st_upsilon mass

Heavy Flavor phone meeting: 12/4/2007

  • We ran over ~4.1 million st_upsilon events. With this we make like sign and unlike sign pairs.

  • Event Selection and Particle ID (i.e. preselection or "Base Cuts"):

    1) E > 1
    2) 1 < p < 50
    3) nHits >= 20
    4) -30 < vertex_z < 30
    5) 0.5 < p/E < 2.5
    6) 3.3E-6 < dE/dx < 5.0E-6
    7) nHitsFit/nHitsPossible > 0.52
    8) Removed Ahmed's hot towers


  • "Pseudo-trigger cuts" of E_Tow1 > 2.5 GeV and E_Tow2 > 4.0 GeV
      and that tracks are separated by angle θ >= π/3 radians.

  • In addition to "pseudo-trigger cuts" we apply the cuts of p/E and dEdx particle identification to our pairs:
    1) 0.5 < p/E < 2.0
    2) 3.4E-6 < dE/dx < 4.7E-6

  • A comparison to Debasish's upsilon mass peak study:
    1) we made similar momentum requirement p > 3
    2) structure looks somewhat similar... but still need to study things

  • "Pseudo-trigger" cuts and particle ID methods with a gauss fit (naturally, this is an admittedly zero'th order peek at the "signal")
    1) 0.5 < p/E < 2.0 AND 3.4E-6 < dE/dx < 4.7E-6

  • We plan to study the mass distribution using the preshower.
    comments made in meeting...
    1) to more realistically match L2 trigger decision, need "online" adc and gains. SEE 2007 (RUN7) TRIGGER FAQ
    2) error on subtracted plots is too low (we just let root do it... we will fix by using quadrature.)


    brief update: 12/11/2007

  • Error fixed in "subtracted plots": Error_quad = sqrt(errA^2 + errB^2)

  • After Manuel's comment that there must be additional care taken to make the cuts closer to what the trigger did, we looked at the pedestal subtracted ADC spectrum for towerA and towerB. Below we plot the (ADC-PED) distributions for towerA (black) overlayed with towerB (red). We have not yet read in the online gain values, so we cannot effectively match the online trigger cut. The "smooth" distribution is the L0 trigger, while the sharp rise in counts at ~285 ADC is the L2 trigger condition.
    (L2 trigger counts sit on top of the L0 trigger counts)

  • Instead of using the "pseudo-trigger" cuts on energy, we now looked at the effect of an (ADC-PED) cut on the mass distribution. There is a tight p/E cut applied, as well as tight dE/dx and a requirement that particles have p > 3. Fig. 1 has nothing to do with ADC spectrum. The Fig.2 has (ADC-PED) > 200 and nothing to do with a "pseudo-trigger" energy cut.
    1) 0.5 < p/E < 2.0 (tight)
    2) 3.4E-6 < dE/dx < 4.7E-6 (tight)
    3) p > 3
    "Pseudo-trigger" energy cut and 0.5 < p/E <2 and tight dE/dx
    Fig. 1
    Cuts described above and a cut of (ADC-PED) > 200
    Fig. 2

    While nothing conclusive can be said about the mass distribution, we do see the L2 trigger in the (ADC-PED) spectrum. Still need to untangle the L2 and L0 triggers, which we are currently working on.


    update: 12/23/2007

  • We have read/saved the information from the 2007 L2 trigger result in the L2 array. Refer to the 2007 Upsilon L2 trigger: Online monitoring for information about L2 parameters. An example of the typical mass and cosine distributions we get from L2 are shown below:

    • L2 Mass plot
    L2 Cosine plot

  • The type of information accessed from L2:

    • tag: 1
    time: 279
    Events seen: 4070
    Events accepted: 1613
    triggerTowerL0: 327
    triggerTowerL2: 3135
    Energy of L0 cluster [MeV]: 6176
    Energy of L2 cluster [MeV]: 3596
    invMass [MeV]: 9254
    cosTheta: -0.92792
    vertexZ [cm]: 0
    numberOfHotTowers: 25


    update: 2/4/2007

  • We have accessed the online results for ADC, Pedestal and Status. We do not yet simulate the trigger decision (still working on it).

    We also plotted the background mass distribution for e+e+ + e-e-, then separated the background into e+e+ and e-e- components, as shown below:

    •
    20 bin plot from (5-15) GeV
    8 bin plot from (8-12) GeV


    update: 2/19/2007

  • We have read/saved the 2007 L2 trigger result. We are studying the effectiveness of the L0/L2 trigger decision relative to our "mimicking of the trigger decision" (by using the online database results for the adc, ped, gains, etc.). Having access at the Event and Track level, we look at 1) how the L0/L2 mimic affects the tracks and 2) how the mimic affects the paired tracks:
    •

    L0/L2 trigger results vs. online mimic of the L0/L2 triggers
    Event/Track Condition NO cuts p, E, dE/dx dE/dx cut only p cut only E cut only cuts p, E, dE/dx

    L0 condition: (adc>>4) ≥ 19

    282

    *

    **

    £

    30

    Track matching: tower = L0_tower

    269

    *

    **

    £

    30

    Track matching: tower = L2_tower

    351

    *

    **

    £

    5

    L2 condition: ((adc-ped)>>4) ≥ 5

    10207

    *

    **

    £

    387

    Total number events

    779

    *

    **

    £

    779

    Total number tracks

    440163

    *

    **

    £

    970

    Total im a place holder for something

    *

    *

    **

    £

    *



    Heavy Flavor phone meeting: 2/26/2007

  • Rory is documenting this work on our other page (see below)
  • Study of the 2007 L0/L2 trigger result on un-paired tracks:Upsilon Trigger Investigation
    •


    Heavy Flavor phone meeting: 3/4/2007

  • Rory is documenting this work on our other page (see below)
  • 2007 L0/L2 trigger "mimic" for paired tracks:Upsilon Trigger and Preshower Investigation
    •

    created by rfc and mcc...
    (updated by mcc on 3/6/08)