Figure 2 shows the effects of removing the lead and neutron shields and of removing the shields and the nosecone. The portions of the histograms that correspond to valid trigger-roads are shown cross-hatched. Note that 69 of the 100 events produce no tracks to plane 3 and 22 events produce only one track to plane 3.
As indicated by the arrows in the figure, removing the shields has only a slight effect, increasing the trigger rate from 9 to 10 %. However, removing the nosecone has a significant effect, increasing the trigger rate to 35 % or, assuming a Au + Au central collision rate of 1.4 kHz, it gives a trigger rate of 490 Hz.
Figure 3: Distribution of trigger roads for 500 UA1 Events.
The histograms on the left and right are the same, except that the
one on the right shows only the cases where the number of trigger
roads equals at least 2. Both the lead and neutron shields were removed
for the PISA run used to generate this data.
Figure 3 shows the effects of turning ``on'' the hodoscopic
ambiguities. This is shown for 500
UA1 events run with the nosecone installed
but with both shields omitted. The results for ambiguities ``off'' is shown
by the solid lines and the results for ambiguities ``on'' is shown by the
dashed lines. The set of histograms on the left and right hand sides of the
figure are the same except the right hand side shows the distribution of
roads for which the number of roads is 
. For the 500 event case
the number of trigger roads with ambiguities ``off'' is 9 % and with
ambiguities ``on'' the rate increases only slightly to approximately 11 %.
Figure 4: Distribution of trigger roads for 100
J/Psi events.
The two upper histograms show the same results except that the one on the
right shows only the distribution where there are at least two roads.
A similar remark applies to the two lower histograms. In all cases the
nosecone is installed and hodoscopic ambiguities are ``on''.
The results obtained for 100 J/Psi events is shown in Fig. 4. The momentum and angle cuts specified for the event generator should guarantee 100% trigger efficiency. However, for the case where the shields are installed, two of the 100 events produce only one track reaching to plane 3, presumably because multiple scattering has caused two of the tracks to be deflected to the piston or the lampshade magnet before reaching the third plane of the muon identifier. Therefore, with the shields installed the trigger rate is 98% as indicated in the upper right-hand section of Fig. 4.
When the shields are removed, however, all particles produce tracks reaching to plane 3 so that the trigger rate is 100% as indicated in the lower right hand section of the figure.
In Fig. 5 we show the result of superimposing the UA1 and J/Psi events using the MERGE facility of PISORP. Now, the superposition of the UA1 and J/Psi tracks produces a trigger rate of 100% in all cases.
Figure 5: Same as Fig. 3 but with each event being a superposition
of one UA1 and one
J/Psi event. The dashed histograms are the UA1
event distributions. They are identical to the superimposed
distributions but
shifted to the left by two roads.
Hodoscopic ambiguities should of course always normally be ``on''. We have looked at the ``off'' case to show how performance would vary if outside information (such as timing across a tube to provide crude horizontal resolution) were used to resolve hodoscopic ambiguities. Of course, such information would not be available to a LVL-1 trigger. Moreover, hodoscopic ambiguities will be ``broken'' or resolved by means of extrapolation from the muon tracker both offline and in the LVL-2 trigger.
Finally, it is recognized that this study effectively utilizes on the order of 10,000 trigger road-starting x-y intersections. This is too costly and not necessary. Subsequent simulations have shown (at the October 6 Muon Arm meeting, for instance) that requiring at least two x-view trigger roads and at least two y-view trigger roads provides a more feasible LVL-1 algorithm requiring only 744 trigger roads.