Ralph Minehart opened the meeting at 9:00 AM. He stated that he was especialy
concerned about the track detection efficiency and the status of the time-base
tracking. He stated that the main question facing the group was a determination
of readiness for data taking in December. During the meeting, these and many
other important issues were discussed.

A number of presentations were given on the status of the Sep/Oct run analysis.

Jim Mueller discussed the status of hit-based tracking in RECSIS. He stated
that the RECSIS tracking was diverging from the SDA program. Important
questions for the hit-based tracking were associated with pattern recoginition,
specifically, segment finding, segment linking and the fit to wire positions.
He reported a big improvement between July and September, largely due to fixes
of hardware. As a result, the efficiency in track finding has gone from 80% to 92%.
This is not as good as we need, but the progress is encouraging. T
here are still peculiarities in the soft-ware, such as some unexplained large
chi-squares. He also reported that the efficiency for finding the proton track
after making cuts on elastic electrons was 98% for 1.6 GeV runs.
 

Franz Klein reported on the status of time-based-tracking in RECSIS.
One problem that he discussed was that we need to identify an electron
to provide the start time for the time-based tracking. At the present
time something like 2/3 of the events are lost due to lack of an
identified electron. In the event sample he has studied, 16% of
the events have an identified electron, whereas the fraction
should be 40-50%. This problem might be fixed by using fast pions
to provide the start signal. He also reported that the resolution of
the drift chambers in his analysis is not very good yet, in the range
of 600 to 1200 microns, whereas the design goal is 200 microns.
This is probably not an important worry as he is using one time
curve for all 6 super layers. Improvements are expected in the near future.
 

Dan Carman reported on the status of the drift chamber system.
In general the situation looks good, and he expects that by December,
the number of problem areas will be reduced to 2-3% level in each of
the three regions. The problems are in high voltage supplies, high voltage
disconnects, electrical connections, low voltage supplies, missing wires,
etc. He also presented results of luminosity studies in the form of drift
chamber occupancies as a function of beam current. These studies were run
over a period of a few hours with stable acclerator and detectors.
The occupancies are linear with current, and is less than 3% at
luminosity of 10**34 for all regions. At the high beam current non-track
hits are a big fraction of the occupancy.
 

Cole Smith reported on the status of the EC and its calibration.
He is currently running cosmic ray tests almost every night, and
these are leading to steady improvements in the gain settings.
Several bars require repair work, (such as replacement of the photo-tube),
and he will do this before the December run, especially for those bars that
are in the small angle region. He reported on studies of calibration using
pi-zero signals, a pair of showers in the EC with no track pointing to them.
Such events may be useful for on-line calibration. He also reported on
studies of the EC trigger efficiency. He found that sectors 4 and 5 provided
systematically fewer triggers than the others. This could be due to a
misalignment of the EC relative to the beam line. He is continuing to study this.
He also reported that the fraction of triggers at 4 GeV associated with
negative particles was around 20 %. These runs were made with the threshold
at half the beam energy. Data runs will probably require even lower thresholds,
so it is clear that we must have the Cerenkov counter in the trigger.
 
 

Volker Burkert reported on the results of Alex's Cherenkov counter efficiency studies.
Sector 1 has an efficiency problem in the center, and the reason is not known.
It is possibly due to a mirror misalignment. Sectors 4 and 5 have good efficiency.
Sector 6 shows very poor efficiency over most of its area. This is attributed
to the discovery after the run that someone had cut the gas line to this chamber
in order to steal a gauge in the line. Thus the chamber was drawing in air during
the entire run. He also reported that windows for the remaining two sectors will
arrive on Nov. 11, and it is possible that if things go well another chamber
can be installed before the December run.

