Difference between revisions of "G2p Analysis Minutes"

Revision as of 13:43, 17 May 2013

Minutes of the weekly analysis meetings

Agenda

5/17/2013

Present: JP, Kalyan, Jixie, Karl, Min, Alex
By-Phone: Pengjia, Toby, Ryan, Elena

General discussion:

NEW meeting time: We will meet at Wednesday 10:30 from next week.
Poster competition abstracts due on today, May 17th. Please submit an abstract ASAP!

Feature presentations:

Ryan
- Working on Ytg calibration. Ryan present the Ytg calibration result for LHRS using

a run without target field and also that for RHRS but with 5T longitudinal target field.
It looks like that the beam position is not stable at all during the course we took optics data.
If one apply a cut on the beam position the optimized Ytg will achieve 0.5 mm resolution.
Otherwise it will be about 1.1 mm. Details can be seen here.

5/10/2013

Present: JP, Vince, Kalyan, Jixie, Jie, Min, Melissa
By-Phone: Pengjia, Toby, Ryan, Ellie, Karl, Alex

General discussion:

If you haven't yet filled out the Doodle poll for a new meeting day/time, please do so ASAP!

Feature presentations:

Melissa
- Working on a cross-check of beam charge asymmetries. Showed stability of up/down BCMs. For some energy
  settings there is a drift in one of the BCMs.
- Also showed results for charge asymmetries, but there may be a problem with helicity decoding (shifted by one bit).
  She will talk to Pengjia/Chao about this and fix results for next time. Full details can be seen here.

Pengjia
- Showed an update on beam charge asymmetries. He used Chao's helicity decoder and his own as a double
  check, and saw similar results using both decoders. Full details can be seen here
- Still working on BPM calibrations (just finished copying over data)

Min
- Showed an update on SNAKE matching data, this time using a run where the beam was better centered.
  Since the new run was taken with the foil in LHe, the energy loss from the LHe will have to be accounted for.
- The current result for the difference in y between data and SNAKE is ~3.3mm, with delta-y = ~3.2mm.
  Full details can be seen here

5/03/2013

Present: JP, Vince, Min, Jixie, Kalyan, Jie, Melissa
By-Phone: Pengjia, Toby, Ryan

General discussion:

Feature presentations:

Toby
- Showed an update of study to determine systematic uncertainty for TE measurements. He used 1st-4th order
  polynomials and compared the quality of the fits. 2nd, 3rd and 4th order polynomials seem to give good results,
  but the 1st order polynomial is much worse. To determine the uncertainty, he chose the two fists with the most
  agreement and looked at the variance between the two fits. This resulted in a relative uncertainty that
  varied from 2-30%.
- JP/Vince suggested trying a different function to fit the baseline subtracted signal, since the fluxation in
  uncertainty suggests that the polynomial fit is not stable. Full details can be found here

Ryan
- Looking at how y-target values are calculated with the optimization code. It looks like the code uses 3
  different values for real y-target while it runs, which results in a "shoulder" in the reconstructed y-target
  variable. He will look at BPM information to see if multiple peaks are visible.

Pengjia
- Showed results from using helcity information to determine the fast clock rate. Because the fast clock is
  not stable, there is a shift in time in the phase fit for the slow raster ADC data. Pengjia is working on how
  to account for this.
- Also working on beam charge asymmetries. He first determined them using Chao's decoder, but will check using
  his own decoder. His presentation can be found here

Jie
- Showed his procedure for determining uncertainties for muti-track efficiencies. For 2 and 3 track
  events, two methods were used to determine the systematic uncertainty. Method 1 considered tracks that
  were within one block of each other. Method 2 loosened the cut on the distance between blocks (two block
  widths), but additionaly required that at least one track satisfy a set of acceptance cuts. JP suggested
  that method 2 is the better method, but method 1 could provide an upper limit on the systematic uncertainty.
- Also working on data quality checks for optics variables (t0 and tracking variables). Looks good
  so far with a few problem runs. Full details can be found here

