-
Notifications
You must be signed in to change notification settings - Fork 0
/
pff_short_conflict-20170228-183539.tex
682 lines (581 loc) · 28.5 KB
/
pff_short_conflict-20170228-183539.tex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
% ****** Start of file apssamp.tex ******
%
% This file is part of the APS files in the REVTeX 4.1 distribution.
% Version 4.1r of REVTeX, August 2010
%
% Copyright (c) 2009, 2010 The American Physical Society.
%
% See the REVTeX 4 README file for restrictions and more information.
%
% TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
% as well as the rest of the prerequisites for REVTeX 4.1
%
% See the REVTeX 4 README file
% It also requires running BibTeX. The commands are as follows:
%
% 1) latex apssamp.tex
% 2) bibtex apssamp
% 3) latex apssamp.tex
% 4) latex apssamp.tex
%
\documentclass[%
reprint,
superscriptaddress,
%groupedaddress,
%unsortedaddress,
%runinaddress,
%frontmatterverbose,
%preprint,
%showpacs,preprintnumbers,
%nofootinbib,
%nobibnotes,
%bibnotes,
amsmath,amssymb,
prl,
% aps,
%pra,
%prb,
%rmp,
%prstab,
%prstper,
%floatfix,
]{revtex4-1}
\usepackage{graphicx}% Include figure files
\usepackage{dcolumn}% Align table columns on decimal point
\usepackage{bm}% bold math
\usepackage{color}
\usepackage{verbatim}
%\usepackage{hyperref}% add hypertext capabilities
%\usepackage[mathlines]{lineno}% Enable numbering of text and display math
%\linenumbers\relax % Commence numbering lines
%\usepackage[showframe,%Uncomment any one of the following lines to test
%%scale=0.7, marginratio={1:1, 2:3}, ignoreall,% default settings
%%text={7in,10in},centering,
%%margin=1.5in,
%%total={6.5in,8.75in}, top=1.2in, left=0.9in, includefoot,
%%height=10in,a5paper,hmargin={3cm,0.8in},
%]{geometry}
\begin{document}
%\preprint{APS/123-QED}
\title{High Bandwidth Arrival Time Stabilisation of an Electron Beam at
the
50~fs Level}
%\thanks{A footnote to the article title}%
\author{J.~Roberts}
\email{Corresponding author Jack.Roberts@cern.ch}
\affiliation{John Adams Institute,University of Oxford}
\affiliation{CERN, Geneva}
%\altaffiliation{CERN, Geneva.}
\author{P.~Burrows}
\affiliation{John Adams Institute,University of Oxford}
\author{G.~Christian}
\affiliation{John Adams Institute,University of Oxford}
\author{R.~Corsini}
\affiliation{CERN, Geneva}
\author{A.~Ghigo}
\affiliation{INFN/LNF, Frascati}
\author{F.~Marcellini}
\affiliation{INFN/LNF, Frascati}
\author{C.~Perry}
\affiliation{John Adams Institute,University of Oxford}
\author{P.~Skowronski}
\affiliation{CERN, Geneva}
\date{\today}
\begin{abstract}
CLIC, a proposed future linear electron-positron collider, and other machines
such as XFELs, place tight tolerances on the phase stabilities of their beams.
CLIC proposes the use of a novel, high bandwidth and low latency, `phase
feedforward' system required to achieve a phase stability of
\(0.2^\circ\)~at~12~GHz, or
about 50~fs. This work documents the results from operation of a prototype
phase
feedforward system at the CLIC test facility CTF3, with 30~MHz bandwidth and a
total hardware latency of 100~ns. New phase monitors with
30~fs resolution, 20~kW amplifiers with 47~MHz bandwidth, and electromagnetic
kickers have been designed and installed for the system. The system utilises a
dog-leg chicane in the beamline, for which a dedicated optics have been created
and
commissioned. The prototype has demonstrated CLIC-level phase stability,
reducing an initial rms phase variation of \(0.92\pm0.04^\circ\) to
\(0.20\pm0.01^\circ\) across a duration of 10~minutes.
\end{abstract}
%\pacs{Valid PACS appear here}% PACS, the Physics and Astronomy
% % Classification Scheme.
