I think I found one problem with the simulation/reconstruction. The code requires
(fabs(bsut0-bsuiter->getHit()->getTDCTime()) < delta)where bsut0 is generally equal to 165 nsec, delta is about 17 nsec, and, for simulation, the TDC time averages 186 with an RMS of 4.4
Clearly we lose about half that way.
I'm not interested in increasing delta, since that seems to be working OK with real data. I need to pull the simulated BSU time back down to something sensible. Since the BSU and TSU go together, the TSU is probably a little out of line also.
Raw data peaks at about 210, with a broad peak at 70 and a spike at about 165. I think that spike is the real thing, though statistics are still low. That means that I need to reduce the values coming out of simulation.
A little more detail shows what looks like a double peak between 160 and 165 in real data (Bob found that East and West are a little different).
I put a hack in MuonSim/src/ImuEventdata.cc to reduce the BSU and TSU raw times by 23 nsec. (188->165) That should bring the distribution down near the t0. Not tested yet.
10:17 13-March-2007
I ran 200 events with multiple muons with the original timing, and got 31 events with stublets, with a total of 39 stublets.
I then ran 200 events with the same multiple muons, subtracting 23 nsec in the simulation. The reconstruction found 115 events with stublets, from which 167 stublets were found. No eta information included here.
This change was committed to the repository.
Modified 13-March-2007 at 10:19
http://hep.physics.wisc.edu/~jnb/imu/12Mar2007