PBG1 Trigger Efficiencies
The PBG1 part of the neutral trigger is designed to tag events with large
back to back energy deposits in the CCAL. The signals from the 1280 lead glass
blocks first go through the level 1 summers where for each ring, nine
consecutive blocks' signals are summed. The first and last block in the sum are
treated as overlap blocks, i.e. they are counted in the sum of two adjacent
sections of nine blocks. For example, in ring 1, the signals from blocks 1 to 9
are summed, as are 9 to 17, etc., and finally, blocks 57 to 64 back to (and
including) block 1. Thus, 160 signals leave the level one summers (8 from each
ring). These signals then split; one set forms the MLU strobe and the other
set of 160 signals enters the level 2 summers.
At the level 2 summers, the signal from each octant of blocks is summed
with the same octant in five other predetermined rings. For example, octant 1
(wedges 1 to 9) from rings 1 to 4 are summed, from 4 to 8, etc., and finally,
from 16 to 20. Notice that the first sum has only 4 signals and that there is
again an overlap (of rings this time). The 40 signals that leave the level 2
summers are the signals for the 40 Super Blocks of the CCAL (table 1).
Table 1
Super Rings
|
---|
Super Ring # | Ring #'s |
---|
1 | 1 to 4
|
2 | 4 to 8
|
3 | 8 to 12
|
4 | 12 to 16
|
5 | 16 to 20
|
Super Wedges
|
---|
Super Wedge # | Wedge #'s |
---|
1 | 1 to 9
|
2 | 9 to 17
|
3 | 17 to 25
|
4 | 25 to 33
|
5 | 33 to 41
|
6 | 41 to 49
|
7 | 49 to 57
|
8 | 57 to 64 to 1
|
After level 2 the 40 signals split. One set goes to ADC's (hereafter
called PBG ADC's) and the other set goes to an integrator with a clipping
cable. After the integrator the signals enter a discriminator with thresholds
set according to pi0pi0 two-body decays. The thresholds for each
Super Ring are set to 60% of
the lab energy that a pi0 from ppbar -> pi0pi0 would have in that Super Ring.
The 40 signals split again after the discriminator; one set going to TDC's
(PBG TDC's) and the other set going to the PBG OR. The PBG OR looks for signals
(that have passed the hardware thresholds cut) within a Super Wedge. If at
least one of the five Super Blocks' signals in a Super Wedge has passed the
threshold cut, a signal is sent to the Neutral Memory Lookup Unit (NMLU). The
NMLU has bits 0 to 7 reserved for the outputs of the PBG OR. The NMLU then
looks for signals in back to back Super Wedges, i.e. bits 0 and 4, 1 and 5, 2
and 6, 3 and 7. If one or more is found, and the bits for H1*H2' OR and Fch OR from
the charged trigger are not set (neutral veto), the PBG1 bit is set in the Master
Memory Lookup Unit (MMLU).
To calculate the PBG1 efficiencies, the signals from the 1280 CCAL blocks
are used to duplicate offline what happens in the hardware during data taking.
Only Etotlo data was used.
The first task is to sum the energies from the 1280 blocks into the 40 Super
Blocks (as is done by the level 1 and 2 summers). At this point the threshold
values are calculated as a check on the hardware set value as well as to check
for consistency among Super Blocks in as Super Ring. Dividing the Super Block's
energy distribution when there is a PBG TDC hit (i.e. the signal has passed the
discriminator) by the complete energy distribution gives a threshold curve
(figures 1-4). The curve is fit to a hyperbolic tangeant function of the form,
y = 0.5 + 0.5 tanh(par(2)*(x-par(1))), where par(1) is the mean. The error bars
are based on a binomial distribution, error = sqrt(np(1-p)). The set
discriminator threshold value is compared to the generated curve. Generally it
is considered acceptable if it falls within the 10% and 90% levels of the
threshold curve.
Figures 1-4
For efficiency calculations real physics events must be used and
accordingly, pi0pi0 events from the Etotlo data stream were chosen. The events
were chosen with the following selection:
Once the pi0pi0 events were selected the
procedure outlined above for threshold determination was followed. That is, the
energies for the Super Blocks were found and compared with the set
discriminator value if the value was in agreement with the calculated value
(note that there were no major discrepencies for any runs). If the energy for
some Super Block in a Super Wedge was greater than the threshold for its Super
Ring and the opposing Super Wedge had a Super Block above its Super Ring
threshold, a pbg1 software trigger was recorded. A hardware trigger is present
when the MMLU input bit 1 was set. The calculated efficiency is a ratio of the number
of events with both the hardware and software trigger to the number of events
with a software trigger only.