T0/TOF Analysis Update  February 6, 2003

	T0 time resolution and collision time determination:
		Looks OK, but more complicated than Au + Au.
		Expect a significant background due toÓsecondary-onlyÓ events which are difficult to distinguish from ÒprimaryÓ events.
	Processing time calibrator information:
		Monitor stability.
		How do we implement the stability corrections?
	TOF calibration needs:
		What needs to be done?
		What do we need?
		When will it be done?

T0 Time Resolution

	Direct determination of the time resolution of the T0 counters has been complicated by the difference in response for d + Au compared to Au + Au.
	For Au + Au, the slewing parameters were optimized by looking at the time difference between counters in the same array for good collision events.
	This procedure fails for d + Au.
	For collision events, there appears to be at most one prompt T0.

T0 Time Resolution

	For beam-gas events, there appear to be many prompt T0s.
	The raw time difference distributions are 240 ps wide.
	Applying the slewing corrections from PR01 (for those PMTs that were used in PR01) reduces the width to 150 ps or 110 ps for each PMT.
	Since the HV on the PMTs are different in PR03, compared to PR01, we will need to optimize the slewing parameters.

T0 Time Resolution

	The lack of prompt T0s for collision events has a number of consequences:
		The T0 vertex resolution does not improve when the multiplicity of the event increases.
		The collision time resolution does not improve when the multiplicity of the event increases.
		The only approach we can take in selecting the prompt T0 counter is to take the first T0 counter that fires in a particular event.

T0 Time Resolution

	The observations in the data are consistent with the results of the MC simulations.
	These simulations show that most hits are not in prompt time coincidence.
	We will need to study how to extract a valid collision time by using events with good vertex resolution (RMSSelVertex).

T0 Time Resolution

	The collision time depends on the average between the time of T0N and T0P: tcol = (T0N + T0P)/2 - distance/c.
	Based on the observed resolution in the time difference spectra we expect that the time resolution of prompt-prompt T0 coincidences is 75 ps.
	But we do not have any way of determining whether the first hit is a prompt hit.   We need to understand how this will contaminate the time spectra.

Stability of time delay. Big change between PR01 and PR03.

Time Calibrator Information

	Procedure to use time calibrator information:
		Time calibrator events are stored in DV monitoring histograms.
		On a sequence-by-sequence basis we determine the shift required to bring the time calibrator peak backs to their ÒstandardÓ position.
		The position of the prompt timing peak with respect to the time calibrator peak will be determined once and for all and will be assumed to be constant.
		We will be using the T0 information to test this procedure.

Time Calibrator Information

	We are using run 10920 (with its 60 sequences) to develop the procedure to make the drift corrections.
	Conclusion on how well this procedure works will be available next week (we are currently generating the DSTs for this run which contain only the relevant plastic information to carry out this study).

Up to TOF PID.  What is needed?

	T0 collision time procedure finalized (2/14).
	PR03 TOF DV (especially the time calibration data) included in the DV process.
	Generate track DSTs (on a sequence by sequence basis) to be able to apply TOF and T0 calibrations on a sequence-by-sequence basis.
	Sufficient TOF tracks to optimize the slewing corrections for each scintillator (detailed numbers presented on the next slide).

TOF Calibration Needs

	Optimize slew parameters for each scintillator:
		Need to determine the TOF vs E curve for each PMT.
		This will require 100 - 200 hits per scintillator.
		Require 12,000 - 24,000 hits per wall.
		Require 24,000 - 48,000 TOF tracks.
	Track assumptions:
		There are 50 curved tracks per 1000 events.
		A small fraction of these (1%?) will be aimed at the TOF.
		Required number of minimum bias events = 24,000 x 20 x 100 = 48,000,000.

Time line

	Assuming we have the required number of events to do the TOF calibrations:
		Determine the time delay calibration parameters (based on DV output): 2 days.
		Create the track DSTs: ? days.
		First order slew optimization: 1 week.
	Major concerns:
		Collision time resolution.
		Time structure in TOF looks similar to structure in T0.  How many prompt hits do we see?