G. Van Buren - BNL
24 Jan 2007

A discussion of the probability calculations that were used can be found here. A discussion of the parameters (e.g. values of σ_x and σ_y) as well as the "GN" and "IK" pad layout options can be found here. Note that these simulations assume no non-track sources of hits, nor any inefficiency in reconstructing hits in the IST detectors. Some discussion of the latter can be found here.

In the below plots, I show the performance as a function of occupancy (hit density, hits/cm^2) for several optional layouts using my probability formulas.

• BLACK: long strips only
• GREEN: half-length strips only
• RED: GN pad layout
• BLUE: IK pad layout

Plot of probability logarithmic in occupancy occupancy to emphasize low-medium occupancy performance:

What we see in the above plots is:

1. The pads+strips show some improvement over the strips alone.
2. The IK pad layout improves hit matching with respect to the GN pad layout.
3. The half-length strips are better than all other proposed options (essentially an approach towards a pixel detector).

Given point 2 above, I have also done a study of what pad configuration gives the best performance. I define f to be an integer describing how many pad lengths it takes to equal one strip length. The minimum for f is set to be such that the pad width is larger than the track projection resolution in the r-phi direction, because I do not trust my calculation at or below that value. The result is that the probability functions prefer the smallest f possible (the probabilities are monotonically descreasing with increasing f at a fixed occupancy), but that the probabilities do not exceed the half-length strip probabilities unless I take the calculations to small enough f that I no longer trust them using my pad+strip formula.