[Olympus] OLYMPUS main tracker upgrade

[Alexander Kisselev]

  Dear colleagues,

  Doug pointed out to me yesterday (I admit I overlooked 
this issue in the TRD draft) that main tracker upgrade will 
consist of a *single* 2D XY-plane GEM in each arm, meaning 0
redundancy design. Also, if I get it right, foils will be 
patchy, meaning some (small?) dead zones. 

  Such low-redundancy setups may become problematic in many 
respects (although perhaps not all of them apply to GEM/APV 
configuration, I have no experience with this hardware):

  - track finding and ghost track rejection is very sensitive to 
    background hits; in case of noisy/dead channels/planes it 
    needs to fall back to less efficient "low-resolution" mode 
    provided by other detectors (existing wire chambers in our case);

  - tracking resolutions for cases when GEM detectors have no hits
    in X-layers will be as bad as before, so data analysis will have 
    to combine both low- and high-resolution tracks, which is a pain;

  In my view, if we really tend to use this upgrade to reduce operating 
costs (obtain much better resolution --> suppress background --> lower 
down magnet current at least for the last 2 months of operation), we 
should better take care that there are usable hits in GEM tracker for 
~all tracks under any conditions.

  Ideally we would probably need to at least double number of 
registering planes within the existing design. If this is excluded 
for financial reasons, perhaps consider to install say 2x2D planes in 
each arm, but with bigger strip spacing (thus maintain comparable number
of channels). Or at the very least to install suggested single 2D layers 
at +/-45 degrees instead of XY-orientation, so that each side (being 
combined with the beam line constraint) delivers a measurement in 
horizontal plane, where exactly we have a problem.

  Regards,
    Alexander.