GSI Forum: Event Generators » New DPMgen results disagree with earlier simulations

Home » PANDA » PandaRoot » Event Generators » New DPMgen results disagree with earlier simulations

Show: Today's Messages :: Polls :: Message Navigator

New DPMgen results disagree with earlier simulations [message #10071]

Tue, 02 February 2010 17:55

Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich

first-grade participant

From: *e18.physik.tu-muenchen.de

Dear colleagues,

I just tried to re-do some results from last year for the TPC space charge simulations. I need simulations for panda background, including elastic and inelastic processes. On first glance I noticed that the new simulation results look very different to what I obtained last summer.

Attached you can find 2 plots of the MCTrack phase space of 50.000 events as given by the DPM generator. Both have been simulated with elastic and inelastic scattering turned on in the DPM settings (option "1") and with a very recent checkout of the pandaroot framework (svn rev. 7557). The "old" results correspond to rev. 4226 of the DPM folder (with everything else unchanged), that is a version before the changes in September '09.

As you can see, the very prominent band of elastic scattering almost disappeared. The spectrum seems no longer to be dominated by elastic processes. Can somebody from the DPM experts please comment on this?

Kind Regards,

Felix

https://www.e18.physik.tu-muenchen.de/~fboehmer/phasespace_newDPM.jpg

https://www.e18.physik.tu-muenchen.de/~fboehmer/phasespace_oldDPM.jpg

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10087 is a reply to message #10071]

Wed, 03 February 2010 16:14

Ralf Kliemt
Messages: 507
Registered: May 2007
Location: GSI, Darmstadt

first-grade participant

From: *cb.uni-bonn.de

Hello Felix,

As far as I know we have the option to choose what DPM should include.

"pandaroot/pgenerators/PndDpmDirect.h"

/** Standard constructor
* @param Mom in GeV/C
* @param Mode = 0. - No elastic scattering, only inelastic
* @param Mode = 1. - Elastic and inelastic interactions
* @param Mode = 2. - Only elastic scattering, no inelastic one
**/
PndDpmDirect(Double_t Mom, Int_t Mode);

Does that help?

Greetings to Munich, Ralf

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10089 is a reply to message #10087]

Wed, 03 February 2010 16:38

Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich

first-grade participant

From: *e18.physik.tu-muenchen.de

Hello Ralf,

as I wrote, both runs have been simulated with OPTION 1 (elastic and inelastic).

Cheers

Felix

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10091 is a reply to message #10089]

Wed, 03 February 2010 16:48

Tobias Stockmanns
Messages: 489
Registered: May 2007

first-grade participant

From: *ikp.kfa-juelich.de

Dear Felix,

in the old DPM generator the elastic option only included the simulation of the hadronic part of the elastic scattering. I know that Aida has extended the generator to include Coulomb and interference term. Maybe this is the reason for the different results.

Cheers,

Tobias

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10092 is a reply to message #10091]

Wed, 03 February 2010 16:53

Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich

first-grade participant

From: *e18.physik.tu-muenchen.de

Hi Tobias,

this would make sense if the elastic part would have increased, but it is significantly lower now. Apparently something substantial has changed? Can they interfere somehow? *scratching my head*

Kind Regards

Felix

[Updated on: Wed, 03 February 2010 17:02]

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10106 is a reply to message #10071]

Fri, 05 February 2010 10:33

Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich

first-grade participant

From: *e18.physik.tu-muenchen.de

On second thought the effect could be explained if the relative size of the hadronic and Coulombic cross sections fits accordingly...

The effect, however, is drastic: deposited energy in the TPC is reduced by a factor of 20!
I NEED confirmation from a DPMGen expert.

Regards

Felix

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10108 is a reply to message #10106]

Fri, 05 February 2010 11:07

StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino

first-grade participant

From: *to.infn.it

Whe don't you check an older version of dpm package, when you have done your first test, and check if it is a problem coming from the generator or from the remaining code? I think this test could be helpful.

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10113 is a reply to message #10108]

Fri, 05 February 2010 12:01

Sebastian Neubert
Messages: 282
Registered: March 2006
Location: Munich

first-grade participant

From: *natpool.mwn.de

Hi Stefano!

As Felix has written in his previous message, that was one of the first checks he did.

Cheers! Sebastian.

Sebastian Neubert
Technische Universität München
Department Physik E18
sneubert@e18.physik.tu-muenchen.de
tel: +49-8928912592

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10114 is a reply to message #10113]

Fri, 05 February 2010 12:02

StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino

first-grade participant

From: *to.infn.it

If you write the svn release you are using, one can see which are the changes in the code.

