From: Alberto Fasso' (fasso@SLAC.Stanford.EDU)
Date: Sun Jun 13 2004 - 02:22:39 CEST
Hi Paolo,
your problem comes from a misunderstanding of how Fluka records energy
deposition, which is different (and more accurate) than in other
programs where the energy deposited by ionization is "dumped" in the middle
(or at the beginning, or at the end) of a charged particle step.
As a matter of fact, there are two kinds of energy depositions in Fluka:
1) energy deposited by ionization by charged particles (dE/dx).
This energy is NOT deposited at a point, but over a "step" (a trajectory
between two points).
This kind of energy deposition is interfaced to the MGDRAW entry, and is
accessible as an array DTRACK(I),I=1,MTRACK (in common TRACKR).
At each MGDRAW call, there are NTRACK substeps (usually NTRACK = 1, except
in magnetic fields, where the trajectory is divided in several substeps),
and corresponding MTRACK deposition events (MTRACK = 0 means no energy
deposition, i.e. vacuum or neutral particle, otherwise MTRACK = 1, i.e.
dE/dx energy deposition). This energy must be considered as DISTRIBUTED OVER
THE WHOLE STEP.
If instead of doing your own scoring, you were using the standard Fluka
Fluka scoring via USRBIN, you would get the energy loss in a step
apportioned to each bin traversed by the step, according to the fraction of
the path length contained in each bin.
2) energy deposition at a point. This can happen for elastic or
inelastic recoils which are not transported and for low-energy
neutron kerma (generic average energy of charged particles which are not
necessarily identified). It can happen also when a particle is qstopping
(but generally particles below energy cutoff are "ranged out" to rest,
so their residual energy is deposited according to 1).
This kind of energy deposition is interfaced to entry ENDRAW of the
mgdraw subroutine. Note that this kind of "point" energy deposition
generally concerns some approximation to the real physical reality
(also a recoil deposits its energy over a path, but it is often so short
that we can approximate it as deposition at a point).
So, to answer specifically your question, the fact that the step ends at a
point P (whether on a boundary or not) does not mean that the corresponding
energy is deposited at that point.
The energy deposited is deposited between the points
XTRACK(0),YTRACK(0),ZTRACK(0) and XTRACK(NTRACK),YTRACK(NTRACK),ZTRACK(NTRACK)
(i.a. over a a segment if NTRACK=1, otherwise over a curved path).
I would avise you against using STEPSIZE, unless you are working with magnetic
fields. You have the impression to getter better results, but you still are
getting wrong results if you think that the energy is lost at
XTRACK(NTRACK),YTRACK(NTRACK),ZTRACK(NTRACK). You have made the step smaller,
so the error made assuming a point deposition is also smaller: but it is still
an error. And STEPSIZE breaks the step optimization which is one of the best
feature of Fluka.
You will get much more accurate and fast results using the built-in
USRBIN facility.
Alberto
On Sat, 12 Jun 2004, Paolo MAESTRO wrote:
>
> Hi Fluka users,
> I simulated a sampling calorimeter
> made of tungsten layers (0.35 cm thick)
> alternated with planes made of 0.05 cm thick scintillator.
> I noticed that protons (150 GeV kin. energy),
> passing through the detector as minimum ionizing particles,
> deposit energy in each scintillator plane only at the end of
> the considered region at the boundary with the next region.
> To explain better, if a scintillator plane extends
> from z= -0.35cm to z=-0.4 cm
> , dumping the energy deposit in MGDRAW & ENDRAW
> subroutine of mgdraw.f, it turns out that
> the deposit is recorded at z=-0.4 cm, i.e.
> at the boundary with the next tungsten layer.
> This has no physical meaning, and I guess it depends from
> the fact that the step size is bigger than the considered region
> thickness. I tried to set by hand a step size for the scintillators planes
> using the STEPSIZE options and it seems working well.
> But I wonder:
> - which is the empirical rule to set the min and max stepsize ?
> (I tried min = 1/50 scintillator thickness and max = 1/10)
> - in my case, do I need to set STEPSIZE also for tungsten layers?
> - is this procedure to avoid the energy deposit at boundaries correct ?
>
>
> Many thanks for your help.
> Best regards
>
> Paolo Maestro
-- Alberto Fassò SLAC-RP, MS 48, 2575 Sand Hill Road, Menlo Park CA 94025 Phone: (1 650) 926 4762 Fax: (1 650) 926 3569 fasso@slac.stanford.edu
This archive was generated by hypermail 2.1.6 : Sun Jun 13 2004 - 13:59:46 CEST