Transcript RF Limitations - Indico

Collected by E. Jensen, BE-RF
ATOP days 4.-6. 3. 2009
RF limitations
1
Introduction
 RF limitations:
 Voltage necessary to create a certain accelerating
bucket area – this becomes often a power limit
 Beam loading (BL) will increase the power needs by (at
least) the power transferred to the beam (good for
higher efficiency)
 But BL also leads to an induced voltage that interacts
with the various beam and cavity control loops, which
may become unstable.
 High intensity beams will get closer to a number of
stability limits (or instability thresholds). This is RF
related since the system dealing with these are often
RF systems.
ATOP days 4.-6. 3. 2009
RF limitations
2
ATOP days 4.-6. 3. 2009
RF limitations
3
Linac 2
 Linac 2 can produce 180 mA (190 mA in
TRA10 were obtained in MD’s in 1994)
 It produces regularly 150 – 170 mA for all
requested intensity ranges.
 Rien à signaler …
M. Vretenar
ATOP days 4.-6. 3. 2009
RF limitations
4
ATOP days 4.-6. 3. 2009
RF limitations
5
Name
harmonic h
f range [MHz]
peak voltage
remarks:
C02
1
0.6 … 1.8
8 kV
acceleration
C04
2
1.2 … 3.9
8 kV
acceleration, bunch shaping
C16
8, 9
6 … 17
6 kV
controlled long. BU
+ transverse damper
•
The PSB regularly produces very high intensity beams
for Isolde, (3.5 · 1013) not limited by the RF systems.
Krusche, Paoluzzi: http://cern.ch/AccelConf/e98/PAPERS/TUP03H.PDF
ATOP days 4.-6. 3. 2009
RF limitations
6
PSB RF limitations at high intensity
 C04 power limitations (for faster cycle)
 C04: “High Loss Mode” (not intensity related)
 Instability at low C04 voltage and high BL
 “Ring 4 problem” (transverse plane)
 C02 beam loading
ATOP days 4.-6. 3. 2009
RF limitations
7
PSB: Power limitations
• During the tests for 900 ms cycling (2005),
power limits were observed in the
C04 system.
• It was concluded that – for reliable operation
with 900 ms – the C04 system would require
a fundamental upgrade (water-cooling). For
even shorter cycles, this would become
mandatory.
Haase, Paoluzzi: http://cdsweb.cern.ch/record/877814/files/ab-note-2005-037.pdf
ATOP days 4.-6. 3. 2009
RF limitations
8
PSB C04: “High Loss Mode” Limit
(for completeness – not intensity related)
 The C04 were constructed from the old C08
system – the ferrites are not optimal.
 When too long
, the
ferrites go into “High Loss Mode”.
 This figures shows
the time for the
losses to increase
by 20%.
Paoluzzi: https://edms.cern.ch/file/593255/1/RFN2000-013.pdf
ATOP days 4.-6. 3. 2009
RF limitations
9
PSB: Instability at low C04 voltage and high BL
In MD’s in 2007 and 2008, a stability threshold was observed when increasing the
beam-loading (BL) by lowering the C04 voltage. This leads to phase and amplitude
coupling and a complex interplay between C02 and C04 control loops.
Note in preparation: Blas, Findlay, Haase, Paoluzzi, Pedersen
ATOP days 4.-6. 3. 2009
RF limitations
10
Strong detuning by BL
locus of cavity admittance
Y
IG/V
seen by tuning loop: φL
I0/V
cavity tuning angle: φZ
IB/V
If IB /I0 becomes large, small perturbations of IB will have large effects.
ATOP days 4.-6. 3. 2009
RF limitations
11
“Ring 4 problem”
 Only in outer rings, mainly ring 4, at high intensity on the flat top,
one observes:

shaving of trailing edge of the bunch,
 large transverse intra-bunch excursions,
 often with synchro-loop transients and during extraction bump (5 ms
before extraction),
 beam loss.
:

Influence of transverse feedback settings checked,
 voltages and loops behaviour checked and excluded
 A likely explanation:

