Agenda:

• status of LUT for hitfilter[Bill Lee]
• news on ordering CPU's [Bill Lee]
• STT infrastructure needs [Bill Lee]
• STT simulator status [Silvia Tentindo-Repond]
• update on beam spot monitoring [Yi Dai]
• AoB

Minutes:

1. status of LUT for hitfilter [Bill Lee]
   (see Bill's transparencies ppt, pdf)
   Bill has now code that
   • creates LUT for all detectors but takes a long time to generate LUT for all detectors )loops over 2^17 tracks and 432 detectors)
   • has improved output file
   • can read in from file
   To do:
   • increase speed
     (now takes a long time to generate LUT for all detectors -- loops over 2^17 tracks and 432 detectors)
   • commit code
   • make binary LUT file
   • implement into L2STTHitsinRoadFactory
   • fix allowed CTT sectors (talk to Brian Connolly)
   • pt sign change
   • determine roads

2. news on ordering CPU's [Bill Lee]
   
   • have quote from Arrow ($3015.--), on stock
     (note quote from Avnet much worse -- $1000.-- more and 9 weeks delivery)
   • purchase through FNAL would cost $2450.-- (plus 10% overhead)
   • quote for Vxworks runtime licence coming (Monday?), expected cost $400.--
     this is transferable licence
   • also will need to order adapter board supplies

3. STT infrastructure needs [Bill Lee]
   Bill met with Jerry Blazey a week ago;
   • one crate moved from 206 to 203, one still to be moved;
   • cables installed: serial, ethernet connection for CPU's in all crates (including overflow crate),
     SCL connection for all crates except overflow crate
     (Bill will check with Jerry about that).
   • additional things needed for MCH:
     • power supplies -- need to specify power needs (in priciple info exists)
       note that Fermilab is lately very picky with safety for power supplies -- can cause long delay!
     • internal cabling scheme for all cables, including STT board-to-board connections
     • cables (and labels for them)
       • to teststand
       • from L1CTT
       • from SMT
       • for VBD
       • to L2
     • need to understand our plans for how to use L2 teststand

   • STT simulator status [Silvia Tentindo-Repond]
     (see Silvia's transparencies ppt, pdf)
     have now working simulator version, even though there are still some problems to be cured.
     Tracking efficiency for single muon events looks OK.

   • update on beam spot monitoring [Yi Dai]
     (see Yi Dai's transparencies ppt, pdf).

     Below is the plan I've come up so far for the database related part of the beam spot monitoring/control. The basic idea is as follows:

     (1). The initial beam positions are used as a "static" parameter for each run. These parameters can be obtained from a special run such as a calibration run or some thing like that. Such a calibration might be needed for each new store of beam. The parameters should be stored in the online database and avaiable when a new run starts.

     (2). The beam positions are constantly calculated by beam vertex EXAMINE during the data taking. Evert a few minutes(to be determined, depending on the statistics needed) the deviations relattive to the "static parameter are constantly monitored. Monitoring histograms or plots will be provided by EXAMINE for shifters and accelerator people.

     (3). When the deviations exceed the threshold values, a special trigger is issued. This trigger activates the feedback to the accelerator control for the beam position adjustment; it also enables the entry of the new beam position information (after adjustment) into the database. This is the "dynamic" part.

     I think if the above scheme works. The crucial question is how frequent the "dynamic" part is, and whether the "dynamic" info should be enetered in the same online database as the "static" parameters or some other database.

     From the information I've collected and the understanding I've established so far, The occurrence of the "dynamic" trigger might not be high since the beam position fluctuation during the same beam store is expected to be well within +-100 microns. It is possible the trigger is issued every a few 10-minutes on average.

     I should also point out that the above scheme is based on the assumption that the correction control is responsive. For example, a correction of 100 microns for beam control should take no more than a few minutes.

     discussion:
     Uli: what is the dynamic range of the correction dipoles?
     answer: +- 1mm in position, +- 1.5 mrad in angle, provided that beam is on nominal orbit and correction elements run at 0 A initially.