I have used the low luminosity Run XIV data AuAu200 for days 84-90 (RFF) reconstructed with new (12/14, time stamp 20120101.001454) TPC super sector alignmnet.
I compare global coordinate of hit and prediction for primary tracks with pT > 0.5 GeV versus  pixel sector (1-10), pixel half detector (11-12), pixel as whole (13), ist (14), sst(15 - place holder), and 
HFT as whole (16). Average differences for Δx, Δy, Δz, α, β and γ are show below.

1. It is supprised me a big variation of shifts and rotations versus Pixel sector. 

2. I do see a particlur sensor position variation in dX versus Z ~100 μm.  I did naively expect that it should be ~ 40 μm.
More of that you can find at http://www.usatlas.bnl.gov/~fisyak/star/Tpc/Alignment/Hft2Tpc/HftTreesP7_plots/Global.html
Can we check that the HFT survery parameters are properly loaded in Data Base ? 

2.