Present: Erik Katsavounidis, Massimo Orsini, Valeriano Marrelli,
Sophia Kyriazopoulou, Lori Gray, Chris Orth, Hwi Kim(week 2), Nicola
Zaccheo (week 2)
Week one was pretty quiet. Monday morning the acquisition crashed for
some unknown reason - microvax one was filled with highly informative
"NO-NAME" error messages. But we caught it quickly, even before
Erik's monitors knew something was wrong, and the run restarted
easily. Unfortunately the run soon lost contact with microvax 2 and
had to be restarted again. During week 2 the biggest problems were
the frequent shutoff of the gas system, and both EVD and DGN stopped
working on the main console due to some yet unfixed problem with
creating GIF output files.
The hot PHRASE channel from last week (1T05) "mysteriously" fixed
itself, though most of us suspect remote interaction was taken from
our friends at Pisa.
Similarly a dead LIP channel, 5E13, fixed itself when Nicola wiggled
its cables during the second week of this shift.
Chuck emailed us about a CSPAM inefficiency in the interSM events
during the first week of this shift. Erik believes that these may
have been due to microvax 3 being out of acquisition during runs the
week before that were probably used in Chuck's efficiency analysis.
For several run the WFDs on SM2 were producing a couple of events per
run that looked like some of the odd events Erik showed at the Boston
meeting - pedestal droop at the level of -4V at peak, and several usec
long. This is most likely due to an actual "signal" going into the
waveforms. The pedestal droop is expected because of the
differentiating nature of the WFD discriminator (for a wide enough
signal, the WFD cannot discriminate over the duration, so the signal
appears to be "pedestal".) By taking a quick survey of the channels
producing errors in Ed Kearn's WFD monitoring code, we found that only
half of the PMT signals on SM2 are involved. These channels exactly
coincide with with one PMT fanout crate (of the 2 crates, one has
channels that are all "good" and the other has channels that are all
"bad). Unfortunately this is not conclusive, since the 2 LeCroy HV
modules on SM2 follow the same division. Nicola showed events from a
long time ago that also appeared on the same channels. So this
wednesday the LeCroy supplies will be swapped. If the problem appears
again in the same place, the PMT Fanout crate is suspect. If the
problem moves, the Lecroy supply is suspect.
Who's betting that the problem never appears again?
On Wednesday calibrations could not be performed do to some missing
system libraries. Sandro Marini discovered that the problem was due
to a filled system disk on the calibration VAX. Apparently when the
VAX was rebooted, VMS noticed the lack of space and sandbagged some
files that were essential to calibration acquisition. Wednesday
afternoon the technicians performed a test run and verified our backup
acquisition system was restored.
The second week, calibrations were more productive. The entire
calibration system has almost been moved to the spare microvax on uvax
2 from the main uvax nearby. The advantages to this move are
numerous. Calibrations will not be subject to delays in other
acquisition going on in uvax and vice versa. We expect an increase in
the efficiency with which we can do time consuming calibrations like
LAMOSSKA calibrations and TOHM/LI sensitivity measurements. Also, a
new scheme was developed for these calibrations whereby the pulsers
are triggered by the turn off of the Computer Busy signal plus a
delay. A new calibration module is being continually designed by
Massimo Orsini to form this trigger to the pulsers, and to switch
easily from the old to new pulser configurations. Some minor problems
(causing mysterious asymmetries between end 0 and end 1 calibrations
in the configuration were fixed. One problem was that Massimo's
trigger module couldn't drive the 1 MOhm input impedance of the
pulsers.
Chris Orth performed a special Laser calibration with WFD readout of 2
channels per SM enabled (one horizontal and one vertical). This data
will be used to compare WFD to ERP response, and to test the WFD
response for fast monopoles.
Erik has been doing a lot of work lately creating a Web interface into
not only his infamous "stat" files (a.k.a "Extended" LGB), but also
Paolo Saggese's "MSP", a graphical representation of these text
files. Now any MACRO user can get clickable up-to-the-run information
on all of the scintillator systems. Many other files have been made
web available as well, including but not limited to Rare DST
summaries, macro data locations, and tutorials for the monitor jobs.
Work is underway toward processing all existing MACRO data into Rare
DSTs. Soon we will have batch jobs that copy and convert MACRO Zebra
data back into MACRO raw data, and then launch parallel versions of
Erik's automatic jobs. A side effect of this is that in addition to
regenerating raw and zebra (wfd-combined) versions of the rare dsts,
we will also get, almost for "free", statistics files for these runs,
as well as other "DSTs", including the "lamosska" DST (containing all
lamosska events, mostly showers), the "ERPGC" DST (which has PHRASE
and WFD data stripped, giving 90% reduction), and hopefully an "UPMU"
DST, containing all of the data (including WFDs) for upmu events "in
the tails" as well as within the 1/beta=-1 peak. So far we have
specific requests for event from F. Cei (for monopole search related
studies), F. Ronga and A. Surdo.
While testing the job that converts Zebra data back into raw
data (thanks go to Alec Habig on this code), Erik discovered that some
events are being thrown away during the standard MACRO Raw to ZEBRA
production that is performed for all MACRO data. By looking at the
kept logfiles, we know that since Aug 1995 we have lost about 250
events. In order to check for systematic (are the lost events all
upwards muons?) effects, we have been looking into events as the
problems occur, while we still have raw data on disk that we can go
back to. By modifying DREVEN a bit, we have obtained raw dumps for
five events so far. 2 of these are ERP only events. In one the
decoding apparently failed during STAS readout (the last successfully
found equipment). In the other the failure was during WFD readout.
Another event was a tohm only event with readout problems again in the
WFD, while another event was an ERP GC only event. There is not
enough statistics to say anything else meaningful yet, but so far
things seem random.
On a final personal note, congratulations to *Dr.* Ioannis
Katsavounidis. For MACRO's sake let's hope the job offers don't come
too quickly.
And as for L'Aquila these days... "Wow what a difference!"
That building with a large blue awning with ugly yellow block letters
between the Superals is, indeed, a Blockbuster Video. And Teatro Rex
will be stomped out by the real King of American food-culture -
McDonalds (unless Alessandra DiCredico can rally enough support
against it). I'm with her on this one.
Chris Orth
Boston University