Whaddaya mean "High Density Tracking"?

Moving the EOS TPC from the Bevalac (~1 A GeV beams) to the (2-10 A GeV beams) means an increase in track density. Indeed, the track densities in this experiment surpass those of STAR, which has events that look like this!.
This is because E895 is fixed-target and the TPC sits closer to the interaction.

Well, the hardware came "free" (once it was built for the EOS experiment), so it is up to the software to worry about doing a more difficult job.

Some relevant numbers

The "standard approach" to TPC event reconstruction is to find "hits" on parallel planes defined by a padrow and the time direction. These planes are oriented so that tracks go roughly perpendicular to them.

Each "plane" is the thickness of a pad, which is 1.2 cm for the EOS TPC. The planes are built of 2d pixels of size 0.55 cm x 0.8 cm, or 2.27 pixel/cm^2. This size is based on the intrinsic "size" of the hit, and how it grows with diffusion.

The design of the 2d hitfinder in E895 is such that a hit starts with a "significant peak," and a peak's significance is based on a peak-valley cut. Thus, there must be a valley in the 2d pixel space between two hits-- the closest hit cannot be "on" the next pad or time bucket. (Of course, hits are not "on" pads or buckets; see the other 2dh pages or Stony Brook pages for details.) This means that the maximum hit density is 1 hit/((4 x 0.55 cm) x (4 x 0.8 cm)) = 0.14 hits/cm^2. Keep in mind that even this modest density is achieved only in the theoretical limit where the tracks conspire to arrange themselves perfectly. For the first few padrows in particular, this value is strongly exceeded in the zero-degree cone.

Track and hit densities encountered in E895

Click here for a tour of the 2 A GeV track (top pictures) and hit (bottom pictures) densities, in 1/cm^2.
Clearly, even past the center of the TPC, the track density is higher than a hitfinder can handle, with our pixel size.

Where does our Hitfinder poop out?

This is an important question, because if the track density reaches an order of magnitude higher than the theoretical maximum hit density (see above), then we are only asking the tracker to screw up by giving it essentially random hits in the high density cone.
Therefore, we will want to define a high-density cone of exclusion where hits are thrown away before going to tracking. It is like the old streamer chamber trick of masking off part of the detector from the cameras.

To decide, look at these pictures. They show contour plots of the track density and hit density overlaid, with contour lines in different color. You cannot tell which is the track density and which is the hit, but you can see that the two start to differ at track densities around or a little higher than 0.05 tracks/cm**2 (which is the green contour). E.g. although the track density hits 0.1 (blue lines), the hit density never gets that high (see also the "tour" above).

Back to E895 Hitfinder Main Page