In the following simulation analysis discussion, the charge integration methodology has been utilized in determinig the multiplicity. Figure 3 and Figure 4 display the detector resolutions obtained from the detector response simulations for the CTB and MWC, respectively. Here resolution is defined as the sigma of the distribution of multiplicity determined from the detector response simulation divided by the multiplicity obtained directly from GEANT hits. These have been obtained from event generation involving 4000 FRITIOF events at 200 GeV/u for Au+Au, with an impact parameter of b=0-3 fm. These have been processed through the full STAR detector geometry utilizing GEANT. The results indicate that we can expect a 2.2% resolution for the CTB and greater than a 3% resolution for charge integrating in the MWC.
In hardware, Level 0 analysis is planned to be performed using a tree of FPGA's (Field Programmable Gate Array). This is simulated with a series of summing and comparing algorithms. In the first stage of the FPGA trees, the ADC multiplicities are summed so that the 336 input data words (240 from the CTB, 96 from the MWC) or pixels in eta-phi space, are compressed to 32 pixels with 1/2 unit coverage in eta and 90 degree coverage in phi per pixel. These 32 data words are then passed to the Level 1 simulator. At Level 0 the 32 eta-phi pixel array is further compressed to total multiplicity while comparisons of coarse distributions in eta-phi space are performed. The effectiveness of very simple analysis on these coarse multiplicity distributions, as well as total multiplicity, in selecting events of potentially interesting physics at Level 0 has been investigated.
Total multiplicity results from our Level 0 simulation utilizing the previously mentioned 4000 FRITIOF events are shown in Figure 5. The fluctuation displayed is merely a result from the conversion of total multiplicity to an 8 bit number and the number of bins utilized in the figure. FRITIOF predicts a nearly linear relationship between charged particle multiplicity and the number of participants in the reaction as shown in Figure 6 . One preliminary requirement for Level 0 is to select the top 3% of the Au+Au reaction cross section which we can in turn determine from the multiplicity distribution. Applying this constraint to the present simulated data, events with multiplicity > 4200 would be tagged as interesting events and passed to Level 1, which is shown in the blue shaded area.
In order to test the sensitivity to selecting unusual events at Level 0, 400 events with artificial plasma bubbles injected into a standard HIJET event generator background have been produced. Within these 400 events there are 4 separate sub-classes of event types based on plasma bubble dynamical evolution, these include Plasma (spherical isotropic bubble expansion), Smoke (torroidal ring expansion) and Chiral (spherical bubbles of pions at T=70 MeV). Figure 7 displays the total multiplicity results from these simulations. Clearly a threshold of 4200 is much too restrictive and would only allow the passage of a small fraction of these unusual plasma events. Thus total multiplicity alone may be ineffective in selection of interesting plasma events.