The curious case of the Phi Structure
Soon after the L1 alignment was begining to be understood the attention turned to TRT-Only tracks.
Looking at the residuals on a module level the following strange pattern was seen.
This effect which was seen is using no alignment constants.
At the time there was only L1 constants and the tracks here are TRT-Only.
Clues
Facts.
-This was not seen in combined tracks
-Seen as far back as the monitoring was run
-with B-Field
-without B-Field
-with Xe
-with Ar
-In M6
-In M8
-In first fast-or running
Why is it not in combined tracks? Is the Si information removing some bais in the TRT?
A: No. A D0 cut on TRT-Only tracks removes this structure:
Is there something special about the upper and lower modules such that they do not give rise to this structure?
A: No. A reproducing the plot in bins of track phi0 shows that the structure is there, independent of module, for non-collision-like tracks:
See slides 10-16 in here, for details.
Does the L2 Alignment fix this?
A: Nope.
Is this seen in MC?
A: Nope.
Why doesn't the L2 Alignment fix this?
Is it a bug in the L2 alignment?
Is there something sick about the geometry that there is a local minima that can't be gotten out of?
A: Nope. It is a problem with the monitoring plots.
The solution to Curious Case of the Phi Structure
Exectutive Summary:
-Local straw coordinate frames are different on A and C sides. See this.
-When you plot global A and C residuals together there needs to be a realative minus sign between the two sides.
-When the monitoring plots are changed the residual after alignment does go to zero.
Details here.
... however now the residuals on the A and C side seperately are not zero. Need more DoF?