G2p Analysis Minutes

Minutes of the weekly analysis meetings

---

Agenda

3/25/2015

Present: Jie, Min, Vince, JP, Chao
By Phone: Ryan, Karl, Toby, Alexandre, Pengjia

Feature Presentations:

• Jie
  • Gave an update on simulation package. The elastic radiation tail include soft photon contribution is compared with
    the result from the predication calculated using Mo/Tsai. The tail matches at ΔE~6MeV. He also looked at the effect
    of the beam position on the acceptance. The simulation covers the full angle acceptance to test the aperture cut
    in the SNAKE model. The simulation shows that φ acceptance has large correlation with the horizontal beam position.
    JP suggests the aperture setting in the model may be not sufficient to represent the real situation. Chao suggests
    that to use the simulation to determine which aperture stops most particles. More details can be seen in his slides here.

• Chao
  • Gave an update on the optics calibration. As Jie points out before, the reconstructed dp does not agree with the
    prediction very well when horizontal beam position changes. The beam position scan runs are used to fit a correction
    to the dp reconstruction matrix. The correction could correct the dp reconstruction result for optics runs. However, the
    reason of this correction is still not clear. After discussion, the suggestion is to add the correction back to the dp matrix
    and redo the fit of higher order terms. More details can be seen in his slides here.

• Toby
  • Gave an update on dilution study. He showed dilution results for all completed settings. The tech note of the dilution
    study is in progress. He also mentioned an alternative way to calculate the packing fraction. The method only works in
    DIS region but a 0th order polynomial fit at large ν can be used to find the packing fraction everywhere. All packing
    fractions seem about 5-10% larger than typical values. More details can be seen in his slides here.

• Pengjia
  • Presented some preliminary results of asymmetry and yield for RHRS. He compared the right arm asymmetry and yield
    with left arm for two kinematic settings. The asymmetry results agree quite well but the yields has some deviation for
    the 2.2GeV, 5T longitudinal field setting, perhaps caused by acceptance effect. More details can be seen in his slides here.

3/18/2015

Present: Jixie, Chao, Jie, Min, Kalyan, JP, Melissa
By Phone: Ryan, Karl, Toby, Alexandre, Vince, Pengjia

Feature Presentations:

• Melissa
  • Gave an update on packing fraction analysis. For each setting, she tried to identify ammonia runs that could
    be used for the analysis, ideally runs that were taken in close proximity to the dilution (dummy/empty) runs, or
    at least where the beam position is similar. In addition, the fitting routine (to understand the level of
    contamination from the 2nd peak) was adjusted for each energy setting. The 2.2 GeV, 5T settings (longitudinal
    and transverse) have ammonia runs taken close in time to the runs to helium. The 1.7 GeV setting has some
    variation in the ammonia yields, but they are not too large. The 1.2 GeV energy setting has significant variation
    in the elastic yields; she will look more into these runs to see if there is an obvious reason for the variation.
    She also showed the fit results for the different energy settings; JP suggested checking the hydrogen fit, as it
    seems a bit wide. More details can be seen in her slides here.

• Ryan
  • Working on an automated model tune, which will be used to determine the ratio of carbon to nitrogen for dilution
    analysis. Last time, he showed an automated tune of the Bosted model, which assumed that the quasi-elastic and
    inelastic channels are independent one another and can be optimized independently. The caveat to this is that the
    tune is only valid in the Q² range and for the target type of the input data. For this time, he updated the
    automated tune procedure to include an adjustment for the "dip" region. The Bosted model parametrization of the
    dip region is a Gaussian that is calculated as part of the inelastic channel. Including the dip region improves the fit
    to carbon data for each individual setting, however the overall fit it worse. JP suggested that fitting the dip region
    first may be the problem; Ryan will try to swapping the order of the fitting routine to see if it improves the fit.
    More details can be seen in his slides here.