%\keywords{Suggested keywords}%Use showkeys class option if keyword
%display desired
\maketitle
%\tableofcontents
%\section{\label{s:intro}Introduction}
The Compat Linear Collider, CLIC, \cite{CLICCDR} is a proposal for a future
linear electron--positron collider. It uses a novel two beam acceleration
concept to achieve a high accelerating gradient of 100~MV/m
and a collision energy of up to 3 TeV. In this concept the 12~GHz RF power used
to accelerate each high energy colliding beam is extracted and transferred from
a high intensity drive beam in 24 decelerator sectors. \textcolor{red}{The
drive beams are generated by compressing an initial
\(140~\mathrm{\mu s}\) beam pulse bunched at 0.5~GHz into 24 shorter 240~ns
beam pulses bunched at 12~GHz, in a bunch recombination process using a
sequence
of combiner rings and delay loops [REF]. }
CLIC's luminosity quickly drops if the drive beam phase, or arrival time,
jitters with respect to the colliding beams, causing energy errors and
subsequent beam size growth at the interaction point. The drive beam phase
stability must be \(0.2^\circ\)~at~12~GHz (around 50~fs) rms or better to limit
the luminosity loss to below 1\% \cite{CLICCDR}. However, the drive beam phase
stability cannot be guaranteed to be better than \(2^\circ\)~at~12~GHz [REF]. A
mechanism to improve the drive beam phase stability by an order of magnitude is
therefore required.
\textcolor{red}{Other machines, such as XFELs, have similar beam phase
stability
requirements to CLIC. At FLASH, DESY, these requirements have been met
using an RF phase
and power feedback based on the measurement of electro-optic beam arrival
time monitors [REF]. However, the CLIC drive beam presents a different set
of challenges. In particular, FLASH has 1~MHz bunch spacing and a 500~ms
beam pulse, whereas the CLIC drive beam has 12~GHz bunch spacing and 240~ns
pulse length. A feedback with a latency of several microseconds is
therefore not suitable for CLIC.}
CLIC instead proposes a drive beam ``phase feedforward'' (PFF)
system.
A prototype PFF system has been designed, commissioned and operated at
the CLIC test facility CTF3, at CERN, to prove its feasibility. The prototype
system follows the same concept as the CLIC scheme, and is the focus
of this work. CTF3 provides a 135~MeV electron beam bunched at 3~GHz with a
pulse length of
1.2~\(\mathrm{\mu s}\) and a pulse repetition rate of 0.8~Hz [REF]. All phases
quoted in the paper are given in degrees at 12~GHz, as relevant for CLIC.
%\section{\label{s:ctfLayout}System Design}
\begin{figure*}
\includegraphics[width=\textwidth]{figs/ctfpffLayout}% Here is how to
%import EPS art
\caption{\label{fig:pffLayout}Schematic of the PFF prototype at CTF3,
showing the approximate location of the phase monitors (\(\phi_1\) ,
\(\phi_2\) and \(\phi_3\)) and
the kickers (K1 and K2). The black box “PFF” represents the calculation
and output of the correction, including the phase monitor
electronics, feedforward controller and kicker amplifiers. A bunch
arriving early at \(\phi_1\) is directed on to a longer path in the TL2
chicane
using the kickers (blue trajectory), whereas a bunch arriving late will
be directed on to a shorter path (red trajectory). }
\end{figure*}
A schematic of the prototype PFF system is shown in Fig.~\ref{fig:pffLayout}.
The system corrects the phase using two electromagnetic kickers installed
before the first and last dipole in a four bend, dog-leg shaped chicane. The
beam's path length through the chicane depends on the magnitude and polarity of
the voltage applied to the kickers. The phase is measured using a monitor
upstream of the chicane, and then corrected by setting the kicker voltage to
deflect bunches arriving early at the phase monitor on to longer trajectories
in the chicane, and bunches arriving late on to shorter trajectories.
Downstream of the chicane another phase monitor is placed to measure the
effects of the correction.
The beam time of flight between the upstream phase monitor and the first kicker
in the chicane is 380~ns. By bypassing the \textcolor{red}{combiner ring (CR)
and TL1 transfer
line} (see Fig.~\ref{fig:pffLayout}) the total cable length required to
transport signals between the monitor and kickers is shorter, approximately
250~ns. The PFF correction in the chicane can therefore be applied to the same
bunch initially measured at the phase monitor, providing the total system
hardware latency is less than 130~ns.
