DAQ and Trigger (Karen) --
Missing ep and en coincidence events from yesterday (including empty
target runs) was solved. Right sector paired TOFs were always TRUE going
into MLU. In diagnosing this, had to disconnect a number of cables. After
reconnecting everything, the problem was gone (although all the cables had
appeared properly seated to begin with).
A logic error was introduced into the trigger settings on 2/25/04 by
Karen, when attempting to remove the logic that said "electron is TOF AND
Cerenkov or TOF AND neutron". Instead, since then a sector hit has been
"TOF AND Cerenkov AND maybe neutron" OR "TOF AND no Cerenkov and no
neutron". This is a problem anywhere we have neutron detectors behind the
TOFs (most of the right sector, and the back left sector). Effectively we
have been discriminating against pions in those regions, also against
higher energy protons or deuterons that don't get stopped in the TOFs. For
electrons, we have essentially been requiring the Cerenkov in the trigger
except where the TOF and neutron bar acceptances don't match. This
explains the high R Cerenkov efficiencies seen after adding L shielding on
2/24. It also might explain the low electron cross section at forward
angles in the Right sector, if the RC0 efficiency is low. The error should
not cause any asymmetries. Karen will fix it today.
Vitaliy notes many events in the upstream end of the Ohio walls for the en
trigger. Suspect this end is not covered by the TOF, so you can't use the
TOF as a veto there in the hardware trigger.
Our deadtime comes from several sources. It takes 800usec to read out an
event. We do NOT use the buffering feature in the ADCs and TDCs, so we
have a deadtime that is linear with event rate. There is also a non-linear
part as the rate increases, and the probability of getting a second event
in that 800usec goes up. We could experiment with buffering during the
spring shutdown; you want to be careful that your event fragments match
up. Finally, there is the wall we hit due to the Ethernet bandwidth (at
about 340 events/sec). That wall can be moved to higher rates by
introducing ADC readout thresholds, or by buying new PowerPC readout
controllers.
ABS (Genya) --
On 3/22, found that the top half of the transverse holding field coil was
shorted out. This explains the measured tensor asymmetries over the past
10 days. The measured transverse and longitudinal fields aren't balanced
now, but the probe isn't in the ideal location.
Genya is away 3/26-3/29. Vitaliy is the target expert for that period.
The intensity is down 20% since the nozzle warmup. It has been taking
longer after each warmup to regain intensity. This is a sign the nozzle
should be replaced; will happen 3/30 when we're open to fill the
solenoids. At least a full day job.
The tensor elastic asymmetry from overnight is nearly consistent with zero
(-2 +-4 and 6+-4, should be +12 and +30%). Vector qe looks fine. Could
the tensor - state really be vector? Some checks will be made today. We
may try pure longitudinal spin again.
Quasielastic tensor statistics are still limited, can't say anything about
the asymmetry there (Vitaliy).
Neutron Bars (Michael via email) --
The time calibration macros are almost final, also the macros to find the
TDC offsets. Eugene, Sebastian, Vitaliy and Michael are working on
this. All timing parameters are collected in blast.sc_cal, to be used when
crunching.
The timing walk has been measured on all LADS except the 15cm "right"
wall. Ohio also still needs to be done. The technique varies the flasher
input to the splitter boxes, and Ohio&L15R are in the top box with the
photodiode. Varying the input means the photodiode doesn't fire the
trigger. Will work on that next Tuesday during the regular hall access.
Also next Tuesday, will try the RC circuit to get rid of the ringing in
the phototube pulse that makes the time walk correction discontinuous.
Cerenkov (Baris) --
RC0 ADC looks narrow, seems to be a lower than expected number of
photoelectrons in that box.
One possibly bad tube (LC1 top?) to diagnose and replace next Tuesday.
Beam (Shannon, Townsend) --
Tried a lower injection rate last night for the fill (2Hz). Fills slower
(10 seconds more) but better injection efficiency (by a factor of 2). No
effect on detector rates, but may be better for wire chambers since more
beam gets into the ring instead of being splashed around.
If the beam tune didn't have to satisfy the Compton, BQM rates could be a
factor of 3 lower with the slits out. When we are done changing the
transverse holding field, Ops needs 1 day to tune best beam.
Longitudinal holding field voltage is oscillating +-10% since
yesterday. Townsend will investigate.
Compton (Bill) --
Circular light polarization correction is bigger than though (Pockels Cell
is damaged). Electron polarization is a few % bigger than reported, closer
to agreement with the transmission polarimeter at the front end (75%
yesterday). Adjusted transmission of laser light yesterday so that only
central part sees the electron beam, polarization is 99+%. During
shutdown, will replace the Pockels Cell.
The beam tune to satisfy Compton and BLAST makes compromises in the
vertical steering (compensation for the transverse holding field). Perhaps
adding steering coils will help; Townsend and Shannon will look into that.
General --
Collaboration meeting 4/2.
Analysis meeting 4/1? Some outstanding analysis issues:
1) Tosca for Cerenkov shielding (Vitaliy)
2) Beam energy reconstruction -- do we get the beam energy? Same in L and R?
3) Resolution?
Karen
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:30 EST