Diffractive Triggers

1
Diffractive Triggers
Andrew Brandt, U. Texas at Arlington
Trigger Integration November 18, 2003
2
GapJet Triggers
JT_15TT_GAPN or S CJT(2,3)GAP L3(1,15)
Prescaled, currently .3-.4 Hz each Llt40E30
lt.1Hz at 40 E30 e80-l40-n30? JT_15TT_GAPSN
CJT(2,3)GAPS L3(1,15) Currently unprescaled
0.6 Hz at 40E30 expect a prescale at higher
luminosity unless can improve trigger (or
natural prescale from double gap takes
hold) JT_45TT_GAPN or S CJT(2,5)GAP L3(1,45)
Currently prescaled by 2 at 40E30 .08 Hz each
would like to keep unprescaled at least to 60E30
(like highest inclusive ET trigger). Also will
have natural SI prescale (note s80-u60 in
prescale file!) JT_45TT_GAPSN CJT(2,5)GAPS
L3(1,45) Unprescaled at all luminosity 0.03 at
40E30 Also have 3 zero-biasgap triggers
These triggers are used for single diffractive
and double pomeron jet physics (Gaps and FPD)
3
J/?? Gaps
J/?? Gap 2MT1_2TRK_GAPN 2MT1_C_2L2L_2TRK
ALMNorthv 2MT1_2TRK_GAPS 2MT1_C_2L2L_2TRK
ALMSouthv Before shutdown prescale of 4 at
30E30, 200 at 40E30 (prescale now
lowered) Unprescaled rate .25 Hz _at_20E30 .5
Hz_at_40E30 e80-l60-n40 would be good Could be
unprescaled at all lum, when low PT track match
works J/?? Gaps 2MT1_2TRK_GAPSN
2MT1_C_2L2L_2TRK ALMSouthvALMNorthv lt.01 Hz
at 20E30 lt.04 at 80E30 should be unprescaled at
all luminosity These triggers are being used to
search for exclusive J/?? and ?C , a key
step towards validating diffractive Higgs models

4
FPD DAQ
5
FPD L1 Trigger

• Modelled after CTT (central fiber tracker
  trigger) but with fiber detectors read out by
  multi-anode phototubes (MAPMTs) instead of
  VLPCs
• Requires a transition board (TPP) to shape
  signals and discard excess charge for use with
  AFEs (Analog Front End boards) which receive
  signals, record analog values, and discriminate
• DFE (Digital Front End boards) receive digital
  signals from AFEs and apply tracking firmware to
  perform hit multiplicity cut (veto events with
  excessive multiplicity), divide tracks in
  momentum and angle bins
• LM TDC boards used to perform timing from hits in
  scintillators from the Luminosity Monitor
  sub-detector as well as the FPD trigger
  scintillators
• FPD timing and LM information will feed into FPD
  trigger manager (TM)
• along with DFE information to form FPD AND/OR
  terms

6
Trigger Manager Inputs

• FPD_LM
• Information on which detectors are hit and
  halo
• 2) LM pass through 16 LM and/or terms
• includes GapN GapS GAPSN SI etc.
• (who in LM group is doing this? ) mostly
  can get this info separately
• DFE information from scintillating fiber
  detectors,
• ?, t, multiplicity

7
FPD Trigger Status

• All 18 FPD detectors are in readout, trigger
  scintillators in readout soon.(!)
• Currently can only trigger using special runs
  (NIM to generate AND/OR terms).
• Need LM vertex board (lt6 months, according to
  Brendan) combined with Trigger Manager
  (commissioning starting in next few weeks) for
  scintillator based triggers.
• Could have DFE-based triggers through TM earlier
• Will need some resources, bandwidth, exposure
  groups, trigger bits
• Proposed strategery add high rate monitor and
  calibration triggers (elastic, inclusive
  diffractive) in separate global run not to be
  recoed.
• Use TM to define an FPD exposure group (gaps 3-gt1
  or 2). A few global run triggers for double
  pomeron, J/??, and dijets. Other diffractive
  physics symbiotically.

