ACD Geometry

Contents

1. Features of ACD geometry in recent GlastRelease tags
2. To-do list
3. Pointers to other documents

Features of ACD geometry in use (xmlGeoDbs v1r46p2)

• Bent tiles are modeled with two boxes
• Tile sizes and shapes are accurate except that trapezoidal side tiles are approximated by inscribed boxes
• Screws and screw holes (accurate positions) are implemented for top tiles
• Shingling (overlap; orientation angle for side rows) is accurately modeled.

• Ribbon sizes, shapes and positions are reasonably accurate except that, of course, each ribbon has to be modeled as many pieces

To-do

• Trapezoidal prisms for side tiles in rows 1 and 2.  The geometry description is done (as of tag xmlGeoDbs  v1r47, created 29 October) and Acd analysis code has been upgraded to support it, but it has not been systematically tested. The xmlGeoDbs tag has not yet been incorporated in a GlastRelease HEAD build.
• Screws and screw holes for side tiles.  Awaiting accurate position information
• Add significant missing mass for BEA

Other documents

See also links to various information concerning the ACD geometry and child pages below, especially the most recent ones.

Geometry status as of January 2006

Geometry status and tests of simulation before DC2 (Oct. 2005)

Old Home of ACD Geometry containing very old information. 

Known differences between ACD model in GlastRelease and Reality

• Overlaps between top tiles go in the wrong direction
• Screws are not simulated
• Curvature of top tiles is not modeled. We have tentatively agreed with the ACD to simulate this through the introduction of a vertical box.
• Shingling along the sides is not included.

Testing the leakage of the ACD geometry as currently simulated in GlastRelease - Oct2005

Outcome of conference call between Bill Atwood, Heather Kelly, Alex Moiseev, and Steve Ritz on Oct 19, 2005

Before proceeding with large scale background rejection studies, Bill needs to be assured that the ACD geometry model as simulated in GlastRelease is worst case in terms of leakage. It is strongly believed that Alex's Geant3 simulation is more accurate. Interested in demonstrating that when the ACD model in GlastRelease is "improved" to adhere to the full design, that new sources of leakage will not be introduced.

The Plan

Provide a direct comparison between Alex's Geant3 simulation and what is currently in use in GlastRelease. If we can demonstrate that the GlastRelease model is the same or slightly worse than Alex's Geant3 simulation, Bill will be satisfied. We spent the rest of the call outlining the details of a run to be performed using both simulations. Alex will forward a final set of details via email.