There exists a transverse impedance, particularly high in ring 4, which
the transverse damper can cope with only marginally.
 The additional power demand due to the perturbations (synchro-loop,
ejection bump) leads to saturation of the transverse damper amplifier.
Blas, Chanel, Findlay, Hanke, Mikulec, Quatraro, Rumolo
ATOP days 4.-6. 3. 2009
RF limitations
12
PSB: Beam loading
In the recent “Review on PS Booster with Linac4”, A. Blas investigated
the feasibility of
. He found that
 The current limitation of the C02 system would lead
to a maximum ΦS of 18°, whereas 22° would be
needed for the present cycle. This would indicate an
intensity limit of
.
 Recent results with the new digital beam control
tested in ring 4 are very encouraging. The upgrade to
D-LLRF is essential to eventually get closer to this
intensity limit ( see M.E. Angoletta’s APC presentation 27/2/09).
 Also the beam stability with low C04 voltage/high BL
should profit from this upgrade.
http://indico.cern.ch/materialDisplay.py?contribId=8&materialId=slides&confId=46255
ATOP days 4.-6. 3. 2009
RF limitations
13
ATOP days 4.-6. 3. 2009
RF limitations
14
PS RF systems
Name
harmonic h
count
C10
7, 10, …, 21
10+1
C20
28, 42
1+1
C40
84
1+1
C80
168
2+1
C200
420, 433
4+2
frequency range peak voltage
remarks:
[MHz]:
2.7 … 10.01
1 … 20 kV
acceleration, RF
“gymnastics”
13 or 20
15 kV
LHC 75 ns & 50 ns
bunching
40
3 … 350 kV
LHC bunching, bunch
compression
80
350 kV
LHC bunch
compression
200
30 kV
re-bunching,
controlled long. BU
+ transverse damper (2 x 6 kW)
ATOP days 4.-6. 3. 2009
RF limitations
15
PS RF limitations at high intensity
 Phase drift at low voltages due to beam loading
 Over-current at γtr crossing with high BL
 Beam losses at γtr (partly RF related)
 Insufficient transient beam-loading
compensation for an asymmetrically filled
machine
 Coupled bunch instabilities above γtr. (For LHC
25 and 50 ns beams). Ok up to nominal.
 Bunch lengthening due to residual impedance of
40 and 80 MHz cavities.
ATOP days 4.-6. 3. 2009
RF limitations
16
Phase drift at low voltages due to BL
voltage
voltage
radial position
phase
phase
radial position
beam
beam
When reducing the voltage, the relative BL gets stronger. It leads to the slow phase
drift, which the AC coupled phase loop cannot correct. An intensity dependent phase
drift is observed, which in turn acts also on the radial position.
H. Damerau
ATOP days 4.-6. 3. 2009
RF limitations
17
Over-current at γtr crossing with high BL
 A phase switch is necessary at γtr .
 In the presence of strong BL, this fast phase
change requires large power, resulting in a
significant, but short increase in anode
current, which made the tube protection trip.
 In 2008, the tube protection circuit was
modified to cope (increased lag time from
500 µs to a few ms) – should be OK now.
C. Rossi
ATOP days 4.-6. 3. 2009
RF limitations
18
Beam losses at γtr (RF related ?)
Observed: a large excursion
of MRP when crossing γtr
(and switching RF phase),
too fast for the radial loop
to correct.
MRP vs. C-timing
Possible explanation: The
PU’s see not only the mean
radial position, but also the
orbit distortions.
An additional PU has
recently been installed in
SS76 –
S. Aumon, S. Giladroni, J. Belleman
ATOP days 4.-6. 3. 2009
RF limitations
19
PS: Coupled bunch instability
 Observed on LHC type beams.
 At present, two C10 cavities are
transition
beam peak detected
used as longitudinal kicker in the
CB feedback; this works OK up
to nominal intensity.
 Most probable source: the
impedance of the C10 cavities
themselves.
Beam PU signal around 5th
harmonic, down-converted
H. Damerau et al. : http://cdsweb.cern.ch/record/1055555/files/ab-2007-051.pdf
ATOP days 4.-6. 3. 2009
RF limitations
20
PS: Transverse damper
 System dimensioned for injection error damping
(6 kW per plane) – OK during the cycle.
 Somewhat counter-intuitively (for me), higher
brilliance (same intensity with smaller ε) will slow
down incoherent betatron motion and thus leave
more time for the transverse damper before