8
Special Run Trigger
In-time hits in AU-PD detectors, no early time
hits, or LM or veto counter hits
Current NIM logic allows us to form several
elastic and diffractive triggers for special
runs using trigger scintillators (in parallel
information from scintillators is sent to TDCs
for commissioning FPD_LM system, and CAMAC
scalars), veto counters, and LM. Can trivially
switch from elastic to double pomeron (Aup-Pup
for example)
9
FPD_LM

• TDC Details
• Step I TDC in readout
• 3 TDC boards for FPD compared to 6 LM boards
• 1 board consists of Px1, and Ax1 where
  xUp,Down, In,
• Out pots (first quadrupole castle on
  either side of IP)
• 2 board consists of Px2, and Ax2
• (second quadrupole castle)
• 2 dipoles 4 veto counter signals
• Step II Pass info to vertex board and add scalar
  info
• Allowed 80 bits from each side to vertex board
• Each board should set bits at two times,
  corresponding
• to in-time hits (time of flight from IP)
  and to early-time
• hits from halo (time of flight from IP
  earlier than
• interaction) Dipole 1 halo bit
  (T-D2ltT-D1), VC none?
• For each TDC give in-time bit 0 or 1
  16 (6) bits
• For each TDC give halo bit 0 or 1
  16 (1) bits
• TOTAL
  32 (7) bits

10
More TDC Details

• 1 board should send 8 in-time and 8 halo bits to
  2 board
• which should receive these bits and send 16
  in-time bits and 16
• halo bits to vertex board
• DV board should send 6 in-time and 1 halo bit to
  Vertex board
• Note cable lengths have been fixed such that 1
  in-time/halo signals arrive
• at 1 board at a fixed time (594 nsec/438), and
  all 2 signals arrive at 2 board
• at a different fixed time (625/407) , and DV
  signals arrive at a 3rd
• fixed time (825).
• Manchester group on the job, plan to have TDCs
  operational soon
• Steps to operational FPD TDC boards?
• Scalar on TDC board or vertex boards?
• Vertex Schedule?
• Update next week.

11
Vertex Board Details

• Vertex board should receive 22 in-time bits and
  17 halo bits from TDC boards.
• It could form 4 bit words for each of the 8
  non-dipole spectrometers
• 0no coincidence of 12 pots for that
  spectrometer
• 1coincidence of 12 pots for that spectrometer
• 2coincidence but either one of diagonally
  opposite spectrometer pots has halo bit set
• 3coincidence but both of diag opposite pots has
  halo bit set
• Examples PU1.and.PU2 no early time hits in AD1
  or AD2 1
• PI1.and.PI2 with early time hits
  in AO1 and AO2 3
• OR it could just pass the info to TM, depending
  on Firmware space on Vertex or TM
• OR BOTH
• Vertex board passes information to TM
• Can pass up to 96 bits to TM in 16 bit bursts
  (separated by
• 18 nsec) ,74 bits if send processed and raw
  info
• 1) Spec 1-8 4bit words (32 bits)
• Spec 9 3 bit word, 2 in-time bits from spec. 9,
  1 halo bit from spec. 9
• (note for dipole halo bits are formed from
  order of dipole times D1gtD2 -gthalo) (3 bits)
• 3) 4 veto counter in-time, 4 VC halo bits? Like
  lm halo? (4 bits)

12
Previous Plan for FPD Triggers

• The FPD Trigger Manager allows cuts on ?1-?p/p
  and t, and
• also incorporates information from the
  trigger scintillator via
• the LM boards.
• A track is defined as two detector hits in any
  spectrometer with
• a valid x and t, a trigger scint. confirm, and
  no halo veto set.
• AND-OR term definitions (13 used of 16 allowed)
• RTK track in any spectrometer, (D veto on
  halo)
• RPT proton track RAT anti-proton track
• RTK(1) x gt 0.99, all t
• RTK(2) 0.99 gt x gt 0.9 all t
• RTK(3) x gt 0.9 all t, no halo veto
• RTK(4) x gt 0.9, tgt1 GeV2
• RTK(5) x gt 0.9, all t
• REL Elastic (diagonally opposite p and )
• ROV Overconstrained track (DQ proton tracks)
• REL(1) x gt 0.99, all t REL(2) x gt
  0.99, t gt 1 GeV2
• ROV(1) x gt 0.90, all t ROV(2) x gt
  0.90, t gt 1 GeV2)

13
New Trigger Plan

• Input information
• Currently no global run trigger capability
• Vertex board is delayed
• DFE boards work and TM ready to be commissioned
• Main background not from pileup (multiple
  interactions) but from halo spray
• New strategy
• Instead of calculating bin of ? and t, use fiber
  hit patterns to demand 2 or 3 out
• of 3 planes of each detector are hit.
  Replaces trigger scintillator, simpler algo
• Use multiplicity cut to reject halo spray, code
  several multiplicity levels
• NOTE fiber ADC threshold must be high enough to
  avoid noise, low enough
• to retain efficiency and allow vetoing of
  halo
• One advantage is pot positions not needed at
  trigger level
• Issues
• Setting ADC threshold, need special run (and
  analysis)
• Dealing with noisy channels, variable means,
  could initially set threshold high,
• later load in mean pattern, known hot
  channels
• Need to settle on bit pattern to proceed with TM
  logic

