Shift summary for owl 1/14/00

Subject: Shift summary for owl 1/14/00
From: Vivian O'Dell (
Date: Fri Feb 04 2000 - 15:23:39 CST

             Shift Summary for owl shift, 1/14/00

On Shift: M. Barrio, J. Hamm

Hours of beam available at all intensities: 7
Hours of beam available at nominal intensity: 7
Hours of E799 data (special runs not included): 0
Hours of special runs (laser scans, muon runs...): 5.5

Accounting of lost time: Start/stop run, kfrending, kfrending, kfrending,
kfrending, ....

Runs ended this shift: 15508, 15509 (short reg)

Tapes brought to FCC this shift: None
Accesses during shift: None
Overview of shift: Kfrending interrupted by occiasional short reg/H data
Description of Shift:

run spill time comment
              0:05 l3 monitoring shows that a cpu on ktev2 is taking
                        too long to process; this leads to frequent daq_
                        dead alarms (plane 2 memories are not able to
                        empty by the start of the spill).
                        Stop run 15508.

              0:13 Start run 15509. l3 monitoring looks better, no
                        daq_dead alarms. However, the trigger is now
                        hanging every 15 minutes or so due to a DYC full
                        in crate 11 (which contains ADCs for RC, BAG,
                        UCAL,SA). We try turning off triggers 7 and 15
                        (which contain UCAL and BAG), but this doesn't
                        help. For now, we just kfrend each time
                        this happens.

              4:05 Veritable onslaught of front-end problems. Crate
                        11 hangs almost every spill; when it doesn't, we
                         get busies or ktevana errors from a host of other
                        sources. After repeatedly kfrending, cycling
                        repeaters, initializing the pipeline, we fall into
                        a state of despair and page our run manager so
                        that he, too, may feel our pain.

                        Leo advises us to try pushing the "reset reg adc
                        threshold" button on the ktrig window after
                        kfrending. This button, which apparently
                        spontaneously appeared without warning sometime
                        over the past few days, will sparsify some
                        channels in crate 11, leading to fewer DYC fulls.

                        This works. The lesson to be learned is: if
                        you're having intractable problems, and there's
                        a new button somewhere, try pushing it.

            4:45 Already annoyed and frustrated shift person (OK,
                        it was me) discovers there are NO CLEAN COFFEE
                        CUPS. Morale plummets.

            5:00 We surpass short reg run's goal of 700K reg
                        2pi events. Now, if we can only meet 799's goal
                        of 2 x '97....

            6:25 Leo calls and instructs us to take a run in the
                        BEAM_MU_SHORT configuration, so we stop run 15909.
                        The DAQ really didn't want to stop, though -
                        triggers kept going, and the stop process hung.
                        Fresh start, change tapes.

            6:55 Start muon run 15512.

                        Bob T., upon hearing about our troubles, digs out
                        a post-it note that had been hidden, upside-down,
                        under the ktrig terminal keyboard,
                        which reads: 'After every kfrend, please also hit
                        the reset reg adc threshold button'.

                        It's a good thing it's time to go home.

Follow up Notes on Previous Shift

Items to Follow up on Next Shift

Is it the 17th yet?

This archive was generated by hypermail 2b27 : Fri Feb 04 2000 - 15:23:41 CST