Multi-Bunch for LCLS

Report
Multi-Bunch Operation for LCLS
Franz-Josef Decker
March 17, 2010
1.
Definitions and goals
multi-bunch within a pulse (vs different pulses [beam code])
a few bunches (vs >1000 bunches)
2.
History of SLAC’s multi-bunch running
phase manipulation
high charge
low charge
3.
Multi-bunch scenarios for LCLS
phase and amplitude manipulation for Gun, L0, L1, L2, L3
4.
Transverse consideration: kickers …
Multi-Bunch Operation for LCLS
1
1
Franz-Josef Decker
[email protected]
Definitions and Goals
Multi-bunch:
few (2, 3, …) to many bunches (1400) within a pulse
SLC: 3 bunches, 60 ns apart
E-158: >1000 bunches every RF period, 6.5E11 particles (>100 nC)
[100pC/bunch]
E-155: >1400 bunches, 2.9E9 particles (effectively zero) (<500 pC)
Different pulses on different beam codes
e.g.: 60 Hz here and 60 Hz there separated by pulsed magnets, quads, …
or: 59.5 Hz to experiment and 0.5 Hz to BYKICK dump
Goals for LCLS (-I and/or -II):
Mainly a few bunches up to 8 (420ns/60ns +1) for different
undulators after separation by kicker magnets
Shorter separation possible if RF deflection is used (TCAV 4, 5, 6, …)
More bunches for one experiment (seeding, high power, target hit rate improved?)
Two bunches 105 mm apart + delay one (like sail-boat chicane for FACET)
Multi-Bunch Operation for LCLS
2
2
Franz-Josef Decker
[email protected]
RF Amplitude and Phase Change
SLC 1990: Energy spread
for e-scav was not compensated:
RF phase was changed fast (30 ns)
by 90 in compressor cavity to achieve
about 6 integrated phase change in 60 ns:
Multi-Bunch Operation for LCLS
3
3
Franz-Josef Decker
[email protected]
RF Phase Manipulation
SLC 1990: 6.8 in 60 ns in NRTL
LCLS 2009: 4.0 difference between
over- and under-compression in L2
with SLED: 1/2 at input (1/3 at output)
and 800ns fill time (600 ns NRTL):
we need about 100 ns for 4 change
Multi-Bunch Operation for LCLS
4
4
Franz-Josef Decker
[email protected]
High power (600 kW) E-158 Beam
Energy of the different SLAC beams
Conclusion:
Transverse betatron lattices can be
designed for beams with different
energies like a factor of 2 in sector 4
or 1.8 in sector 10.
The E-158 beam (blue, __) is at 45 (48) GeV at the end of the
linac (3000 m) and has the highest energy along the linac.
The scavenger beam (magenta, --) ends at 25.5 GeV, the
PEPII beams for filling HER (red, -.) and LER (green, --) have
an end-energy of 9.0 and 3.11 GeV.
Multi-Bunch Operation for LCLS
5
5
Franz-Josef Decker
[email protected]
Synchrotron Light on Gated Camera (60 ns)
<6.5E11 particles
=5m
E
16%
Multi-Bunch Operation for LCLS
6
6
Franz-Josef Decker
[email protected]
Start of Pulse like Electron in BPM, Tail like Positron
Charge jitter caused energy jitter
Linac phase offset reduced energy jitter 
Multi-Bunch Operation for LCLS
7
7
Franz-Josef Decker
[email protected]
Transverse jitter of beam tail
 Transverse (dispersive) jitter at end of pulse
could be reduced from 300 m (shown) to
25 m rms by launching oscillations early in
the linac
Jittery setup
Jitter of tail reduced, but “bowed” trajectory?
Multi-Bunch Operation for LCLS
8
8
Franz-Josef Decker
[email protected]
RF Tricks for Low Current (E-155)
Instead of one 180 switch PSK to make a SLED RF pulse use three, or five:
RF energy gain
RF SLED pulse with notch
»
Multi-Bunch Operation for LCLS
9
9
Franz-Josef Decker
[email protected]
Double Notch in SLED Pulse
0.5% energy variation  0.1%
Multi-Bunch Operation for LCLS
10 10
Franz-Josef Decker
[email protected]
Multi-Bunch Scenarios for LCLS
Five different longitudinal regions for LCLS (same parameter):
RF-Gun: effectively standing wave:
L0: traveling wave section no SLED
L1: traveling wave section with SLED
(lower RF ampl. to keep peak, more power)
(keep RF flat at low charge)
(we have Ampl/Ph control for E and chirp)
L2: traveling wave section with old SLED (needs Ampl/Ph control, PSK for few bunches)
L3: traveling wave section with old SLED (needs Ampl control, PSK mostly o.k)
TCAV3 is an indication what is possible with EPICS RF system:
 L1S
 TCAV3
PSK
A/P change
180
to compensate
temperature,
[300 ns gives
90, TCAV3
separates
multi bunches]
Multi-Bunch Operation for LCLS
100=1000ns
11
11
Franz-Josef Decker
[email protected]
Energy Change after BC2
If different energies are
desired: L3 can give
E = -5 GeV in 350 ns.
(needs big notch for all
SBSTs)
Multi-Bunch Operation for LCLS
12 12
Franz-Josef Decker
[email protected]
Transverse Effects
We saw mismatches with up to twice in beam size (Li01: 30% energy difference),
so local compensation will be necessary for LCLS at least before BC2.
Separating the bunches to different beam lines (stability?)
a) By kickers and septum/Lambertson magnets
b) RF separators (TCAV)
Fast quadrupoles (kicker type) might not be necessary (Lattice for 2 * E is o.k.)
…
Multi-Bunch Operation for LCLS
13 13
Franz-Josef Decker
[email protected]
Diagnostics and Controls
BPM at 140 MHz: 2856/140 = 20.4 times 5 give 102 buckets or 35.7 ns,
at this bunch delay the BPMs will see the sum of two bunches,
at about 400 ns EPICS might be able to separate two signals.
BLM ?
Wires: 60 ns gate was used during SLC
Screens and cameras: 60 ns gated camera (maybe faster these days)
Feedbacks
Laser “stacks” now already two pulse, just delay one.
More ideas: “just” ask Bill White
…
Multi-Bunch Operation for LCLS
14 14
Franz-Josef Decker
[email protected]
Summary
Multi bunch operation is possible with the SLAC Linac even for LCLS
Rule of thumb: RF can change: 10% in 80 ns (Tf = 800 ns), with SLED half
Energy difference of up to 30% have been equalized to 0.1 % historically.
Transverse: One lattice can accommodate different energies
Kickers and multiple undulators are necessary for additional beam lines
Two bunches 105 mm apart and delay the first one in chicane
eliminates timing jitter difference from two linacs
Multi-Bunch Operation for LCLS
15 15
Franz-Josef Decker
[email protected]

similar documents