Since the most challenging charged particle search is for charge +1 objects, we
simulate the apparatus with the spectrometer magnets set to +1.5T (positive
strangelet search).
To study the calorimeter, we begin with events generated as described
above in the section on HIJET/GEANT. To measure the background rejection,
8000 full central collisions with particles from the shielding overlayed
were used. To study the efficiency, 15 GeV/c 
strangelets
produced according to the single particle production model described above
were overlayed on a small sample of these events.
The impact point, time of flight and
four vector of every particle striking the calorimeter was available.
The calorimeter response was simulated in two ways. First we used the measured (average) transverse shape of hadronic showers in the SPACAL calorimeter and an analytic expression for the longitudinal shower shape. These average distributions were used in a way that simulated some of the shower fluctuations about the average behavior. We do not discuss this in detail because the second approach is easier to explain and gave essentially the same results.
The second method used the program CALSIM to simulate showers in a simplified
calorimeter geometry (transverse lead and scintillator plates with
the correct volume ratio of 4 to 1, and with the 10 cm by 10 cm tower geometry
we will use). In addition, the timing of the calorimeter signal at each
phototube was calculated as if the scintillators were running in the beam
direction. This simulation also made other simplifications such as ignoring
energy deposit mechanisms other than fast particle energy loss and stopping
protons. Thresholds for neutral particle tracking were 2.5 MeV for gammas and
10 MeV for neutrons. Energy deposits from particles below the tracking
threshold account for 15% of the shower energy. The resulting simulation is
certainly not adequate for designing calorimeters, but it should be quite
adequate for evaluating the effects of resolution and fluctuations. The
simulation showed a resolution of 
.
A library of 300 shower simulations (100 at each of three energies: 2.5 GeV, 5.0 GeV, and 10.0 GeV for the incident neutrons) was generated and used to calculate the calorimeter phototube signals from the entire array of particles in a HIJET/GEANT event striking the calorimeter. For any given hadron energy the nearest library energy was chosen and the library outputs were scaled to give the same total energy deposit as the given hadron would have done. Before using the signals in the analysis, the pulse heights were smeared by an additional 4%. This was done to lessen any effects due to the repeated use of the same showers from the library. It was assumed that the contributions to the signal of any tower due to more than one shower simply added linearly.
For each event we generated a set of tower energy deposits and a set of times (a ``software'' discriminator was attached to each tower and the time at which the phototube detected an energy deposit greater than .33 GeV was determined). The timing was accurately simulated, taking into account the propagation time of light in the scintillator, the arrival times of all the particles striking the tower, the velocities of the shower particles, etc. All the particles striking the calorimeter contributed to the set of tower energies and times.