A RIKEN BNL Research Center Workshop

1
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
SPIN'04
p-Carbon CNI Polarimeters at BNL the Fastest
Physics Setup in the World.
I.Alekseev1 A.Bravar2 G.Bunce2,3 S.Dhawan4
R.Gill2 H.Huang2 W.Haeberli5 G.Igo6
O.Jinnouchi3 V.Kanavets1 K.Kurita7
A.Khodinov8 Z.Li2 Y.Makdisi2 A.Nass2
W.Lozowski9 W.W.MacKay2 H.Okada10 S.Rescia2
T.Roser2 N.Saito10 H.Spinka11
E.Stephenson6 D.Svirida1 D.Underwood11
C.Witten6 T.Wise9 J.Wood6 A.Zelenski2
ITEP1 BNL2 RBRC3 Yale U4 Wisconsin
U5 UCLA6 Rikkyo U7 StonyBrook U8
Indiana U9 Kyoto U10 Argonne U11
D.Svirida (ITEP/BNL)
2
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
Preface
SPIN'04

• The goal of this talk
• is to show, how the modern achievements of
  technology and electronics help to solve physics
  problems, which look hopeless at the beginning.
• Demand
• The research program with polarized protons at
  BNL collider-accelerator complex put out the
  demand for a fast and cost effective tool for
  beam polarization measurements in a wide energy
  range with a few accuracy.
• History
• In 1998 the experiment E950 at AGS proved the
  general possibility to use pC elastic scattering
  in the CNI region for proton polarization
  measurements and measured the analyzing power of
  the process.
• 3 days of data taking to achieve reasonable
  accuracy in AN.
• Yet very promising process because of small
  energy dependence of AN predicted by theory.

D.Svirida (ITEP/BNL)
3
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
E950 Experience. Challenge.
SPIN'04

• Small CNI analyzing power of few
• Need huge statistics ?2107 events for a single
  polarization measurement
• Fortunately large CNI crossection
• Allows to get the statistics in seconds, but at
  the event rate of 107/sec
• Necessary recoil carbon identification for
  background suppression
• Have to know the momentum transferred because of
  strong t dependence of AN
• Signal parameters measurement, TOF energy,
  usually slow
• Looks impossible to have just counting
  experiment
• Unpredictable correlated deadtime may cause
  false asymmetries
• Need very thin carbon target to avoid its
  destruction, light emission and beam emittance
  blowup

D.Svirida (ITEP/BNL)
4
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
Carbon Target
SPIN'04

• The first wonder
• Thickness 35 ?g/cm2 (15 nm, only ?100 atomic
  layers)
• Width 100600 ?m with 50 mm length
• Dimension ratio is 1 10000 3000000
• Solid state fixed target at a collider store
• Nearly infinite target lifetime
• No light emission
• Very small effect on the stored beam lifetime,
  allows tens of minutes for measurements in a
  single store
• Event rate up to ?1 per RHIC bunch crossing
  (?110 ns) at design luminosity
• Advanced technique using carbon spraying and the
  substrate etching (Indiana University)

Just imagine A scaled copy made of paper would
be 1 m wide and 300 m long
D.Svirida (ITEP/BNL)
5
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
Why WFD ? Whats WFD ?...
SPIN'04

• Need deadtimeless intellectual DAQ with carbon
  eventselection ON-THE-FLIGHT
• E950 DAQ version with FERA ADC/TDC would have
  had gt95 dead time _at_ RHIC design luminosity
• Have a powerful WFD module design by Yale
  University
• Only need FPGA design appropriate to
  polarimetry requirements
• Whats WFD ?
• 4 independent channels per single width CAMAC
  module
• Digitize input signal shape _at_ 420 MS/s 8 bit
• Powerful FPGA chip for fast signal processing
  and extreme flexibility
• 32(64) MB on-board DRAM for the result storage
• Commercially available components (no ASIC) ?
  relatively low cost

FPGA!
D.Svirida (ITEP/BNL)
6
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
Hows WFD ?
SPIN'04

