p1 ] [ 8.3.p2 ] [ 9.0 ] [ 9.0.p1 ] [ 9.0.p2 ] [ 9.1 ] [ 9.1.p1 ] [ 9.1.p2 ] [ 9.1.p3 ] [ 9.2 ] [ 9.2.p1 ] [ 9.2.p2 ] [ 9.2.p3 ] [ 9.2.p4 ] [ 9.3 ] [ 9.3.p1 ] [ 9.3.p2 ] [ 9.4 ] [ 9.4.p1 ] [ 9.4.p2 ] [ 9.4.p3 ] [ 9.4.p4 ] [ 9.5 ] [ 9.5.p1 ] [ 9.5.p2 ] [ 9.6 ] [ 9.6.p1 ] [ 9.6.p2 ] [ 9.6.p3 ] [ 9.6.p4 ] [ 10.0 ] [ 10.0.p1 ] [ 10.0.p2 ] [ 10.0.p3 ] [ 10.0.p4 ] [ 10.1 ] [ 10.1.p1 ] [ 10.1.p2 ] [ 10.1.p3 ] [ 10.2 ] [ 10.2.p1 ] [ 10.2.p2 ] [ 10.2.p3 ] [ 10.3 ] [ 10.3.p1 ] [ 10.3.p2 ] [ 10.3.p3 ] [ 10.4 ] [ 10.4.p1 ] [ 10.4.p2 ] [ 10.4.p3 ] [ 10.5 ] [ 10.5.p1 ] [ 10.6 ] [ 10.6.p1 ] [ 10.6.p2 ] [ 10.6.p3 ] [ 10.7 ] [ 10.7.p1 ] [ 10.7.p2 ] [ 10.7.p3 ] [ 10.7.p4 ] [ 11.0 ] [ 11.0.p1 ] [ 11.0.p2 ] [ 11.0.p3, ] [ 11.0.p4 ] [ 11.1 ] [ 11.1.1 ] [ 11.1.2 ] [ 11.1.3 ] [ 11.2 ] [ 11.2.1 ] [ 11.2.2 ] [ 11.3.0 ]

Diff markup

Differences between /processes/electromagnetic/adjoint/src/G4ContinuousGainOfEnergy.cc (Version 11.3.0) and /processes/electromagnetic/adjoint/src/G4ContinuousGainOfEnergy.cc (Version 10.2.p1)


  1 //                                                  1 //
  2 // *******************************************      2 // ********************************************************************
  3 // * License and Disclaimer                         3 // * License and Disclaimer                                           *
  4 // *                                                4 // *                                                                  *
  5 // * The  Geant4 software  is  copyright of th      5 // * The  Geant4 software  is  copyright of the Copyright Holders  of *
  6 // * the Geant4 Collaboration.  It is provided      6 // * the Geant4 Collaboration.  It is provided  under  the terms  and *
  7 // * conditions of the Geant4 Software License      7 // * conditions of the Geant4 Software License,  included in the file *
  8 // * LICENSE and available at  http://cern.ch/      8 // * LICENSE and available at  http://cern.ch/geant4/license .  These *
  9 // * include a list of copyright holders.           9 // * include a list of copyright holders.                             *
 10 // *                                               10 // *                                                                  *
 11 // * Neither the authors of this software syst     11 // * Neither the authors of this software system, nor their employing *
 12 // * institutes,nor the agencies providing fin     12 // * institutes,nor the agencies providing financial support for this *
 13 // * work  make  any representation or  warran     13 // * work  make  any representation or  warranty, express or implied, *
 14 // * regarding  this  software system or assum     14 // * regarding  this  software system or assume any liability for its *
 15 // * use.  Please see the license in the file      15 // * use.  Please see the license in the file  LICENSE  and URL above *
 16 // * for the full disclaimer and the limitatio     16 // * for the full disclaimer and the limitation of liability.         *
 17 // *                                               17 // *                                                                  *
 18 // * This  code  implementation is the result      18 // * This  code  implementation is the result of  the  scientific and *
 19 // * technical work of the GEANT4 collaboratio     19 // * technical work of the GEANT4 collaboration.                      *
 20 // * By using,  copying,  modifying or  distri     20 // * By using,  copying,  modifying or  distributing the software (or *
 21 // * any work based  on the software)  you  ag     21 // * any work based  on the software)  you  agree  to acknowledge its *
 22 // * use  in  resulting  scientific  publicati     22 // * use  in  resulting  scientific  publications,  and indicate your *
 23 // * acceptance of all terms of the Geant4 Sof     23 // * acceptance of all terms of the Geant4 Software license.          *
 24 // *******************************************     24 // ********************************************************************
 25 //                                                 25 //
                                                   >>  26 // $Id: G4ContinuousGainOfEnergy.cc 91870 2015-08-07 15:21:40Z gcosmo $
 26 //                                                 27 //
 27                                                    28 
 28 #include "G4ContinuousGainOfEnergy.hh"             29 #include "G4ContinuousGainOfEnergy.hh"
 29                                                    30 
 30 #include "G4EmCorrections.hh"                  << 
 31 #include "G4LossTableManager.hh"               << 
 32 #include "G4Material.hh"                       << 
 33 #include "G4MaterialCutsCouple.hh"             << 
 34 #include "G4ParticleChange.hh"                 << 
