From heintz@bu.edu Sun Jun 6 13:05:42 1999 Date: Sun, 22 Nov 1998 10:56:29 -0500 From: Ulrich Heintz To: heintz@bu.edu, Fred Borcherding , Harold Evans , John Hobbs , Marvin Johnson , Meenakshi Narain , Dave Toback , Horst Wahl , Kin Yip Subject: L1CFT output format Hi, in response to Dave's proposal to give H layer hit index and pT of L1CFT tracks, I have made a few calculations. The results are as follows: In order to convert the L1CFT information that you propose to send (H layer, pT, sign) into a road in the SMT, I look at the family of all possible tracks that go through the specified H layer fiber and have pT within the specified pT bin (assuming the bins you quoted in your first message). I do not think that it creates a bias to assume the track came from the origin at this point, because that was your assumption in computing the pT L1CFT. That defines a range of A layer hit positions, compatible with the supplied Level 1 information. The following table gives the width of that range in degrees for each of the 4 groups of pT bins (the numbers are identical for bins spanning the same range in 1/pT). The last column gives the corresponding numbers if the A layer hit index is known directly. pT bin 1.5-1.7 3.0-3.3 5.0-5.7 10-13.3 A layer hit index A layer pos 0.90 0.42 0.34 0.34 0.28 I then define the road in the SMT by the family of all tracks with pT>1.5 GeV that pass through this range in A layer positions and the specified H layer fiber. The widths of the roads (in cm) in the inner and outermost SMT layers, assuming all tracks come from the origin, are as follows. pT bin 1.5-1.7 3.0-3.3 5.0-5.7 10-13.3 A layer hit index inner layer 0.067 0.033 0.027 0.027 0.023 outer layer 0.21 0.10 0.083 0.083 0.069 If I allow tracks with impact parameters up to 2 mm, I get: pT bin 1.5-1.7 3.0-3.3 5.0-5.7 10-13.3 A layer hit index inner layer 0.15 0.12 0.11 0.11 0.11 outer layer 0.48 0.37 0.35 0.35 0.34 So, the widths of the roads are dominated by the impact parameter variation in both schemes, (H,pT) and (H,A). It is however not true that H,pT specifies the A layer hit position more precisely than the A layer index. Thus the roads are somewhat wider if I use (H,pT). There is another important point, though: we are not really comparing apples with apples here. In the (H,A) scheme, the SMT roads have 100% geometrical acceptance for the tracks that go through the specified H and A layer fibers. The way I defined the roads for the above calculations, this is not true. Your pT bins presumably refer to the mean pT of the L1CFT patterns that end up in the bin. That means that there can be tracks with lower or higher pT than the bin label in any bin. These are not guaranteed to lie within my roads. I would have to base my roads on the lowest and highest possible track pT within each bin to achieve 100% acceptance. In this scheme your bins would overlap. I don't have the necessary information to perform this calculation, but the roads would certainly be wider. I'd be happy to plug the numbers in if you send them to me. My conclusion is that the L2STT is better off with the (H,A) scheme. greetings, Ulrich -- Ulrich Heintz Department of Physics, Boston University phone: (617) 353 9057 590 Commonwealth Avenue fax: (617) 353 9393 Boston, MA 02215 e-mail: heintz@bu.edu