• Hows WFD ?
• Digitize short-shaped Si signal
• 40 ns FW signal 16 points enough for good
  signal measurement
• Specific signal processing 6 points in
  parallel
• Define signal time (CFD), amplitude integral
  (TOFEnergy)
• Filter the values through the lookup tables to
  select CARBON events (kinematic correlation)
  and suppress doublepulsed events
• Bunch polarization pattern account
• Increment ON-BOARD scalers/histograms according
  to the result
• No data transfer to the host computer during the
  measurement time
• Everything is done on-the-flight, NO DEAD TIME
• Small amount of resulting data fast readout
  after the measurement

D.Svirida (ITEP/BNL)
7
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
CNI WFD Block Diagram
SPIN'04

• 4 modes of operation
• 8-bit amplitude, integral
• 1.2 ns time resolution
• Bunch , revolution
• Windowed sensitivity
• Baseline subtraction
• Noise reduction
• Time/amplitude LUT
• Integral/ampl LUT
• LUTs programmable trough CAMAC
• 5 always accessible scalers (P,P,P0, 2BG)
• 120 ch bunch distribution
• 3128 ch amplitude distributions for P,P,P0
• 3232 Time vs Ampl
• Crate or external INH
• Data delimiter
• 32 (64) MB data storage
• Zero deadtime
• FASTCAMAC option

CH1
BUFF
Virtex FPGA
Virtex FPGA
CAMAC Databus
Virtex FPGA
DRAM CNTR
Virtex FPGA
Control CPLD
CH2
CH3
DRAM 32 MB
CH4
140 MHz
420 MS/s/ch 2.4 ns/S
Mode FIFO, ev Bytes/ev Max
ev/s Note RAW 16 96
104 Debug/JET WAVEFORM 16 96 104 Debug/JET RESULT
256 6 1.5105 Test/Run BOTH 1 96 104 Test SCALER
S 0 107/ch Run Expect
ing rates 107/72 chans
D.Svirida (ITEP/BNL)
8
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
FPGA Block Diagram V9
SPIN'04
MUX
ADC
Virtex FPGA
3pt Filter
Waveform FIFO / Delay
Baseline Subtraction
140 MHz
1/4
Time CFD, TMAX
Amplitude
2
Baseline Calculation
Control Logic
Integral
Result FIFO
70 MHz
Level Trigger
Delim
CAMAC readout
Int/Ampl LUT
Time/Ampl LUT
Bunch , Rev.
Bunch 0
SYNC and Window Logic
AND
Bunch Polarization Pattern
1
Scalershistograms
Bunch Distr
Energy P0 Distr
Special Scalers P,P,P0, BG always accessible
INH
Energy P Distr
Energy P Distr
2-dim T(E) Histo
Ext INH
D.Svirida (ITEP/BNL)
9
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
WFD test with pulser signal
SPIN'04
Internal BunchHistogram
Raw signal on ch.1
Timereading
Amplitude andintegral readings
InternalAmplitudeHistogram
Bunchreading
Baseline subtracted on ch.2
D.Svirida (ITEP/BNL)
10
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
On-Board Histograms (old)
SPIN'04
Time, ns
Energy, KeV
ENERGY SPECTRA SORTED BY POLARIZATION

• Programmable LOOKUP tables allow to select and
  count CARBON events on-the-flight
• Coarse histograms due to limited amount of
  memory in V5

D.Svirida (ITEP/BNL)
11
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
Off-Line Processing
SPIN'04
3s Mass cut
carbon
non-relativistic kinematics
Time of Flight (ns)
MC 11.17 GeV sM 1.5 GeV
prompts
alpha
Invariant Mass
Energy (keV)

• Data taken in RESULT mode
• CARBON selection through the banana cut
• Mass spectrum

