FIRST ATTEMPT AT FINDING SVT TRACKS

I perform here two analysis: first assuming no a priori knowledge on detector geometry, then using the nominal SVXII geometry. For simplicity I did this in an interactive Mathematica session.

Initialization

Read in data(2382 candidate tracks in 2 wedges)

dat=Table[Read["float_102831.txt",Real],{i,1,2382},{j,1,5}];Close["float_102831.txt"]

I am going to use wedge 3 only, because it has more data

As with the previous smaller sample, the smallest eigenvalue is too big to be just resolution

0.1877440906638008`	0.31033913234003846`	0.5019103800074148`	0.7851928105886267`
			0.45820164928869167`
0.5831392870921455`			0.3918209464881586`
	0.7471787327501067`		0.14142080710492277`

I can, however, now plot the distribution of its projection since I have some more data

The above distribution has a narrower central component over a broader background; this is evidence for some real tracks in the sample !
Its width is about 100 microns.

Now select the cleaner sample in the central peak:


		0.4754713729161473`
0.6070203400012769`		0.3766605208496238`
	0.7516288324672467`	0.13457025381347892`

This is (intuitively) the best candidate for the impact parameter wrt to the beam for this sample

Plot its distribution:

This is rather wide, but we used tracks of all Pts. Let's try to use high-Pt only

This is the best candidate for curvature:

Select tracks of high Pt (unclear how much, at this point). Find 123 tracks.

Now plot "impact parameter" for those:

ANALYSIS BASED ON NOMINAL GEOMETRY

Constraint and parameter expressions, generated from nominal geometry (CORRGEN program)

Now plot the (single, because we lack XFT) constraint:

This looks vaguely similar to the one obtained "in the dark" , but much cleaner ! Note it comes out centered at zero without ad-hoc corrections.
NO DOUBT THERE ARE TRACKS HERE !!!!

now select the central peak: