Using HF for Luminosity

1
Using HF for Luminosity

• Jim Freeman
• FNAL
• June 14, 2001

2
Overview

• HF is finely segmented and very fast. Sensitive
  to single crossings.
• Monte Carlo studies show HF has low occupancy
  even at 1034 luminosity.
• ? can count empty towers as measure of
  luminosity.
• ? Use muon feature bit in HF to flag empty
  tower.
• ? Would like to monitor luminosity for each
  bucket
• Ideally, count towers with lt single pe. This
  measurement is stable against energy scale
  shifts,
• ? need single pe well separated from pedestal in
  normal data-taking mode
• ?Do monte carlo study to see if possible

3
Lego of crossing, ltMBgt17
4
Lego of crossing, ltMBgt17
5
Some numbers

• PMT Gain 4X104
• Front end input noise 6000e-
• QIE bin 17,000 e- (non-inverting input)
• Lightyield 0.25 pe/GeV
• PMT pe resolution 50/sqrt(Npe)

6
Toy Monte Carlo

• Model
• Inject energy into tower (multiple MB events,
  using Sarahs histograms as MB Et model)
• HF light yield GeV to pes (quantize nr pes)
• Pmt gain
• Pmt resolution for photoelectron
• Front end noise (rms6000e-)
• QIE quantization (17000 e- / bin)
• Then study results

7
Eta4, ltMBgt1, 1 physical tower
Quantization on X axis is ADC binning converted
back into GEV. Recall 4 GeV ADC bin 3.35 Note
that single pe should also appear as 4 GeV. (0.25
pe/GeV)
8
Eta4, ltMBgt 10, physical tower
9
Eta4, ltMBgt 20, physical tower
An empty tower (zero tower) means ADC count 0
Single PE peak
10
Fraction zeros vs ltmbgt for physical tower
Very linear relationship between ltMBgt and zero
fraction. ? good for luminosity measurement
11
Higher PMT gain doesnt hurt
Things get easier at higher pmt gain. Perhaps can
run lower eta pmts at higher gain?
Better separation at pmt gain 105
12
HF Trigger Segmentation
HF trigger tower is 6 physical towers
13
Eta4, ltMBgt 20, trigger tower1 physical tower
Simulate trigger tower by summing ADC values of
physical towers inside of trigger tower, making
cut on sum. X axis of histogram is ADC sum for
trigger tower.
14
Eta4, ltMBgt 20, trigger tower2 physical towers
15
Eta4, ltMBgt 20, trigger tower6 physical towers
16
Fraction zeros vs trigger tower size, ltMBgt 20
17
Fraction zeros vs ltMBgt, trigger tower 6
physical towers
18
Problem with Large trigger towers

• 6-to-1 physical-to-trigger towers looks like a
  problem. There is substantial saturation of
  occupancy.
• A possible solution
• Associate muon feature bit with one of the six
  towers (arbritrary). Then lose 6X statistics but
• One orbit is 88 ms. So 1000 measurements of each
  crossing bucket in 88 ms.
• 144 trigger towers. So 144 independent
  measurements per crossing.
• We have seen that average physical tower zero
  occupancy is 30 at 1034. So measurement is
  Sqrt(40,000)/40000 0.005
• ? in 88 ms, can measure each crossing bucket
  luminosity to 0.5 accuracy

19
Conclusions (tentative)

• HF occupancy .ge. single photoelectron seems to
  be low, even at full luminosity
• Counting zeros on a physical tower basis should
  work well. Even better at pmt gain 105.
• Losing a factor of 6 due to trigger tower
  granularity not a problem.
• Counting zeros looks like a promising way to
  measure instantaneous luminosities of each
  bucket. (Accurate measurements each 88ms).
• Should investigate feasibility of doing counting,
  histogramming in Level 1 Global Trigger. Can use
  otherwise unused HF muon feature bit to transmit
  information.