• Main INDEX
  • Monthly INDEX
  • PREV
  • NEXT

    User name R. Michaels

    Log entry time 22:29:47 on July 5,2000

    Entry number 45330

    Followups:

    keyword=Trigger Modifications

    During the July 4 shutdown the following modifications were made
    to the trigger, with implications noted.
    The two purposes: 1) Implement a pulser to measure electronics
    deadtime; and 2) Include the Aerogel detector on H-arm.
    DEADTIME -->
    The pulser goes into a fan-in together with the scintillator signal.
    The 4 PMTs involved are S1-1-L, S1-1-R, S2-1-L, and S2-1-R.
    The fan-in plus cabling introduces delay: 9 nsec on E-arm and
    14 on H-arm (longer cables). This is compensated by:
    a) increase all delays on s12e(h)_L(R)_delay channels except for
    the channels to fan-in. And b) decrease h_delay_1 by 8 nsec
    with compensation by 8 nsec on h_delay_2.
    The pulser introduces noise: about +/- 15 mV on H-arm (worst
    case) and much less on E-arm. This is caused by pickup on long
    cable, poor grounds etc. It broadens the pedestal a factor of two.
    Gate is shifted on E-arm, but has no affect on amplitude of
    ADCs for scintillators, preshower, shower detectors.
    A pulser flag, for signal above 45 mV (the same threshold used
    by the CAMAC discriminators) goes into two TDC channels
    ROC1, slot 19, channel 84 (numbering starts at zero).
    ROC2, slot 9, channel 84.
    The pulser flag also goes into two scaler channels
    Channel 15 (numb. starts at 1) on E-arm normalization scaler
    Channel 22 (" ") on H-arm normalization scaler.
    Each pulser will be a T1, T3, and (for infinite momentum) a T5.
    At some point, below momentum setting approx. 800 MeV, the
    pulser will cease to make T5, but will always make T1,T3.
    The pulser rate is 6.8 Hz (fixed rate, not related to beam current).
    Analysis of the 4-second updates of scalers under stable beam
    will lead to deadtime correction. Relative accuracy depends on
    deadtime. Example in 1/2 hour: 0.25+/-0.05%, or 1.00+/-0.10%.
    AEROGEL -->
    Aerogel was plugged into MLU2 on the 4th input.
    New definition of T3 was desired and is now the default:
    T3 = S-Ray and "NOT" Aerogel (selects protons or otherwise no Aerogel signal)
    New trigger T6 is "S-Ray and don't care about Aerogel" (it can be on or off).
    Consequently, all T3 are also T6. Therefore we delay T6 relative to
    T3 at trigger supervisor input (100 nsec). Also, to correctly estimate the
    "trigger averaged DAQ deadtime" (displayed by datamon, though it is
    not a truly meaningful quantity; better is the deadtime for a particular trigger),
    we subtract T3/p3 ratio from T6.
    One danger: Since MLU2 is in transparent mode, fluctuations in
    the timing or width of Aerogel signal can cause fluctuations in the
    width of T3. This was minimized by making a good overlap (for cosmics),
    but could be a problem with beam. One should look at coincidence time
    spectra for T5: Are the randoms flat ?
    Also, it may be wise to study the T3's without requiring Aero.
    (An aside: it was impractical to make a T7 from T1.and.T6.
    So we did not do it.)
    Here is how to switch from caring to not caring about Aerogel:
    First type "gotrigger" to go to correct directory on adaq account, then
    cp mlu2_hadron_requireNoAero.data mlu2_hadron.data (DEFAULT)
    cp mlu2_hadron_dontCareAero.data mlu2_hadron.data (never mind Aero)
    and then download trigger again.
    T6 plugged into scaler 21 on H-arm normalization scaler.