• Main INDEX
  • Monthly INDEX
  • PREV
  • NEXT

    User name R. Michaels

    Log entry time 09:42:26 on March15,2005

    Entry number 141578

    keyword=deadtime studies

    Since we have no beam, I did some cosmics runs to check about
    the "negative" deadtimes. It was a sequence of runs in which
    each trigger takes a turn having its prescale factor 1 while the
    others have infinite ps factor.

    First, the "bottom line": With cosmics I see normal (i.e. zero)
    deadtime for each trigger. Exception: T5 couldn't be tested
    unless we went in the hall to plug in pulser (T5 is very low rate).
    Exception#2 : T7 is 7500 Hz, so only has zero deadtime if the
    PS factor is high enough (obviously).

    Another issue is the halog entries for "livetime": For a
    typical production run, the livetime is wacky, e.g. 160%.
    This is because T7 has an extremely high rate and the scaler
    over-flows. It may be a negative number or some wrap-around
    value, which throws off the calculation. I would expect that
    when we calculate livetime for each trigger using the class
    THaNormAna, all will be fine. Haven't done it, though.
    A test one could do: Do a production run with PS for T7 = 99999.
    This turns it off and the halog deadtime should be reasonable.

    Here are the runs:
    For runs prior to 2685, I had PS for T7 = 9999 which was not
    turning it completely off (but didn't matter).

    Run.....Trig.....PS-factor....Deadtime.......Halog Livetime
    ..............................(10 sec avg)

    2678 .... 1 ....... 1 ........... 0 ........ 99
    2679 .... 2 ....... 1 ........... 0 ........ 90
    2681 .... 3 ....... 1 ........... 0 ........ 92
    2682 .... 4 ....... 1 ........... 0 ........ 98
    2683 .... 5 ....... 1 ........... 0 ........ 98
    2684 .... 6 ....... 1 ........... 0 ........ 97
    2685 .... 6 ....... 1
    2687 .... 5 ....... 1
    2688 .... 4 ....... 1 ........... 0 ........ 92
    2689 .... 7 ....... 100 ......... 0 ........ 95
    2690 .... 8 ....... 20 .......... 0 ........ 95
    2691 .... 7 ....... 5 ........... 15 ....... 78 .... (ok)

    I suspect the problem with datamon during production arises
    from the extremely high rate of T7. Otherwise, I see no problem !