The Broad RAnge Hadron Magnetic Spectrometer experiment (BRAHMS) is designed to measure charged hadrons over a wide range of rapidity and transverse momentum for all available beams and energies in experiments at RHIC (Relativistic Heavy Ion Collider), Brookhaven National Laboratory. Figure 1 shows a perspective view of BRAHMS detector system. One set of these detectors, called Beam--Beam Counters is employed to characterize collisions from a global perspective. As one can see on this picture, they are located some distance backwards and forwards of the assumed collision vertex (marked pink on this picture). Beam-beam Counters are designed to provide an initial trigger, permit a rapid estimate for the collision point (vertex location) and serve as a first probe of charged multiplicity.

First, detectors were designed and tested using computer simulations with the help of the CERN code GEANT. FRITIOFF and Venus models were used as event generators, in order to study detectors response with respect to different parameters. Later detector's performance was tested in the Commissioning Run of Summer'99. All these studies can be summed up as follows.

Figure 2 shows the actual view of the two Beam--Beam Counters arrays. Each detector consists of Cherenkov radiators, glued to the photomultiplier tube. Phototubes diameters are 19mm or 51mm with 4cm long or 3cm long radiators respectively. As it can be seen on the bottom picture, the Right Array has reduced azimuthal coverage because of the magnet D1. It is partially compensated by higher spatial density of modules in the Right Array. Also, several counters in both arrays are located at the same distance from the beam-pipe, which should provide better precision and might help in case of any malfunction diagnostics.

Basic experiment logic and outline is given on Figure 3. A module will trigger when a Coincidence Unit receives three signals at the same time: from two OR units, related to each of the two arrays (Left and Right); and a signal from the RHIC Clock, saying that beam entered the Interaction Region. The OR unit can be set to click if some desired number of tubes in the Left and in the Right Array are hit simultaneously. This number depends on the kind of events one wants to detect. As for the good timing of an interesting high multiplicity event, it will be produced if at least one charged particle with beta=1 strikes the active area of the module in every array and arrives within 30psec of the arrival time expected for it. Tubes alignment and an intrinsic time resolution for each tube of 50psec, will make possible fast coincidence between arrays which will serve to identify the reaction vertex to +- 15mm.

Good trigger must be provided not only fast enough, but also has to be able to distinguish collisions between two gold nuclei from the possible collisions of a beam with a residual gas (say, Oxigen) in a beam-pipe. In order to answer that question we used some computer simulations and simple calculations. Figure 4shows results of GEANT simulations of 3 kinds of collisions:

• Central collision of two Au beams.
• Collision of 2 Au beams with minimum bias.
• Collision of one Au beam with the residual Oxigen in the beam-pipe.

Figure 5: RHIC Commissioning Run in Summer'99 provided us with a real test of our detector's system. No real collisions were achieved during this time, but instead two independent particle beams were circulating the collider's rings. `` Blue beam" of golden nuclei was moving in the direction of our Right Array and `` Yellow Beam'' in the direction of our Left Array. As a result, data saved during some runs had been identified as a ``good runs'', meaning: worth looking at, mainly by the number of events in them and by trigger configurations. These good runs were merged and they served as a major data source for our analysis. Since no real collisions were achieved, only ADC spectrum analysis is worth mentioning here.

Figure 6 shows raw ADC spectra of all modules in the Right Array for the Blue Beam. Vertical axes shows number of hits and horizontal gives ADC channels. Main conclusion we make out of this picture, is that some particular modules in the Right Array were hit more than others. Mainly: Right Modules ## 1-3, 6, 7, 10, 12, 13, etc. If we take a look at the modules location scheme on Figure 7, we can easily see, that these modules are located in the upper-left corner of the Right Array (marked navy on the picture). Basically, it means, that beam was not properly centered and all detected particles mainly have come from the beam scraping the beam-pipe somewhere far behind the Left Array. It was confirmed later by the results and analysis of the Multiplicity Tiles, located cylindrically around the vertex.

Figure 8 shows raw ADC spectra of all modules in the Left Array for the Yellow Beam and again demonstrates ability of the Beam-Beam Counters to provide some information regarding centrality of the beam.

These pictures by them selves give very little information, so these spectra had to be calibrated by the following scheme:

• only events with Beam-Beam Trigger were selected;
• pedestal was fitted by Gauss and subtracted;
• ADC values were divided by the Gain values from the Lab studies for every particular module.

Figure 9 shows calibrated ADC spectra for some modules in the Right Array for the Blue Beam. It demonstrates, that some tubes have really experienced 2 or 3 hits. These hits had same amount of energy deposited for different tubes.

Figure 10 shows the same information for the Left Array hit by the Yellow Beam. Pictures, showed above demonstrate our ability to provide some rough charged multiplicity estimates for the gold-gold collisions in BRAHMS experiment at RHIC.

Thus, for the Conclusion we can say, that from the performance during the Commissioning Run of Summer'99 it is clear, that Beam--Beam Counters will be able to serve as a zero--level trigger for BRAHMS, to identify collisions for RHIC and also to provide some rough primary estimate of a charged multiplicity. As of today, all phototubes have been tested again and are ready for the major work. More over, it is possible, that they are taking some data right now, as we speak, during RHIC initial stage of operation.

• Tubes timing in a test-beam. Figure 11 shows real-time studies which were made with two assembled phototubes, (with good wrapping and UVT attached) in a test beam of pions. Tubes were tested in two different configurations: facing the beam and sideways to it. TDC and ADC spectra are represented here. TDC spectrum is given without slewing correction. As one can see, width of 2 channels, assuming time resolution of 50psec-per-channel will give us required resolution of 50psec after walk correction. Bottom row shows ADC spectra, pretty clean and giving us a resolution of somewhat 16%.