[Prev][Next][Index][Thread]

Re: New NM decay generator (A. Olin)


   Dear Art, thank you for your useful comments. 

   I still need some explanation on the problem of IHADR flag in your
   code. The problem I see is that when the K- stops the hypernucleus
   is generated AND the FLUKA code is called, which is not what we want.
   Isn't it? I have not yet checked the program flow with the debugger
   but this is what I read from the routine GUHADR.

   Concerning the time-space relationships, in vers4.0 we have kept only
   the original code where the drift velocity was considered constant.
   The improvements of the drift velocity (two velocities) was implemented
   as a test code by A. Panzarasa and P. Cerello and was not included in
   the released version. On next Monday Alberto will come in Brescia
   and we will produce a cradle with the code giving the best results
   on the last cosmic ray data. In the following we plan to insert in the
   code time-space relationships obtained by the simulation of the cell
   by garfield. We will made available the corresponding cradles on the
   web.

   We used MIGRAD from MINUIT since we were used to it and we obtained
   good results on Monte carlo events. Do you think that we could use
   something better? 

   Best regards    aldo

On Tue, 11 Jan 2000, Art Olin wrote:

> Hi, Aldo. I'd be happy to test the new NM version of the code. It's been a
> long time since I looked at it, but I likely have some old results to
> test against, or I can run both versions. Funny how often mistakes are 
> spotted going through this exercise.
> 
> If I remember the HADR flag problem, it was the following. For the
> nuclear muon capture I use explicitly Fermi distribution and nuclear
> evaporation codes from Fluka. So it must be loaded - and I probably force
> the initialization if it is not done - in that part of the code.
> 
> I think I ran into a problem when I forgot about this and loaded GEISHA
> with the HADR flag, and this screwed up the linking. That's why I always
> set the flag for Fluka.
> 
> Regarding the sequences problem, the geant321 source files (.F) are always
> distributed, so the GCDES is all that is required.
> 
> I've started looking through the reconstruction program, and the
> space-time relation coded in v4.0 appears to me to just use a constant
> drift velocity. How did you code the improvements to this that was
> discussed at our meeting? I also notice that the minimization engine that
> we use is migrad. While this should give excellent results, it's not
> known for it's efficiency. Have other more efficient minimization routines
> been explored or do we need the full power and versitility of MIGRAD?
> 
> Ciao, Art
> ***********************************************************
> Art Olin                        4004 Wesbrook Mall
> Senior Research Scientist       Vancouver BC V6T 2A3
> TRIUMF                          CANADA
> Adjunct Professor               work phone: (604)222 7349
> Dept of Physics and Astronomy   home phone: (604)222 0650
> University of Victoria          fax:(604)222 1074
> ***********************************************************
> 
> 
> On Tue, 11 Jan 2000, Aldo Zenoni wrote:
> 
> > 
> >    Dear Art, we are implementig in FIDAMC the new NM decay generator
> >    you have written on the base of the Ramos' work. It seems to me
> >    really useful and well inserted in the previous logic of the code.
> > 
> >    I have only one remarks. I saw that you request the set of the flag
> >    IHADR=4 for the running of the procedure. However, looking at the
> >    new GUHADR routine it appears to me that, in case of stopping
> >    particles, the FLUKA code is always called, also for K-. 
> >    Isn't it? Or have I misunderstood something in your code?
> > 
> >    Let's proceed as follows. We will prepare a 4.00 version of FIDAMC
> >    containing your mods and make it available on the web. Could you
> >    please download it and make a test? I would prefer that you provide
> >    a check since you prepared it a long ago and we have not the time to
> >    implement it in previous versions.  Gervasio will send you a message
> >    as soon as the .cmz will be ready.
> > 
> >    Concerning your suggestions sent in a previous mail we need the
> >    GEANT321 and GEANG321 .CMZ file to load the sequences contained and
> >    sometimes to extract routines for debugging purpose. Anyway I will
> >    take into account the possibility to include the GEANT,GEANG sequences
> >    in FIDAMC/RC in order to avoid the distribution of the .cmz to users.
> > 
> >    Best regards   Aldo
> > 
>