CSC Multimuons Alignment

Data

Multimuons

• Files on CASTOR:

  /castor/cern.ch/user/a/atlidali/data/csc/multimuons

• Files on the MPI file server:

  blphmpi00:/net/fsphmpi01/mnt/data/id_align/data/csc/multimuons

CSC preAlignment results

Track Parameters

• Reconstructed Q/Pt for the multimuons sample. The Q/Pt vs phi0 distribution can be used to detect some of the detector deformations. In particular it can show us evidence of sagitta and clocking effect (see figures for fun). As one can see in the scatter plot (iPatRec , GlobalX2Fitter ) of the Q/Pt versus Phi0, one can distinguish a sine waweform that corresponds to a sagitta effect. It is also possible to distinguish an assymetry between the momentum reconstructed for positive and negative particles as the black bands are not equally distributed around zero. This assymetry is due to a clocking effect. Positive charged particles are reconstructed systematically with lower momentum, whilst negative charged particles are reconstructed systematically with a higher momentum.

• Perigee and Beam Spot: CSC multimuons have been generated at (0,0,0) whith some spread, while the detector (PIX and SCT) has been shifted. All these conditions have an impact on the impact parameter and perigee of the reconstructed tracks. In the reconstruction the perigee is computed w.r.t. the origin (0,0,0) instead of w.r.t. the primary vertex.
  • Ideal case (perfect alignment though modules not at nominal position):
  • Nominal case (ignoting the misalignment, placing modules at nominal positions):
    • d0 cartesian view:
    • z0 cartesian view:
    • cartesian view:
    • polar coordinates view:

Results with the TrkValidation/InDetAlignNt ntuples

Results with the TrkValidation ntuple with ID tracks and with SiOnly tracks. Futhermore, track parameters and residuals comparation has been done using the InDetAlignNt ntuple (Unbiased residuals):

ID Tracks Vs SiOnly Tracks Plots
Setup	Detectors	# events	Tracker	Summary plots
Nominal	Pixel+SCT+TRT	10k	KalmanFitter
Nominal	Pixel+SCT	10k	KalmanFitter
Comparison plots (from the InDetAlignNt): Trk Params: , Pixel/SCT bias res:

Results comparing the differences between the use of the ErrorScaling in each case:

ErrorScaling Plots
Setup	# events	Tracker	ErrorScaling	Summary plots
Nominal	50k	iPatRec	No
Nominal	25k	iPatRec	Yes
L1L2L3 Corrected	25k	iPatRec	No
L1L2L3 Corrected	25k	iPatRec	Yes*
L1L2L3 Corrected	50k	GlobalChi2Fitter	Yes*

Results with the GlobalChi2AlignAlg ntuple

Results with our ntuple (biased residuals and SiOnly tracks). Here, the DoubleCone geometry (2328 barrel modules - 13968 DoFs) is used:

GlobalChi2AlignAlg ntuple
Setup	# events	# tracks	# hits	Tracker	ErrorScaling	starting IDAlignVis	LocalPos	GlobalPos	Summary plots
Nominal	25k	233779	2179873	iPatRec	Yes*	-	-	-
L1 Corrected	25k	232516	2356026	iPatRec	Yes	()	()	()
L2 Corrected	25k	237241	2381481	iPatRec	Yes	()	()	()
L1L2 Corrected	25k	241449	2262378	iPatRec	No	()	()	()
L3 Corrected	25k	234609	2186907	iPatRec	Yes	()	() () ()	()
L1L2L3 Corrected	25k	235266	2721301	iPatRec	No	-	-	-

rel12 Vs rel13

Comparation between results with rel 12.0.4 and rel 13.0.10. In this case we have used Si-Only tracks. The DoubleCone geometry has been use to produce the plots labeled with processed in the title.

rel12 Vs rel13
release	Setup	# events	# tracks	# hits	# processed tracks	# processed hits	Tracker	ErrorScaling	VtxFit	Summary plots
12.0.4	L1L2L3 Corrected	25k	235266	2721301	100895	1107087	iPatRec	No	Yes
12.0.6	L1L2L3 Corrected	25k	235266	2721301	100895	1107087	iPatRec	No	Yes
13.0.10	L1L2L3 Corrected	25k	234204	2698346	99469	1091847	iPatRec	No	Yes
13.0.10*	L1L2L3 Corrected	25k	234194	2697968	99469	1091847	iPatRec	No	Yes
12.0.4	Nominal	25k	233779	2179873	99622	940220	iPatRec	Yes	Yes
12.0.6	Nominal	25k	233779	2179869	99622	940220	iPatRec	Yes	Yes
13.0.10	Nominal	25k	-	-	-	-	iPatRec	Yes	Yes
13.0.10	Nominal	25k	213739	1325465	10559	88515	iPatRec	Yes**	Yes

(*) Using the iPatRec bugfixes from Alan Poppleton
(**) Error Scaling tunned by Wolfgang for rel 12.0.4

ID tracks, SCTOnly tracks and PixelOnly tracks

Studies done using the RefitSiOnlyTool to refit tracks in rel13:

RefitSiOnlyTool
Setup	Tracks	# events	# tracks	# hits	Tracker	ErrorScaling	VtxFit	Summary plots
L1L2L3 Corrected	Pixel-Only	5k	-	-	GlobalChi2Fitter	No	Yes
L1L2L3 Corrected	SCT-Only	5k	-	-	GlobalChi2Fitter	No	Yes