4/26/2013

Present: JP, Vince, Kalyan, Jixie, Alex, Jie, Min, Melissa
By-Phone: Elena, Toby, Ryan, Karl, Pengjia

General discussion:
In regards to status of data quality checks:

List of variables can be found here
- PID detectors (Cherenkov, lead glass): pretty much complete
- VDC: Jie is making good progress
- Optics: Some variables will have to wait until optics is more complete. For now, Jie will start looking at
  focal plane variables, which don't require optics.
- Helicity: Pengjia/Chao will start working on this
- EPICS variables: Ryan will take a look at this.
- Scintillators: Trigger efficiencies and deadtime are done. Ryan will look at helicity gated deadtime.

Feature presentations:

Melissa
- Showed results from data quality checks for RHRS PID detectors. Found a total of 46 runs with "issues".
  Full details can be found here.
- Will compile results from LHRS and RHRS into a document.

Min
- Showed updates on SNAKE tuning. Described her plan for improving results shown previously; first she will
  look at a run with better centered beam (but carbon in LHe). Then she will try to tune the Q1 field. If this
  doesn't work she will adjust other fields/currents in the SNAKE model to match. JP suggested talking to
  John Lerose to understand how well different parameters of the system are known, and adjust the model appropriately.
  The effect of the septum field on Q1 is still the biggest unknown. Her presentation can be found here

4/19/2013

Present: Vince, Chao, Jie, Melissa, Jixie, Min, Kalyan
By-Phone: Elena, Toby, Ryan, Karl, Alex, Pengjia

General discussion

Feature presentations:

Toby
- Showed an update on target uncertainty analysis. The uncertainty in the TE measurements seems to
  be very small (~10^-5), however, there might be an error in the calculations.
- Also showed the results of using different order polynomials to fit the baseline subtracted
  NMR signals. He used 1^st-4^th order polynomials and found a large variance of uncertainty (70%)
  between fits. Next time he will show more details of the procedure to understand this large variance.
  Full details can be found here

Jie
- Working on determining the uncertainty for the multi-track efficiency. For zero track events,
  the systematic uncertainty was determined by counting events that survive the PID cuts, but have
  no track in the VDC. For 2/3 track events, two methods were used to determine the systematic
  uncertainty. Method 1 considered tracks that were within 14.5 cm (the width of one block) of each
  other, so it was not possible to distinguish between them. Method 2 loosened the cut on the
  distance between blocks (29cm, two block widths), but additionaly required that at least one
  track satisfy a set of acceptance cuts. Method 1 gives a larger systematic uncertainty. Full
  details can be found here. Next time Jie will show results for the "worst case" kinematic setting.

Pengjia
- Showed the status of the raster calibration. Previously, he used information from both BPM A and B
  to determine the raster size at the target. This led to a huge uncertainty (1mm), so instead
  he will use only information from BPM A.
- Also showed new results for fitting the slow raster phase - fit looks much better! However, it seems
  that the fast clock rate appears to be unstable on a run to run basis. He will check this using other
  copies of the fast clock. For the fast raster phase reconstruction he will first use the ADC value directly.
- Next he will work on a BPM noise study and improving the BPM resolution. His presentation can be found here

Ryan
- Working on y-target optimization. Found that optimization didn't rely on BPM information. It's possible
  that the two peak issue is a result of an issue with the BPM reconstruction, specifically that the reconstruction
  doesn't take into account the foil location. Ryan will discuss with Pengjia to figure out the problem.
  Full details can be found here

4/5/2013

Present: Karl, Vince, Jixie, Kalyan, Chao, Jie, Min, Melissa
By-Phone: Elena, Toby, Ryan, Alex

General discussion

Practice talks for Chao/Min will be Thursday, April 11th
An e-mail will be sent out with the time/location

There will be no meeting next week, as many people will be gone for GHP/APS meetings.