%\textcolor{red}{Some details on CLIC system differences...One PFF system will
%be installed
%in each decelerator sector (24 PFF
% systems per drive beam)...C-shaped chicane instead...}
%\textcolor{red}{CLIC hardware, bandwidth requirements and how they compare to
%CTF3. The PFF system poses many challenges, particularly in terms of the
%hardware
% bandwidth (\textgreater17.5~MHz \cite{Gerber2015}), power (500~kW
% amplifiers)
% and latency requirements.}
The PFF system presents a significant hardware challenge, in particular in
terms of the power and bandwidth requirements for the kicker amplifiers, and
the resolution of the phase monitors. As well as the kickers themselves, the
system utilises a low latency digitiser and feedforward controller. New
components have been designed and built for the prototype at CTF3, and a
summary of their parameters in comparison to the CLIC requirements is shown in
\textcolor{red}{Table~x}.
\textcolor{red}{Table:
No. Systems
No. Kickers
Phase Monitor Resolution
Amplifier Power
Angular Deflection @ Kicker
Correction Bandwidth}
\textcolor{red}{The main differences between the CTF3 and CLIC systems:
48 systems at CLIC (one per decelerator sector) - synchronisation between
systems not addressed at CTF3.
C-shaped correction chicane with 16 kickers
CLIC drive beam much higher energy, so much higher power requirements}
\textcolor{red}{Correction bandwidth requirement of 17.5~MHz derived from
\cite{Gerber2015}...
Higher frequency errors damped by combination process and accelerating
cavities...}
%\subsection{\label{ss:hardware}Hardware}
%The PFF system uses three phase monitors, two electromagnetic kickers, kicker
amplifiers and a digitiser/feedforward controller.
The three phase monitors used at CTF3 \cite{phMonEuCard} are designed and built
by INFN
Frascati, with the
associated electronics built by CERN. The monitors are 12~GHz resonating
cavities with a dipole and monopole mode present. The output from opposing
vertical pairs of feedthroughs are summed in hybrids to create a position
independent signal. This signal is split and mixed with a reference 12~GHz
signal in eight separate mixers. The output from the eight mixers is combined,
allowing a resolution of \(0.12^\circ\) to be
achieved whilst maintaining linearity between \(\pm70^\circ\) [REF]. The quoted
resolution
is determined by
comparing the measurements of the two adjacent upstream monitors (installed in
the CT
line, see Fig.~\ref{fig:pffLayout}). \textcolor{red}{More details: cylindrical
cavities so big with notch filters, mixed with 3GHz reference multiplied to
12GHz etc.}
The two electromagnetic stripline kickers \cite{kickerIPAC11} were also
designed and built by INFN
Frascati,
and are based on the DAFNE design [REF]. Each kicker is approximately 1~m in
length,
with a horizontal strip separation of 40~mm. A voltage of 1.26~kV applied to
the
downstream end of the kicker strips yields a horizontal deflection of 1~mrad
for the 135~MeV CTF3 beam. \textcolor{red}{more details - tapering? Move after
amplifier and state actual max deflection?}
The kicker amplifiers \cite{RobertsThesis} have been designed and built by the
John Adams
Institue/Oxford University. The 20~kW amplifiers consist of low voltage Si FETs
driving high voltage SiC FETs, and for
an input voltage of \(\pm2\)~V give an output of up to \(\pm700\)~V. The
amplifier response is linear within
3\% for input voltages between \(\pm1.2\)~V, then starts to saturate. The
output
has a bandwidth of 47~MHz for small signal variations up to 20\% max output,
and is slew rate limited for larger variations.
Finally, the Feedforward digitiser and controller (FONT5a board)
\cite{RobertsThesis} was also
designed and built by John Adams Institute/Oxford University. This digitises
the
processed phase monitor signals and then calculates and outputs the appropriate
voltage with which to drive the amplifier in order to correct the phase. The
board consists of a Virtex-5 field programmable gate array (FPGA), nine 14-bit
analogue to digital converters (ADCs) clocked at 357~MHz, and four digital to
analogue converters (DACs). The parameters of the correction, such as the
system timing and gain, are controlled on the board via a LabVIEW data
acquisition and control system. \textcolor{red}{mention board latency? ref a
font paper/ip feedback?}
The combined hardware latency for the PFF system is approximately 100~ns. The
output from the FONT5a board is delayed by 30~ns so that the drive voltage from
the amplifiers and the beam arrive at the kickers synchronously.