 35 #include "G4ParticleDefinition.hh"             << 
 36 #include "G4PhysicalConstants.hh"                  31 #include "G4PhysicalConstants.hh"
 37 #include "G4Step.hh"                           << 
 38 #include "G4SystemOfUnits.hh"                      32 #include "G4SystemOfUnits.hh"
 39 #include "G4VEmFluctuationModel.hh"            <<  33 #include "G4Step.hh"
                                                   >>  34 #include "G4ParticleDefinition.hh"
 40 #include "G4VEmModel.hh"                           35 #include "G4VEmModel.hh"
 41 #include "G4VEnergyLossProcess.hh"             <<  36 #include "G4VEmFluctuationModel.hh"
 42 #include "G4VParticleChange.hh"                    37 #include "G4VParticleChange.hh"
                                                   >>  38 #include "G4AdjointCSManager.hh"
                                                   >>  39 #include "G4LossTableManager.hh"
                                                   >>  40 #include "G4SystemOfUnits.hh"
                                                   >>  41 #include "G4PhysicalConstants.hh"
 43                                                    42 
 44 //////////////////////////////////////////////     43 ///////////////////////////////////////////////////////
 45 G4ContinuousGainOfEnergy::G4ContinuousGainOfEn <<  44 //
 46                                                <<  45 G4ContinuousGainOfEnergy::G4ContinuousGainOfEnergy(const G4String& name, 
 47   : G4VContinuousProcess(name, type)           <<  46   G4ProcessType type): G4VContinuousProcess(name, type)
 48 {}                                             <<  47 {
 49                                                    48 
 50 ////////////////////////////////////////////// <<  49 
 51 G4ContinuousGainOfEnergy::~G4ContinuousGainOfE <<  50   linLossLimit=0.05;
                                                   >>  51   lossFluctuationArePossible =true;
                                                   >>  52   lossFluctuationFlag=true;
                                                   >>  53   is_integral = false;
                                                   >>  54   
                                                   >>  55   //Will be properly set in SetDirectParticle()
                                                   >>  56   IsIon=false;
                                                   >>  57   massRatio =1.;
                                                   >>  58   chargeSqRatio=1.;
                                                   >>  59   preStepChargeSqRatio=1.;
                                                   >>  60   
                                                   >>  61   //Some initialization
                                                   >>  62   currentCoupleIndex=9999999;
                                                   >>  63   currentCutInRange=0.;
                                                   >>  64   currentMaterialIndex=9999999;
                                                   >>  65   currentTcut=0.;
                                                   >>  66   preStepKinEnergy=0.;
                                                   >>  67   preStepRange=0.;
                                                   >>  68   preStepScaledKinEnergy=0.;
                                                   >>  69   
                                                   >>  70   currentCouple=0;  
                                                   >>  71 }
 52                                                    72 
 53 //////////////////////////////////////////////     73 ///////////////////////////////////////////////////////
 54 void G4ContinuousGainOfEnergy::ProcessDescript <<  74 //
                                                   >>  75 G4ContinuousGainOfEnergy::~G4ContinuousGainOfEnergy()
 55 {                                                  76 {
 56   out << "Continuous process acting on adjoint <<  77  
 57          "continuous gain of energy of charged <<  78 }
 58          "tracked back.\n";                    <<  79 ///////////////////////////////////////////////////////
                                                   >>  80 //
                                                   >>  81 
                                                   >>  82 void G4ContinuousGainOfEnergy::PreparePhysicsTable(
                                                   >>  83      const G4ParticleDefinition& )
                                                   >>  84 {//theDirectEnergyLossProcess->PreparePhysicsTable(part);
                                                   >>  85 
                                                   >>  86 ; 
 59 }                                                  87 }
 60                                                    88 