Feature presentations:

Chao
- Working on updating the simulation package. Showed a comparison of the 2 carbon elastic
  models. There seems to be good agreement for the region where data is available. Also
  showed tests of the similuation for scattering angle and delta-p. Full details can be
  found here

Min
- Working on tuning the SNAKE model. Showed an updated table of uncertainty from first order
  matrix, this time with y0 and phi0 as lineraly correlated variables. Working on tuning the Q1 field,
  possibly by studying the sieve hole projection closer to the focal plane center. Full details
  can be found here

Melissa
- Showed the results of data quality checks for the LHRS. Found a total of 48 runs with "issues".
  Also had to update PID cuts for runs taken with a short ammonia cell. Full details can be found here

3/29/2013

Present: Vince, Jixie, Kalyan, Chao, Jie, Melissa
By-Phone: Pengjia, Min, Karl, Elena, Toby, Ryan, Alex

General discussion

We will have practice talks starting next week for students going to GHP/APS meetings:
Toby/Melissa: Thursday, April 4th
Chao/Min: Thursday, April 11th

An e-mail will be sent out with the time/location.

Feature presentations:

Pengjia
- Showed the status of the raster calibration. He is going to use the position transport function to calulate
  the size of the raster at the target. The problem with this method is it will increase the uncertainty because
  of the BPM.
- Also working on slow raster shape resonstruction to minimize uncertainty caused by the accuracy limit of the ADC.
  He will try using a successful fit (entries 2000-2100) and extrapolating to another region (entries 7000-7100).
  Also, he needs to determine the actual rate of the fast clock from scaler data. Full details can be found here

Jie
- Working on multi-track efficiencies. He extended his previous analysis to include 7-track events. Next he will
  determine how this contributes to the overall systematic error. His presentation can be found here

Toby
- Working on the target polarization uncertanties. The leading uncertainty comes from the fit to the wings of
  the baseline subtracted signal. He will check the fit with different functions to determine an uncertainty of the fit.
  . His presentation can be found here

Ryan
- Gave an update on RHRS longitudinal optics for the "very bad" septa configuration. He showed the results for the y-target
  calibration after applying a tight cut on BPM x-values. The results seem to agree with values predicted by the formula from
  Min's technote. Full details can be seen here

3/21/2013

Present: JP, Jixie, Kalyan, Jie, Melissa, Chao, Min
By-Phone: Karl, Elena, Ryan, Toby, Alex

General discussion

Feature presentations:

Melissa
- Showed the procedure for data quality checks using PID detectors. See details here

Chao
- Working on updating the simulation package. Tested the cross section weighted distribution

with P. Bosted model. Currently working on the elastic carbon-12 model. Full details can be found here

Min
- Working on tuning SNAKE model. Trying to find the center sieve hole coordinates at various end plates.

Currently able to account for only half of the effect (either phi or y).
Will work on adjusting the septum current and HRS offset to optimize SNAKE model with data. Full presentation can be found here

Kalyan
- Showed a list of variables being used for data quality checks.

3/15/2013

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

General discussion Next meeting will take place in Thursday 14:30. Vince will reserve a room and
send notification through email.

Feature presentation:

Jie:
- Updated the multiple track analysis.
- The discrepancy between L-HRS and R-HRS in multiple track events was solved after updating the database

which define the geometry.
Studied the faction of at least one good track in all events for all runs. After carefully reconstruction,
the multiple track events, we can find at least one good electron in 98% of total triggers, with about
1% systematic error. Jie will provide the detail number later. See details here

- Will do the same analysis to the right HRS.

Toby:
- Working on target polarization uncertainty
- Preparing the presentation for GHP.

Ryan:
- Rerun replay for 5T longitudinal field Optics runs with the CORRECT BPM calibration constants.

Used a wrong BPM settings in previous replay and can not get reasonable beam position. Got reasonable BPM information
at this trial but beam position covers a wide range of 2 mm, which means the beam was drifting way slowly.

- Did a Y_tg optimization. The result is about 2 to 3 mm away from what it should be. JP suggested to make cuts

on beam position at the target (mainly horizontal) then try again.

Pengjia:
- BPM calibration status: most run is ready, but the part between April 26 to 30 is not ready yet.
- Working on raster calibration. See Pengjia's talk here.