• Min
  • Gave an update on her acceptance study. To improve the simulation, she included data from the windows into the
    calibration and updated the results for the delta scan runs. This new function was applied to both horizontal and
    vertical beam scan runs. This improved the agreement between the data and simulation, especially for the horizontal
    beam scan run. She also looked at beam scan runs with the target field on, this time adding an additional constant
    on y and φ. For the fit of the focal plane, JP cautioned not to use too high-order of a fit, as it can become
    non-physical. For next time, she will adjust the fitting function. More details can be seen in her slides here.

• Pengjia
  • Looked at the effect of raster cuts on the yield. He looked at several kinematic settings with consistent yields
    and a applied a raster cut with a diameter of 6mm. He found that applying the raster cut had a negative effect;
    meaning the spread in the yields became larger after the cut was applied. He also showed the x and y positions at
    BPM A and B for each set of runs. In some cases, the x-position at BPMA is stable, but it varies at BPMB, which
    could be the source of the variation in the calculated scattering angle. It may be more useful to do a relative
    study, (similar to what Jie has shown previously) for individual runs than trying to compare separate runs. More
    details can be seen in his slides here.

3/11/2015

Present: Chao, Jie, Min, JP, Melissa
By Phone: Vince, Toby, Ryan, Karl, Alexandre

Feature Presentations:

• Jie
  • Working on updating the energy loss model in the simulation package. Previously it was found that the energy loss
    model in g2psim does not match the prediction calculated using the Mo/Tsai formulation. He found that the soft photon
    contribution was missing. This term could have a large contribution, up to 20%. He also looked at the effect of
    changing the scattering angle on the dp distribution. Over an 80mrad change in the scattering angle, he saw only a
    ~0.16% change in the dp distribution. He also looked at 5 different areas of the beam spot, to see how the dp changed
    in that region. The simulation results show no change in the dp distribution in the different regions, but the data
    shows that the dp distribution shifts as you move around the beam spot. More details can be seen in his slides here.

General Discussion:

• Ryan sent around an abstract for the upcoming Chiral Dynamics workshop, comments are appreciated.
• JP sent around the latest version of Pengjia's BPM paper, feedback is appreciated.

3/4/2015

Present: Chao, Jie, Min, JP, Melissa
By Phone: Vince, Toby, Ryan, Karl, Pengjia

Feature Presentations:

• Toby
  • Gave an update on dilution analysis. He finished tuning the Bosted model in order to obtain cross section ratios used
    for radiation length matching for dilution runs (for example, scaling a helium run to match the radiation length of a carbon
    run). He showed the calculated background over the full kinematic range for the 2.2 GeV, 5T, Transverse setting. The
    dilution factor can then be extracted using the ratio of the background yield to the total yield. Eventually he will be able
    to update the dilution factor using material specific packing fraction values (when they are ready), and an improved tune
    of the XS model. Next he will work on other settings and compiling a technote. More details can be seen in his slides here.

• Ryan
  • Working on a method to automate model tuning for the QFS and Bosted models. For dilution analysis it is necessary
    to scale carbon data as a replacement for nitrogen yields. Since there is no nitrogen and carbon cross sections available
    from the same experiment, it is necessary to use to model tune of saGDH data for nitrogen, and a model tune of data
    from UVa archives for carbon. Previously, this tuning was done "by hand", with a reduced χ² to determine the
    quality of the fit. A problem with this method is that it could result in inconsistent results for different people tuning the
    same data set; an automated method could solve this problem. JP commented that a possible problem with an automated
    method is that the physics is simplified. He showed results for tuning the Bosted model to carbon data, next he will work
    on saGDH data, and then will repeat the procedure for the QFS model. More details, including a description of the code
    structure, can be seen in his slides here.

• Melissa
  • Gave an update on packing fraction analysis. Previously, she described an update to her method to include cross section
    input. Elastic form factor models were used to determine the cross section ratio σN/σHe. To obtain radiated cross sections
    two methods were used. First, using the g2psim energy loss model, and second, using the Mo/Tsai formalism to calculate
    the correction to the elastic peak. There is a discrepancy between these methods, specifically that the results suggest that
    the correction is smaller than it should be. This will have to be resolved to get an accurate value for the cross section ratio.
    She is also working on fitting the data for dilution and production runs for other energy settings to extract packing fraction
    values for other materials. More details can be seen in her slides here.

