Today's Progress 5. July. 2006

Tune of PB time walk by (K^-,X) fixing PA time (3)

Since we have determined T0/PA offset well, finalize PB offset tune here.

Method

We try to follow the PB time walk again by delayed-event, defined by the condition,

deltaT'''''(T0->PA) = Tpa''' - Tt0''' -TOFkstop -TOFsec'' > 0.8 (nsec) .

In order to do that, updated residual, delta T'''''(PA->PB) defined by

delta T'''' (PA->PB) = Tpb'' - Tpa'' -TOFkmu2

->

delta T''''' (PA->PB) = Tpb'' - Tpa'''-TOFkmu2 ,

is studied run-by-run/arm-by-arm, firstly.

Result

The run-by-run variation of the gaussian center of delta T'''''(PA->PB) for L(black) and R(red) arms.
time walk of PB
The run-by-run variation of the gaussian width of delta T''''' (PA->PB) for L(black) and R(red) arms.
variation of PA-PB TOF resolution

We activate the third-stage run-by-run, but arm-by-arm PB offset, PBOFFSET3(idrun), and calculate the 1/beta by

1/beta (PA->PB) = (Tpb'''-Tpa''')*c/L_TOF

, where

Tpb''' = Tpb'' - PBOFFSET3(idrun).

PBOFFSET3(idrun) is defined as the Gaussian center of the distribution of delta T''''' (PA->PB) as is for PBOFFSET(idrun) and PBOFFSET2(idrun).

Further investigation of the PA-PB TOF analysis

Adopting the most-updated PA-PB TOF analysis described avobe, the monochromatic peaks from meta-stable states are shown and fitted with 3 gaussians and a 3rd order polinomial background below, with 100% statistics of E549 DAQ duration.
  • StopK ID function > -1. MeVee
  • Reaction time, i.e. delta T'''''(T0->PA) > 0.8 nsec
    • are doubly applied for the Kstop-charged-triggered events with their rection vertex within 4He target (fiducial volume cut).

    Kmu2/Kpi2/Ke3 peaks from delayed events by PA/PB Larm. run 25-306 are all summed up.
    Kmu2/Mpi2 peaks by PA-PB Larm
    Kmu2/Kpi2/Ke3 peaks from delayed events by PA/PB Rarm. run 25-306 are all summed up.
    Kmu2/Mpi2 peaks by PA-PB Rarm

    The overall PA/PB beta-inverse resolution is 0.0227/0.0189 for L/R arms, respectively, at this moment, although they may weakely affected by unknown background shape. Here, we expect substantial segment-dependent time walk which broaden the peaks, hence the resolution can be improved more. For the purpose, whole production runs are divided into 18 parts (each parts containes 10~12 runs), and segment-by-segment tune is performed for each of parts. The definition of the part is tabulatted below.

    Definition of run part. E549 run 25~306 are divided into 18 parts. Calibration runs (Cosmic/K^+/pi^- runs) are treated separately.
    Part 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
    Run 25~46 47~6669~89 91~105106~122 123~135 142~155156~168 169~181182~195 196~208 209~222223~236 237~253 254~266267~281 282~294295~306
    Kmu2/Kpi2/Ke3 peaks from delayed events by PA/PB Larm after sgment-by-segment tune. run 25-306 are all summed up.
    Kmu2/Mpi2 peaks by PA-PB Larm
    Kmu2/Kpi2/Ke3 peaks from delayed events by PA/PB Rarm after segment-by-segment tune. run 25-306 are all summed up.
    Kmu2/Mpi2 peaks by PA-PB Rarm

    The resulting resolution is now 0.0221/0.0182, to which the resolution is improved more by 3~4 %.

    The part-by-part stability of the Kmu2 peak position and width is exhibitted, with following two fitting methods:

    1. Fits all 3 monochromatic peaks simultaneously with a global 3rd-order-polinomial background.

    2. Fits Kmu2 peak with a local linear background.

    The stability of the 1/beta peak center for Kmu2 (black/red for L/R arms) by method 1.
    center of inverse-beta for Kmu2 (1)(part-by-part)
    The stability of the 1/beta gaussian sigma for Kmu2 (black/red for L/R arms) by method 1.
    width of inverse-beta for Kmu2 (1)(part-by-part)
    The stability of the 1/beta peak center for Kmu2 (black/red for L/R arms) by method 2.
    center of inverse-beta for Kmu2 (2)(part-by-part)
    The stability of the 1/beta gaussian sigma for Kmu2 (black/red for L/R arms) by method 2.
    width of inverse-beta for Kmu2 (2)(part-by-part)

    Those two methods gave consistent results. It should be noted that the resolution plot for Larm is fairly synchronizing with that of T0->PA TOF resolution(i.e. bad PA resolution, maybe due to large noise on the analogue lines, breaks T0->PA and PA->PB simultaneously).