 61 //////////////////////////////////////////////     89 ///////////////////////////////////////////////////////
 62 void G4ContinuousGainOfEnergy::SetDirectPartic <<  90 //
 63 {                                              <<  91 
 64   fDirectPartDef = p;                          <<  92 void G4ContinuousGainOfEnergy::BuildPhysicsTable(const G4ParticleDefinition&)
 65   if(fDirectPartDef->GetParticleType() == "nuc <<  93 {//theDirectEnergyLossProcess->BuildPhysicsTable(part);
 66   {                                            <<  94 ;
 67     fIsIon     = true;                         <<  95 }
 68     fMassRatio = proton_mass_c2 / fDirectPartD <<  96 
 69   }                                            <<  97 ///////////////////////////////////////////////////////
                                                   >>  98 //
                                                   >>  99 void  G4ContinuousGainOfEnergy::SetDirectParticle(G4ParticleDefinition* p)
                                                   >> 100 {theDirectPartDef=p;
                                                   >> 101  if (theDirectPartDef->GetParticleType()== "nucleus") {
                                                   >> 102    IsIon=true;
                                                   >> 103    massRatio = proton_mass_c2/theDirectPartDef->GetPDGMass();
                                                   >> 104    G4double q=theDirectPartDef->GetPDGCharge();
                                                   >> 105    chargeSqRatio=q*q;
                                                   >> 106   
                                                   >> 107    
                                                   >> 108  }
                                                   >> 109  
 70 }                                                 110 }
 71                                                   111 
 72 //////////////////////////////////////////////    112 ///////////////////////////////////////////////////////
                                                   >> 113 //
                                                   >> 114 // 
 73 G4VParticleChange* G4ContinuousGainOfEnergy::A    115 G4VParticleChange* G4ContinuousGainOfEnergy::AlongStepDoIt(const G4Track& track,
 74                                                << 116                                                        const G4Step& step)
 75 {                                                 117 {
 76   // Caution in this method the step length sh << 118    
 77   // A problem is that this is computed by the << 119   //Caution in this method the  step length should be the true step length
 78   // not know the energy at the end of the adj << 120   // A problem is that this is compute by the multiple scattering that does not know the energy at the end of the adjoint step. This energy is used during the 
 79   // during the forward sim. Nothing we can re << 121   //Forward sim. Nothing we can really do against that at this time. This is inherent to the MS method
 80   // time. This is inherent to the MS method   << 122   //
 81                                                << 123   
                                                   >> 124   
                                                   >> 125   
 82   aParticleChange.Initialize(track);              126   aParticleChange.Initialize(track);
 83                                                << 127   
 84   // Get the actual (true) Step length            128   // Get the actual (true) Step length
                                                   >> 129   //----------------------------------
 85   G4double length = step.GetStepLength();         130   G4double length = step.GetStepLength();
 86   G4double degain = 0.0;                       << 131   G4double degain  = 0.0;
 87                                                << 132   
                                                   >> 133   
                                                   >> 134  
 88   // Compute this for weight change after cont    135   // Compute this for weight change after continuous energy loss
 89   G4double DEDX_before =                       << 136   //-------------------------------------------------------------
 90     fDirectEnergyLossProcess->GetDEDX(fPreStep << 137   G4double DEDX_before = theDirectEnergyLossProcess->GetDEDX(preStepKinEnergy, currentCouple);
 91                                                << 138    
 92   // For the fluctuation we generate a new dyn << 139   
 93   // = preEnergy+egain and then compute the fl << 140   
 94   // case.                                     << 141   // For the fluctuation we generate a new dynamic particle with energy =preEnergy+egain
                                                   >> 142   // and then compute the fluctuation given in  the direct case.
