ER August 28 update - sm

Plots use Inkyu's newly processed tracks

no cuts (other than demanding single track ... n_trks=1), just processed tracks, look at normalized energy deposited in layer 2 vs. layer 1

same data, no cuts, look at x residual as fn of local x,y and local x vs. local y

note: good tracks hit top of layer three in y

same data no cuts,  look at y residual as fun of local x,y and local x vs. local y (repeat plot)

Now demand prob>0.0001 and theta<1.5, plot x residual as fn of local x,y and local x vs. local y

Do we see a shift in the position of the x residual as a function of local x??? 

Same as above with surface plot on top and angle of viewing adjusted.  There are a couple of places along local x where the residual seems to shift or diffuse out a little ... at local x = 4 and just above 2.5.  I see this a bit more when I make the scatter plot with the full data ... not shown here.  I wonder if this is a silicon effect?

Fit the full x residual with a single gaussian from -0.04 to 0.04.  Looks like the residual wants more than one gaussian and the fit range is important for the final answer.   The parameters returned from the fit (not what is below in the box) gives a mean of -0.0043 and a sigma of 0.028. (this is a 40 micron shift ... smaller than what we saw with the other track sample)

Same as above, now use range -0.02 to +0.02.  Fit gives mean of -0.0047 and sigma of 0.021.

Now look at x residual for local x<1.5.  Use -0.04 to +0.04 range and get mean of -0.0036 and sigma of 0.026

look at x residual for x>3.0 in same range.  Get a mean of 0.0053 ... actually looks like the fit is too low given the high peak to the left of the fit peak.

So maybe we do see an x residual shift as a function of local x ... perhaps implying a rotation in the phobos x,z plane.  It's interesting that the shift seems to have become smaller with the new tracking.  I guess we have a rotation between planes in the x-z plane by design ... need to think if this shift is consistent with what I would expect.