- Fatiha ...
-
... spoke about spectrometer pointing and particle identification.
Spectrometer Pointing
She concluded the pointing is in good shape.
There are three methods to determine the spectrometer pointing:
- survey by the JLab alignment group.
They shoot several points on the spectrometer and the floor and determine the spectrometer position.
For each point, the precision is about 0.5 mm, this gives about 0.003° (0.05 mrad) for the spectrometer central angle.
- LVDT method.
These are two pins which measure the horizontal and vertical distance of the spectrometer versus a plate attached to the scattering chamber.
The reliability of this method fails if someone bumps the system, which happens quite often.
Also, only the horizontal system was installed during E89044.
- pointing method.
This uses the beam spot picture in y_tg to determine the horizontal spectrometer offset.
It requires precise knowledge of beam and target position.
Fatiha compared all three methods.
Whenever survey and pointing were done at the same setting, they agreed.
Also LVDT is in agreement most of the time, but sometimes it is totally of.
The conclusion is that survey should be taken, if no survey is available pointing is the choice.
There was a discussion about whether we can distinguish between a shift of the spectrometer axis and a rotation around the backward point at the floor marks.
Marty rose the question how much the uncertainty in the spectrometer pointing affects the effective target length.
Moreover, how the effective target lengths of both spectrometers match.
It seems only a simulation can answer this question.
It was mentioned that a stack of Carbon foils will be mounted on the cryo target, for acceptance studies.
Also N->Delta needs precise acceptances, we may should coordinate efforts with them.
It was also mentioned that MCEEP may use empirical apertures to match the simulation to the data.
Particle Identification
HRSH
Fatiha compared beta for the Sigma1 kinematics.
At pmiss = 0 MeV/c, there is only one peak, towards higher pmiss, two more peaks show up.
Using the momentum, she could identify the peaks as proton, deuteron and triton.
There was some discussion about why there are more deuterons and tritons at high pmiss.
Bill mentioned this is not neccessarily true, because they might be hidden at low pmiss by the huge proton peak.
Marat mentioned he checked online, the tritions originate from the target cell walls.
Fatiha can check this by a reactz%beta scatter plot.
Fatiha also showed a scatter plot of energy loss in the scintillator versus beta.
The particles with low beta have higher energy loss, another prove that the interpretation from beta is correct.
HRSE
Electrons have to be distiguished from pi-.
This is possible with the shower/preshower detectors.
However, in the shower (was it shower, what was plotted vertically?) some blocks seem not to work properly, or are not correctly calibrated.
In the Cerenkov, in a single mirror spectrum one can identify one-, two, and three-photo-electron peak, before it smears into a continuum.
She also showed a spectrum summed over all the mirrors, uncorrected, and corrected for gains and pedestals.
She applied a cut above the one-photo-electron peak, and deduced an electron efficiency of about 99.5%, consistent with the findings of other experiments.
There was a discussion how to interpret the corrected sum, it has to be checked how ESPACE calculates it.
- Nilanga ...
-
... reported that the Carbon pointing method is reliable, however it measures only the horizontal pointing.
He said whenever it was checked (by survey???), the vertical pointing was always within 1 mm.
John said he would be surprised, because the spectrometer makes a depression in the floor, and there is nothing that would force it to stay within such a small range.
Nilanga also spoke about data base optimization.
He mentioned nobody of E89044 showed up to learn how to use his code, but Zhengwei did it for the December data.
Whoever is interested, can come to Nilanga's office to learn it.
Documentation can be found at here comes a link.
Zhengwei also did a bpm calibration, he found different bpm offsets than those Michael gave to Marat.
Arun said he has the newest coefficients.
Thus, Zhengwei and Arun have to sit together to check if both results are consistent.
Nilanga also mentioned briefly Paul's and his method to determine experimentally beam energy and spectrometer central angles.
He said that this method favors a higher beam energy than the old ARC energy measurement.
The spectrometer angles seem to be off by 0.6 mrad for HRSE and 0.2 for HRSH, but this is within the error of 0.6 mrad he estimated for this method.
Bill was raising the question if a by 10-3 wrong beam energy would affect the experimental error bars, e.g. for sigmaLT.
He bets $1½ that the error bars are insignificant.
- Arun ...
-
... spoke about the current status of the energy measurements.
A typo was found in the ARC analysis code which made a +10-3 correction to the ARC energy.
There is still a discrepancy to the ep measurements of about 0.7·10-3, at quoted errors of about 0.2·10-3.
But overall, it looks more consistent now, except for energy measurements around 3 GeV.
Here, strip #5 is used, which is suspected to be shifted.
It will be checked in the August/September downtime if this strip is at the assumed location.
He showed a complete recompilation of the energy measurements.
However, he will send it out after Pascal Vernin has applied another correction coming from the quadrupoles, which is expected to be on the 10-4 level.
- Marat ...
-
... shows results of the optics optimization.
He showed react_z spectra with all the foil targets (Carbon and the three dummies) added up.