3/8/2013

Present: Jixie, Jie, Min, Melissa, Chao, JP, Kalyan
By-Phone: Ryan, Toby, Elena Long, Karl, Pengjia

General discussion

Feature presentation:

Melissa:
- Moller write up (By Sasha) is ready in the wiki. (G2p wiki-> Analysis ->Moller)
- Beam polarization run by run result is ready. The bleed through from hall C in terms of beam ratio

was less than 1%. The beam polarization for the bleed though was as small as several percents. We can ignore
this part in the analysis. (Melissa is very kind to provide a run by run table.)

- Tech note for colorimeter and Cherenkov efficiency is ready to review.

Chao:
- Working on the simulation packages. Optimized the drift in field module by dynamically changing the

step length.

- Present the idea that how we determine the vertical interaction position which is needed to do reconstruction.
- Check how large the distribution center will move after weight by cross section. See details here.

Min:
- Kept tuning snake model to match real spectrometer. Made the following changes.

A) Solved a few bugs of position offsets.
B) Extended septum box to cover more field region, aiming to reduce the BdL lost
C) Adding HRS offsets using survey results
D) Using the real Q1 currents

- With this update, the center hole hits on focal plane have about 8.7 mm shift in y position if match the phi angle

or have 8.5 mrad shift in phi angle if match the y position. See details here.

Pengjia:
- Present a slice for the beam position for the optics data that were taken on March 14, 2012. It turned out that

the beam position was not centered. It had 4.4 mm shift to the left and 1.5 shift to the up.

- The slice also indicated that BPM B is not stable. Pengjia agreed to find a way to avoid unnecessary fluctuation.

Jixie:
- Present the study of how well can we determine the central value for a distribution.
- There is no good event by event resolution for the scattering angle when the target field is on. We will reconstruct

FP hits only to the target plane. By comparing the central value for the scattering angle and also the cross section distribution
between cross section weighted or unweighted situation, we can tell how well we can determine the central value.
Jixie did this comparison in several HRS momentum settings. The work show that we can determine the central value better than
1% using snake model. To reach this precision in real data, we have to calibration the reconstruction to avoid any offsets and bias.
See details here.

3/1/2013

Present: Jixie, Jie, Min, Melissa, Chao, Karl, Vince, Kalyan
By-Phone: Ryan, Toby, Elena Long, Alex

General discussion

Feature presentation:

Toby:
- Analyzed the target polarization uncertainty.

PengJia:
- Merged HAPPEX decoder to g2plib. The beam packages will use HAPPEX decoder by default. Before it is

fastbus. HAPPEX has much better resolution than fastbus if fastbus is using auto gain. Pengjia compared
the resolution of these 2 daqs at various situations. See details here.

- Found an optics run (22259) that happex and fastbus did not agree with each other. PengJia agree to

to check all optics runs to see if this problem occur or not.

Ryan:
- Tried to use PengJia's BPM package. Need some more work to get it right.
- Practiced Ytg optimization.

Jie:
- Updated the multiple tracks efficiency analysis.
- The goal is to find the ratio of at-least-one-good-track event in two-track events under normal cuts.
- There is a large deviation between L-HRS and R-HRS in previous talk. Jie found that it was because the

database of the prl1 and prl2 geometry was wrong. After manually correct the data, L-HRS and R-HRS agree
well to each other. See his nice work here.

2/22/2013

Present: Jixie, Jie, Min, Kalyan, Melissa, Chao, JP, Alex
By-Phone: Ryan, Toby, Elena Long, Karl

General discussion

Feature presentation:

Melissa:
- Got the asymmetry result run by run from QWEAK Compton polarimetry. Match it

to the g2p run by run according to the time stamp. Will use this asymmetry table to
to check if there is any fluctuation in the beam polarization. See details here.

Chao:
- Working on building the single arm simulation program, testSNAKE. Add field and field integration modules.
- Test the method of "Match reconstructed trajectory to Beam trajectory to get the interaction point".

The results shows that this method does not have a good event by event reconstruction at all.
See the slices for details.