Kyungseon Joo reported on trigger studies. He showed the EC threshold energy
calibration based on three different threshold settings (150 mV, 250 mV and
350 mV) by looking at the histogram of reconstructed electron momenta.
He showed that at 150 mvolts discriminator setting for 1.6 GeV run we see
the second resonance region, whereas this is mostly cut out with a 250 mV
setting. He analyzed the runs, 6079 at 1.6 GeV and 6389 at 2.4 GeV. Run 6079
used high threshold EC trigger and run 6389 used two triggers, EC and EC*CC
triggers. For run 6079, he studied CC efficiencies by looking at the
electron x and y hits in the EC with/without matching the drift chamber and
CC information for each track for electron identification. For this analysis,
he used inclusive and exclusive data for elastic and inelastic proton
scattering (e,e'pi+) to study a wide range of electron momentum. T
he results showed an overall efficiency of about 77% with either inclusive
or exclusive data. There were also local inefficiencies due to the difference
in geometric acceptance between the EC and CC. Inelastic scattering data also
show similar results. For run 6389, he analyzed two types of triggers separately.
One is evtype=1 (CC) and the other one is evtype=14 (CC .and. EC).
The spectrum of the hadron invariant mass W for evtype=14 is missing the second
resonance peak. This shows that we need a better understanding of our
trigger types. More studies are underway.
 

Simon Taylor reported on the status of the TOF system calibration.
The process is still going on using cosmic ray runs. He noted that
he can't do all the sectors until the detector system is closed up.
The time-walk correction is being done with two different functions
for two regions of pulse height, close to threshold where the walk
changes rapidly with pulse height, and farther from threshold where
the correction is linear. The functions are matched at their connecting
point. He reported that the time resolution of the scintillators
obtained with laser runs was approximately 100 ps, except for a
few peculiar bars.
 

Latifa Elouadrhiri reported on her work to incorporate energy loss
in the scintillators into the particle identification. The difficulty
is that this requires complete calibration of the TOF bars.
By selecting certain combinations of pairs associated with the
electron and proton in elastic scattering, she has been able
to calibrate a sub-set of the bars for the 1.6 GeV runs, and shows
excellent results in proton and pion separation on a delta-E vs.
momentum(from wires) plot.
 

After lunch we had a series of reports on data acquisition, monitoring
and beam-line instrumentation.
 

Dieter Cords reported on the status of the on-line system He displayed
a block diagram of the data acquisiton hardware and mentioned that we
now have two RAID disks for data storage. He reported that the data
acquisition rates can be summarized as follows. Sept. 97 December, 97
(projected) Trigger; 500-600 Hz 700 Hz to 1 kHz Data : ~1.6 Mb/s 2.5 Mb/s

Event Size: ~3 kBytes (4.5 kBytes at 4 GeV) File sizes are limited
to 2 Gbytes,so that new files will be created about every 20 minutes.
With 30 GByte disk partitions, the disk will need to be dumped about
once or twice per shift. He also showed us the new start-of-run GUI,
which, hopefully, will make the run starts simpler and more fool-proof.
 

Mark Ito discussed the status of run monitoring and controls.
He presented a check list that was developed mainly by Dave Tedeschi.
He emphasized that further development and maintenance of the
documentation should not be dumped entirely on Dave. There is a document
that is keyed to the check-list, but the connections need to be improved.
People should go through this documentation to make sure it is correct
and comprehensible. It should be possible for individuals to take on the
tasks of upgrading individual sections. Please contact Mark for more details.
The sstem check list can be found at http://clon00.cebaf.gov/clas/
 

Arne Freyberger discussed the beamline instrumentation for e1.
The beam position and beam current monitors will be put into the
data stream. Harp scans do not fit naturally into the data stream,
and will be maintained as separate files, as in the past. There is a
new GUI for the harp scan beam profile, using GNUPLOT because that is
more natural for the accelerator operators. A thin optical transition
radiation (OTR) monitor will be installed immediately in front of our
target. This will be viewed by a mirror and camera system, and will
require commissioning. It is likely that the accelerator group will
ask for some part of our scheduled time to do this task. It is also
hoped that the target cell for the December run will be a 100% plastic cell.
The first two tested had leaks, but the leaks were due to design features
that could be easily changed. A third cell is being tested now.

