GEn Analysis Meeting for Feb. 21, 2007


Main meeting upload page

Follow the link below to access the plots provided for the meeting:

http://hallaweb.jlab.org/experiment/E02-013/mtg/2007feb21

Meeting minutes and contributions
Present: Sergey Abrahamyan, Rob Feuerbach, Aidan Kelleher, Jon Miller, Bogdan Wojtsekhowski; and by phone Ameya Kolarkar, Nilanga Liyanage, Tim Ngo, Seamus Riordan.

Collaboration meeting is scheduled for March 13 at JLab

Nilanga will send out a message tomorrow announcing it.

Brandon Craver (left a report)

I am currently working to finalize the raster calibration. The GenBeam? class has been altered to perform
a running average of beam position as reported by the BPM’s. This average position, along with the beam
shift reported by the GenRaster? class provides an event by event beam position. The evaluation of the
performance of this code will be completed and posted within a week.

Seamus Riordan

Plots at http://hallaweb.jlab.org/experiment/E02-013/mtg/2007feb21/riordan/riordan-20070221pres.ps.

The last two weeks were spent working on improving the measurement for GEn
and improving the optics code.

For GEn I continued looking into the background by using the event offset
method suggested by Gregg. A number of changes to how my code performed were
required to do this type of analysis. A list of cuts and runs, as well as
other plots from this analysis are found in the subdirectory of where you
found this file.

The basic results are as follows:

page 1-3:

These are the pmisspar, pmissperp, and W plots for the measured counts of
total (black), charged (red), and uncharged (blue), in the upper left. The
background is the upper right, and the background subtracted in the lower
left. In the lower right is the asymmetry for the charged (red) and uncharged
(blue).


I did some measurements of the background by taking things out of time
and calculated what percentage the background is compared to the analysis:

Neutral background ratio = 0.414 +/- 0.002
Charged background ratio = 0.067 +/- 0.000

So, it was determined that about 40% of the neutral events are background in
our cuts and 7% of the charged events are background.

Working this into my analysis with a 2.5% proton leakage rate I get an
uncharged to charged ratio:

Unch/ch = 0.083 +/- 0.000

Which, if 0.4x_s(p) = x_s(n), a 30% neutron detection efficiency and a
70% proton detection efficiency, we expect:

Unch/ch(expected) = x_s(n)*0.3/(2*x_s(p)*0.7) = 0.086

So that puts us in the right ballpark. GEn was:

Page 4:
GEn(Q2=1.69GeV2) = XXXX +/- XXXX (see original file for value)

There is a considerable amount of other things to try to try and understand
the background. I think looking in other sideband areas, such as things out
of time, out of space, etc could be very useful. Also, looking at other event
types would be very interesting, such as T2s. I believe that the areas that
go negative in the background subtraction are significant candidates for
looking at other physical processes that are showing up in the background
which may not be correlated with quasielastic events (and therefore should not
be subtracted).

Also, contributions to the error right now need to be studied. I have not
looked into this yet, though I believe that a considerable amount of error is
currently coming from applying false asymmetry calculations. I need to make
some minor modifications to my code to study this, though my focus right now
is not on GEn.

The rest of my time has been spent looking at the optics code for BigBite. I
have spent some time implementing a new version that moves away from our
effecive bend plane model and into an effective field boundary model. This
allows us to apply certain types of corrections more naturally.

page 5 (left): effective field boundary model

Essentially, the tracks are traced to the field boundary, still assumming that
there is no change in the non-dispersive direction, so that we can find a
unique point along the beam. First order corrections are applied to the vertex
Z in the LAB coordinates (as we have done in the past). The intersection with
the field boundary closer to the wire chambers is uniquely defined by the
information from the drift chambers. A point on the midplane in the dispersive
direcion is determined by tracing the track to that point. The intersection of
the line connecting the midplane and the beam and the first field boundary
defines the track intersection.

A radius of curvature can then be calculated by enforcing a circle must have
two points intersecting the points on the field boundary and the derivatives of
the circle must match the slopes of the tracks leaving and entering the field
at the points of intersection.

The radius of curvature is proportional to momentum of the particle. H2 data
can then be used to calculate this. First and second order corrections can
then be applied to the momentum reconstruction.

One correction that was used in previous BigBite analysis was, as I call the
Penner correction, to account for fringe field effects. As a charged particle
enters the field non-perpendicularly, there will be some bending in the non
disperive plane.

Page 5 (right)

This correction is of the form:

d(phi) = -y * tan(beta)/R

where phi is the slope in the nondisperive direction, y is the distance
from the symmetry plane (as we call y in target coordinates), beta is the angle
that the track makes with the plane normal, and R is the radius of curvature.
A similar correction is applied with opposite sign as the track leaves the
field area.

