MVD sagging and bowing

We know that the MVD is both sagging (down in the middle) and bowing (out in the middle). Hubert created this page to describe the problem. It includes links to some code to put this into Pisa. Currently, the offline code assumes the whole detector is 5 mm low. The truth is more complex.

Here are two sketches summarizing my notes of measurements of sagging and bowing:

We do not yet have code to correctly deal with this problem in the offline software. For year-3, I would like to minimize the sagging and bowing mechanically. I'm not sure how to do this -- a stiffening bar? Pieces of thread strung to the top of the MVD? This should be done while reassembling the MVD. This postscript plot, taken from Jan's drawings shows an assembly diagram of half the MVD. Something that connected the C-cages to the tube on top of the MVD might be able to reduce the sagging problem.