This page collects together some of my work on the ITk layout.
Traditionally, ATLAS geometries are built up in C++ code creating a GeoModel graph, with dimensions stored in the geometry database. I have developed a way of storing the geometry in an xml file with a new Athena package "GeoModelXml" to parse it and create the GeoModel tree. Dimensions, transformations, materials, etc. are all defined in the xml. GeoModelXml can also deal with sensitive volumes, alignable volumes, and offline identifiers.
The manual for the package GeoModelXml is here.
There is also a guide to developing strip layouts with GeoModelXml.
GeoModelXml is available in gitlab.
The athena package SCT_GeoModelXml uses GeoModelXml to implement the Strip layout.
The xml files are located in the athena package ItkLayouts, wherabouts not known to me.
Services have important effects on the performance of a layout. A services model has been built up using a program called servX0. and the program is available at https://svnweb.cern.ch/trac/atlasusr/browser/hessey/servX0 SVN.
The Strawman layout document, 2007.
The Utopia-layout parameter book, 2010.
I have written up the basic formulae for endcap sensors with stereo angle built in.
It is useful to know how many rows of barrel modules are needed at a given radius to give hermetic cover. This depends on sensor size, tilt angle, and overlap required. I have made a mathematical model with a spreadsheet that applies the model. These can be used as a guide when designing barrel layers.
Similarly it is useful to know how much overlap is needed at a given radius to give hermetic cover in the r-z plane. This depends on sensor size, barrel radius, distance between high and low modules, etc. From knowledge of overlaps for example, the total stave length is determined. Then the endcap disc positions can be designed into the layout. Parametrising these allows changes to be easily explored. The formulae are given here and can be used e.g. in a spreadsheet as a guide when designing barrel layers.
In the end-caps we need overlap in the r-phi direction, to make the detector hermetic for low momentum tracks. The formulae are given here and a there is a spreadsheet implementing the formulae.
I have written a program "idres" to calculate the error matrix for the five helix track parameters, given a (position dependent) magnetic field, range of eta angles, and range of track momenta. The tracker is specified as any number of cylinders and discs, with individually-settable radiation lengths and resolutions (precision direction and second-coordinate). The program currently stores the B, eta, pt, and square-root of the leading-diagonal elements of the error matrix for further processing, by packages such as R or root.
More details here
I have tested idres on the current ATLAS inner tracker; the comparison is very good (typically within 10~\% agreement with full atlas simulation). The report is available.