\subsection{\label{ss:optics}Chicane Optics}
The PFF system places additional constraints on the optics of the correction
chicane, and also on the beam lines between the upstream phase monitor and the
chicane. These ensure the maximum possible phase shift per volt applied to the
kickers without degrading the transverse beam orbit and beam size after the
chicane, and high correlation between the initial (uncorrected) upstream and
downstream phase.
The correction range of the PFF system is defined by the kicker design, the
maximum output voltage of the kicker amplifiers, and the optics transfer matrix
coefficient \(R_{52}\) between the kickers in the chicane, which relates the
change in path length through the chicane per unit
deflection at the first kicker.
For the maximum amplifier output of \(\pm700\)~V the kickers deflect the beam
by \(\pm0.56\)~mrad. Together with \(R_{52} = 0.74\)~m in the chicane optics
these define the system correction range of approximately
\(\pm400~\mathrm{\mu m}\), or \(\pm6^\circ\).
The measured phase shift in the chicane versus the amplifier input voltage is
shown in Fig.~\ref{fig:corrRange}, and agrees with the expected range.
However, the response of the amplifier and therefore the phase shift is
non-linear. The correction algorithm assumes linearity, but this has a
negligible effect compared to the limitations placed by the
upstream-downstream phase correlation and phase monitor resolution.
%\textcolor{red}{MADX units for R52, R56, i.e. conversion between distance and
%phase.}
\begin{figure}
\includegraphics[width=\columnwidth]{figs/corrRange}
\caption{\label{fig:corrRange}Downstream phase vs. the kicker amplifier
input voltage. Standard errors on the measured phase are shown.}
\end{figure}
The PFF system also should not change the beam orbit after the chicane. The
chicane optics are designed so that the second kicker closes the orbit bump
created by the first kicker.
Fig.~\ref{fig:orbClos} shows the horizontal beam orbit in the region
of the chicane for the maximum and minimum kick. The closure in
the BPMs following the chicane is better than 0.1~mm, compared to a maximum
offset of 1.5~mm inside the chicane.
\begin{figure}
\includegraphics[width=\columnwidth]{figs/orbClos}
\caption{\label{fig:orbClos}Horizontal orbit in and around the TL2 chicane
at maximum (blue) and
minimum (red)
input to the kicker amplifiers. Markers show the measured position in beam
position monitors, and dashed lines the predicted orbit using the
CTF3 MADX model and hardware parameters.}
\end{figure}
%\textcolor{red}{All this must be achieved whilst keeping dispersion low,
%matching betas etc.
%within constraints of pre-existing buildings. Achieved R52 0.74m with max
%dispersion 1.16m...}
\subsection{\label{ss:r56} Phase Correlation}
\textcolor{red}{The PFF system acts to subtract the measured upstream phase
(\(\phi_u\)) from
the initial downstream phase (\(\phi_d\)) with a gain factor (\(g\)):
%\begin{equation*}
\(\phi_{\mathrm{PFF}} = \phi_d - g\phi_u\)
%\end{equation*}
, where \(\phi_{\mathrm{PFF}}\) is the corrected downstream phase. The optimal
system gain is given by:
\(g = \rho_{ud} \sigma_d/\sigma_u\)
, where \(\sigma_u\) and \(\sigma_d\) are the initial upstream and downstream
phase jitter respectively, and \(\rho_{ud}\) is the correlation between the
upstream and downstream phase. The theoretical limit on the corrected
downstream phase jitter (\(\sigma_{\mathrm{PFF}}\)) with this gain is given by:
\(\sigma_{\mathrm{PFF}}=\sigma_d \sqrt{1-\rho_{ud}^2}\).}
One of the key challenges in operating the PFF prototype at CTF3 has been
obtaining high correlation between the initial, uncorrected, upstream and
downstream phase. A correlation of 97\% is required to reduce a typical initial
phase jitter of \(0.8^\circ\) at CTF3 to the target of \(0.2^\circ\).