filamentation smears out the transverse beam.
A. Blas, G. Rumolo, E. Benedetto
ATOP days 4.-6. 3. 2009
RF limitations
21
PS: Residual impedance of 80 MHz cavities
 Bunch length (4 σ Gaussian fit) vs. bunch # using 2 (blue)
and 3 (red) 80 MHz cavities.
 Bunches near the end of the batch are longer, due to
residual impedance of 40/80 MHz systems.
 This effect is more pronounced with all 3 cavities.
 Bunches # 50+ are shorter due to coupled-bunch feedback.
Damerau: http://cdsweb.cern.ch/record/1141522/files/AB-Note-2008-052-MD.pdf
ATOP days 4.-6. 3. 2009
RF limitations
22
ATOP days 4.-6. 3. 2009
RF limitations
23
SPS 200 MHz – 4 TW cavities
Each line feeds one cavity
ATOP days 4.-6. 3. 2009
RF limitations
24
The lines of the TWC200 system
…
1 line sketched
Total voltage (2 cavities): 4.1 MV w/o BL,
≈ 3.3 MV with nominal BL (1.15 1011 ppb).
With new tubes: 350 kW/TX
…
1 line sketched
Total voltage (2 cavities): 5.5 MV w/o BL,
≈ 4.2 MV with nominal BL (1.15 1011 ppb).
With new tubes: 400 kW/TX
ATOP days 4.-6. 3. 2009
RF limitations
25
SPS 800 MHz
System is essential at high intensity
to cope with the dominant coupled
bunch instability!
“Valvo” klystrons (YK1198) are very
old. These klystrons have not been
built any more for decades! Of 16
existing tubes, 10 are broken, 6 are
operational.
Equally the transformers in their
power supplies are at the end of
their lifetime. The cavities are OK.
We managed “just” (with a lot of
personal commitment of some
people) to supply the necessary 800
MHz voltage (700 kV) thru 2008.
ATOP days 4.-6. 3. 2009
RF limitations
26
Transverse Damper SPS
• 4 x 25 kW RF
• System runs
stably
• RAS
ATOP days 4.-6. 3. 2009
RF limitations
27
SPS Limitations
 Voltage and power OK up to ultimate LHC
intensity for nominal cycle (7.5 s). For faster
cycles, see next slide for maximum current!
requires 800 MHz
system, which is presently running marginally!
 Power 200 MHz:
 coaxial lines limited to 750 kW
 ageing tetrodes (more wear with CNGS operation!)
 power couplers: ceramics upgraded, but still limited
(transition coupler – cavity).
 Losses at flat bottom at high intensity –
presently not well understood (e-cloud?).
E. Chapochnikova
ATOP days 4.-6. 3. 2009
RF limitations
28
SPS TWC200:
Accelerating voltage in one 5-section cavity
Vacc/MV
nominal LHC
(1.15 1011 ppb, 25 ns)
2.5
2.0
1.5
0.2
0.4
0.6
0.8
1.0
Ib/A (200 MHz component)
Extrapolated from: G. Dôme: “The SPS Acceleration System”, CERN-SPS/ARF/77-11
ATOP days 4.-6. 3. 2009
RF limitations
29
Interventions – let the statistic speak!
# of interventions
per week outside
normal working hours
mentioned in D. Manglunki’s talk
CNGS
Run 2002
Run 2003
Run 2004
Run 2005
Run 2006
Run 2007
Run 2008
weeks with piquet service
•
•
•
From 2005, limited resources forced us to reduce some maintenance work.
CNGS type beams result in more wear and thus reduce tube-lifetime (16! tubes
broken in 2008!)
This is more related to maintenance and high power than high intensity
E. Montesinos
ATOP days 4.-6. 3. 2009
RF limitations
30
SPS 800 MHz system upgrade programme
• The upgrade to more modern
IOT is in full swing (white
paper).
• IOT’s are used widely for digital
TV transmitters (DVB-T).
• At present in the “Market
Survey” state.
• Modular: Each line will be
composed of 4 identical PA
“cubicles”
• A cubicle can produce 60 kW, 4
cubicles will make up one
transmitter, a total of 2
transmitters is needed.
• Present planning: 1 (test) cubicle
in 2009, 4 cubicles in 2011,
completion in 2013.
ATOP days 4.-6. 3. 2009
RF limitations
31
Summary “
”
 All systems OK up to nominal LHC.
 Main Issues:





PSB C04 with strong beam loading at low voltage
PSB transverse damper installed power marginal
PS C10 with strong beam loading at low voltage
PS around transition – under investigation
PS coupled bunch instability and residual
impedances
 SPS 200 MHz: frequent interventions, requires
regular maintenance
 SPS 800 MHz – renovation underway (IOT based)
ATOP days 4.-6. 3. 2009
RF limitations
32