                                                   >> 143   //-----------------------------------------------------------------------
 95   G4DynamicParticle* dynParticle = new G4Dynam    144   G4DynamicParticle* dynParticle = new G4DynamicParticle();
 96   *dynParticle                   = *(track.Get << 145   *dynParticle = *(track.GetDynamicParticle());
 97   dynParticle->SetDefinition(fDirectPartDef);  << 146   dynParticle->SetDefinition(theDirectPartDef);
 98   G4double Tkin = dynParticle->GetKineticEnerg << 147   G4double Tkin = dynParticle->GetKineticEnergy(); 
 99                                                << 148 
100   G4double dlength = length;                   << 149 
101   if(Tkin != fPreStepKinEnergy && fIsIon)      << 150   size_t n=1;
102   {                                            << 151   if (is_integral ) n=10;
103     G4double chargeSqRatio = fCurrentModel->Ge << 152   n=1;
104       fDirectPartDef, fCurrentMaterial, Tkin); << 153   G4double dlength= length/n; 
105     fDirectEnergyLossProcess->SetDynamicMassCh << 154   for (size_t i=0;i<n;i++) {
106   }                                            << 155     if (Tkin != preStepKinEnergy && IsIon) {
107                                                << 156       chargeSqRatio =  currentModel->GetChargeSquareRatio(theDirectPartDef,currentMaterial,Tkin);
108   G4double r = fDirectEnergyLossProcess->GetRa << 157     theDirectEnergyLossProcess->SetDynamicMassCharge(massRatio,chargeSqRatio); 
109   if(dlength <= fLinLossLimit * r)             << 158   
110   {                                            << 159     }
111     degain = DEDX_before * dlength;            << 160   
112   }                                            << 161     G4double r = theDirectEnergyLossProcess->GetRange(Tkin, currentCouple);
113   else                                         << 162     if( dlength <= linLossLimit * r ) {
114   {                                            << 163         degain = DEDX_before*dlength;   
115     G4double x = r + dlength;                  << 164   } 
116     G4double E = fDirectEnergyLossProcess->Get << 165     else {
117     if(fIsIon)                                 << 166         G4double x = r + dlength;
118     {                                          << 167         //degain = theDirectEnergyLossProcess->GetKineticEnergy(x,currentCouple) - theDirectEnergyLossProcess->GetKineticEnergy(r,currentCouple);
119       G4double chargeSqRatio = fCurrentModel-> << 168     G4double E = theDirectEnergyLossProcess->GetKineticEnergy(x,currentCouple);
120         fDirectPartDef, fCurrentMaterial, E);  << 169     if (IsIon){
121       fDirectEnergyLossProcess->SetDynamicMass << 170       chargeSqRatio =  currentModel->GetChargeSquareRatio(theDirectPartDef,currentMaterial,E);
122       G4double x1 = fDirectEnergyLossProcess-> << 171       theDirectEnergyLossProcess->SetDynamicMassCharge(massRatio,chargeSqRatio);
123                                                << 172       G4double x1= theDirectEnergyLossProcess->GetRange(E, currentCouple);
124       G4int ii              = 0;               << 173 
125       constexpr G4int iimax = 100;             << 174                         // Loop checking, 07-Aug-2015, Vladimir Ivanchenko
126       while(std::abs(x - x1) > 0.01 * x)       << 175                         G4int ii=0;
127       {                                        << 176                         const G4int iimax = 100;
128         E = fDirectEnergyLossProcess->GetKinet << 177       while (std::abs(x-x1)>0.01*x) {
129         chargeSqRatio = fCurrentModel->GetChar << 178         E = theDirectEnergyLossProcess->GetKineticEnergy(x,currentCouple);
130           fDirectPartDef, fCurrentMaterial, E) << 179         chargeSqRatio =  currentModel->GetChargeSquareRatio(theDirectPartDef,currentMaterial,E);
131         fDirectEnergyLossProcess->SetDynamicMa << 180         theDirectEnergyLossProcess->SetDynamicMassCharge(massRatio,chargeSqRatio);
132                                                << 181         x1= theDirectEnergyLossProcess->GetRange(E, currentCouple);
133         x1 = fDirectEnergyLossProcess->GetRang << 182         ++ii;
134         ++ii;                                  << 183               if(ii >= iimax) { break; }
135         if(ii >= iimax)                        << 184       } 
136         {                                      << 185     }
137           break;                               << 186     
138         }                                      << 187     degain=E-Tkin;  
139       }                                        << 188     
140     }                                          << 189     
141                                                << 190     
142     degain = E - Tkin;                         << 191     }
