Документ взят из кэша поисковой машины. Адрес оригинального документа : http://zebu.uoregon.edu/~uochep/talks/talks05/srtor01-05.pdf Дата изменения: Tue Jan 11 23:07:37 2005 Дата индексирования: Tue Oct 2 11:22:24 2012 Кодировка: Поисковые слова: m 5

SR Energy Spectrometer

January 8th, 2005 MDI at the ILC Workshop SLAC

Eric Torrence University of Oregon · Extraction Line Energy Spread · Extraction Line Layout · Testbeam Plans · Time dependence and correlations

Eric Torrence

1/15

January 2005


Synchrotron Spectrometer Overview SLC WISRD
Spectrometer Magnet Quadrupole Vertical Doublet eHorizontal Bends for Synchrotron Radiation

Dump Synchrotron Light Monitor

e+

Motivation · BPM measurement looks hard · Alternate, complimentary method with 100 ppm potential · Neutral signal avoids stray field sensitivity · Compact detector plane (micron-level absolute accuracy not crazy), long lever arm · Single spectrometer magnet · Very passive device, high reliability · Potential to measure differential spectrum · Good precision pulse-to-pulse
Eric Torrence 2/15 January 2005


Spectrometer Dilemma

0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0 240 245 250 255 Ebeam (GeV)

0.09

In

0.08 0.07 0.06 0.05 0.04 0.03

Out

+ - E b + E b s

0.02 0.01 0 240 245 255 Ebeam (GeV) 250

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 490 492 494 496 498 500 502 504 Root(s) (GeV)

s distribution

What can a beam energy measurement really tell you?
Eric Torrence 3/15 January 2005


Extraction Line Energy Spread Study Guinea Pig Inputs · Input NLC and TESLA TRC files · Vary vertical offset by ± 5 nm in 1 nm steps 11 x 6 x 6 GP runs per machine Some estimate of `realistic' variations Guinea Pig Outputs · lumi.ee.dat - Lumi weighted s spectrum · beam.dat - Disrupted beam files s is what we care about Disrupted beam is what we can measure Analysis
+ - Fit outgoing E b and E b distribution (best a spectrometer could hope to do)

Correlate s to outgoing beam parameters in truncated range +/- 10 GeV of peak MC Truth exercise Detector resolution will only make this worse
Eric Torrence 4/15 January 2005


Fit Example
400 350 300 250 200 150 100 50 0 240 600 500 400 300 200 100 0 240 242 244 246 248 250 252 254 Beam Energy (GeV) Events / ( 0.075 )

Electrons

242

244

246

248

250

252 254 Beam Energy (GeV)

Events / ( 0.075 )

Positrons

Circe Gaussian P ( E ) = g ( E ;E 0, ) + a where x = E / E'

0



E

x a 1 ( 1 ­ x ) a 2 g ( E ' ) dE '

Total of 5 parameters Implemented in RooFit, easy to make complicated observables

Eric Torrence

5/15

January 2005


Basic Features
Ecm (MeV) Ecm (MeV)

NLC 500
-800

-1000

TESLA 500

-1000

-1200

-1200 -2 -1 0 1 2 Offset (nm)

-1400 -2 -1 0 1 2 Offset (nm)

s variation with y : 70-100 MeV RMS over ± 5 nm (not Lumi weighted) Could almost live with this (< 200 ppm) if you believed the simulation
NLC 500
Ecm (GeV) -2.3 -2.4 -2.5 -2.6 -2.7 -2.8 -2.9 -3 -3.1 -3.2 -4.2 -4 -3.8 -3.6 -3.4 -2.3 -2.4 -2.5 -2.6 -2.7 -2.8 -2.9 -3 -3.1 -3.2 -4.2 -4 -3.8 -3.6 -3.4 + (GeV)

TESLA 500

+ - s vs. E b + E b shows mild negative correlation (cold)
Eric Torrence 6/15 January 2005


Parameter Correlations NLC
Ecm (GeV)
-2.4 -2.6 -2.8 -3 -3.2 -2.4 -2.6 -2.8 -3 -3.2 0.8 0.85 0.9 0.95 1 a0 0.6 0.65 0.7

TESLA

a0
0.75 0.8 a0

s vs. Circe Parameters Nothing stands out clearly
-2.4 -2.6 -2.8 -3 -3.2

-2.4 -2.6 -2.8 -3 -3.2 3 4 5 6 7 8 a1 9 9.5 10

a1
10.5 11 a1

-2.4 -2.6 -2.8 -3

-2.4 -2.6 -2.8 -3 -3.2 -1 -0.8 -0.6 -0.4 -0.2 0 a2 -0.6 -0.55 -0.5

Try general linear combination (J. Kolb, U. Oregon)

-3.2

a2
-0.45 -0.4 a2

i = A ij

j

i - predicted observable (i.e. s ) A ij - linear coefficients (determined from 2 minimization) j - 5 + 5 Circe fit parameters (spect. observables)
Eric Torrence 7/15 January 2005


