May 24, 2007

For 5/24/2007 Spin PWG meeting

BBC - ZDC comparison

The ratio of the ZDC coincidence and the BBC coincidence is not flat in several ways.

  • It is not flat over the time.
  • It is not flat over the bxing.

The plots show that the ratio is not flat over the time. It jumps during a fill.

However, this jump is not the reason why R3(BBC) - R3(ZDC) doesn't have peaks at zero for some fill.

  • The fill 7887 doesn't have a jump but R3(BBC) is systematically smaller than R3(ZDC).
  • The fill 7896 has a jump but R3(BBC) - R3(ZDC) has the peak at zero.

This shows ZDC/BBC over the bxing for the fill 7887. This is not flat. The ratio is large for the first about ten bunches for the both yellow beam and blue beam.

This effect can be seen in all 200 GeV fills.

I computed R3 without those twenty bunches. R3(BBC) - R3(ZDC) gets closer to zero in the fill 7887.

But it is not true for all fills. It becomes worse for the fill 7896.

The relatively large ratio in the first ten bunches for the yellow and blue beam is not the only phenomenon that can be obesrved about the raio. This plot shows the ratio for one of the scaler files from the run 7140042 fill 7887.

Three phenomena can be observed.

  • Leading bunch effect
    The ratio is large for the first ten bunches for the yellow and blue beam.
  • Ripple effect
    The ratio oscillates with low frequency.
  • Sawtooth effect
    The ratio oscillates with high frequency.

I plot BBC and ZDC separately. It is not clear which one causes above three phenomena.


For 5/10/2007 Spin PWG meeting

R3 -- BBC - ZDC comparison --

I compared R3 from BBC and ZDC for the Board 5 and 6.

  • R3(BBC) was computed by integrating over all the timebin.
  • R3(ZDC) was obtained from the bit 10 and 11.

In the following image, R3's from both BBC and ZDC are plotted.

  • No major discrepancy is observed.

This is a histogram of the R3(BBC) - R3(ZDC) for the 62 GeV runs. One entry is from one scaler file.

This is the same for the 200 GeV runs.

This shows the mean of the above histograms for each fill for only the 200 GeV runs. I fit with the constant.

  • The red line is the fitted constant.
  • The value is -0.001448.
  • chi^2/ndf is 125.9 / 77.

R3 -- Fill 7729 --

The R3 changed substantially during the fill 7729. It started around 1.100 and ended around 1.056.

This image shows how the BBC coincidence rates changed for each bunch during this fill.

  • Each bunch has its own panel.
  • Panels with blueish color are for the spin parallel (spin4 = 5, 10).
  • Panels with orangeish color are for the spin anti-parallel (spin4 = 6, 9).
  • Grays are not filled.

I am studying if particular bunches caused this substantial change of R3 during this fill.

This plot shows how the raio of the maximum rate and the minimum rate for each bunch distributed for each spin bit. It is not the case that the bunches with specific spin bits have outliers. It seems that the spin bit 9 tend to have larger values in this plot. Since the spin bit 9 comes in the denominator for the R3, the R3 becomes smaller as time passes.

I will continue to study this fill. But it is unlikely that this problem will be attributed to few bunches.

For 5/3/2007 Spin PWG meeting

Summary up to now

  1. QA on the BBC coincidence rates
    • The following boards and bits were not rejected by the QA on the BBC coincidence rates.
      • Board 5 timebin
      • Board 6 timebin
      • Board 11 bit16
      • Board 12 bit16
  2. The fill patterns
    • There were unfilled bunches which were not kicked bunches or the bunches in the abort gaps.
    • The fill pattern was determined from CDEV for each RHIC fill.
    • The patterns were consistent with the BBC coincidence rates.
    • All unfilled bunches will be masked for the spin analyses.
    • tai_070412_SpinPWG_Scaler.pdf

R3 - overall view -

I computed R3 for the Board 5 timebin, the Board 6 timebin, the Board 11 bit16, and the Board 12 bit16 with the fill patterns which mask all unfilled bunches.

This plot shows the overall view of R3 for 2006.

R3 - board comparison -

The plot on the right is to show that R3 from the four boards are consistent.

  • There is a panel for each fill.
  • The y-axes are sliced, meaning in each panel, the y-axis covers the same size of the range.
  • The x-axes are free, meaning in each panel, the x-axis is determined to cover the all points for the fill.
  • It is deceiving in the sense that R3, for some fill, appears not to have changed much during the fill when it did.

R3 - vertex dependency -

The plot on the right shows how R3 changes when the timbins that are included are widened.

  • The y-axes are sliced.
  • The x-axes are free.

R3 - time dependency -

The plot on the right shows how R3 changed during the fill.

  • The y-axes are same for all panels.
  • The x-axes are sliced.
  • The same run has the same color.
  • In 2006, for some runs, R3 changed a lot during the run.
  • It is important to understand the effect of this change on the spin asymmetries.
  • It is not obvious which number should be used for R3 for the runs during which R3 changed significantly.
  • For the runs in which R3 changed little, it is probably safe to use R3 from the board 5.
  • It could be possible to parametrize R3 as a function of time and use R3 for the time at which each event occured.
  • It might be a good idea to use only the fills for which R3 changed little for the first iteration of the spin analyses.
  • Then later add the other fills and see the impact on the spin asymmetries.

QA plans within scaler files

  1. Comparison with ZDC
  2. Background study

QA plans in coordination with each final state analysis

  1. Single spin asymmetry

File Format for Providing Relative Luminosity for Spin Analyses

I am planing to provide the final numbers in the same format Murad did for 2005 so that individual analysis can keep using StTamuRelLum as long as possible. This may change if it becomes necessary to subtract background, parametrize the relative luminoisty, or do some other operation on the values.