T0-by-padrow calibrations were performed for Run 9 using run 10118050 for RFF and run 10184053 for FF. After calibrating by padrow internally within each sector, a "super-sector" T0 calibration (shifting entire sectors) was performed using signed DCAs in Z of global tracks (with primary partners) to the highest ranked primary vertex. The results results from the two runs (two fields) can be compared by subtracting the T0s found in the RFF data from those found in the FF data. Here we find an interesting pattern shown versus sector and padrow (left) and versus phi (right: black is west, red is east TPC):

Time scales are in microseconds, and please note that certain problematic regions of the TPC have been excluded (and are blank) to avoid distraction from the overall patterns.

Summarizing, I see two artifacts I do not yet understand:

1. An overall shift of over 30-40 nanoseconds between the runs.
2. Some structure atop that shift, most prominently on the west side of order 10-20 nanoseconds. The structure appears nearly sinusoidal as sectors 9-12 reach a ~50 nanosecond shift, while (azimuthally) neighboring sectors 1 and 8 are near ~40 nanoseconds.

It may be helpful to note that with a drift velocity of ~5.5 cm per microsecond, a 30 nanosecond interval corresponds to ~1.65 mm in Z. Shifting both halves of the TPC by such an interval separates the data by over 3 mm between the halves.

Any comments or insight into understanding these artifacts is appreciated.

-Gene