Double counting of energy deposit

[Stewart Boogert]

In EM and hadronic process examples. There is a collimator which should absorb all the energy so if N particles are generated with energy E. There should be a total of N*E energy deposited in the collimators. So for the EM process examples

N = 100 E = 250 GeV

So a total of 25000 GeV should be deposited. The integral of ElossHisto in this case is

ElossHisto->Integral() 50000 GeV

It appears we are double counting the energy of energy deposits all of the time.

[Laurie Nevay]

Unable to duplicate. Can you provide an exact example please. I tried several examples em / copper and stainless steel with various numbers of primaries and all conserve energy to the sub-percent level. Below are a few test cases from features/processes

First column is integral E(GeV), second is fraction of simulated energy ( Integral(E) / N*E(i) )

Copper 1 particle 2.49958604315511565e+02 -> 9.99834417262046249e-01 2 particles 4.99918907100187596e+02 -> 9.99837814200375208e-01 10 particles 2.49944340709228027e+03 -> 9.99777362836912098e-01

Stainless Steel 1 particle 2.49983466389756359e+02 -> 9.99933865559025414e-01 2 particles 4.99981709479116148e+02 -> 9.99963418958232264e-01 10 particles 2.49984975808408808e+03 -> 9.99939903233635241e-01 100 particles 2.49985880888789688e+04 -> 9.99943523555158809e-01

[Stewart Boogert]

This was a problem with robdsimAnal

A file was used to generate the original histogram list and the histograms were double counted.

[Jochem Snuverink]

Due to this fix the _ref.root files are no longer valid (since they had double counting in them) and need to be replaced.