GlobalChi2 Alignment

Alignment Results using SiOnly tracks and the DoubleCone geometry (2328 barrel modules - 13968 DoFs). Plots from our ntuple (biased residuals):

Perfect Geometry (Misalignment Known)

Perfect Geometry - L1L2L3 Corrected (18.02.2007) - Release 12.0.4
Iteration	# events	# tracks	# hits	Tracker	ErrorScaling	VtxFit	LocalPos	GlobalPos	Summary plots	Alignment Constants
0	99150	399937	4387874	iPatRec	No	Yes	()	()		(pool.root) (db) ()
1	99150	399938	4388049	iPatRec	No	Yes	()	()		(pool.root) (db) ()
2	25000	100897	1107125	iPatRec	No	Yes	()	()		-
0	99150	400496	4402638	GX2Fitter	No	Yes	()	()		(pool.root) ()
1	25000	101035	1110782	GX2Fitter	No	Yes	()	()		-
Comparison plots of Iterations 0, 1 and 2 for iPatRec
Only for tracks laying within the DoubleCone acceptance region
- Number of PIX and SCT hits associated per track:
- Reconstructed track parameters:
- Impact parameted error and d0 vs phi0:
- Q/p plots in various formats for systematic studies:
- Pixel phi residuals:
- Pixel eta residuals:
- SCT residuals:

Perfect Geometry - L1L2L3 Corrected (17.09.2007) - Release 13.0.20
Iteration	# events	# tracks	# hits	Tracker	ErrorScaling	VtxFit	LocalPos	GlobalPos	Summary plots	Alignment Constants
0	50k	-	-	GX2F	No	Yes	()	()		(-)

Nominal Geometry (Misalignment Unknown)

Misalignment Unknown - Nominal Case - Release 12.0.4 (07.04.2007)
Iteration	# events	# tracks	# hits	Tracker	ErrorScaling	VtxFit	LocalPos	GlobalPos	Summary plots	Alignment Constants
0	99150	394995	3728284	iPatRec	-	-	-	-		-
1	99150	398967	4316909	iPatRec	-	-	-	-		-
2	99150	399110	4373332	iPatRec	-	-	-	-		-
3	10000	40296	441927	iPatRec	-	-	-	-		-
Comparison plots of Iterations 0, 1, 2 for iPatRec
Only for tracks laying within the DoubleCone acceptance region. The ideal case is also included and serves as reference
So far, in all iterations, error scaling has been applied
- Track parameters:

In the reconstructed Pt, one may observe that the Q+Q- assymmetry decreases slihgtly after each iterarion. It does not clear, but the trend is a monotonical drecrease. Whilst in the equivalent plots extracted using newTracking (GlobalX2Fitter) it steadily grows.

Misalignment Unknown - Nominal Case - Release 12.0.4 (21 Feb 2007)
Iteration	# events	# tracks	# hits	Tracker	ErrorScaling	VtxFit	LocalPos	GlobalPos	Summary plots	Alignment Constants
0	99150	341604	2952281	GlobalChi2Fitter	Yes	-	-	-		-
1	99150	385028	3910740	GlobalChi2Fitter	Yes	-	-	-		-
2	99150	394033	4339267	GlobalChi2Fitter	Yes	-	-	-		-
3	99100	394606	4353202	GlobalChi2Fitter	Yes	-	-	-		-
4	10000	4043	44451	GlobalChi2Fitter	Yes	-	-	-		-
Comparison plots of Iterations 0, 1, 2, 3 and 4 for GlobalChi2Fitter
Only for tracks laying within the DoubleCone acceptance region. The ideal case is also included and serves as reference
- Track parameters: (More tracks are reconstructed after each iter)
- Track params: Reco-Truth:
- Hits associated to tracks: More hits are associated to the tracks. SCT is back to 8 and PIX to 3.
- Pixel phi biased residuals: All residuals become more centered/narrower after each iter
- Pixel eta biased residuals:
- SCT biased residuals:
- SCT biased residuals (wider range):
- Tranverse plane Impact parameter (d0) error and its phi0 dependence:
- Accepted tracks per events (so tracks within double cone acceptance):
- Q/p plots in various formats for systematic studies:

L3 Misalignment Unknown (L1L2 Corrected)

L3 Misalignment Unknown - L1L2 Corrected - Release 12.0.4
Iteration	# events	# tracks	# hits	Tracker	ErrorScaling	VtxFit	LocalPos	GlobalPos	Summary plots	Alignment Constants
0	-	-	-	-	-	-	-	-	-	(pool.root) ()
1	99150	331970	3033087	iPatRec	No	Yes	-	-		-
1	99150	396769	4184346	GlobalChi2Fitter	Yes	Yes	()	()		(pool.root) (db) ()
2	99150	399842	4396654	GlobalChi2Fitter	Yes	Yes	()	()		(pool.root) (db) ()
3	99150	39743	4388241	GlobalChi2Fitter	No	Yes	()	()		(pool.root) (db) ()
4	25000	100855	1109255	GlobalChi2Fitter	No	Yes	()	()		-

-- CarlosEscobar - 19 Apr 2007