Online Delayed-Coincidence bit in UMC
				Takeshi K. Komatsubara (KEK)
				latest update: 2002-Nov-20


 **********************************************************************
  In UMC, the online Delco bit does **NOT** work as a cut on the tpi-tk
  time difference but as an **IC fiducial cut**, whicn follows
  if an event satisfies the Tdot2 requirement
  the event should certainly satisfies the online Delco in the trigger
  (regardless of the kaon decay time).
  All of the DELCO acceptance are obtained from real data, not from MC. 
 **********************************************************************



 In trigl0.F, 

C
C   Delayed coincidence was actually between CK and the analog sum of IC's.
C   Here, I will use a coincidence of B4 and individual I-counters.
C   (This was later modified to use t=0 as a reference.  NOTE: use of real
C   beam will give a different DELCO calibration.)
C
      IF(.NOT.B4OR)GO TO 1600
      IF(EICTRS .LE. 0.)GO TO 1600
      DO 1520 I=1,6
        IF(EICTR(I) .LE. 0.)GO TO 1520
        CALL TDGET( 8, I, 1, -1, NTD, ETD, TTD )
        DO 1510 J=1,NTD
C          IF(ETD(J) .LT. THRIC)GO TO 1510
          IF(ETD(J) .LT. ETHIC)GO TO 1510
          IF((TTD(J).LT.DELTIM)
     $         .OR. (TTD(J).GT.WIDTHK))GO TO 1510
          DELCON = .TRUE.
          NDELCO = NDELCO + 1
          GO TO 1600
 1510   CONTINUE
 1520 CONTINUE
 1600 CONTINUE
C

 
 In alldat.F

      DATA L0TRG, L0TRG2, DELTIM, WIDTHK, TWIDTH
     +   /     1,    0,     -1.,    50.,    20. /
C  L0TRG:  Level 0 trigger (turned on), using base 10 bits
C                1=PNN, 2=KP2, 3=KM2, 4=PN2, 5=PMM, 6=PGG, 8=ALL triggers
C                e.g. 1101 turns on PNN, KM2, and PN2
C                NOTE: L0TRG = 8 turns on all triggers in BOTH L0TRG and L0TRG2
C  L0TRG2: continues:
C                1=KP22, 2=KP23, 3=KM22, 4=PNNold, 5=min bias, 6=Kshort
C                e.g. L0TRG = 1 and L0TRG2 = 100 turns on PNN and KM22
C  DELTIM: Delayed coincidence (nsec).  Note, we usually leave this set
C          negative (i.e., impose IC geometry requirement only) and demand
C          true stopped K using IKSTOP (see above).  We then get all of the
C          DELCO acceptance from data.  This is not ideal, since some 
C          reconstruction code may benefit from having a minimum K-pi 
C          time separation.   Since the code works off real times, it isn't
C          always clear what DELTIM setting corresponds to, say, a DELCO
C          setting of 1.5 ns.  With the UMC beam (IBMMOD = 0), I found prompt
C          to be 3.5 ns, hence DELTIM=5.0 for a DELCO setting of 1.5 ns.
C          For IBMMOD = 1-3, prompt shold be approximately 0.0 ns, and for
C          IBMMOD = 4-5, prompt shold be exactly 0.0 ns
C  WIDTHK: Width of the KB pulse.  (Determines max K lifetime in trigger.)
C  TWIDTH  Half width of trigger coincidence time.

----