• Min
  • Gave an update on her acceptance study. Last time she found that the focal plane variables for production runs didn't
    match well with the simulation. Production runs have a rastered beam, as opposed to the point beam used in optics
    calibration runs, so this effect must be corrected for. For this week, she looked at events at the target window in the focal
    plane, and used these events to improve the fitting at the virtual plane. She also looked at events from the carbon foil at
    the focal plane, and saw a reasonable match. Next she will work on further improving the fitting and incorporating the
    results into g2psim. More details can be seen in her slides here.

2/25/2015

Present: Chao, Jie, Min, JP, Melissa
By Phone: Toby, Ryan, Ellie, Pengjia

Feature Presentations:

• Jie
  • Working on understanding the yield drifts due to changes in the beam position. He used beam position information
    as input to the simulation, and applied loose acceptance cuts to determine the yield. He looked at several momentum
    settings that had a drift in yields >3% in the 1.7 and 1.1 GeV energy settings. In general he found that the effect was
    opposite of what is seen in the data. JP suggested doing the study independent of data, so as not to bias the results,
    and to show the dependence of the yield on each parameter (x,y,θ,φ) individually. He also suggested checking to
    make sure the acceptance effects are included properly, as it is difficult to get them to agree at the boundaries. Jie will
    keep working on this, and will start working on the septum SNAKE model for the 403216 septum configuration. More
    details can be seen in his slides here.

• Pengjia
  • Trying to determine if applying a raster cut to the data will fix the spread in yields. The difficulty is in determining
    the total charge after this cut is applied. To determine the charge, he used the calibrated charge info from both the
    happex and fastbus, and applied the same raster cut to each. He was then able to use the happex raster cut info to
    get the charge from the happex bcm. JP pointed out that, since the happex and fastbus aren't synchronized, this
    method may only work if the beam is stable, so it is important to quantify an uncertainty on the charge. He suggested
    trying this method again with a much smaller raster cut, so that whatever effect this method will be magnified. Pengjia
    will first look at a setting where the yields are stable to see the effect of this cut, then will move on to settings that have
    issues. The raster cut library and testcode is available on the work disk, see his slides here for more details.

• Chao
  • Posted the draft of a technote on target field mapping here. Any comments would be appreciated!

2/18/2015

Present: Chao, Jie, Kalyan, Melissa
By Phone: Vince, Ryan, Toby, Min

Feature Presentations:

• Min
  • Gave an update on her acceptance study. She showed a comparison of simulation results with data from a production
    run. As was suggested previously, she expanded the acceptance range when generating events. She found that the
    focal plane variables still don't match very well. One possibility is that the production runs have a raster beam, but the
    optics runs she had looked at previously used a point beam, so this effect needs to be corrected for somehow. More
    details can be seen in her slides here.

General Discussion:

• Chao mentioned that a problem was found in the most recent BPM package, and it may have caused a problem
  in the replay. Toby will discuss this with Pengjia offline.

2/11/2015

Present: Chao, Jie, Min, Melissa
By Phone: Pengjia, Toby, Ryan, Ellie, Vince

Feature Presentations:

• Toby
  • Gave an update on dilution analysis. The dilution factor is define as one minus the ratio of background yield over
    total yield. The background yield includes contributions from nitrogen, helium and the foil endcap on the target cell.
    To extract the dilution factor he must tune the Bosted model (for helium and carbon) to match cross section data,
    radiate the simulation results and take ratios to obtain necessary scaling constants, and then apply these scaling
    constants to calculate the background yield. For tuning the Bosted model, he is currently using cross section data
    from the UVa quasi-elastic database. The kinematics do not match ours exactly, but this will work until g2p cross
    sections are available. He used a χ-squared minimization routine to determine the scaling factor for each kinematic.
    He also showed the resulting cross sections ratios, which are used to determine the scaling factors α, β and γ.
    Next he will apply the cross section ratios to yields from dilution runs to calculate the background yields and dilution
    factor for the 2.2 GeV, 5T transverse setting, after which he will move on to other settings. More details can be seen
    in his slides here.

