• Main INDEX
  • Monthly INDEX
  • PREV
  • NEXT

    User name R. Michaels

    Log entry time 11:59:14 on February25,2006

    Entry number 163813

    Followups:

    keyword=bad7 versus threshold (DC)

    Regarding the DC readout problem (status2 chip error).
    We see some interesting structure to these "bad7" events:

    1. Usually several TDCs are affected in one event, then its
    quiet for a few thousand events (this is for cosmics). This
    cross-TDC correlation suggests noise on all inputs. Another
    structure we see is sometimes two consecutive events have multiple
    TDCs with "bad7" counts.

    2. On a scope we do see such correlated noise, looking at two
    inputs on two seperate TDCs we see occassional "glitches" of
    long or multiple pulses. Not yet proven these cause the bad7.
    Sometimes the glitches appear in rapid succession (hence consec-
    utive events) then go quiet for several seconds (1000+ events).

    3. This glitch rate appears to depend on threshold. We first
    verified this on a scope (counting glitches in 2 min):
    threshold .... glitches
    .. 3 ........ 12
    .. 4 ........ 14
    .. 9 (high).. 4

    This glitch rate is consistent with the bad7 rate for cosmics.

    Next we do the same counting of bad7 in data
    Thresh. ... Num bad7 events
    3 ....... 49 in 50K events (about 4 min)   run 2263
    9 ....... 1 (only one lonely event out of 50K, #2616).   run 2264

    The tentative conclusion is that some noise on the input is
    paralyzing the TDC chips. Next question is what to do about it ?
    Thinking .......