143   }                                            << 192   //G4cout<<degain<<G4endl;
144   G4double tmax = fCurrentModel->MaxSecondaryK << 193     G4double tmax = currentModel->MaxSecondaryKinEnergy(dynParticle);
145   fCurrentTcut = std::min(fCurrentTcut, tmax); << 194   tmax = std::min(tmax,currentTcut);
146                                                << 195   
147   dynParticle->SetKineticEnergy(Tkin + degain) << 196   
148                                                << 197   dynParticle->SetKineticEnergy(Tkin+degain);
149   // Corrections, which cannot be tabulated fo << 198 
150   fCurrentModel->CorrectionsAlongStep(fCurrent << 199   // Corrections, which cannot be tabulated for ions
151                                                << 200   //----------------------------------------
152   // Sample fluctuations                       << 201   G4double esecdep=0;//not used in most models
153   G4double deltaE = 0.;                        << 202   currentModel->CorrectionsAlongStep(currentCouple, dynParticle, degain,esecdep, dlength); 
154   if(fLossFluctuationFlag)                     << 203 
155   {                                            << 204     // Sample fluctuations
156     deltaE = fCurrentModel->GetModelOfFluctuat << 205     //-------------------
157       fCurrentCouple, dynParticle, fCurrentTcu << 206   
158       - degain;                                << 207     
159   }                                            << 208   G4double deltaE =0.;
160                                                << 209     if (lossFluctuationFlag ) {
161   G4double egain = degain + deltaE;            << 210           deltaE = currentModel->GetModelOfFluctuations()->
162   if(egain <= 0.)                              << 211             SampleFluctuations(currentCouple,dynParticle,tmax,dlength,degain)-degain;
163     egain = degain;                            << 212     }
164   Tkin += egain;                               << 213   
165   dynParticle->SetKineticEnergy(Tkin);         << 214   G4double egain=degain+deltaE;
                                                   >> 215   if (egain <=0) egain=degain;
                                                   >> 216   Tkin+=egain;
                                                   >> 217   dynParticle->SetKineticEnergy(Tkin);
                                                   >> 218  }
                                                   >> 219  
                                                   >> 220  
                                                   >> 221   
166                                                   222 
                                                   >> 223   
167   delete dynParticle;                             224   delete dynParticle;
168                                                << 225  
169   if(fIsIon)                                   << 226   if (IsIon){
170   {                                            << 227   chargeSqRatio =  currentModel->GetChargeSquareRatio(theDirectPartDef,currentMaterial,Tkin);
171     G4double chargeSqRatio = fCurrentModel->Ge << 228   theDirectEnergyLossProcess->SetDynamicMassCharge(massRatio,chargeSqRatio);
172       fDirectPartDef, fCurrentMaterial, Tkin); << 229     
173     fDirectEnergyLossProcess->SetDynamicMassCh << 
174   }                                               230   }
175                                                << 231   
176   G4double DEDX_after = fDirectEnergyLossProce << 232   G4double DEDX_after = theDirectEnergyLossProcess->GetDEDX(Tkin, currentCouple);
177   G4double weight_correction = DEDX_after / DE << 233   
178                                                << 234   
                                                   >> 235   G4double weight_correction=DEDX_after/DEDX_before;
                                                   >> 236  
                                                   >> 237   
179   aParticleChange.ProposeEnergy(Tkin);            238   aParticleChange.ProposeEnergy(Tkin);
180                                                   239 
181   // Caution!!! It is important to select the  << 
182   // as the current weight and not the weight  << 
183   // the track is changed after having applied << 
184                                                   240 
185   G4double new_weight =                        << 241   //Caution!!!
186     weight_correction * step.GetPostStepPoint( << 242   // It is important  to select the weight of the post_step_point
                                                   >> 243   // as the current weight and not the weight of the track, as t
                                                   >> 244   // the  weight of the track is changed after having applied all
                                                   >> 245   // the along_step_do_it.
