• Main INDEX
  • Monthly INDEX
  • PREV
  • NEXT

    User name R. Michaels

    Log entry time 23:52:33 on May 4,2003

    Entry number 99968

    keyword=deadtime

    Someone asked if a tiny deadtime was normal for 11 kHz if prescale is 5.
    The short answer is that it looks ok, however some of the halog end-run
    entries look wrong, e.g. run 1631 has livetime > 100 %. Probably a script bug.

    For a prescale factor K, the probability to have >= K events in time T at rate R
    is = Sum_(n>=K) (u^n exp(-u)/n!) where u = RT and T is approx 25 microsec
    in present setup. Note, if K =1 then deadtime = 1 - exp(-RT) ~ RT. So, for run
    1607 at 4.7 kHz with K=1 you got 11% DT, and for recent runs with K=5 the
    DT is much smaller. Its a little more complicated since there is also a readout
    time (T ~ 80 usec) which is buffered by the 25 usec frontend. But 1% seems ok.
    Intuitively, large K means the trigger becomes like a pulser (no deadtime).