TOF/T0 Status
March 28, 2003


	T0 Arrays:
		T0 beam-gas studies
		Vertex studies: why do we care?
		T0 correlations
		Results of MC studies
		Do we see correlations for Au Au?

	TOF detectors:
		Track matching
		Particle identification

	Outlook for pp

	Summary

Beam-gas Studies.


	Beam-gas events are the only events in which we observe a strong time correlation between the T0 counters in each array.
	The beam-gas data are being used to optimize the slewing parameters for the T0 counters.
	We have obtained a first set of slewing parameters for T0N and are working now on T0P. Slewing corrections have brought the T0 resolution to around 100 ps (1s).

Beam-gas Studies


	A dedicated run (11508) was carried out in which the beams were tuned for no collisions.
	A minimum bias data set was collected in order to collect sufficient data for slewing studies.
	Despite being tuned for no collision, the overall rate was still dominated by collisions.
	The strong correlations seen in the collision events needs to be understood.

T0 Vertex Studies:
The best way to determine time resolution.


	The collision time, obtained from the T0 array, is equal to

		t_col = (t_T0N + t_T0P)/2

	The resolution of the collision time is equal to

		s_Tcol = s_T0/Ã2

	The vertex position, obtained from the T0 array, is equal to

		z_vertex = c (t_T0N - t_T0P)/2

	The resolution of the collision vertex is equal to

		s_vertex = c s_T0/Ã2 = c s_col

	We use s_vertex to determine s_Tcol.

T0 Vertex Studies:
Comparing Spec Vertex with T0 OR Vertex.


	The T0 OR Vertex is obtained from the OR information obtained using TDC 8:

		T0 OR Vertex = 29.979*((tExtra[19] - 311.87) - (tExtra[20] - 302.65))/2.0

	The vertex resolution is 4.8 cm (1s).
	This implies a resolution of the T0 ORs of 225 ps (1s).
	The collision time resolution obtained from the OR would be 160 ps (1s).

T0 Vertex Studies:
Comparing Spec Vertex with T0 Fast Vertex.


	If all T0 counters are properly aligned in time, we expect that by using the time of the first T0 to fire we should be able to exceed the performance of the T0 OR vertex.

	The results obtained with these first T0 signals are very poor, suggesting that the T0 counters are not properly aligned in time.

T0 Vertex Studies.
Check relative timing of T0N detectors.


	In order to verify the alignment of the T0 counters, we look at all possible combinations between T0N and T0P0.

	We expect that in each case the vertex resolution is the same, but the centroid of the vertex difference distribution may be different as a result of errors in the software alignment of the T0 counters.

T0 Vertex Studies:
T0N5/T0P5 Vertex vs Spec Vertex

T0 Studies:
Run 10920: #T0N=1 and #T0P=1


	What do we learn from this analysis?
		If we require a multiplicity of 1 in either array, we rule out contributions from cross talk within the cables.
		The vertex resolution for all combinations of detectors is between 2.5 and 3.5 cm.
		The different detectors are clearly not aligned in time (a 1 ns shift is required to move the vertex by 15 cm).
		The centroids of the vertex difference distributions can be used to realign the individual T0 counters on either side.

T0 Studies:
Run 10920: #T0N=1 and #T0P=1


	After correcting the delays of the T0 channels for the relative time delays obtained by from the previous analysis, all vertex difference distributions (calculated with respect to sensor T0N2 and T0P2) are aligned.
	The width of the vertex difference distribution is 3.2 cm (1s).
	This corresponds to a T0 collision time resolution of 100 ps (without slewing corrections).

T0 Studies:
Run 10920: #T0N=1 and #T0P=1


	What did we learn from these studies:
		Because of the requirement that only one T0 counter on each side fires, crosstalk within the cable is not an effect here.
		Crosstalk between P and N cables is possible, but requires that for example pair 8 in the P cable has an effect on pair 0 on the N cable.
		The widths of the vertex difference distributions are narrow and correspond to a collision time resolution of 100 ps (uncorrected).
		All these events can be used, but we need 10 different sets of calibration constants.

T0 Studies:
Correlations in Au Au


	The observations made for d+Au lead us to investigate whether or not similar effects were observed in PR01 (Au+Au).

