Following up on the questionable behaviour Jan & I observed when using our corrections to GridLeak for anode voltages with Run 12 pp510 data (see page 10), I took a look at the luminosity dependence using the most current calibration for SpaceCharge & GridLeak:
    sc = (4.373e-08)*(vpdx-(255200))+(4.13e-08)*(bbcbb)
    GL = 10.26

Here are the west and east mean signed DCAs (dcaw and dcae respectively) as a function of phi and luminosity (bbcx):

Features on the west are sector 3 at the edges (gets worse with luminosity) and sector 6 (seems worst at low luminosity).

Features on the east are sectors 19 (low), 20 (mildly high), and 21 (at the edges, very low).

The other sectors show a gentle rise in <sDCA> with luminosity.

I am attaching to this page the codes I used to make these plots.
 

-Gene