14
TM Algorithm
DFE will pass one word for each spectrometer
indicating coincidence of two detectors and
multiplicity level, use this to define a track
(could have different multiplicity level in
different spectrometers). At TM we would form
terms DIFFQany quadrupole spectrometer track
(could be false if gt1 or 2 on A or P
side) DIFFDdipole track ELASAU-PD or AD-PU or
AI-PO or AO-PI DPOMAU-PU or AD-PD or AI-PI or
AO-PO OVERAU-DI or AD-DI or AO-DI or AI-DI
(over-constrained track for alignment)
15
FPD Trigger List
Tentative L1 FPD trigger list. V13 (mid-January)
possible? 1) elastic (diag opposite
spectrometers) GAPSN 2) soft diffraction (single
spectrometers)GAPS or GAPN 3) overconstrained
track (pbar in quadrupole dipole
spectrometers)GAPN 4) double pom (up-up, dn-dn
etc.)GAPSN if needed 5) CJT(2,3) FPD Track
(DIFFQ or DIFFS) GAPS or GAPN if needed 6)
CEM(1,3)? FPD track GAP? 7) TTK(1,?) FPD track
GAP? 8) MU(1,x) FPD track GAP? Monitors may
be necessary, will need to study with
special runs. Also rates are unknown. If 4
(dpom) it is not low enough to run unprescaled we
would need to repeat 5-8 with two FPD tracks. We
think triggers 1-3any monitors are best done in
a separate global run since they will not need
general farm reconstruction.
16
Trigger Strategy I

• Write out FPD, LM for every DØ event (may need to
  strip LM info
• from Raw data?)
• -cross section determined using standard
  methods x fraction
• that are diffractive corrected for
  acceptanceefficiency
• 2) Trigger on events that would not get written
  otherwise
• (Ex. single diffractive, elastic, double
  pomeron)
• using FPD track AND/OR terms, sometimes
  combined
• with gaps (veto on LM N or S)
• -high rate processes, cross sections will
  be measured in special
• run at some point (along with total cross
  section?)
• -these triggers will be in global list (or
  2nd global run) to measure ? and t
• distributions, also be used as monitors and
  for alignment,
• calibration, and efficiency studies
• -double pomeron lower rate, will require
  extra thought for cross
• section, may be able to tie to elastics, or
  work backwards
• from double pomeron object triggers

17
Trigger Strategy II
3) Trigger on events that would otherwise be
too heavily prescaled (Ex. Jets, J/?, and
then add gap(s) and/or track(s)) -cross
sections determined using object trigger that
does not have diffractive conditions and
then bootstrapping
18
Different Exposure Terms
FPD track(s)0 with no vetos on halo,
multiplicity, veto counter, or LM -standard
exposure group? Monitor (halo veto requires
vertex board) FPD track(s)1 (includes FPD halo
vetosmultiplicity cuts) (probably only use in
special run) FPD track(s)2 (FPD track1 a veto
counter) VCNbar or VCSbar5.2lt?lt5.9
(standard inclusive SD term) 2 separate groups
(VC requires vertex board) FPD track(s)3 (FPD
track1 a veto counterLM gap on same side)
(standard inclusive SD term, gap may be necessary
to control rate) 2 groups FPD tracks4 (2 FPD
tracks3) (elastic or double pomeron term) GAPN,
GAPS, GAPSN Global list FPD track0, FPD track2
or 3 (and also object), FPD tracks4 (and also
object), LM gap object, LM gapsobject
Currently have 3 Gap exposure groups in global
list out of 8 maximum. I assume 3 is limit for
diffraction. Propose a single LM gap term and
an FPD term.
I count 10 different diffractive exposure groups,
7 needed in global list!
19
Trigger Work
Short term trigger work DFE algorithms Mario,
Wagner Data analysis inputs from special runs
Molina, Mike, James, Renata DFE firmware
Ricardo, Daniel CTS tests Daniel DFE
examine? TM Daniel, Mario MC? Trigger
database L2 Gap ? L3 SI, Gap, PLtot,
FPDreco?
With detectors in readout, next key is maximizing
useful data sample getting L1 trigger online is
vital to FPD physics success