This bug has been corrected in the latest
(Feb 12, 2010) distribution of the EGS5 code. (Version 1.0.4)
In an egs5 simulation using mesh tallies, Takeshi Sugita (SSL) noticed
that for electrons, the sum of tvstep was 8 times greater than sum of ustep,
which was not expected for small meshes.
Upon inspection of egs5_electr.f, Namito noted that ustep is the quantity
which must be used to score absorbed dose, etc., and that tvstep was used
only to define the multiple scattering step. Namito also noted that
egs5_photon.f contains the line "tvstep=ustep" just before the call to
ausgab(0), so that ustep and tvstep agree for photons.
The issue here is that because of the introduction of the random hinge
transport mechanics for electrons in egs5, none of the previous (egs4)
relationships between the variables tstep, tustep, ustep, vstep, and
tvstep hold. Many of these variables are redundant in egs5 and have
been removed. The manual, however, was not adequately updated to
explain the changes, and in fact, had retained language from the egs4
manual that was misleading and inaccurate. For instance, table B.6 of
SLAC-730/KEK 2005-8 stated:
USTEP: User (straight line) step length requested and granted.
TVSTEP: Actual total Multiple scattering step to be transported.
To address the discrepancy between the manual and the source code and
to retain some of the egs4 usage of the electron step variable tvstep,
we have modified the code as follows:
1. Renamed previous egs5 variable "tvstep" as "tmstep", since it is used
only as the step in multiple scattering.
2. Equivalenced the variables "ustep" and "tvstep".
After this modification "tvstep" is just a dummy variable, but, as in
egs4, it can be used in egs5 to score dose, fluence, etc. (Note that
"ustep" can also be used for scoring in egs5.)
We have also updated SLAC-730/KEK 2005-8 to reflect these changes and
to better explain the relationships between the electron step size
variables in egs5 vs. in egs4.