2/4/2015

Present: Chao, JP, Jie, Min, Melissa
By Phone: Toby, Ryan, Pengjia, Karl

Feature Presentations:

• Chao
  • Gave a status update on optics analysis. For the 2.2 GeV, 5T settings (longitudinal and transverse) there was
    limited optics data taken. Instead, the 1.7 GeV optics matrix matrix was used for this setting, with a 0th order
    correction for the longitudinal setting, and a 1st order correction for the transverse setting. He showed the
    longitudinal setting as an example, applying the 1.7 GeV matrix and including the 0th order correction to center
    the sieve holes. JP commented that, if the offset is large, there may be higher order effects to consider. There
    are no optics runs for the 3.3 GeV setting, so he plans to use g2psim to determine the optics matrix. He is
    working on an optics technote, and a technote summarizing the last target field measurement. More details can
    be seen in his slides here.

• Min
  • Gave an update on her acceptance study. For this time, she improved the fitting at the virtual plane (for both
    x/y and , and θ/φ) and projected the results onto the focal plane. The corrections on the forward transport
    function were incorporated into g2psim. For the reverse transport function, the fit was done using all dp scan runs.
    The simulation does not match with the data at the edges; this is due to the boundary conditions in the simulation.
    Min will try expanding the boundary to see if it better matches the data. More details can be seen in her slides here.

• Melissa
  • Updated the data quality check for the 2.2 GeV, 2.5T setting to include RHRS runs. In general the problem
    settings on the RHRS are consistent with the LHRS. She also gave an update on the packing fraction analysis;
    there was a problem in the previous definition of the packing fraction. The updated method requires some input
    from cross section models. The resulting packing fraction is slightly larger than the result from the previous
    method. Next she will work on extracting packing fraction values for other settings. More details can be seen
    in her slides here.

• Jie
  • Calculated the energy loss using two methods; using his energy loss model (Monte Carlo, step by step), and
    using a radiated cross section model (from Stein). He used a fixed initial energy and fixed scattering angle for
    a carbon target, and didn't apply any acceptance cuts. He compared the dp spectra for the two methods and
    found that his energy loss model agrees well with the radiation model. Next he will work on a full simulation
    for a beam dependence study (once the acceptance is ready). More details can be seen here.

General Discussion:

• Starting next week we will change the meeting start time to 9:30am.

1/28/2015

Present: Chao, JP, Jie, Melissa
By Phone: Vince, Ryan, Toby, Min, Pengjia, Ellie, Alexandre

Feature Presentations:

• Ryan
  • Gave an update on data quality checks for the 1.1 GeV setting. He found that the spread in yields ranges
    from 0.01% - 6.1%, with an average of ~2%. For this setting we also took runs on a short target cell (material
    14). For settings with a large spread, there was a correlation between beam drift and yield drift. Overall, this
    setting seems ok. Next he will move back to working on polarized radiative corrections. More details can be
    in his slides here, and a detailed summary
    of the data quality check can be seen here.

• Pengjia
  • Gave an update on data quality checks for the 1.7 GeV setting. For settings with a large spread in yields,
    he looked at the correlation with the calculated scattering angle. He found that in some cases the change in
    yield was proportional to the change in scattering angle, but for some settings it was inverse proportional.
    He also showed an example of a setting where the yield was drifting, but there was no change in the scattering
    angle. He tried applying raster cuts to see the effect on the yield, but first needs to find a way to calculate the
    charge while including the raster cut. There was some discussion how this could be done; it may be possible to
    use the Happex DAQ, as it is triggered by helicity. More details can be seen in his slides here.

General Discussion:

• Toby will start replaying RHRS runs today, hopefully they will be done by the end of the week.

1/21/2015

Present: Jie, Min, Chao, JP, Kalyan, Melissa
By Phone: Pengjia, Vince, Ryan, Toby, Karl

Feature Presentations:

• Melissa
  • Gave a summary of data quality checks for the 2.2 GeV, 2.5T, transverse setting. There are several momentum
    settings with a significant spread (>2%) in the yields. In most cases, the spread in yields seems to be correlated
    with a change in beam position, which often corresponds to beam down time during the run period. Examples of
    problem settings can be seen in her slides here, and a summary of all settings is available here.

