Performance

High SVD strip yields and good S/N ratios are needed to ensure the efficient matching between tracks detected by CDC and clusters detected by SVD. Distributions of normalized cluster energies for minimum ionizing particles (MIPs) from hadronic events are shown in Fig. , where the cluster energy is normalized to the same track path length in DSSD (300 $\mu$m). The measured most-probable peak height is approximately 19,000 e$^-$. The noise level and the S/N ratio for each DSSD side ($p$, $n$ and $p$-$n$) are summarized in Table . Strip yields, which are defined to be the fraction of channels with S/N ratios larger than 10:1, are measured to be 98.8% on layer 1, 96.3% on layer 2, and 93.5% on layer 3 [28].

Figure: Cluster energy distribution in which the cluster energy is normalized by the track path length in DSSD.

The track-matching efficiency is defined as the probability that a CDC track within the SVD acceptance has associated SVD hits in at least two layers, and in at least one layer with both the $r$-$\phi$ and $r$-$z$ information. Tracks from $K_{S}$ decays were excluded since these tracks did not necessarily go through SVD. Fig. shows the SVD-CDC track matching efficiency for hadronic events as a function of time. The average matching efficiency is better than 98.7%, although we observe slight degradation after one year operation as a result of the gain loss of VA1 from radiation damage  [28]. The momentum and angular dependence of the impact parameter resolution are shown in Fig. and well represented by the following formula:
$\sigma_{xy}=19 \oplus 50/(p\beta\sin^{3/2}\theta)$ $\mu$m and $\sigma_{z}=36 \oplus 42/(p\beta\sin^{5/2}\theta)$ $\mu$m. We performed the same study for the MC sample and obtained an IP resolution of $\sigma_{xy} = 15 \oplus 49/(p\beta\sin^{3/2}\theta)$ $\mu$m and $\sigma_{z}=28 \oplus 41/(p\beta\sin^{5/2}\theta)$ $\mu$m. This indicates that the IP resolution term dominated by the scattering is well simulated by the MC data. The flat components are larger in the real data than in the MC simulation, which indicates the amount of remaining misalignment errors. Detailed description of Belle SVD performance studies can be found elsewhere [29].

Figure: SVD-CDC track matching efficiency as a function of the date of data taking.

Figure: Impact parameter resolution.