Arne also reported that he is trying to inocorporate two additional
16-channel VME scalers into the EPICS system. Among other things this
expanded scaler capability will allow us to monitor the rates in each
of the six EC sectors, and the rates in six inner TOF rings.

Arne hopes to get 16 hours of beam time between Nov. 17 and Dec.4 to
perform system checks. If this is not possible, the system checks will
be done during the December run, or possibly not at all.
 

Kyungseon Joo showed his analysis on elastic scattering and inelastic
scattering using 1.6 GeV. He showed several kinematic cross checks
using scattering angles and momenta from electrons and protons.
The results are in the right ballpark and the disprepancy can be
interpreted in terms of chamber misalignment, magnetic field and
energy losses. He also showed his analysis on inelastic scattering.
He explained his analysis of particle identification to separate
protons and positive pions in terms of E-Delta E in the TOF counters.
since complete calibration of TOF bars are not complete, he used a
simpler but incomplete way to approximate the calibration.
Despite this being a "quick and dirty" method, the results were
quite good and certainly help to give us confidence in the quality
of the 1.6 GeV data.

Joe Manak reported that the study of the elastic ep rates is now
producing a cross section in reasonably good agreement
with the known cross section. Typically the measured cross sections
are about 60-70% of what they should be. Part of this discrepancy
will be resolved by radiative corrections (which I think are
typically of the order of 1.1 for the cut he made on t
he elastic peak), and a better handle on our efficiencies, etc.

Marco Battaglieri presented results of systematic studies of
data related to the performance of the large angle EC.
This report requires the slides that he showed. Basically
their analysis of the detector performance at 1.6 GeV is consistent
with expectations.
 

Ralph Minehart reported that he has programs to calculate the expected
inclusive electron rates, both for elastic and inelastic scattering.
The inelastic rates are based on the Brasse parametrization.
An approximation of the CLAS acceptance is included in the calculations.
Anyone wishing to use either the results or the programs
 

Joe Manak gave a report on the plans for data analysis.
He said that at the moment on the CEBAF CPU's RECSIS requires
50 ms/evt, so that 20 events per second can be analyzed. With
ten CPU's this translates to 200 evt/sec, which for the Sept/Oct
data means that it takes about twice as long to analyze data as
it does to acquire it. For the E1 run, he estimated that it will
take about 6 times as long. This is not a major concern,
however, since CPU's are cheap and getting cheaper. The real
concern is the quality of the analysis program. He thinks that
a significant time is required to get the program into proper shape.
Some specific problems that he listed were:

1. The analysis code ocassionally dies in BOS.

2. Time based tracking is up and down.

3. Particle I.D. is still a concern. This involves the integrity

of the SC, EC, CC and pid codes)

4. Calibration is not complete.

Volker Burkert presented an analysis of luminosities and trigger
rates and the expected time needed for data at energies of 1.6
and 2.4 GeV in December. He proposed that we run for 8 days at 1.6
(4 days times an efficiency factor of 2), and spend the remainder
of the December run at 2.4 GeV. There seemed to be general agreement
that since no real analysis of 4 GeV data had yet taken place,
it was not possible to make any sensible choice of trigger and
luminosity for this energy. Furthermore, it seems rather inefficient
to make an energy change when only two or three days of 4 GeV
running would be possible in any case. The next meeting will be
on Sunday, Jan. 11, 1998, following the CLAS Collaboration Meeting.
The meeting was adjourned rather informally at about 5:30 PM.
 

Attendance:

Marco Battaglieri
 

Volker Burkert
William Bertozzi
Dan Carman
Dieter Cords
Hovanes Egiyan
Latifa Elouodrhiri
Arne Freyberger
Herbert Funsten
Mark Ito Kyungseon
Joo Franz Klein
Sebastian Kuhn
Joe Manak
Bernhard Mecking
Ralph Minehart
Jim Mueller
Bogdan Niczyporuk
John Price
David Rowntree
Cole Smith
Simon Taylor