• Chao
  • Gave an update on the status of optics for the RHRS. For the longitudinal and transverse settings at 5T, only a
    few optics runs were taken, so the plan is to use the 1.7 GeV optics matrix for these settings. He also included a
    summary of the RMS value of each kinematic variable (δ, θ and φ) for each setting, and commented that
    the RMS value for y_tg is around 3-4mm, but this is a rough estimation as the fitting is not as good for this variable.
    Next he will test 2.5T transverse optics matrix on the 5T settings to see if it works. His slides are available here.
  • The simulation package has been updated to cover the RHRS and include some upstream geometries. The most
    recent version of g2psim is available through github here: https://github.com/asymmetry/g2psim

• Min
  • Gave an update on her acceptance study. For this time, she combined all dp scan runs in order to do the
    corrections at the virtual plane, which was then projected onto the focal plane. The agreement at the virtual
    plane looks good, but there is still some disagreement at the focal plane. Next she will fit the reverse transport
    functions and compare target plane variables. More details are available in her slides here.

1/14/2015

Present: Chao, Jie, JP, Melissa
By Phone: Pengjia, Ryan, Toby, Ellie, Karl, Min, Vince, Alexandre

Feature Presentations:

• Jie
  • Gave an update on the dp simulation, specifically trying to understand why the simulation result for elastic dp was
    narrower than the data. Comparing the data (with just PID cuts) to the simulation result (using full acceptance, and
    a raster diameter of 2cm), the simulation is narrower than the data. However, if he applies acceptance cuts to the data
    and simulation, and uses a beam size of 4mm (which matches with the data), the results show a better match between
    data and simulation. He also tried dividing the beam spot into 5 different zones, and looked at the resulting dp
    distribution for each. He did see a shift in the central value of dp between zones, however this shift was not seen in the
    simulation. He also compared the shift in θ and φ for the 5 zones. JP commented that the θ distribution in
    the data is symmetric and centered around zero, while the simulation result is not. It is possible that this is due to the
    fact that Jie is using the old SNAKE model; he will get the updated version from Min and see if it makes a difference.
    More details can be see in his slides here.

• Pengjia
  • Posted his false asymmetry results to the ELOG.
  • Working on data quality check for the 1.7 GeV, 2.5T setting. It looks like the spread in yields in some settings is
    strongly dependent on the scattering angle. JP suggested looking at beam position or other raw variables, since the
    scattering angle is a calculated value and depends on optics. It seems most settings with problems correspond to a
    break in time when the runs were taken, which may suggest a problem with the spectrometer. Pengjia will look into
    this next. More details can be seen in his slides here.

• Toby
  • Gave an update on his yield drift study for the 3.3 GeV setting. For this time, he looked at the target field (central
    NMR value) over the 2.34 GeV momentum setting, a period of about 6 hours. While there are some jumps in the signal,
    there isn't a consistent drift like is seen in the yields. However, there is a period where the signal jumps significantly; he
    will look into a possible cause for this. He also looked at the scaler BCM counts and trigger rates for these runs. There
    was no noticeable drift in the scaler rates, but the beam current was not stable during this time.. He also computed the
    "scaler yield" by dividing the total triggers by the BCM counts and saw a drift similar to what is seen in the normalized
    standard yield. Previous to these runs being taken, there was a flood in the hall resulting in a power outage. It's possible
    that the drifting is a result of systems not being stable while these runs were taken. More details can be seen in his
    slides here.

• Ryan
  • Working on a data quality check for the 1.1 GeV, 2.5T setting. In general the setting looks pretty good; the maximum
    spread is 10% and the minimum is 0.01%, with an average of ~2% spread. Fewer runs were taken in this setting, and the
    runs were taken consecutively, resulting in smaller spreads overall. In addition, there was data taken at this setting at
    3 different momentum settings, all looking at elastic, and all on the same material. Next he will try to determine the
    problem for the settings with a large spread (~10%). More details can be seen in his slides here.

