Performance

The Belle experiment has been taking data at the $\Upsilon$(4S) resonance energy since June 1999. In order to operate CDC in the high beam-background environment, we readjusted high voltages and electronics parameters such as the bias voltage of the preamplifiers. These changes reduce cross-talk noise hits caused by the very large signals due to background particles spiraling around sense wires. These changes also reduced the preamplifier gains and slowed their response times. The calibration constants were recalculated using the beam data. In addition, corrections for the additional magnetic field non-uniformity caused by the accelerator magnet components near the interaction region were introduced into the Kalman filter using a magnetic field map data measured with these magnets in place. We again obtained 130 $\mu$m overall spatial resolution, although the residual distribution has considerable non-Gaussian tails. We analysed $e^+e^- \rightarrow \mu^+\mu^-$ events, detected as two charged tracks by CDC and identified as muons by the outer detectors, to extract the $p_t$ resolution of CDC. We can calculate the $p_t$ resolution using the fact that each muon track has the same momentum in the cm system. The $p_t$ resolution measured is 1.64 $\pm$ 0.04 % in the $p_t$ range from 4 to 5.2 GeV/c, which is imposed by kinematics and the acceptance of the TOF system used for triggering. This result is somewhat worse than the resolution of 1.38 % expected from Monte Carlo simulations. The $\phi$ dependence of the $\Delta p_t$ resolution has been checked with $e^+e^- \rightarrow \mu^+\mu^-$ events. No significant $\phi$-dependent systematic effects were observed. The $K_S^0$ mass was reconstructed from $K_S^0 \rightarrow \pi^+\pi^-$ decays in hadronic events in order to check the $p_t$ resolution at low momenta. Most of the decay pions have momenta below 1 GeV/c as shown in Fig. . Fig. shows a $\pi^+\pi^-$ invariant mass distribution. The FWHM of the distribution is 7.7 MeV/$c^2$, which is slightly worse than the idealized Monte Carlo prediction of 6.9 MeV/$c^2$.

Figure: Transverse momentum distributions for pions from $K^0_S$ decays. The solid and dashed histograms correspond to $\pi^-$ and $\pi^+$ tracks, respectively.

Figure: Inclusive $K^0_S \rightarrow \pi^+\pi^-$ mass distribution for multi hadronic events.

The truncated-mean method was employed to estimate the most probable energy loss. The largest 20 % of measured $dE/dx$ values for each track were discarded and the remaining data were averaged in order to minimize occasional large fluctuations in the Landau tail of the $dE/dx$ distribution. As described in Section of the beam test [36], the $<dE/dx>$ thus obtained is expected to give a resolution of 5 %. A scatter plot of measured $<dE/dx>$ and particle momentum is shown in Fig. , together with the expected mean energy losses for different particle species. Populations of pions, kaons, protons, and electrons can be clearly seen. The normalized $<dE/dx>$ distribution for minimum ionizing pions from $K^0_S$ decays is shown in Fig. . The $<dE/dx>$ resolution was measured to be 7.8 % in the momentum range from 0.4 to 0.6 GeV/c, while the resolution for Bhabha and $\mu$-pair events was about 6 %. These results are somewhat worse than the Monte Carlo results. At present the resolutions obtained from the beam data seem to be slightly worse than those expected from the cosmic ray data and Monte Carlo simulations. Although the present performance level of CDC provides the momentum resolution sufficient for the goals of the Belle experiment, efforts to further improve CDC performance are being made.

Figure: Truncated mean of $dE/dx$ versus momentum observed in collision data.

Figure: Distribution of $<dE/dx>/<dE/dx>_{exp}$ for pions from $K^0_S$ decays.