Problem 1163

Summary:	G4Navigator issues Track stuck warnings in G4Tubs and G4Polycon
Product:	Geant4	Reporter:	Steffen Hauf <hauf>
Component:	geometry/solids	Assignee:	tatiana.nikitina
Status:	RESOLVED INVALID
Severity:	minor	CC:	Gabriele.Cosmo
Priority:	P5
Version:	9.3
Hardware:	PC
OS:	Linux
Attachments:	A test scenario for the "track stuck" warnings I encounter

Description Steffen Hauf 2011-01-26 11:59:06 CET

Through the scintillation process I am creating a large number of optical photons. I recieve G4Navigator::ComputeStep() - "Track stuck - not moving" Warning on average every second event. The photons are being reflected and transported in a G4Box and a G4Tubs cylindrical volume. The warnings only occur for the G4Tubs volume. Replacing the G4Tubs with a G4Polycon reduced the ammount of volumes, but did not get rid of them.

Comment 1 Gabriele Cosmo 2011-01-26 12:44:09 CET

Have you verified that your geometry is NOT affected by overlaps ?
Can you please provide a simple test case geometry where this problem happens?
Tubs and pcones are regularly heavily tested in many configurations also in complex setups and do not show problems of this kind. In order to start any effective investigation, it would be useful if you could provide a test case as simple as possible that reproduces the problem.

Comment 2 Steffen Hauf 2011-01-26 15:44:07 CET

Created attachment 99 [details]
A test scenario for the "track stuck" warnings I encounter

Comment 3 Steffen Hauf 2011-01-26 15:46:41 CET

Comment on attachment 99 [details]
A test scenario for the "track stuck" warnings I encounter

The test scenario consists of a NaI Box in which scintillation occurs and and attached christal disc, which is bedded into a steel plate using boolean subtraction. It can be run by executing test.mac

I have checked that no overlaps occur:

Checking overlaps for volume steelPMT1 ... OK! 
Checking overlaps for volume NaI ... OK! 
Checking overlaps for volume TeflonWrapping ... OK! 
LUT - data file: /usr/local/geant4.9.3/geant4.9.3.p02/data/RealSurface1.0/GroundTeflon.dat read in! 
Checking overlaps for volume Quartz Plate ... OK! 
LUT - data file: /usr/local/geant4.9.3/geant4.9.3.p02/data/RealSurface1.0/PolishedLumirror.dat read in! 

Still I get the following warnings:

WARNING - G4Navigator::ComputeStep()
          Track stuck, not moving for 10 steps
          in volume -steelPMT1- at point (-23.80517832,39.03372753,58.43582486)
          direction: (-0.3084794058,-0.1881297977,-0.9324417598).
          Potential geometry or navigation problem !
          Trying pushing it of 9e-10 mm ...
WARNING - G4Navigator::ComputeStep()
          Track stuck, not moving for 10 steps
          in volume -steelPMT1- at point (-40.24395688,21.6965973,61.528325)
          direction: (0.8927798999,-0.3691115235,0.2582648517).
          Potential geometry or navigation problem !
          Trying pushing it of 9e-10 mm ...

Comment 4 tatiana.nikitina 2011-01-27 10:12:27 CET

Dear Steffen,

Can you add the data files needed to run the testscase(LUT...)?

Thank you,

 Tatiana Nikitina

Comment 5 Steffen Hauf 2011-01-27 10:23:20 CET

The LUT files are from the optical surface data files:

http://www.geant4.org/geant4/support/download.shtml

In my installation I extracted them into the standard geant/data directory.

Cheers,

Steffen

Comment 6 tatiana.nikitina 2011-02-01 09:41:24 CET

Dear Steffen,

By looking in the 'DetectorConstruction' code, I found that you use Boolean operations with shared surfaces.  Most probably this gives the warning messages.
When Boolean operations are used for solids with the sharing surfaces, it can cause different problems in tracking, visualization ...(See the Application Developers Guide, chapter 4.1.2.2).To avoid sharing surfaces in boolean operations, you can make subtraction of the solid, bigger by a small value. 

Tatiana Nikitina