Futher validate the L2 alignment
Alignment with more runs, estimate uncertianties.
B On Vs Off
Have run the L2 with 3Dof both Combined and TRTOnly with field off, agreement with field on is good.
Running the L2 with 3Dof TRT Only tracks.
Futher validate the L2 alignment
Alignment with more runs, estimate uncertianties.
B On Vs Off
Have run the L2 with 3Dof both Combined and TRTOnly with field off, agreement with field on is good.
Running the L2 with 3Dof TRT Only tracks.
Found that the error scaling that I was using has large errors for Si and TRT. tag 'IndetTrkErrorScaling-002-00' is recommended by Grant. Need to switch to this tag and see if there is an impact.
This is running now.
This actually looks like a big problem, the residuals I'm now finding are "split" in two.
It turns out that runing the reco with no error scaling before the L2 alignment the residual plot had a double peak.
If you then run the alignment on (with no error scaling), the result is a geometry in which there is a double peak.
If however you start with no TRT L2 alignemnt, run the L2 with error scaling, then run reco with the resulting geometry with out error scaling the double peak is NOT present.
Running the L2 alignmetn from this geometry with no error scaling improves the combined residuals slightly.
Danielle:
If there is a coherent distortion of Si should manifest as a phi structure (in relative p-scale say)
Momentum scale of the TRT wrt to the Si, If there is a radial misalignment residuals will not show discrepancies for combined tracks, but there will be a shift in the pscale
Following plots are from Manuel
Compare TRT to Si.
Old here is 1st reprocessing
New here is 2nd reprocessing + TRT L2