Results from O. Jinnouchi
D.Svirida (ITEP/BNL)
12
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
RHIC Polarization History in pp Run04
SPIN'04
Change in Si dead layer parameters
ONLINE
little loss or none at the ramp
switch to horizontal target
Polarization
Results from O. Jinnouchi
Days from April 1st
Polarization
Data Points Black 24GeV Color 100GeV
Days from April 1st
D.Svirida (ITEP/BNL)
13
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
RHIC CNI measurements within a fill
SPIN'04
ONLINE
P(RHIC) / P(AGS)
injection
in the last week of run
1.001-0.016 (expect 0.974), pol.48.7 ?0.5
0.952-0.019 (expect 0.993), pol.48.6 ?0.7
injection

• Pol. Size is confirmed at RHIC AGS CNI
  (assuming same AN values)
• 56 bunch mode (alternating spin pattern)
• 4x1012 proton /ring

Results from O. Jinnouchi
D.Svirida (ITEP/BNL)
14
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
RHIC Ramp measurements
SPIN'04
the last and the best one
May 12th fill 5332
Results from O. Jinnouchi
Beam Energy
Event rate

• 5 ramp measurements taken in Run-04 (Fill
  5159,5169,5170,5199,5332)
• Larger on-board memory, larger statistics
• All of them suffered large profile effects due to
  the beam motion along the ramp
• Found the depolarization point prior to the
  energy ramp (200MHz RF cavity magnet ramp)

D.Svirida (ITEP/BNL)
15
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
RHIC Issues on polarization profile
SPIN'04
BLUE H scan
Yellow V scan

• Polarization
• Beam Intensity
• H Scan --horizontal scan with vertical target
• V Scan vertical scan with horizontal target

3mm
3mm
BLUE V scan
Yellow V scan
3mm
3mm
Results from O. Jinnouchi

• Large polarization profile in vertical (small
  profile in horizontal)
• Observed position dependent fluctuation in pol
  measurements

D.Svirida (ITEP/BNL)
16
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
RHIC bunch by bunch azimuth fit
SPIN'04
ONLINE
Amplitude
q
56th
4th
3rd
2nd
1st
typical run (not best)
cos q
Expected

• Bunch by bunch polarization fit feasible

Results from O. Jinnouchi
D.Svirida (ITEP/BNL)
17
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
AGS Asymmetry during the Ramp
SPIN'04
0.015
1 ms bin width
0.01
Raw asymmetry AN ? Pbeam
0.005
0
-0.005
-0.01
-0.015
G?
25
30
35
40
45
Ebeam
12
14
16
18
20
22
24
Asymmetry flips sign at every G? n
Results from J. Wood
D.Svirida (ITEP/BNL)
18
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
AGS RHIC And Physics Too
SPIN'04
momentum transfer t (GeV2/c2)
p 3.9 GeV
CNI peak 4
AN ()
áPBñ 73
preliminary
statistical errors only
p 6.5 GeV
áPBñ 65
p 9.7 GeV
áPBñ 60
p 21.7 GeV
áPBñ 47
recoil Carbon energy (keV)

• pC Asymmetry dependence on t at 100 GeV RHIC

• pC AN(t) dependence at different AGS beam
  energies

Results from O. Jinnouchi and J.Wood
D.Svirida (ITEP/BNL)
19
A RIKEN BNL Research Center Workshop A RIKEN BNL
Research Center Workshop A RIKEN BNL Research
Center Workshop
WONDER !!!
SPIN'04

• WONDER with WFD technique
• Up to 1M events/s seen in RUN2004
• 500M events/s generally achievable by DAQ in
  scaler mode
• 20 60 s target-in-the-beam time for 20M carbon
  events
• lt 10 s for readout and asymmetry calculations
  (scaler mode)
• 1 run is statistically equivalent to 3 days of
  E950 data taking
• Allows to measure polarization several times in
  one RHIC fill
• About 1000 polarization measurements in a 5 week
  run
• Lots of various beam polarization studies in
  RHIC and AGS
• Interesting physics results, important for
  hadron spin flip contribution understanding
• Excellent performance, meeting all reqirements
• Plan even faster readout with FASTCAMAC
  technology or gigabit serial interface for
  result mode

D.Svirida (ITEP/BNL)