B/C ratio changes depending on Energy Range
Page 1 of 2

Author:  childers [ Sun Jun 17, 2007 12:47 pm ]
Post subject:  B/C ratio changes depending on Energy Range

I have taken your default GALDEF file, 599278, and made just a change to the energy limits. The default limits are 1e1 - 1e7 MeV/n.

I have tried 3e1 - 1e9 and 1e1 - 1e8 MeV/n, and gotten really nonsensical results.

I have also tried 1e3 - 1e9 and 1e2 - 1e8 MeV/n, and these produced good output that agrees with data when plotted on the same graph. The default limits also worked as expected.

Any idea why changing the energy limits would cause problems?

Taylor Childers
University of Minnesota

Author:  strong [ Mon Jun 18, 2007 12:59 am ]
Post subject:  energy limits

I can see no obvious reason, will try it myself.

Author:  childers [ Mon Jun 18, 2007 6:46 am ]
Post subject: 

Here is a plot that I made of the data with the default file 599278 and the same file with a modified energy range 600003.

I've included the galdef file for you to look over. ... 600003.txt

Thanks for you help,

Author:  strong [ Wed Jun 20, 2007 2:07 am ]
Post subject:  B/C

could you plot the B and C separately, to see which one is causing the problem
in the ratio

Author:  childers [ Wed Jun 20, 2007 6:58 am ]
Post subject:  B and C spectra

Here they are: ... ctrumA.jpg ... ctrumA.jpg

Author:  strong [ Wed Jun 20, 2007 7:09 am ]
Post subject:  B/C

so B is OK
but C is going to zero at below 10 GeV
.. very odd

Author:  strong [ Wed Jun 20, 2007 9:38 am ]
Post subject:  test C

I ran 600003 and C looks normal.
which deepens the mystery

The only thing I changed was Z up to 6 only , and no electrons etc, to make it go faster.
Maybe you can try that, just in case.

Author:  childers [ Thu Jun 21, 2007 9:21 am ]
Post subject:  New B and C plots

I did as you said, running 600003 without tracking electrons, positrons, etc. and also setting all "use_Z_*" = 0 for Z > 6. Here are the results, they still don't look good. ... ctrumB.jpg ... ctrumB.jpg

Author:  strong [ Thu Jun 21, 2007 9:48 am ]
Post subject:  puzzled

indeed I did not expect any change, was just a check.
I admit I'm stumped.
Must be something obscure connected with the compiler or CPU. ? Unexpected.

Tomorrow I will work out how to investigate using the log file - maybe switching
on a debug option.

Meanwhile can you also plot some more - p, He etc to see if there's a similar effect?

Author:  strong [ Sun Jun 24, 2007 11:26 pm ]
Post subject:  B/C

no progress yet.
I asked Igor to run it too.
Meanwhile can you see if it is only C or are any other nuclei affected?

Author:  imos [ Wed Jun 27, 2007 9:24 am ]
Post subject: 

Sorry, I am very busy preparing for ICRC.
I will try to do it this week.

Author:  childers [ Thu Jun 28, 2007 4:30 pm ]
Post subject:  Other Nuclei Plots

Here are those other nuclei plots you asked for: ... eiPlot.jpg

proton is red
helium is green
lithium is blue
beryllium is yellow
boron is cyan
carbon is light blue

Sorry this took so long, I've been getting ready for the ICRC, too. Perhaps I'll see you there.


Author:  childers [ Thu Jun 28, 2007 4:39 pm ]
Post subject:  B/C diffusion constant

I've also been trying to reproduce your diffusion coefficient work from your 1998 ApJ paper.

If I take 599278 and change its parameters to match those of your run with the index 01000.I don't get the same results

Here is my galdef file ... 50p_600025

here is a plot of its output:

As you can see, it doesn't reproduce your fitting results of the HEAO-3 data.


Author:  imos [ Thu Jun 28, 2007 6:51 pm ]
Post subject: 

The plots look terrible!
Something is definitively wrong.

Author:  imos [ Thu Jun 28, 2007 6:54 pm ]
Post subject: 

The paper of 1998 had used old cross sections.
I suggest you to try to use the parameters of Ptuskin et al. (2006) or
from Strong et al. (2004).

I am currently thinking of cancelling my travel to Merida, but if I go
there we can certainly solve the problem.

Page 1 of 2 All times are UTC - 8 hours [ DST ]
Powered by phpBB® Forum Software © phpBB Group