The achievable correlation depends on the phase monitor resolution and any
additional phase jitter introduced in the beam lines between the upstream and
downstream phase monitors. The phase monitor resolution of \(0.12^\circ\)
limits the maximum upstream-downstream phase correlation to
\(98\%\) in typical conditions, and places a theoretical limit of
\(0.17^\circ\) on the \textcolor{red}{measured corrected downstream phase
jitter.}
Any beam jitter that changes the time of flight of bunches influences the
resulting downstream phase stability and upstream-downstream phase correlation.
The dominant source of uncorrelated downstream phase jitter at CTF3 is beam
energy jitter being transformed in to phase jitter in the transfer lines
between the upstream and downstream phase monitors.
\begin{figure}
\includegraphics[width=\columnwidth]{figs/r56Scan}% Here is how to import
%EPS art
\caption{\label{fig:r56Scan}Downstream (red points) and upstream (blue
points) phase jitter vs. the \(R_{56}\) value in the set TL1 optics.
}
\end{figure}
The first order phase-energy dependence can be described via the optics
transfer matrix coefficient \(R_{56}\):
\(\phi_d = \phi_u + R_{56}(\Delta p / p)\)
, where \(\Delta p / p\) is the relative beam energy offset.
Optimal conditions for the PFF system are obtained when the total \(R_{56}\)
between the upstream and downstream monitors is zero.
The \(R_{56}\) value in one of the transfer lines at CTF3, TL1, has been tuned
in order to achieve this, compensating for \(R_{56}\) terms in other beam lines.
Fig.~\ref{fig:r56Scan} shows that with an \(R_{56}\) of around 10~cm in TL1 the
downstream phase jitter is reduced to the same level as the upstream jitter.
The upstream-downstream phase correlation is also increased to above 95\%.
However, a large second order phase-energy dependence was also identified and
this remains uncorrected. This leads to a degradation in upstream-downstream
phase correlation if there are drifts in beam energy. Energy drifts resulting
from klystron trips and RF power drifts at CTF3 have made it difficult to
maintain high phase correlations for timescales longer than 10~minutes as a
result.
\section{\label{s:results}Results}
\subsection{\label{ss:gScan}Gain Scan}
With the optimal gain the PFF correction acts to remove all correlation between
the upstream and downstream phase, reducing the downstream phase jitter. If the
gain is too small some residual correlation will remain, and if it is too large
the correlation will flip sign.
The optimal system gain can be derived empirically by observing the dependence
of the downstream phase on the upstream phase with the correction on, as seen
in Fig.~\ref{fig:gScan}. Optimal gain values for the system are typically in
the range 1.0--1.5, being larger than unity when the downstream jitter is
larger than upstream, as per the predicted theoretical values.
\begin{figure}
\includegraphics[width=\columnwidth]{figs/gScan}% Here is how to import EPS art
\caption{\label{fig:gScan}Downstream phase jitter with the PFF system on at
different gains. Markers show the measured phase jitter with standard error
bars. The shaded red region shows the expected performance given the initial
jitter and correlation with the PFF system off.}
\end{figure}
\subsection{\label{ss:shape}Intra-Pulse Phase Variations}
The PFF correction is shaped to remove phase variations along the
1.2~\(\mathrm{\mu s}\) CTF3 beam pulse. The predominant intra-pulse feature at
CTF3
is a roughly parabolic ``phase sag'' of \(40^\circ\) peak-to-peak, resulting
from the use of RF pulse compression. As this is much larger than the
\(\pm 6^\circ\) range of the PFF system, only approximately a 400~ns portion of
the pulse can be optimally corrected. The phase sag would not be present at
CLIC, where in any case the drive beam pulse length is less than 400~ns.
%2015: (Peak-to-peak variation of 5.76 degrees in initial phase reduced to
%0.65 degrees in corrected phase -- OR -- standard deviation of phases reduced
%from 1.68 to 0.26 degrees...).
