STT NEEDS
    =========

    initial alignment of the detector with the beam:

    detector limit          determining factor
    STT      offset<1 mm    acceptance loss due to tracks that cross sectors
    STT      angle<300 urad z correction based on barrel number
    CFT      offset<few mm  can be corrected in equation table
    CFT      angle<25 urad  <50 um twist over 2 m length

    stability of this alignment over time:

    The offset needs to be stable to 30 um else the STT translation tables need
    to be recomputed and downloaded. The CFT can probably tolerate larger
    variations (0.5 mm) without having to adjust the equation tables 

    We expect relative alignment between CFT and SMT to 50 um.

    EXPECTED VARIATIONS
    ===================

    Based on the CDF Run I experience we expect the beam to vary store-to-store
    by 100-150 um (offset) and 200 urad (angle). Within a store we expect the
    offset to be stable to 40 um and the angle not to exhibit measureable
    variations.

    ADJUSTMENT CAPABILITIES OF THE ACCELERATOR
    ==========================================

    During the initial orbit tuning the beam offset can be adjusted over a 2 mm
    range and the angle over a 1 mrad range. These numbers are estimates by John
    Mariner and could be too large by a factor two. A feedback mechanism during
    the run could compensate for position variations with a dynamic range of 
    1 mm with a precision of 5 um (offset) and 1.5 mrad with a precision of 
    50 urad (angle) (These numbers need to be verified).

    CONCLUSIONS
    ===========

    Upon roll-in, we need to place the detector with respect to the beam to
    a precision of 500 um (offset) and 250 urad (angle).
    => need to know the beam trajectory to sufficient precision. This may 
       requires a survey program similar to that planned by CDF.
    => understand that we can place the entire 5000 ton detector this precisely
       or allow for position adjustment of the tracker.

    In order to maintain this alignment we will implement a feedback
    mechanism a la CDF. In run I, the dynamic range of the relative
    accelerator-detector motion was too small. It appears that in Run II, the
    dynamic range of the beam position control is larger and may be sufficient
    for most of our needs. However, caution may be advisable since these numbers
    are estimates which have undergone substantial variation. To be on the safe
    side, we have to also foresee the possibility that the central tracker has
    to be repositioned during the run.
   
    REQUIRED ACTION
    ===============
    Need to make preparations which will allow positioning of the central
    tracker to within the required precision, and also think of a scheme which
    will allow repositioning later in the run if needed.