The energy, range, and decay sequence of charged particles are measured
in the range stack.
Charged particles which emerge from the drift chamber outer circumference are
stopped in a cylindrical array of layers of plastic scintillators
segmented azimuthally into 24 sectors.
Each range-stack sector consists of 20 layers of
1.82-m-long 
19.05-mm-thick plastic
scintillation counters which were fabricated of BC408 by Bicron Corp.
(Newbury OH, USA). An additional innermost
layer, 520-mm long 
6.35-mm thick (T counter),
is used in the trigger to define
the 
acceptance of the range stack as viewed from the target,
coinciding roughly with
that of the drift chamber. The scintillator
material was chosen for its good light output,
intrinsic short pulse width (
ns),
and reasonable attenuation length.
Two layers of range-stack multi-wire proportional chambers
(RSPC) [14] are located following layers 10 and 14, as shown in
Fig. 10 in order to obtain tracking information within the range stack.
Figure 10: Schematic end view of one range-stack sector with the
corresponding barrel photon veto modules at the outer radius (top). The
A, B, C superlayers and the positions of the proportional chambers
are indicated.
Overall, the range-stack sectors fill the radial region between 451 mm and 896 mm.
The range-stack system is supported in the magnet by a pair of welded stainless-steel web-like frames at the upstream and downstream ends of the magnet volume, fastened to the center section of magnet yoke indirectly through the support frame for the barrel photon veto. Each frame extends towards the center plane of the detector up to the axial extent of the drift chamber, approximately, at the inner diameter, and tapers back in a conical shape towards the outer diameter. This design provides adequate support without introducing inert material into the fiducial region.
The twenty 19-mm-thick layers form the stopping region of the range stack.
The nine innermost counters are grouped into superlayers A (4-fold),
B (3-fold), and C (2-fold), as shown in Fig. 10, each viewed by a
single PMT at each end. This arrangement simplified the problem of
fitting the PMTs into the limited space, but at the cost of range resolution
for charged particles that stop in this region.
In the strategy to restrict the search for 
to the kinematic region
above the 
peak, where the 
s of interest
have a range in scintillator of >25 g cm
,
range resolution is not critical in the inner 
cm. The remaining eleven
layers are each viewed by a PMT at each end, for a total of
720 PMTs on the entire range stack, including the T-layer.
Scintillation light is brought out of the magnet
at both ends of the detector by UVT acrylic lightguides
made in two sections. The 900-mm-long inner straight sections were glued
to the scintillator layers before they were stacked,
with 25-
m-thick Al foil between layers,
in radial groups to provide for the gaps to be occupied by
the RSPCs. Each radial
group was wrapped tightly with Kapton tape to support it as a unit to be
slid into the frame. The lightguide sections were wrapped loosely
with black polyester tape and fitted with shims for alignment and support.
The inner sections pass through the gaps between
the flux-return spokes of the magnet
endplates and each is joined to the outer section through an air gap.
The outer section of each lightguide is curved to join with
the PMT array where it is glued to the PMT.
EMI 9954KB PMTs are used for the stopping layers and 28-mm-diameter
Hamamatsu R1398 for the T layer.
The light output at each end produces approximately 15 photoelectrons
in the PMT per MeV deposited in the scintillator.
PMT signals are passively split and recorded by
ADCs, TDCs, and by the TD system. The latter provides a
pulse-shape history in each tube as well as precision timing
(
ps). The signals fed to the TDs from the A, B, C
counters are
processed through amplifiers with logarithmic transfer functions
in order to compress the larger dynamic range of analog pulses
from these thick layers.
Signals from the range stack form the basis for most low-level trigger
decisions.
Individually discriminated range-stack PMT signals
provide information for fast pattern recognition.
For example, the stopping layer determined from the deepest
layer struck provides a coarse measure of the
range of a charged track. About 80% of 
events are rejected
by vetoing on the long range muon which tends to penetrate deeper than
pions from the 
endpoint. In addition, the pulse heights from the
range stack are summed in hextants;
discriminated sums from hextants not
involved in a charged-particle track are used to derive veto signals
to augment the photon detection systems.
The range-stack energy calibration was done for each end separately, using
muons from 
triggers, by comparison to a Monte Carlo simulation.
Attenuation effects were determined from
cosmic-ray data taken with zero magnetic field.
Drift-chamber tracks were extrapolated into
the range stack to get the z position and the attenuation length
was adjusted so that the z positions determined
from them agree with the DC extrapolation. The average attenuation length
determined this way was 2.07 m. However, the effects of the exponential
attenuation can be eliminated by taking the geometric mean of the
pulse heights from both ends of the counter.
The range-stack calibration was monitored
using a prescaled sample of 
triggers taken during running,
and high voltages were adjusted as necessary to compensate for changes.
From the calibrated ADC information, energy and range measurements
for 
from 
and 
from 
decays
are shown in Fig. 11.
Figure 11: The (a) range and (b) energy for 
and
events from the
range stack, corrected for losses in the target and other materials in
the trajectory.
The range determination uses polar-angle information from the drift chamber and a correction based on the energy measurement in the last counter hit to establish the range contribution from the stopping layer. Both energy and range are corrected for the target contribution.
The two layers of multi-wire proportional chambers (RSPC) in each
sector are located radially from the detector axis at approximately 635 mm and
720 mm. These chambers measure the axial (z) and
azimuthal (
) positions of the track
in order to refine the range measurement for pions in the off-line analysis.
In addition, the z information from the inner RSPC layer is
used to give the slope
of the track in order to improve the range determination
in the second trigger level for online event selection.
Fig. 12
Figure 12: Schematic showing the serpentine cathode and anode wire layout
of the RSPCs. The dimensions shown are approximate for the inner layer.
is a schematic of the design of one of the chambers.
To minimize the introduction of inert material and maintain
full azimuthal coverage, the RSPC frames were constructed
of very light, thin boxes molded from Kevlar cloth in epoxy.
Each chamber is approximately 11-mm thick
with a total material thickness of 0.315 g cm
.
The active area of the inner (outer) chambers is 926 (1067)-mm long
by 160 (180)-mm wide.
Electrical components are epoxied and soldered in place. The 25-
m-diameter
Au-plated tungsten anode
wires are strung 6-mm apart along the z direction.
The 
measurement is obtained
from the position of the struck wire using a coupled
delay-line readout scheme, similar to that of ref. [15]. The maximum drift time
in the 9-mm gap is 
ns using a gas mixture of
argon : isobutane : methylal
of 65.2 : 26.4 : 8.4 and high voltage of 2.5 kV.
The z measurement is derived from the
end-to-end time difference of signals from the serpetine-patterned copper
cathode etched onto G10, making up a 150-ns-long transmission line.
The system of 48 chambers is read out by two 96-channel
LRS 1879 Fastbus TDC modules
in addition to the fast digitizing second-level trigger.
The position resolutions (
) are roughly 10 mm in z and
2 mm in 
.