The knowledge of the origin of a decay in the 
or 
beam relies on the information provided by the tagging counter
located upstream of the target. The or assignment
is obtained by measuring the time difference between the passage
of the proton in the tagging counter and the event time in the detector. Events
with a time difference inside a given interval are called " ", any
other events are called " ". Inefficiencies in the tagging counter
such as misalignment and deadtime cause decays to be identified
as . On the other hand, any accidental hit in the tagging counter
produces a to transition. Since these misidentifications
are decay-mode independent, they only lead to a dilution of
.
The proton tagging detector has been
designed to cope with proton rates above
10MHz and to provide a detection efficiency close to
100 
combined with a time resolution better than 500ps[2]. It
consists of two sets of twelve staggered scintillator foils arranged
alternately in the horizontal and vertical directions. All foils have a
thickness of 4mm. The width of the foils vary between 200 
m near
the beam axis up to 3000 m for the scintillators located at the beam edges
so that each counter sees only a fraction of the beam.
To ensure excellent geometrical efficiency, an overlap of 50 m is kept
from one counter to the next one in each projection. The horizontal foils
are made of NE-102a scintillator whereas the vertical ones are made
of BC-418 scintillator.
The scintillator foils are mounted on a carbon fiber
support structure which has a mechanical precision better than 40 m.
The scintillation light is transmitted to R-2076 Hamamatsu
photomultipliers via alumnized mylar envelopes and lucite light guides. The
light yield obtained with beam particles varies between 300 and 400 photo-
electrons.
The precise positioning of the tagging system with respect to the beam axis
is achieved by using
four independent remote-controlled motors for horizontal and
vertical displacements and for rotations around the horizontal
and vertical axis. Alignment accuracy of 200 rad can be reached
during beam scans
by looking at the tagger multiplicity
distribution normalized to the number of protons that hit only one counter.
The signals of the individual counters are digitized by two 8-bit 480Mz
FADC chips running in interleaved mode. The average time resolution of a
single
counter was measured to be about 180ps and the double-pulse separation
is below 5ns, yielding at 20MHz a dead time of less than 1 for the proton
detection.
Figure 2 shows the time difference distribution of two horizontal counters
when the tagging counter is slightly tilted to let the proton beam hit
several foils.
Figure 2: Time difference distribution between tagging counters H1 and H2.