%2016: Std \(0.960\pm0.003^\circ\) reduced to \(0.285\pm0.004^\circ\) across
%440~ns portion of pulse. Worse absolute but better removal small features
%compared to 2015.
\begin{figure}
\includegraphics[width=\columnwidth]{figs/shape}% Here is how to import EPS
%art
\caption{\label{fig:shape}Effect of the PFF system on intra-pulse phase
variations. The pulse shape upstream (green), and downstream with the
PFF
system off (blue) and on (red) is shown.}
\end{figure}
\begin{figure}
\includegraphics[width=\columnwidth]{figs/flatness}% Here is how to import
%EPS art
\caption{\label{fig:flatness}Distribution of downstream rms phase values,
here
referred to as the
``pulse flatness'', for each beam pulse with the PFF system off (blue)
and
on (red).}
\end{figure}
Fig.~\ref{fig:shape} shows the effect of the PFF system on the intra-pulse
phase variations. \textcolor{red}{The convention at CTF3 is to operate the PFF
system in
interleaved mode, with
the correction applied to alternating pulses only. This allows a measurement of
the initial (`PFF Off') and corrected (`PFF On') downstream phase to be
performed concurrently.} The upstream (PFF input) phase is also shown for
comparison. Vertical dashed lines mark a 440~ns portion of the pulse where the
correction is optimal, and this range is used to calculate statistics on the
effect of the system.
In this range the PFF system flattens the phase,
and almost all variations are removed. Residual offsets in the phase are still
present where there are small uncorrelated differences between the shape of the
initial upstream and downstream phase. Fig.~\ref{fig:flatness} shows the rms
phase variation within the 440~ns range
for each beam pulse in the dataset, with the PFF system on and off. The PFF off
pulses have an rms of \(0.960\pm0.003^\circ\) on average, and this is reduced
to \(0.285\pm0.004^\circ\) by the PFF system.
The PFF system at CTF3 has been verified to reduce the amplitude of
phase errors up to a frequency of 25~MHz, exceeding the CLIC requirements.
%Limited by variations in phase propagation along the pulse (energy differences
%etc.).
\subsection{\label{ss:meanJit}Pulse-to-pulse Jitter}
As well as removing intra-pulse phase variations the PFF system simultaneously
corrects offsets in the overall mean phase, i.e. any pulse-to-pulse jitter. The
mean phase of each beam pulse is calculated across the 440~ns range in the
central portion of the pulse, as shown before in Fig.~\ref{fig:shape}.
\begin{figure}
\includegraphics[width=\columnwidth]{figs/meanJit}% Here is how to import
%EPS
%art
\caption{\label{fig:meanJit}Distribution of the mean downstream phase with
the
PFF system off (blue) and on (red).}
\end{figure}
Fig.~\ref{fig:meanJit} shows the effect of the PFF system on the pulse-to-pulse
stability across a dataset around ten minutes in length. An
initial mean downstream phase jitter of \(0.92\pm0.04^\circ\) is reduced to \(0.20\pm0.01^\circ\) by the PFF
correction. All correlation between the upstream and downstream jitter is removed, from
\(96\pm2\%\) to \(0\pm7\%\). The achieved stability is consistent with the theoretical prediction (considering the initial correlation and jitter) of \(0.26\pm0.06^\circ\) within error bars.
%\textcolor{red}{NB: upstream PFF off is \(0.76\pm0.03^\circ\), but
%\(0.68\pm0.03^\circ\) for
%PFF on. Helps to explain why achieved is better than predicted.}
This level of stability could not be maintained for longer periods due to
CTF3's drifting RF sources, eventually leading to degraded
upstream-downstream phase correlation and phase drifts outside the PFF
correction range, as previously mentioned. \(0.30^\circ\) phase jitter has been
achieved in 20~minute datasets. With suitable feedbacks to keep the phase
within the correction range, and a reduction of the higher order phase-energy
dependences in the machine optics, the PFF system could achieve CLIC-level
phase stability continuously.