                                                   >> 246 
                                                   >> 247   // G4double new_weight=weight_correction*track.GetWeight(); //old
                                                   >> 248   G4double new_weight=weight_correction*step.GetPostStepPoint()->GetWeight();
187   aParticleChange.SetParentWeightByProcess(fal    249   aParticleChange.SetParentWeightByProcess(false);
188   aParticleChange.ProposeParentWeight(new_weig    250   aParticleChange.ProposeParentWeight(new_weight);
189                                                   251 
                                                   >> 252 
190   return &aParticleChange;                        253   return &aParticleChange;
191 }                                              << 
192                                                   254 
                                                   >> 255 }
193 //////////////////////////////////////////////    256 ///////////////////////////////////////////////////////
                                                   >> 257 //
194 void G4ContinuousGainOfEnergy::SetLossFluctuat    258 void G4ContinuousGainOfEnergy::SetLossFluctuations(G4bool val)
195 {                                                 259 {
196   if(val && !fLossFluctuationArePossible)      << 260   if(val && !lossFluctuationArePossible) return;
197     return;                                    << 261   lossFluctuationFlag = val;
198   fLossFluctuationFlag = val;                  << 
199 }                                                 262 }
200                                                << 
201 //////////////////////////////////////////////    263 ///////////////////////////////////////////////////////
202 G4double G4ContinuousGainOfEnergy::GetContinuo << 264 //
203                                                << 
204                                                << 
205 {                                              << 
206   DefineMaterial(track.GetMaterialCutsCouple() << 
207                                                << 
208   fPreStepKinEnergy = track.GetKineticEnergy() << 
209   fCurrentModel     = fDirectEnergyLossProcess << 
210     track.GetKineticEnergy() * fMassRatio, fCu << 
211   G4double emax_model           = fCurrentMode << 
212   G4double preStepChargeSqRatio = 0.;          << 
213   if(fIsIon)                                   << 
214   {                                            << 
215     G4double chargeSqRatio = fCurrentModel->Ge << 
216       fDirectPartDef, fCurrentMaterial, fPreSt << 
217     preStepChargeSqRatio = chargeSqRatio;      << 
218     fDirectEnergyLossProcess->SetDynamicMassCh << 
219                                                << 
220   }                                            << 
221                                                << 
222   G4double maxE = 1.1 * fPreStepKinEnergy;     << 
223                                                << 
224   if(fPreStepKinEnergy < fCurrentTcut)         << 
225     maxE = std::min(fCurrentTcut, maxE);       << 
226                                                << 
227   maxE = std::min(emax_model * 1.001, maxE);   << 
228                                                << 
229   G4double preStepRange =                      << 
230     fDirectEnergyLossProcess->GetRange(fPreSte << 
231                                                << 
232   if(fIsIon)                                   << 
233   {                                            << 
234     G4double chargeSqRatioAtEmax = fCurrentMod << 
235       fDirectPartDef, fCurrentMaterial, maxE); << 
236     fDirectEnergyLossProcess->SetDynamicMassCh << 
237                                                << 
238   }                                            << 
239                                                   265 
240   G4double r1 = fDirectEnergyLossProcess->GetR << 
241                                                   266 
242   if(fIsIon)                                   << 
243     fDirectEnergyLossProcess->SetDynamicMassCh << 
244                                                << 
245                                                   267 
246   return std::max(r1 - preStepRange, 0.001 * m << 268 G4double G4ContinuousGainOfEnergy::GetContinuousStepLimit(const G4Track& track,
                                                   >> 269                 G4double , G4double , G4double& )
                                                   >> 270 { 
                                                   >> 271   G4double x = DBL_MAX;
                                                   >> 272   x=.1*mm;
                                                   >> 273  
                                                   >> 274  
                                                   >> 275   DefineMaterial(track.GetMaterialCutsCouple());
                                                   >> 276  
