Thoughts on the Start Counter

[Elton Smith <elton@jlab.org>]

Hi Andi and Werner,

Before I completely forget all the discussion we had about the start
counter, I wanted to put them down on paper. Here are some of my thoughts
on the matter. Perhaps they can steer us toward design decisions.

Requirement                  Issue

1. Tracking near target      improve momentum resolution due to additional
                             track length (deltap ~ L**2)

2. Tracking in cone          improve momentum resolution especially for
                             tracks that miss the CDC

3. Sub-ns timing offline     Resolve RF bucket for pid
(delta t ~ 300 ps)

4. ns timing in trigger      Eliminate accidentals in Level 1 trigger
(delta t ~ 10 ns)

5. Multiplicity count        Use multiplicity in trigger to eliminate
                             single-prong events

In order to satisfy all these requirements, a counter such as the one
proposed is needed:

- Vertex detector is required by 1) and 2)
- (Scintillator) Start counter required by 3)
- Fiber readout of vertex detector required by 5)
- Requirement 4) requires start counter or fiber readout

The penalty in the present design to satisfy all requirements is 1.7 cm of
plastic, or 1.7 g/cm2 and 0.043 radlen, corresponding to a multiple
scattering angle of about 2.8 mrad. Andi and Werner have convinced me that
the amount of material cannot be substantially reduced (though certainly
it can be optimized) without compromising capability.

Studies
-------

1. Understand quantitatively how the degradation in angular resolution
affects our physics, e.g. missing mass resolution, even though momentum
resolution is improved.

Curtis: How much material is presently in MCFast for the vertex/start
counter?

2. Ask the tough questions of relaxing requirements:

a) Eliminate cone
   consequenses
   - degrade momentum resolution for cone tracks
   - loose "hermiticity" for trigger multiplicity
   gain
   - reduce thickness by 6 mm
   - eliminate 2500 electronic channels

b) Eliminate vertex detector
   consequences
   - degrade momentum resolution all tracks
   - loose trigger multiplicity
   gain
   - reduce thickness by 12 mm
   - eliminate 5000 electronic channels

c) Eliminate start counter
   consequences
   - loose good timing offline
   gain
   - reduce thickness by 5 mm

   question: can the vertex detector give rough trigger timing?

Reference

A useful note that investigated the constraints for the CLAS start counter
(CLAS-NOTE 92-001) can be found at

http://www.jlab.org/Hall-B/notes/clas_notes96.html




-- 
Elton Smith
Jefferson Lab
elton@jlab.org
(757) 269-7625