The "story" of the wrongly initialized TPC padrow T0s is documented in this TPC hypernews post.

This wrong padrow T0 table was in place for the Run 14 AuAu15 preview production (May 2014). On 2014-07-22, the Run 13 padrow T0s were copied in the database as a new default for Run 14.

I took the "wrong" and "right" tables, and looked at the mean T0 differences (over all sectors) as a function of padrow, converting padrow approximately into radius, and made a linear fit to give a first order estimation of the average impact on tracks. Here is that fit result:

The units on the axes are [μs] vs. [cm].

Implications:

• The TPC global T0 calibration was affected by only (roughly) ~1.3 ns (intercept of the above fit)
  • This has been confirmed by John Campbell using the TPC global T0 calibration code
• Track projections to the GMT were affected by:
  Δz = (slope of the above fit) * (drift velocity) * (projected radius)
    = (~0.00052 μs/cm) * (~5.5 cm/μs) * (~217 cm)
    = ~0.62 cm
  • Note that this affects east-side and west-side tracks oppositely in reconstructed z, so it would create a splitting of track projections to the GMT modules of 2 x ~0.62 cm = ~1.24 cm.
  • I suspect this is a notable portion of the splitting that Xin Li has been reporting in his GMT studies

-Gene