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10115 is a reply to message #10114]

Fri, 05 February 2010 12:08

Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich

first-grade participant

From: *e18.physik.tu-muenchen.de

This I have also written in the very first message.

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10116 is a reply to message #10115]

Fri, 05 February 2010 14:36

StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino

first-grade participant

From: *to.infn.it

Just to understand,
what have you used exactly? I have tried DPMGen right now and I got a nice elastic peak.

Have you tried PndDpmDirect or the original Dpm? And using which params?

I have tried with the following params:

seed: 1234
p: 15
mode: 1
thetamin: 0.1

I think the problem could stay if you have set thetamin=0 -> divergence of the coulomb part. Which theta min have you used now? Before the coulomb implementation this param was not set before, and I have seen this is not propagated into PndDpmDirect.

Could you please try my settings and let me know?

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10118 is a reply to message #10116]

Fri, 05 February 2010 15:43

Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich

first-grade participant

From: *e18.physik.tu-muenchen.de

Hey Stefano,

you are right! I use the "classic" DPMGen and naively set ThetaMin to zero! I re-tried with your setting of 0.1 and now I have the elastic peak back in principle.

Now I am curious: How was this implemented before? I will have to understand this better to "tune" the Theta parameter...

Many thanks for now!!

Cheers,

Felix

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10120 is a reply to message #10118]

Fri, 05 February 2010 15:58

Tobias Stockmanns
Messages: 489
Registered: May 2007

first-grade participant

From: *netcologne.de

Just in short, why this was introduced:

The cross section for elastic scattering diverges for low scattering angles. Therefore you need to have a cut-off parameter (theta) to define down to which angles you want to simulate it.

The new version of DPM was introduced because there was no EM elastic scattering in the old version but this is needed for the luminosity monitor which relys on this effect to calculate the luminosity.

I hope this helps a bit.

Cheers,

Tobias

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10121 is a reply to message #10120]

Fri, 05 February 2010 16:09

Felix Boehmer
Messages: 149
Registered: May 2007
Location: Munich

first-grade participant

From: *e18.physik.tu-muenchen.de

Hi Tobias,

just so there is no misunderstanding: there was NO elastic EM scattering before, not even some simplified implementation?

Is there some documentation on the DPMGen that I can read?

Cheers

Felix

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10122 is a reply to message #10120]

Fri, 05 February 2010 17:07

StefanoSpataro
Messages: 2736
Registered: June 2005
Location: Torino

first-grade participant

From: *to.infn.it

However,
PndDpmDirect does not propagate anymore this number, which is not so nice and must be fixed.
I am wondering which could be a good value for thetamin, and how much this can change the final results for elastic coulomb, to use it as default value instead of zero.
Do you have some idea on what could be used as default parameter?

Report message to a moderator

Re: New DPMgen results disagree with earlier simulations [message #10124 is a reply to message #10071]

Fri, 05 February 2010 18:37

Aida Galoyan
Messages: 79
Registered: May 2007
Location: Dubna

continuous participant

From: *jinr.ru

Hi all,

When I installed version of DPM with Coulomb, I wrote
(message# 9312 in General)
>>>>
I have commited new version of DPM generator.
Full elastic scattering, Coulomb, interference, hadronic parts are included in the new DPM.

You need to put the minimal angle of scattering -"tetmin"(>0) at DPM running, if you give value of parameter
"Elastic" = 1 (Inelastic with Elastic scatterings)
or = 2 (only Elastic scatterings).

The implementation of full elastic scattering in DPM is
important for Luminosity monitoring and, may be, for estimation of radiation doses in MVD.
>>>>

How I remember, Mohammad also wrote for new PndDpmDirect,
if you choose Elastic = 1 or =2, you need to put tetmin > 0.

The calculations at Elastic 1 or 2 (with Elastic scattering) at tetmin = 0 are wrong, because Coulomb scattering go to infinity
at tetmin=0.

If you want to choose tetmin for your detector (TPC) make some calculations with simulation of only elastic scattering
(Elastic =2) with various values of tetmin. There will be a lot of recoil protons from Coulomb scattering (at small tetmin), but most of them will have low energies and, I think, they will be absorbed by beam pipe or MVD. Only protons at large tetmin will fall in your Detector, because they can penetrate beam pipe and MVD.

Tetmin determines relation between elastic and inelastic events. At increasing tetmin, the Coulomb part of elastic scattering decreases sharply, and number of inelastic events increases.

Best regards,
Aida

Report message to a moderator

Previous Topic:	Random Seed of BoxGenerator
Next Topic:	UrQMD - SMM

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

PDF

]

Current Time: Fri Nov 29 01:01:37 CET 2024

Total time taken to generate the page: 0.00741 seconds