Following up on Flemming's You do not have access to view this node... I looked at TPC clusters from 500 events in run 25121030.

First note, as Flemming pointed out, the wide-in-time clusters seem to be predominantly (but not solely) in the iTPC (inner sector) rows. Here is time bucket width vs. row (all sectors):


I have looked versus x, y, z, phi, time bin... and I see nothing to note here on those. Here is time bucket width when on a reconstructed track (red) and not (black). These large time bucket width hits appear to be on tracks similarly to the smaller hits.



The following plots are for hits on reconstructed tracks, using track-wise quantities, and restricting to the inner TPC clusters.

Time bucket width vs. eta and log₁₀(p_T). The notable thing here is that these are not just high eta tracks for large time bucket width! However, it's a little difficult to understand the differences in the distributions for large versus small time bucket width.



Below I divide the clusters into two groups: time bucket width less than 16 on the left, and greater than 16 on the right. The clusters of size 31 are too few to be consequential.

This plot of eta vs. log₁₀(p_T) does show that there is a difference in distributions, with large time bucket widths not as centralized in eta:



Similarly for eta vs. z of first hit on track: these hits are on a more spread out distribution of tracks in z:



Here is dE/dx vs. log₁₀(p) [NB: not p_T] which shows that the large time bucket width tracks have a greater percentage at high dE/dx, but are not solely at high dE/dx!



Thanks,
-Gene