
Previous page	Up to the index	Next page

BOWING OF THE WHOLE ASSEMBLY

Under the influence of humidity increases, rohacell tends to expand. Since the cages are restrained by the glued-on silicon panels, and not all faces have silicon on them, there is a tendency of the entire 12-cage assembly to bow out in the direction of the non-populated faces (up and out).

In these runs this effect was simulated by placing the sensitive silicon volumes at different radii and also tilting them slightly so they form an arc in the r-z plane. The inner and outer silicon were displaced the same amount, so that their relative spacing was not affected. Because of the Geant geometry structure, each 60-degree sector (for now) gets the same bowing.

(postscript version here)
The picture above shows a edge-on view of one row of outer silicon panels. The y scale is greatly exaggerated to show the curvature. In reality the vertical height of the arc is 1mm and the horizontal dimension is 640mm.
The postscript version shows a little more of this view. On can also see the inner silicon layer, bowed out just like the outer one, and the panels which are not seen edge-on still display some bowing in this projection.

This plot summarizes the vertex-finding efficiency as a function of maximum bow-out of the 12-cage assembly.
Dotted lines indicate a 5% and 10% losses in efficiency, for maximum bowing of 0.1 and 0.2 mm, respectively.
(postscript version)