RUN 2B L1CAL MEETING MINUTES
			     6 March, 2003

Present
o Fermilab:     M.Abolins, G.Cancelo, H.Evans, J.Kotcher, V.Pavlicek,
		R.Partridge 
o MSU:		
o Nevis:        J.Mitrevski
o Northeastern: D.Wood
o Saclay:       J.Bystricky, D.Calvet, P.Demine



Trigger Towers with Splitters: Pavel Demine
------------------------------

o Data Sample (raw data)
  - inclusive   2300
  - zero bias
  - TTs: ieta = -9 - -12
         iphi = 1

o Raw Data Tot Et distributions
  - split towers have ~same distrib's as unsplit
    . some variation in widths of distrib
  - for both inclusive and zero-bias runs

o To Do
  - compare precision Et vs. Trigger Et for split/unsplit towers
  - take splitter out and compare same TTs split/unsplit
    . note: data taken before splitter had different electronics setup
      so direct comparisons are not useful
  - need to think about next step: analyzing the full analog signals



Noise Studies: Jiri Bystricky
--------------

o tsim_l1cal2b: ad hoc noise added at input to TAB
  - Gaussian w/ sigma = 0.25 GeV
  - Jovan's studies based on this

o ADF Simulation: Denis' standalone program
  - white noise added every ns to generated pulse
  - suppressed by analog filter (x8)
  - various noise levels simulated (for no signal)
    . <>=0.25 GeV  --> all FIR output at 0
    . <>=2 GeV     --> <1% FIR output not= 0

o Thermal Noise
  - see Cleland, et al, NIM A338 (1994) 467

o Questions
  - Add random noise at some t: where?
    . generated pulse?      - filter suppresses this
    . after analog filter?
    . to ADC samples?
  - what amplitude?
  - how to propagate: impulse response?
  - can we get information from zero-bias triggers?

o Possibilities
  - use Brown spectrum analyzer to get frequency spectrum of noise
    . currently at HP for repairs


Changes to the GAB: Hal Evans
-------------------

o See slides attached to agenda for details
  - Proposal is to break GAB into three boards corresponding to three
    main functional elements
    1) VME Interface
      . translates A24:D16 VME into serial protocol for transmission
        to TABs and GAB (on 9 ADF-to-TAB style cables)
    2) SCL Interface
      . picks out small subset of SCL signals needed to run TABs and
        GAB
      . again 9 ADF-to-TAB style cables required
    3) Trigger Functions
      . all trigger summing, global algorithms
      . communication with Trigger Framework, L2/L3
      . also deals with L1Error & L1Busy generation
  - This allows us to test the prototype TAB as it arrives
    . keeps us on schedule for July integration test
  - Also simplifies design of system
    . three simple boards vs. one complex board
  - Extra cost from this modification is negligible (~$4K)

o Discussion
  - Can commercial VME-to-Serial interface be used?
    . answer: probably simpler to do this as a custom board because
      the constraints are mainly on the TAB side
    . note: there is a lot of experience at Nevis building these types
      of interfaces - so it should be simple to get it designed
  - Can one (or more) of these boards be farmed out to another group?
    . answer: probably would take more time for a new group to come up
      to speed on system than it would for the designs to be done at
      Nevis 
    . note: design/layout phase for VME Interface and SCL Interface
      should take about 4 weeks each at Nevis