These corrections should be readily apparent in the carbon foil data. As of
now I have the code rewritten and able to run. Analysis is currently being
done.

I plan to continue with studying these effects and seeing what can be done
in the realm of the extreme vertical positions and their corrections.

Nilanga and Bogdan have submitted an abstract for the APS meeting in
Jacksonville, FL in April over BigBite. Gregg and I have determined that it
would be worthwhile for me to give this talk.

Ameya Kolarkar

Ameya showed his progress and near-final results for the dilution factor at his website:
http://www.jlab.org/~kolarkar/gen_analysis/dilution.html.

  1. Nitrogen dilution factor, D, for a few runs is at the end of the dilution page.
  2. Looking into how different momentum cuts (Pperp, etc.) affects the N₂ dilution factor.
  3. The automated script to get the resolutions of the BB Scint. detectors now works in batch mode also - inputs re the start and end runs. This still has a bug - it only works for successive runs. Will try to fix this.
  4. Took a first look at the BB Scint. calibrations for kinematic 3 runs that have been analyzed. Not nearly enough statistics!
  5. A nitrogen dilution automation script is complete. It's still in the semi-automatic phase since scaler information needs to be input. This script takes all temperatures from the RTD EPICS variables from the MySQL database and averages them over the run period. This does not yet incorporate possibly different RTD offsets, beam trips, etc. But I think it's better than just one reading from the start_run halog entry.
  6. Documentation of all the above work is in progress.
  7. Working on the asymmetry analysis for kin. 4.
  8. Speed Determinant: Slowness Factor: None that I could think of. I'll get the target logbook copies when Chiranjib (another Wolfgang's student) comes over to Ky end of Feb. And the D analysis is separate from A analysis (kin. 4). Otherwise, as soon as some kin. 3 runs are analysed, I can check the BB Scint. Calibration. I think there's enough documentation to begin work.

Bogdan

Bogdan mentioned that the present knowledge is the angle of the BB field to be 117.76 deg at the middle of the target, this is ~2 degrees away from the ideal direction.

Tim Ngo

Tim has been working on his calibration quality-check scripts, and has come up with his own set of time-offsets for the neutron detector bars. See the bottom of the page at:
http://www.jlab.org/~ngo/analysis/time_offset/time_offset/time_offset.html

He will:
  1. create plots showing the residual offsets with the predicted time-of-flight effects removed,
  2. send his script for getting the time-offsets to Jon,
  3. expand his method to look at the planes 5-7 as well.

Sergey Abrahamyan

Sergey is working on a Geant4 based MC of a simplified neutron detector, to help with understanding the effects of charge-exchange processes, primarily proton + nucleus -> neutron. A sample event from his MC can be found
at
http://hallaweb.jlab.org/experiment/E02-013/mtg/2007feb21/sergey

Presently the "Physics List" contains only electromagnetic processes; he will be adding in hadronic processes as well as a way to save the events to an ntuple-like format for further processing.

Jon Miller

Jon presented his latest and more complete Veto Efficiency calculation. He finds that the overall efficiency of the veto detectors to detect a proton (using an OR between the layers) is ~96%. He will continue working a little more on this, as well as determining better timing-offsets for the ND.

Aidan Kelleher

Aidan is nearly finished with the draft of the Target document. He is also beginning to use the analyzer and will start by identifying features along the beamline observed in the B.tr.vz distribution, and checking the effectiveness of the collimators.

Rob Feuerbach

I presented a few plots to show where backgrounds in the ND and further work on BB might be needed. Plots can be found under
http://hallaweb.jlab.org/experiment/E02-013/mtg/2007feb21/feuerbac

  1. vz_empty_h2_t3.ps Z vertex for empty and H2-filled reference cell: shows the relative contribution of the "glass" (due to halo?) to the hydrogen events.
  2. thdiff_vs_pdiff.ps Difference between the predicted and observed proton angle in the ND vs. the electron momentum observed and calculated from the scattering angle. Bogdan pointed out and we agreed that this is in the same direction as the radiative tail would be.
  3. Then the other plots of H_Npardiff.ps and H_Cpardiff.ps show the Ppar-q momentum distributions of the neutral and charged events for scattering from a hydrogen target, with the background (green) and subtracted (blue). Here the background is from a side-band in inverse-beta. My statement is that while Seamus's accidental background appears to be well handled via a mixed-event technique, this supposedly also valid sideband-technique does not appear to remove enough from the neutral event sample. Similar plots can be found for the He3 target in the directory.