sPHENIX discussion of software

["Frawley, Anthony" <afrawley AT fsu.edu>]

Hi All,

Sorry to spam the lists with a very specific question, but I need help from a Geant 4 expert.

I have tracked down the problem we have been seeing for the ladder silicon strip intermediate tracker, where we see many cases where the cluster position is wrong by about one strip width. However I don't know how to fix it.

The silicon strips are placed in the sensors in PHG4SiliconTrackerDetector using G4PVParameterised. Then in

PHG4SliconTrackerSteppingAction a G4TouchableHandle is used to get the physical volume for the hit strip from aStep->GetPreStepPoint, and the position is obtained from the GetTranslation method of the strip volume. That works fine if there is only one strip hit by the track.

However if the track passes through two strips (and many tracks do, for the tilted ladders) then the G4TouchableHandle that is returned is always a pointer to the second hit strip. So both hit strips are assigned the same copy number and the same position. I have checked that the strip parameterisation correctly describes the strip positions, and the two hit strips are indeed separate, adjacent strips where the track exits one strip and enters the next one at the same point. The copy number is just incorrect for the first hit strip.

It appears that when aStep->preStepPoint is asked for a touchable volume, it always returns a pointer to the last strip that was hit by this track (which is the correct pointer, if there is only one hit strip).

I have looked at all of the documentation I can find, and have not come up with any information on how to deal with this. I did see a cryptic remark in someone's slides that there is a problem with using touchables for G4PVParameterised volumes, but I can find no further information on that. If so, it seems like something that the User's Guide might mention!

Does anyone have any suggestions with how to proceed?

Thanks

Tony