Jixie:
- Test the trajectory matching method independently. Jixie generated events at 3 2-mm-Z-slices: 10cm upstream,

target center and 10 cm downstream. The results show that the matching method can not find the interaction point
at all. See details here.

Min:
- Continue working on creating a snake model that can represent g2p setting. Phifp and Yfp can not match at the same time.

Need more efforts. See this link for details.

kalyan:
- Updated the analysis status flow chart.

Most of the task is done. BPM and slow raster is in progress. Straight through optics is 80% done.
It depends on BPM calibration.

2/15/2013

Present: Jixie, Jie, Min, Kalyan, Melissa, Chao, JP
By-Phone: Ryan, Toby, Pengjia, Elena Long, Karl , Alex

General discussion The Fall DNP (October 23-26, 2013) for this year will be held at Newport News. Can we have result by then? Answer: Not likely.

Feature presentation:

Toby:
- average N sweep area to get the mean value
- Systematic errors came from Magnetic field drift, temperature fluctuation, NMR drift and so on

will be added to each sweep later.

- Systematic error is about 2-3%. Will details later.
- will go back through the TE measurements and using a more rigorous chi2-minimization technique to exclude poor data

See details here.

Ryan:
- Provided a root library to connect to the mysql database.

Jie:
- Updated on the two-track analysis for both arms. There is about 20% deviation between left and right arm

on the ratio of one good track out of two-track events. This issue is still over there. Jie will keep working
on it. See details here.

Pengjia:
- Generated transport function from BPM to target center for field is on situations.
- Found a better way to determine harp noise.
- Updated BPM packages. Used the uncertainty of each single harp measurement as the weight in the TMinuit fitting.

The BPM uncertainty (preliminary) is about 1.5 mm for position and 1.4 mrad for
angles when target field is off. There will be some room to improve.

2/08/2013

Present: Jixie, Alex, Vince, Jie, Min, Kalyan, Melissa, Chao, JP
By-Phone: Ryan, Toby, Pengjia, Elena Long

General discussion

Feature presentation:

Melissa:
- Finalized the cut and efficiency analysis. The pi/e ratio after the optimized cut are

now 0.4% for both arms. See details here.

- Melissa agreed to write some note about this job.

Toby:
- The previous R-HRS cooking used a wrong calibration file for the preshower. Need to do a recook.

Min:
- Working on tuning the snake to match the real data. Min tried to turn the quadrapole and

dipole such that the simulated variables in FP match the real data. But this is wrong. She will
go back to the original dipole setting and try to tune septum fields.

Chao:
- Gave an update about the simulation.

A ZERO and FIRST order post processing correction have been made to the none raster snake model. The Zero order correction
can remove most of the offsets between real data and snake model. The FIRST order correction can make it event better.
This is just a machinery check to make sure the program work fine. Will present this kind of job once Min finish tuning SNAKE.
From the ZERO order correction we can see that Q1 in the snake model need some tuning.

2/01/2013

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

General discussion

Feature presentation:

Ryan:
- Present the result of quality check:

LHRS: 956 good runs (73%), 104 junks(8%), the others have some
issues which might be useful with extra treating.
R-HRS: 798 runs are good (68%) and 109 are junks(9%)
See details here.

Pengjia:
- Improved the code which is judge whether the beam is drifting.
- Updated MUDIFI program to export c code with error transfer function, refit the BPM transport functions.

See Pengjia's talk here.

Jie:
- Working on the multi-track analysis. Check the status here.

1/25/2013

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

General discussion

Feature presentation:

Chao:
- Add Rotating Coordinate System (RCS) to Focal Plane variable in the code. In the real DB, the reconstruction

takes the variables in the RCS, not the Transport Coordinate System(TCS). With this modification, the reconstruction
between simulation and real DB agree with each other better. But still have noticeable difference.

- The forward transport of the snake model still have some offset to the real detected FP variables. Need to correct

the offsets.

- Checked the first order matrix elements for both real DB and SNAKE backward model. The real DB matrix first

order elements for <theta_tg|y_fp> and <theta_tg|phi_fp> seem too large. See details here.