The react_z resolution for HRSE was about 1 cm FWHM, which results in 2mm for y_tg, as expected.
spec_h.react_z looks a little worse, maybe due to the plastic slate which was covering the spectrometer exit window.
Marat also checked the x_sieve versus y_sieve reconstruction.
He concluded that Zhengwei's optimized database works for the December 845 MeV runs.
However, it has to be checked if it is applicable at higher energies.
He also showed studies where the polarity of the quadrupoles was reversed, cycled, and then after a second reversion moved directly to the desired current,
in contrast to the standard method where on goes first to a high current and then sets the qudrupoles.
As expected, the database doesn't work here.
Marat presented his SQL database.
It can be found at http://marat-rvachev2.jlab.org/php (local access only).
This database contains various information for each data run, i.e. the run-end scaler information.
He showed some examples how to use it.
One topic he discussed is the discrepancy in the charge scaler readings.
This readings are split and put on two separate scalers in both HRSE and HRSH and should give identical results.
E.g., the clock readings agree to a tick, but the charge down-times-one readings differ significantly.
If was critized that invalid runs (short, junk, low current) should be excluded from such a comparison,
but it seems that here is a general discrepancy.
- Michael ...
-
... was showing his transparancies for the Gordon conference and presented the results he obtained for Sigma 1 kinematics.
All good data files are stored on the blue computer which is in Fatihas office.
They were analyzed by Doug using some scripts, and the relevant information was stored in ntuples.
Michael is using a few programs to further condense these files and retrieve the number of good events.
First he showed the cuts he implied to remove the walls and define the acceptances.
Additionally, the signal/noise ratio seems to be good,
thus accidentals can be almost totally removed by cutting on the coincidence peak and the difference of the react_zs.
However, the obtained Emiss resolution of about 4 MeV is much worse than the 1.5 MeV FWHM as obtained online.
A beam_y versus E_miss plot showed a correlation, which he corrected by hand, like Nilanga did it online.
The remaining resolution was still around 2.4 MeV, not allowing to clearly separate two-body and three-body peak.
He fitted the two-body peak and integrated up to two sigma above the peak.
The was a live discussion about whether or not this events are two-body or three-body.
Finally, the general feeling from looking at the trend of the data was that this is likely to be two-body,
and it should be presented at the Gordon conference.
Also kinematics #10, which was affected by the wrong HRSE Q2 setting in February, because it seems to affect primarily y_tg.
In this case no spec_e.react_z cut should be applied.
Michael will modify the transparancies next week and use a realistic acceptance for cross section determination from MCEEP.
Also, some kinematics, especially #1, should be cutted into p_miss slices.
Another suggestion was to mark the expected maximum of the three-body continuum assuming that one nucleon carries all the momentum,
i.e. the photon couples to a correlated nucleon pair.
For the resolution, he verified that the beam position reconstruction works.
Also, this beam position is used by the routine which corrects the momentum for the vertical beam position.
However, changing the matrix elements used for this has no impact on the computed E_miss, implying the correction is not done.
He will continue to check what is going wrong here.
- In the open discussion ...
-
... we tried to break down E89044 into possible thesis topics.
We found:
- 3He(e,e'p)d in perpendicular kinematics. Getting S(pm), RT, RL, RLT
- 3He(e,e'p)np in perpendicular kinematics. Getting S(Em,pm), RT, RL, RLT
- 3He(e,e'p)d in parallel kinematics. Studying the q-dependence, getting RT, RL
Further topics which don't really establish a thesis but can be e.g. a master thesis, are:
- 3He elastic cross section
- 3He(e,e'd)
- 3He(e,e'p) x-dependence
As open questions/problems were identified:
- Emiss resolution
- coincidence timing.
Sometimes the HRSH determines the timing, which causes a distortion of the coincidence time spectrum into two coincidence peaks and some counts in between.
It can be found by plotting any scintillator TDC times versus the coincidence time.
- We need a 4 GeV/c data base.
Michael recalled that these data were taken sometimes in 1999.
Nilanga knows.
- George ...
-
... now also includes all the windows/foils and the collimators into GEANT.
Work just started and is in progress.
- Eric ...
-
... spoke about target density determination from elastic cross section measurements.
First he determined the target pressure and temperature from the readings in the end-of-run statement.
He used readings from temperature sensors which already showed radiation damage.
Michael will circulate the correct readings.
Eric used several cuts to prepare a sample of clean elastic events.
The density he deduces is about 0.03 g/cm3, which is about half of what expected.
He didn't correct for radiative effects.
Also, he may also remove elastic events by cleaning the event sample.
He also uses the cross section at average scattering angle.
Konrad said the acceptance averaged cross section doesn't differ too much.
- Konrad ...
-
... presented the results of his elastic scattering analysis.
They are accessible at the CalState Nuclear Physics web page
http://www.calstatela.edu/academic/nuclear_physics/welcome_np.htm.
He checked runs at different currents, and there doesn't seems to be a beam heating effect.
He gets densities which are about 5%-10% higher than the tabulated densities, but in general his results are pretty close to what it should be.