Linear Combinations
( - 498) GeV ( - 498) GeV 0.9 0.8 0.7 0.6 0.5 0.4 0.9 0.8 0.7 0.6

Caveats · Partly cheating (trained and evaluated on same samples) · Only varying yoffset · Real Eb variation not included

|yoff| <= 1 nm

|yoff| <= 5 nm

± 1 nm
0.4 0.5 0.6 0.7 0.8 0.9 (y - 498) GeV
Entries Mean RMS 108 -0.2826E-03 35.18

0.5 0.4

± 5 nm
0.4 0.5 0.6 0.7 0.8 0.9 (y - 498) GeV
Entries Mean RMS 396 -0.2639 34.11

14 12 10

50 40 30 20 10 0 -100

70 MeV RMS s spread reduced to 35 MeV

8 6 4 2 0 -100 -50 0 50 100 - y (MeV)

-50

0

50 100 - y (MeV)

Want to apply this to some G. White IP feedback data

Eric Torrence

8/15

January 2005


X-line spectrometer layout · Focus at detector plane WISRD-style · Wigglers reduce SyncRad alignment systematics Detector · Wigglers can be turned Plane off for bgd studies · Up/down to maximize y / l (resolution) IP Wigglers cm Split Option Detectors +20 0
1mRad

cm Joined Option Detectors +20 0
1mRad

-20 Large wiggler

-20 Soft Bends

Detect SR using Cherenkov radiation from secondaries in quartz fibers · Avoid RF pickup, inductive cross-talk between wires · Modest energy threshold (Compton electrons, KE > 200 keV) · Simple, cheap, fast readout (multi-anode PMT)

Eric Torrence

9/15

January 2005


Yuri's 20 mRad Extraction Line

Disrupted beta and vertical dispersion.
(m )



x

1/ 2



y

1/ 2

Dy

0.09

1/ 2

2000.

Detector Plane

0.08 0.07 0.06 0.05

Spectrometer
1000.

Polarimeter

0.04 0.03 0.02 0.01

0.0

0.0

20.

40.

60.

80.

100.

120.

140.

160.

180.

0.0 200. s (m)

Working on tracking/SR simulation using GP inputs using Wisrd (Matt Sternberg, U.Oregon) Will switch to Geant4 BDSIM next week... Immediate Goal: Evaluate response/resolution of spectrometer Find tolerance on focus - detector plane distance
Eric Torrence 10/15 January 2005

D (m) y

1/ 2

3000.

0.10


Detector R&D Quartz fiber SR prototype · · · · Intrinsically fast KE > 200 keV threshold multi-anode PMT readout Easy gain adjust

Prototype Geometry 8 x 100 µm fibers (Left) 8 x 600 µm fibers (Right) 1 mm pitch Multi-anode PMT Up to 64 ch. readout Single HV input High gain Other Detector Possibilities Wisrd-style wires Diamond/silicon strips Visible or UV imaging (CCD) ...
Eric Torrence 11/15 January 2005


ESA Beam Test

Exclude

BPM tests

SR tests Use original spear wiggler E c = 0.9 MeV at 28 GeV eE c = 0.1 ­ 0.2 MeV from bends Dominant Cherenkov signal from Compton electrons Produced near/in fibers (TN-2004-7: Ray Arnold) G3 simulation predicts essentially no background at ESA energies Rather inefficient, validate with beam, compare to other options

Eric Torrence

12/15

January 2005


Tracking Energy Variation How precisely do we need to track energy variations?
3500 3000 2500 2000 1500 1000 500 0 204 205 206 207 208 209

2000 LEP2 Collision Energy Luminosity Weighted, in GeV

Only mW cares about details beyond s , and even then... Simple Argument Uncorrected variation in energy should be small compared to other characteristic widths: Ebeam 0.1 %, M / W 2 GeV

Seems modest, but must beware of correlations with other parameters: P e , Luminosity Probably only important in specific cases (ie: top scan)
Eric Torrence 13/15 January 2005


Do we need spectrometers at all? Can we simply use L and s from physics reactions? Contrived Example - Lineshape Scan L s Measured True s N= s



dL ----- ( s ) ds not N = L · ( s ) ds

Need dL / d s spectrum including both inter-bunch and intra-bunch variations Some Observations · No idea if this is significant, real study needed · Depends upon variance in L and s · Need to consider effects of IP feedbacks Fast spectrometer and luminosity monitor (at least relative) are probably essential
Eric Torrence 14/15 January 2005


Summary

MC Truth studies of X-line spectrometer · Some information available to track s · Want to apply to G.White IP feedback files · Will return to this with spectrometer simulation

X-line spectrometer simulation · Work underway, move to BDSIM/G4 · Test detector response, X-line geometry · Background studies

Beam Tests · Simple prototype ready to go · Find alternate SR source?

Really want to get working on integrated collision-energy spectrum measurement with application to physics observable
Eric Torrence 15/15 January 2005