From: Alberto Fasso' (fasso@SLAC.Stanford.EDU)
Date: Fri Jan 26 2007 - 00:12:04 CET
Hi Chris,
the answer is very simple. In FLUKA only right-handed systems are
considered.
Having said that, I would like to remind (not to you, to whom I have said
that many times, but to all FLUKA users) that BOXes, WEDges and ARBs
should NEVER NEVER NEVER be used. They are inefficient and unreliable,
and are still available only for historical reasons. They are in FLUKA
because they were inherited with the original Combinatorial Geometry,
but no requirement or check or anything else has ever been done in
FLUKA concerning them. Probably they should have been eliminated from
the beginning, but in the early times of FLUKA development we were
trying to be as back-compatible as possible. Now I am not so sure
that it was always a good idea.
I understand from your mail that what you have to handle is exactly
that: a historical relic found in an old input. But you can get around it by
replacing the BOX by 6 planes, or maybe by using an RPP and rotating it using
LATTICE.
Ciao
Alberto
On Thu, 25 Jan 2007, Chris Theis wrote:
> Hi Florian,
>
> Thanks for the quick answer, but as discussed the issue is a little more
> subtle than it seems at first glance. I'll try elaborate on it in more
> detail to explain the actual question:
>
> Usually, you define a box with a point (let's call it v) and three
> vectors h1, h2, h3 which all have a certain length. The nice thing about
> this is that you can implicitly define an arbitrary rotation. However,
> this is also where things become more tricky. For example if one wants
> to determine rotational angles from the basis formed by the h vectors it
> matters whether they form a left or a right-handed coordinate system.
>
> Thus, I was wondering whether FLUKA imposes any requirement on the
> specification of the h vectors (e.g. h1 crossprod h2 = h3)? I currently
> have an old input where I found a box with a left-handed coordinate
> system formed by the h vectors and I was not sure what FLUKA does in
> that case and if that poses a problem which is probably not obvious at
> the first look, like for example an RPP where min & max is accidentally
> swapped.
>
> Cheers
> Chris
>
> > Dear Chris!
> >
> > I think there is a misunderstanding. As far as I know the 3
> > vectors do not define height, width and length. Each of these
> > vectors lie on one edge of the box and the length of each
> > vector is the length of the edge.
> > If you look into the manual p255 fig 8.2 (CERN, yellow
> > report) you find a drawing how it works.
> >
> > Hope this helps.
> >
> > Regards, Florian
> >
> >
> >
> > Chris Theis wrote:
> > > Hello,
> > >
> > > I'm currently dealing with an old input file which uses BOX
> > > definitions and came across the following issue, which is
> > not really clear to me.
> > > According to the manual the user has to define an origin of
> > the box +
> > > 3 vectors for height, width & length. Furthermore, it's a
> > requirement
> > > that the user ensures perpendicularity of those vectors. However,
> > > nothing is stated with respect to the orientation of the basis
> > > generated by those vectors. I'm not sure whether it is a
> > requirement
> > > that they form a right-handed coordinate system and what is
> > the effect
> > > if a left-handed system is (accidentally) supplied?
> > Probably somebody
> > > could shed some light on this issue.
> > >
> > > Thanks in advance
> > > Chris
> > >
> > >
> >
> >
> > --
> > ###########################################
> > Florian Sommerer
> > European Organization for Nuclear Research CERN DSU/TT
> > Mailbox Z08700 Geneve 23
> > CH-1211 Switzerland
> > Tel : +41 2276 75582
> > ###########################################
> >
>
>
>
-- 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 : Fri Jan 26 2007 - 09:53:04 CET