sPHENIX tracking discussion

[Anthony Frawley <afrawley AT fsu.edu>]

Hi All,

For the record, Yasser ran some tests that showed the MVTX geometry calculation agrees with the TGeo geometry. I eventually found a problem in my tracking macro with the MVTX setup, and once I fixed that, the Acts geometry transformation agrees with our geometry class.

Sorry for the false alarm.

Tony

---

From: Anthony Frawley
Sent: Sunday, May 31, 2020 12:29 AM
To: sphenix-tracking-l AT lists.bnl.gov <sphenix-tracking-l AT lists.bnl.gov>; sphenix-maps-l <sphenix-maps-l AT lists.bnl.gov>
Cc: Anthony Frawley <afrawley AT fsu.edu>
Subject: Issue with MVTX geometry class?

 

Dear MVTX experts,

While working on the Acts implementation, I turned up a discrepancy between the MVTX geometry extracted by Acts from our G4 model, and the description of that geometry in the geometry class "mvtx/CylinderGeom_Mvtx".

The Acts code analyzes the TGeo objects in the G4 model and extracts a "Surface" corresponding to each active detector volume. The Acts Surface class has a transformation that returns the local position on the surface corresponding to a given world position. Our geometry classes have methods that do the same thing. They get the same answer for the INTT and TPC, but do not get the same answer for the MVTX.

For the MVTX sensors, the local position is (x,z), where x is the (radius*phi) direction. The local z position agrees perfectly, but the local x position is always different by 0.03872 mm. Also, the Acts surface center in world coordinates is slightly different from the sensor center given by CylinderGeom_Mvtx.

I have looked carefully at CylinderGeom_Mvtx and there is no obvious logical error. Doing local-to-world / world-to-local / local-to-world transformations yields the exact starting value. The positioning of the staves in the geometry class seems to be the same as in the MvtxDetector class, which constructs the detector model. The discrepancy seems to be in the location of the sensor in the stave. Since the stave is not built in our software, we import it, I can not easily check that the geometry class has it right.

My specific request to the experts is to check that the parameters in CylinderGeom_Mvtx that position the sensor in the chip, the chip in the module, and the module in the stave are all correct.

Thanks

Tony