	We carried out an analysis similar to the studies we carried out for d+Au and found that the calibration constants that work for all events (dominated by events in which all T0 counters fire) just work as well for events in which only on counter fires on the N side and on the P side.

T0 Vertex Studies:
How to produce correlations?


	Consider collisions produced in the halos.

	The geometry of the collision is such that the hits is the counters oriented at the same position show the highest correlation.

	However, the effect is much smaller than the effect we observe, and it can not explain our observations.

T0 Vertex Studies: MC simulations.






	The strong correlation between T0 counters at the same location (T0Nx and T0Px) is not present in the simulated data.

T0 Vertex Studies: MC simulations.


	The ratio of primaries to secondaries that loose energy in the T0 counters is about 1 to 10.

	Taking into consideration that the T0 are Cherenkov counters, only 1 out of every 3 secondaries will produce a T0 signal.

	The expected ration of primary to secondary T0 signals is about 1 to 3.

T0 Vertex Studies: MC simulations.




	MC studies of the time of primaries and secondaries show a significant probability of a prompt primary in coincidence with a delayed secondary.

	The simulations do not show the observed correlations.

Track Matching d Au


	Track matching to the TOF was studied by comparing the hit position of the extrapolated track with the actual position of the TOF sensors that recorder a MIP or more.

Track Matching d Au


How well do we match tracks?

	Wall TB:
		Centroid = 0.0 cm
		Width = 1.6 cm

	Wall TC:
		Centroid = 0.7 cm
		Width = 1.2 cm

Momentum dependence:
	1 - 2 GeV: centroid = 0.0 ± 1.6 cm
	2 - 3 GeV: centroid = 0.2 ± 1.3 cm
	3 - 4 GeV: centroid = 0.3 ± 1.3 cm
	4 - 5 GeV: centroid = 0.2 ± 1.2 cm

Track Matching d Au



	The momentum distributions of the TOF tracks are obtained by selecting those tracks for which the difference between the predicted hit position and the measured hit position is less than 5 cm.
	The data shown in this figure is contains a few percent of the total data set.

TOF PID


	The TOF particle identification procedure has been optimized for the Au+Au data.

	There are no cuts on time applied to the data in this analysis, except that we require that the hit comes from the correct beam bunch.

	PID spectra are clean, and protons can be identified up to about 2 GeV.

Outlook for p+p


	The correlations between T0 counters observed in d+Au is troublesome. Crosstalk is clearly playing a role, but it is not clear that it can explain all our observations.
	The trouble with individual time signals does not impact the T0 OR, which is what is being used for the trigger,
	The multiplicities in p+p are much lower than the multiplicities in d+Au and for most events there will be only one T0 counter firing. The OR signal will thus carry the time information provided by this counter.
	The individual times should also be OK, except we might need to use different calibration parameters, depending on the channels being hit.

Outlook for beyond p+p.


	We will need to address the T0 cable issue.
	Options to considered:
		Use BNC (required to go from ECL to NIM to the CH and then via a discriminator and a NIM to ECL converter to the TDC).
		Change the relative alignment of cables (in order to ensure that the straight sections of the cables do not overlap).
		Replacing T0 cables with shielded cables.
		Moving FB into the tunnel.
	The cross talk between channels within a cable is easily visible on a scope, but it only becomes serious when the cable is unrolled. Note: in most applications where the twisted pair cables were used to delay ECL channels, the cables are left in their original package and only the two ends are pulled out to be connected to the delay box.

Crosstalk within unrolled cable
Monitor ch 3 for various signals in ch 2, 4.

Possible replacement cables.


Options:

	Shielded ribbon:
		Spectra-Strip:
		Cost: $ 3077/1000Õ ($1538/cable).
		Delivery time: 8 weeks.
		Belden 9736:
		Cost: $ 3107/1000Õ ($1554/cable)
		Delivery time: 3 - 4 weeks.

	Twinax cable:
		Delco Wire and Cable:
		Cost 0.30/1Õ ($1698/cable).
		Delivery time: 2 weeks.

Note: all options/cost are just the cost of the cables itself.

Possible replacements.

Summary.


	The d+Au data revealed strong correlations between the T0 counters on the P side and those located on the N side.
	Study of the time signals revealed crosstalk between the N and P cables.
	This effect in principle effects every detector that generates timing signals.
	This issue needs to be addressed É. and hopefully at that time we will come in for a soft landing.