After integration and test at SLAC and GSFC, BFEM was shipped to the National Scientific Balloon Facility (NSBF) at Palestine, Texas. The experiment was performed on August 4, 2001, and BFEM was successfully launched. When three consecutive pairs of

1
T. Mizuno (Hiroshima University), T. Kamae, H.
Tajima, T. Handa, T. Lindner (SLAC), H.
Mizushima, S. Ogata, Y. Fukazawa (Hiroshima
University), M. Ozaki (ISAS), P. Valtersson, M.
Roterman, N. Karlsson (Royal Institute of
Technology and SLAC), H. Kelly (GSFC) and GLAST
BFEM team
Flight Operation
Figure 5
Figure 6
Figure 4

• After integration and test at SLAC and GSFC, BFEM
  was shipped to the National Scientific Balloon
  Facility (NSBF) at Palestine, Texas. The
  experiment was performed on August 4, 2001, and
  BFEM was successfully launched. When three
  consecutive pairs of x-and-y layers are
  registered to be hit, the event is recorded
  (called Level-1 Trigger).
• BFEM with gondola, hanging from NSBFs Tiny
  Tim, was waiting to be launched.
• BFEM altitude as a function of time. The balloon
  carried the instruments to an altitude of 38 km
  in about two hours, and achieved three hours of
  level flight (until the balloon reached the limit
  of telemetry from NSBF). During the ascent, the
  internal pressure went down to 0.14 atmosphere
  due to a small leak from the PV and we failed to
  record the data to disk after that time. However,
  a random sample of triggered data was
  continuously obtained via telemetry, with about
  200 kbits s-1 or 12 events s-1. We recorded more
  than 105 telemetry events during the level
  flight. Over 14 million events were also
  recorded to disk during the ascent before the
  disk failed.
• Level 1 Trigger rate as a function of atmospheric
  depth (the so-called growth curve). The maximum
  count rate is about 1.2 kHz, still below the
  capacity of the data acquisition system. At the
  float altitude (3.8 g cm-2), the rate was about
  500 Hz.
• The distribution of layers causing the trigger.
  During the data analysis, we found that three of
  26 layers did not participate in the trigger. We
  took this into account in the simulation and well
  reproduced the hit distribution in the TKR for
  charged events (see below).

Figure 7
Comparison of data with simulation
Table 1 Observed event rate and that predicted
by simulation
Proton alpha e-/e gamma mu-/mu total
charged event rate (450Hz observed) 148 Hz 19 Hz 47/29 Hz 52 Hz 76/38 Hz 409 Hz
neutral event rate (50Hz observed) 3.4 Hz 0.0 Hz 6.1/3.9 Hz 35.8 Hz 2.7/1.2 Hz 53.1 Hz
In order to obtain a reliable cosmic-ray
background model, we compared the BFEM data
during the level flight with simulation
prediction. We classified events into two types,
charged and neutral, and made separate detailed
comparisons. Here, charged events are those in
which one or more ACD tiles have energy
deposition above 0.2 MIP. The observed count rate
for charged and neutral events is shown in Table
1 above, and agrees with the simulation
prediction within 10. We also compared the hit
distributions in the TKR (Fig. 8-11). Although
the normalization differs by 10-15 between the
data and simulation, the shape of the count rate
of each layer for charged events is reproduced
well (Fig.8a). We adjusted the cosmic-ray model
within uncertainty, and the hit distributions in
the TKR show good agreement for charged events
(panels b of Fig.8-10). The neutral events,
originating mostly from gamma-rays, still need to
be investigated (Fig. 11).
(b)
(a)
(b)
(a)
Run55(level flight)
Simulation
muon( and -)
gamma
Run55(level flight)
muon( and -)
gamma
e-/e
Simulation
e-/e
alpha
alpha
proton
proton
Top of TKR
Calorimeter

• Figure 8 Distribution of count rate of each
  layer for charged events.
• Comparison between the data and simulation with
  contribution of each particle type. The shape of
  the distribution is well reproduced.
• The same as panel a, but the primary proton flux
  is increased by 15 and the spectral slope of
  e-/e is assumed to be E-1.5 instead of E-1.0
  below 100 MeV.

• Figure 9 Distribution of Top-most hit layer for
  charged events.
• Comparison between the data and simulation with
  contribution of each particle type.
• The same as panel a, but the primary proton flux
  and the spectral slope of e-/e below 100 MeV are
  modified.

(a)
(a)
(b)
(b)
Run55(level flight)
muon( and -)
Simulation
muons
gamma
gamma
Run55(level flight)
e-/e
Simulation
alpha
e-/e
proton
proton

• Figure 10 Distribution of the number of layers
  with hits for charged events.
• Comparison between the data and simulation with
  contribution of each particle type.
• The same as panel a, but the primary proton flux
  and the spectral slope of e-/e below 100 MeV are
  modified.

• Figure 11 Distribution of count rate of each
  layer for neutral events.
• Comparison between the data and simulation with
  contribution of each particle type.
• The same as panel a, but the primary proton flux
  and the spectral slope of e-/e below 100 MeV are
  modified. Discrepancy in layers above 10 needs to
  be resolved.