\textcolor{red}{The PFF system has also been operated
whilst intentionally varying the incoming mean phase, as shown in
Fig.~\ref{fig:wiggle}. The PFF system removes the additional phase variations
and achieves more than a factor 5 reduction in downstream phase jitter, from
\(1.71\pm0.07^\circ\) to \(0.32\pm0.01^\circ\) in this case.}
\begin{figure}
\includegraphics[width=\columnwidth]{figs/wiggle}
\caption{\label{fig:wiggle}Mean downstream phase with the PFF system off
(blue) and on (red) vs. time, with additional phase variations added to
the
incoming phase.}
\end{figure}
%\subsection{\label{ss:pbpJit}Point-by-point Jitter}
%
%\textcolor{red}{I would remove this. Don't think it adds any
%information beyond mean jitter and correction of shape.}
%
%\begin{figure}
%\includegraphics[width=\columnwidth]{figs/BestFF_pbp}% Here is how to import
%%%%EPS art
%\caption{\label{fig:BestFF_pbp}Point-by-point jitter.}
%\end{figure}
%
%Point-by-point jitter of x~degrees achieved across a x~ns portion of the
%pulse, agrees with simulated value...
\section{\label{s:conc}Conclusions}
CLIC requires a PFF system to reduce the drive beam phase jitter by an order of
magnitude, from \(2.0^\circ\) to \(0.2^\circ\)~at~12~GHz, or better than 50~fs
stability. A prototype of the system has been
in operation at the CLIC test facility CTF3, and corrects the beam phase by
varying the path length through a chicane using two electromagnetic kickers.
As
well as the kickers, the system uses newly designed phase monitors with
\(0.12^\circ\) resolution, high bandwidth 20~kW amplifiers and a low latency
digitiser/feedforward controller. The system latency, including hardware and
signal transit times, is less than the 380~ns beam time of flight between the
input phase monitor and the correction chicane. Therefore, the feedforward
correction can
be directly applied to the same bunch initially measured at the monitor.
New optics for the correction chicane and other beam lines at CTF3 have been
developed to yield the desired phase shifting behaviour and ensure high
correlation between the initial upstream and downstream phase.
The prototype system has demonstrated \(0.20\pm0.01^\circ\) pulse-to-pulse
phase jitter on a time scale of ten minutes. It has also been shown to be able
to flatten intra-pulse phase variations up to a frequency of 25~MHz. On longer
timescales the performance of the system is limited by changes to the incoming
beam conditions, in particular beam energy, which would be better controlled in
any future application at CLIC.
%Drifts, in particular in beam energy, degrade the correlation between the
%upstream and downstream phase and prevent this level of stability from being
%demonstrated on longer time scales at CTF3. A key consideration for any future
%system should be to design beam lines and optics with zero phase-energy
%dependence, including non-linear dependencies, to solve this issue.
%\textcolor{red}{Try to apply to XFELs/something else.}
\section{\label{s:ack}Acknowledgements}
\begin{acknowledgments}
We wish to acknowledge everyone involved in the operation of CTF3 for their
help and support in realising the PFF system.
\end{acknowledgments}
\bibliography{pff_short}% Produces the bibliography via BibTeX.
\begin{comment}
\newpage
\pagebreak
\section{Some Figures}
\pagebreak
\begin{figure*}[h]
\includegraphics[width=\textwidth]{figs/ctfpffLayout}% Here is how to
%import EPS art
\caption{}
\end{figure*}
\pagebreak
\begin{figure*}[h]
\includegraphics[width=\textwidth]{figs/alt/ctfpffLayout_alt}
\caption{}
\end{figure*}
\pagebreak
\begin{figure*}[h]
\includegraphics[width=\textwidth]{figs/gScan}% Here is how to import EPS
%art
\caption{}
\end{figure*}
\pagebreak
\begin{figure*}[h]
\includegraphics[width=\textwidth]{figs/alt/gScan_old}
\caption{}
\end{figure*}
\pagebreak
\begin{figure*}[h]
\includegraphics[width=\textwidth]{figs/alt/gScanFull}
\caption{}
\end{figure*}
\pagebreak
\begin{figure*}[h]
\includegraphics[width=\textwidth]{figs/alt/stdMeanPhase}
\caption{}
\end{figure*}
\pagebreak
\begin{figure*}[h]
\includegraphics[width=\textwidth]{figs/alt/FFTUpDown_20161215_2340}
\caption{}
\end{figure*}
\pagebreak
\begin{figure*}[h]
\includegraphics[width=\textwidth]{figs/alt/FFTOnOff_20161215_2340}
\caption{}
\end{figure*}
\end{comment}
\end{document}