1/7/2015

Present: Chao, Jie, JP, Min, Kalyan, Jixie, Melissa
By Phone: Yunxiu, Vince, Pengjia, Toby, Ryan, Ellie, Karl, Todd, Nilanga, Haiyan

Feature Presentations:

Each student gave a short update on their analysis projects since the last collaboration meeting on November 14th:

• Melissa
  • Method for packing fraction analysis is complete; the packing fraction values for each material still need to be
    extracted. This will probably take ~1 week, but the yield discrepancies need to be understood before it can be
    completed. She is also working on data quality checks for yields, specifically the 2.2 GeV, 2.5T setting. After these
    projects are complete she will work on the PbPt check using elastic asymmetries. More details can be seen in her
    slides here.

• Toby
  • Focusing on data quality checks of yields for production runs. He has completed the 2.2 GeV 5T settings, both
    longitudinal and transverse, and the 3.3 GeV setting. The 2.2 GeV 5T settings have few problems, but the 3.3 GeV
    has many momentum settings where the yields drift over time. He sees a similar drift in the left and right arm data.
    He is still trying to figure out the cause for the drift. In addition, he is working on tuning the Bosted model to helium
    and carbon dilution runs to simulate the background for dilution analysis, with a goal of finishing the dilution analysis
    by mid-march. More details can be seen in his slides here.
  • Information on data quality checks can be found on the wiki here.

• Jie
  • Looked at the effect of changing the scattering angle on the cross section for carbon, nitrogen, helium and hydrogen.
    This may help explain the discrepancy we see in the yields, but is problem not the only cause. He is also working on
    finishing the dp simulation, which should be done soon, and will begin working with Min to learn more about the acceptance
    study. More details can be seen in his slides here.

• Chao
  • Gave an update on the status of optics analysis. He is finishing up the 2nd iteration of the optics calibration; the LHRS
    is complete and the RHRS has one setting to go (1.158 GeV). For the 2.2 GeV, 5T, transverse setting, there is no full
    dp scan, so the longitudinal data taken at that energy setting will be used. Once the 2nd iteration of the calibration is
    complete, he and Min will compile a technote on the optics study. He has also modified the geometry part of the g2psim
    package to make it more configurable. More details can be seen in his slides here.

• Min
  • Working on the acceptance study. In the focal plane, there seems to be good agreement between simulation and data.
    In the target plane, in the θ vs φ plot, there seems to be a shift in φ for all dp settings; it seems to be a linear
    correlation with dp. Next she will work on a correction for φ-target vs dp and compare target plane variables. To finish the
    good septum, straight through setting it will probably take 1-2 months, then an additional 1-2 months for the other 2 septum
    settings. More details can be see in her slides here.

• Ryan
  • He has completed a draft for a PRC paper on unpolarized He3 cross section data. He still needs to complete the He3 radiative
    correction analysis for saGDH, but is waiting on updated nitrogen cross sections for saGDH. He is also working on the carbon/
    nitrogen ratio for saGDH, but is waiting on updated analysis from Vince before continuing. He is also working on data quality
    checks for production runs for the 1.1 GeV setting, and is starting to look into polarized radiative corrections. More details can
    be seen in his slides here.

• Pengjia
  • Calculated the false asymmetry by comparing the asymmetry before and after applying detector cuts. He found the false
    asymmetry to be small; less than 200 ppm for all energy settings. He's also working on a data quality check for the 1.7 GeV
    data set. More details can be seen in his slides here.

General Discussion:

• Yields stability/overall data quality check is the main issue at this time.
• Progress of acceptance study looks good, but could become an analysis bottleneck.

---

July-Dec 2014

Minutes_July2014_to_Dec2014

---

Jan-June 2014

Minutes_Jan2014_to_June2014

---

June-Dec 2013

Minutes_June2013_to_Dec2013

---

Jan-May 2013

Minutes_Jan2013_to_May2013

---

April-Dec 2012

Minutes_Apr2012_to_Dec2012

---

Jan-March 2012

Minutes_Jan2012_to_Mar2012

---

July-Dec 2011

Minutes_July2011_to_Dec_2011

---

Jan-June 2011

Minutes_Jan2011_to_June_2011

---