                                                   >> 277   preStepKinEnergy = track.GetKineticEnergy(); 
                                                   >> 278   preStepScaledKinEnergy = track.GetKineticEnergy()*massRatio;
                                                   >> 279   currentModel = theDirectEnergyLossProcess->SelectModelForMaterial(preStepScaledKinEnergy,currentCoupleIndex);
                                                   >> 280   G4double emax_model=currentModel->HighEnergyLimit();
                                                   >> 281   if (IsIon) {
                                                   >> 282     chargeSqRatio =  currentModel->GetChargeSquareRatio(theDirectPartDef,currentMaterial,preStepKinEnergy);
                                                   >> 283   preStepChargeSqRatio = chargeSqRatio;
                                                   >> 284   theDirectEnergyLossProcess->SetDynamicMassCharge(massRatio,preStepChargeSqRatio);
                                                   >> 285   } 
                                                   >> 286   
                                                   >> 287   
                                                   >> 288   G4double maxE =1.1*preStepKinEnergy;
                                                   >> 289   /*if (preStepKinEnergy< 0.05*MeV) maxE =2.*preStepKinEnergy;
                                                   >> 290   else if (preStepKinEnergy< 0.1*MeV) maxE =1.5*preStepKinEnergy;
                                                   >> 291   else if (preStepKinEnergy< 0.5*MeV) maxE =1.25*preStepKinEnergy;*/
                                                   >> 292    
                                                   >> 293   if (preStepKinEnergy < currentTcut) maxE = std::min(currentTcut,maxE);
                                                   >> 294  
                                                   >> 295   maxE=std::min(emax_model*1.001,maxE);
                                                   >> 296     
                                                   >> 297   preStepRange = theDirectEnergyLossProcess->GetRange(preStepKinEnergy, currentCouple);
                                                   >> 298   
                                                   >> 299   if (IsIon) {
                                                   >> 300     G4double chargeSqRatioAtEmax = currentModel->GetChargeSquareRatio(theDirectPartDef,currentMaterial,maxE);
                                                   >> 301   theDirectEnergyLossProcess->SetDynamicMassCharge(massRatio,chargeSqRatioAtEmax);
                                                   >> 302   } 
                                                   >> 303   
                                                   >> 304   G4double r1 = theDirectEnergyLossProcess->GetRange(maxE, currentCouple);
                                                   >> 305   
                                                   >> 306   if (IsIon) theDirectEnergyLossProcess->SetDynamicMassCharge(massRatio,preStepChargeSqRatio);
                                                   >> 307   
                                                   >> 308   
                                                   >> 309 
                                                   >> 310   x=r1-preStepRange;
                                                   >> 311   x=std::max(r1-preStepRange,0.001*mm);
                                                   >> 312  
                                                   >> 313   return x;
                                                   >> 314   
                                                   >> 315  
247 }                                                 316 }
248                                                << 317 #include "G4EmCorrections.hh"
249 //////////////////////////////////////////////    318 ///////////////////////////////////////////////////////
250 void G4ContinuousGainOfEnergy::SetDynamicMassC << 319 //
251                                                << 320 
252 {                                              << 321 void G4ContinuousGainOfEnergy::SetDynamicMassCharge(const G4Track& ,G4double energy)
253   G4double ChargeSqRatio =                     << 322 { 
254     G4LossTableManager::Instance()->EmCorrecti << 323 
255       fDirectPartDef, fCurrentMaterial, energy << 324   G4double ChargeSqRatio= G4LossTableManager::Instance()->EmCorrections()->EffectiveChargeSquareRatio(theDirectPartDef,currentMaterial,energy); 
256   if(fDirectEnergyLossProcess)                 << 325   if (theDirectEnergyLossProcess) theDirectEnergyLossProcess->SetDynamicMassCharge(massRatio,ChargeSqRatio);
257     fDirectEnergyLossProcess->SetDynamicMassCh << 
258 }                                                 326 }
259                                                   327