The goal of this pp510 calibration was to create a calibration based on run12's UU193 barrel calibration and the dE/dx calibration request.  The reason why we want to use the barrel calibration is because it had plenty of statistics(multiplicity+number of events) to perform a cell based calibration on the barrel--also the VPD's(start time) performance is solid in high multiplicity events--not to mention the same running year, hardware, etc. 

To reuse the UU barrel calibration, we need to calibrate the VPD and adjust the global barrel delay timing(T0) correction to match with the new start time+conditions. 

The sample used to produce calibrate was MuDst's created from the .daqs in filelist.txt (the dE/dx calibration request). [BTW, during some of this sample, the VPD's TCD board broke--which would lead to trigger issues.  However, we did not explicitly rely on the triggers to create the ntuples for calibration.  So it should be OK]

Produced ntuples to make the sample more manageable.
Ran and found the timing calibrations for both the East and West sides of the detector.
Then shifted the East timing to align the Vz_vpd with the Vz_tpc.  This timing shift was: (4542.37-2.02792-4.59159+9.59631e-02+9.91213)/c .

The resulting parameters may be found in pvpdCali_4DB.dat , http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0/pp510shiftedchannel20594/pvpdCali_4DB.dat .

With the UU barrel calibrations(t0, tot, and Z) and the vpd calibrations, I ran over the data to find the difference between the expected and actual TOF for pions as a function of tray, on average this number should be 0.  However, its not because of the mixing of calibrations.  The difference was found to be ~.2645ns.  The correction of -.2645ns is then applied in the electronics delay, t0_4DB.dat, calibration parameters.

Basic QA of these parameters may be found at http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0/ and http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0rhic/ .  The former link is the QA done at Rice on the produced ntuples while the latter link is simple QA done with the vpdminbias triggers(380001+380002) and plugging the calibration parameters through the offline software and running in afterburner mode on the MuDsts--a quick way to check that parameters are in the correct format, etc.

Digging into the details of the QA showed a single channel(20594, tray 108) offset by about ~.6ns as shown in the deltaTOF vs channel image below:

To check to see if this behavior was missed in the UU193 calibration, we checked that out too:

But there was no obvious shift in the UU data at that channel.  It did turn out that the same shift was missed and seen in the pp200run12 data sample, which used the UU barrel calibrations too.
The shift in timing was seen using the different macros at Rice and rcas.  The actual time of flights for the channel are quicker than the rest of the detector.  This can be seen in the following image of TOFs for tray 108:

To investigate this further, I double checked the parameters to make sure the global T0 correction was applied to every channel the same, and it was.
I also checked to see if there was a stuck bit for that channel in both pp510 and UU193.  To do this, I looked at the StBTofRawHit information(tdc).  There was no obvious problem with the bits seen.  The readout of tray 108 bits may be found at the following links:
http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0rhic/pp510/Bits.html and http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0rhic/UU193/Bits.html
The channel of interest is channel 50, since we are counting from 0.

After exploring a couple avenues to see where a single channel offset could be made, no source was found.  So I corrected the symptom by shifting the channel; tray 108, board 9, ch3; .6ns from -323.237 to -323.837ns.  The affect of the shift can be seen here:

And a quick QA with the new calibration parameters may be found here: http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0/pp510shiftedchannel20594/

A better investigation on the channel's properties could be made with better statistics, but since this is pp510, that would require many more events and it may not be helpful in the end.

So these are the proposed calibration parameters to be used in the pp510 run 12 production:
vpd: http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0/pp510shiftedchannel20594/pvpdCali_4DB.dat
delay(t0): http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0/pp510shiftedchannel20594/t0_4DB.dat
time over threshold(tot): http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0/pp510shiftedchannel20594/totCali_4DB.dat
local Z(z): http://www.bonner.rice.edu/~jb31/pp510gevrun12correctedt0/pp510shiftedchannel20594/zCali_4DB.dat