    Next, segment-by-segment behavior is exhibitted with E549 100% statistics and compared to that of K^+ calibration runs. Note that i) delta T(T0->PA)>0.8 (nsec) and KstopID > -.1 (MeVee) are imposed commonly for K^+/K^- events.

    The segment-by-segment 1/beta peak position for Kmu2 of stopped K^- events before segment-by-segment correction, and its comparison with Kmu2 events of stopped K^+ events(black/red for L/R arms for K^-, green/magenta for L/Rarms for K^+). The horizontal dashed-line represents the 1/beta peak position to be adjusted, 1.1017.
    peak center of inverse-beta for Kmu2 (3)
    The segment-by-segment 1/beta gaussian sigma for Kmu2 of stopped K^- events before segment-by-segment correction, and its comparison with Kmu2 of stopped K^+ events(black/red for L/R arms for K^-, green/magenta for L/Rarms for K^+).
    width of inverse-beta for Kmu2 (3)
    The segment-by-segment 1/beta peak position for Kmu2 of stopped K^- events after segment-by-segment correction, and its comparison with Kmu2 events of stopped K^+ events(black/red for L/R arms for K^-, green/magenta for L/Rarms for K^+). The horizontal dashed-line represents the 1/beta peak position to be adjusted, 1.1017.
    peak center of inverse-beta for Kmu2 (4)
    The segment-by-segment 1/beta gaussian sigma for Kmu2 of stopped K^- events after segment-by-segment correction, and its comparison with Kmu2 of stopped K^+ events(black/red for L/R arms for K^-, green/magenta for L/Rarms for K^+).
    width of inverse-beta for Kmu2 (4)

    PB Larm ID3 gives bad resolution due to PMTtop trouble, and Rarm ID15 also due to known serious non-linearity of TDC used for upside PMT. These two segment should be omitted from the TOF counter, and used as VETO counters. And, poor overall resolution for Larm is attributed to the poor PA Larm resolution, as has been exhibitted by T0-PA TOF analysis. These troubles had been resolved at the beginning of the E570 October cycle.

    Proton uncorrected momentum spectra by PAPB TOF

    By using the stopK ID function and offset-tuned PAPB, energy-loss uncorrected proton inclusive spectrum from 4He(stoppedK^-,p) reaction with E549 100% statistics is as below. Note that the events form PB Larm segment 3 and Rarm segment 15 are already removed. No clear structure is found.

    Inclusive proton spectrum. About 1,000,000 proton events are accumulated.
    inclusive proton spectrum
    DCA-selected inclusive spectrum. Black -> DCA.le.5. mm, Red -> DCA.gt.5. mm.
    DCA selected proton spectrum

    They will be exclusively studied when prompt-delayed event analysis (T0-PA analysis) would be also enabled for proton.

    T0->PA performance with finalized PA-PB TOF analysis

    Since PA->PB offset tune has been finalized here, T0->PA TOF resolution is checked again with the finalized PA-PB TOF analysis. The run-by-run variation of arm-by-arm T0->PA TOF resolution is shown below. PA Larm seems to be with rather poor resolution compared to Rarm, which may cause the poor resolution of PA-PB TOF analysis on Larm.

    T0->PA residual width for "stop K^- on 4He" events, with PA-PB 1/beta described here.
    T0->PA resolution
    T0->PA Gaussian center for "stop K^- on 4He" events, with PA-PB 1/beta described here. The fit is performed to the region (-0.3, +0.3) for all, hence they are deviated to the positive side due to the hyperon life time. Larm result is stable compared to the Rarm result.
    T0->PA Gaussian Center

    Segment-by-segment results with E549 100% statistics are exhibitted below.

    T0 segment-by-segment T0->PA TOF residual width for "stop K^- on 4He" events. Black/red for PA L/Rarm, respectively.
    T0 by T0 T0->PA resolution
    PA segment-by-segment T0->PA TOF residual width for "stop K^- on 4He" events. Black/red for PA L/Rarm, respectively.
    PA by PA T0->PA resolution