Min:
- Presented a writeup for the uncertainty of the pointing measurement. According to this document, to reach

0.5% relative uncertainty of the scattering angle, the uncertainty of the energy difference between 2 peaks
in the pointing measurement should be 45 keV.

- Present the initial pointing result [1].

Melissa:
- Worked on optimizing the PID cuts for good electrons. She has done that for a few momentum settings. Need to

go through all the momentum settings. See details here.

1/18/2013

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

General discussion

From next week, we will have the weekly analysis meeting at 10 AM every Friday.

Feature presentation:

Ryan:

Presented the optics calibration for 5T longitudinal field data set, R-HRS.

Pengjia:

Presented the status of BPM calibration for straight through data set.

Pengjia was trying to find the time interval that the beam is not stable and treat
them separately. He is using the 1000 events and 7500 events moving RMS to determine
one whole run into several period. See details here.

Jie:

Took an initial look at multiple tracks at the colorimeter. Jie was suggested to add

the electron cut then repeat the analysis. See details here.

1/8/2013

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

General discussion

APS abstract deadline is this Friday. The deadlines for GHP and INPC are also very

close. Those who want to attend these conference please prepare the abstract as early as possible.

Feature presentation:

Min:

Presented the Ytg optimization for R-HRS using NO target field data. See details here.
Prepared the non-raster SNAKE backward model for Chao.

Chao:

Working to compare the 1st order matrix element between REAL DB and SNAKE backward matrix.
Chao has a long term plan to make a simplified version of HRS simulation using existing modules.

Melissa:

Updated the result of Detector efficiency and Cut efficiency for L-HRS. See her latest result here.

Will do the R-HRS next.

Toby:

Working on replay script. The output ntuple is too big. Students need to discuss and determine what variables should

be kept. Jixie suggested to include the unfinished BPM reconstruction in the cooking train at this moment. To save time and
space, do not cook optics runs.

April-Dec 2012

Minutes_Apr2012_to_Dec2012

Jan-March 2012

Minutes_Jan2012_to_Mar2012

July-Dec 2011

Minutes_July2011_to_Dec_2011

Jan-June 2012

Minutes_Jan2011_to_June_2011

@@ Line 2: / Line 2: @@
 ----
 [https://hallaweb.jlab.org/wiki/index.php/G2p_Weekly_Analysis Agenda]
+----
+==5/17/2013==
+Present: JP, Kalyan, Jixie, Karl, Min, Alex  <br>
+By-Phone: Pengjia, Toby, Ryan, Elena <br>
+'''General discussion:''' <br>
+*NEW meeting time:  We will meet at Wednesday 10:30 from next week.
+*Poster competition abstracts due on today, May 17th. Please submit an abstract ASAP!
+'''Feature presentations:'''
+*Ryan
+**Working on Ytg calibration.  Ryan present the Ytg calibration result for LHRS using <br>
+a run without target field and also that for RHRS but with 5T longitudinal target field. <br>
+It looks like that the beam position is not stable at all during the course we took optics data.<br>
+If one apply a cut on the beam position the optimized Ytg will achieve 0.5 mm resolution. <br>
+Otherwise it will be about 1.1 mm.  Details can be seen [https://userweb.jlab.org/~rbziel/Weekly_Meetings/Optics_51613.pdf here].
 ----
@@ Line 27: / Line 48: @@
 ----
 ==5/03/2013==

Difference between revisions of "G2p Analysis Minutes"

Revision as of 13:43, 17 May 2013

Contents

5/17/2013

5/10/2013

5/03/2013

4/26/2013

4/19/2013

4/5/2013

3/29/2013

3/21/2013

3/15/2013

3/8/2013

3/1/2013

2/22/2013

2/15/2013

2/08/2013

2/01/2013

1/25/2013

1/18/2013

1/8/2013

April-Dec 2012

Jan-March 2012

July-Dec 2011

Jan-June 2012

Navigation menu

Views

Personal tools

Navigation

Search

Tools