[SBS_sim] Replay results with tighter cut

[Ole Hansen]

Hi all,

attached is another set of replay results, this time with slightly
different replay parameters and, more importantly, a tight cut to select
a "successfully reconstructed track".  I think the definition of
tracking efficiency I used in the prior message was too loose. A track
is efficiently reconstructed if it matches the true track, as generated
by Monte Carlo, as closely as possible. To select such tracks, I now use
a position cut of 6x6 mm on both the first and the last readout plane,
i.e. a track that, within a generous allowance for resolution, passes
straight through the center of the detector. (Recall that the MC
currently generates all tracks at (0,0) with angles (0,0).)

As you can see, the tracking efficiency defined in this way is lower
than what I quoted yesterday.

Clearly, there is work to be done to track down the source of the
obviously excessive loss of tracks.  My prime suspect is the strip
clustering algorithm in the decoder.

At the meeting this morning, I asked Evaristo to generate (a) another
production with 1/1 background, but 7 x/y GEM planes (evenly spaced),
and (b) a run with NO background (to study the effect of multiple
scattering on a perfectly clean track).  We will also try to include a
more complete set of Monte Carlo truth information into the production
ROOT files to allow us to study the strip clustering and its possible
failure mode in detail.

BTW, we now have a mailing list. This is the first message to the list.

Regards,
Ole
 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: results_v1.txt
Url: https://mailman.jlab.org/pipermail/sbs_sim/attachments/20101207/ae636b99/attachment.txt