1. RF systems for NSLS-II
J. Rose, A. Blednykh, P. Mortazavi and Nathan Towne
BROOKHAVEN SCIENCE ASSOCIATES

2. X-RAY Ring RF requirements
Beam Energy
3 GeV
Beam Current
500 mA
Energy loss/turn (Initial/Upgrade)
0.9 /2.0 MeV
Power to Beam
(Initial/upgrade)
0.45 / 1 MW
p/p acceptance 3%
Momentum
compaction
RF parameters
Frequency
500MHz
Cavity Voltage
3.3 / 4.9 MV
Bucket for 3.3MV

3. NSLS-II RF POWER REQUIREMENTS
Covered in baseline cost
Capability of installed RF
Installing 3rd cavity+300kW
Adding 4th cavity and transmitter
Power
# installed, power in kW
Dipole -,
-, 144
Damping wiggler 3, 4, 8,
Cryo-PMU
3, 38
6, 76 6,
10,
EPU 2,
4, 66 4, 5,
Additional devices ?,
TOTAL
409
545
803
1071
Available RF Power
540
810
1080

4. Cavity Choice High beam currents achieved by B-factories made
CESR-B, KEK-B and PEP-II cavities attractive
Complex SUPERFISH analysis of CESR-B of
HOM impedances used in preliminary CBI growth
rate estimates with ZAP: maximum growth time
of 65ms for 4 cavities, less than the damping time
of 8ms
GdfidL analysis of full 3-D cavity extends analysis
to dipole modes and 3-D effects (fluted beampipe)

5. Cavity Modeling with CFish, GdfidL
GdfidL vs. CFISH
Benchmark ferrite losses,
superconductor surface
resistance
GdfidL vs. measurements
on ferrite loaded pillbox confirm GdfidL model
Full Cavity HOM results used to analyze CB instability up to ~2GHz,
No problems yet but need to extend to beampipe cutoff frequency of 7GHz
Alexei Blednykh

6. CESR-B Cavity chosen for Baseline
Beam energy gain/cav
>2.4 MV
Eacc
>8 MV/m
Unloaded Q
>7108
Standby (static) losses
<30 W
Dynamic + static losses
<120W
Operating Temperature
4.5 K
Max. beam power/cavity
<250 kW
Frequency
500 MHz
SCRF chosen for lower R/Q, highly
damped HOM’s, lower operating
cost and comparable capital cost
Well established commercial
production. Units 15 and 16
now being produced by ACCEL.
In operations at Cornell (4), CLS(2),
Taiwan (2). Being commissioned at
Diamond (3)

7. KEK-B Cavity Parameters
Manufactured by Mitsubishi for KEK at 508 MHz
They have produced one cavity at 500MHz
Frequency (MHz)
500
Energy gain/cav (MeV)
2.5
E-Acc (MV/m)
>12
Unloaded 8MV/m
>109
BeamPower/coupler (kW)
>270
*KEK has recently demonstrated 400kW per coupler
KEK cavity is an option for NSLS-II
*Shinji Mitsunobu SRF2005 ThP52 High Power Test of Input Couplers and HOM dampers  for KEKB Superconducting Cavity

8. Passive Third Harmonic Landau Cavity
A harmonic bunch-lengthening cavity is required to increase
Touschek lifetime. Increases beam stability by increasing
energy-dependent tune spread
~1/3 of 500MHz voltage,
~1 MV, can be met with
one Super3HC cavity
Elletra/SLS Super 3HC cryo-module

9. Harmonic Cavity for Bunch Lengthening
1500MHz “Super-3HC” cavity
Voltage/cell
0.5 MV
Eacc
>5MV/m
Unloaded Q
>7108
Static losses
<50W
Dynamic + static losses
<100W
Operating T.
4.5 K
Frequency
1500 MHz
Work continues on effect of bunch train
transients on bunch lengthening (N. Towne)
required for 3%
Momentum acceptance:
1500MHz requires 3 cavities
N. Towne

10. NSLS-II RF Straight layout
Two 500 MHz cavities
+ one 1500 MHz passive
harmonic cavity fit in
one 8m straight: meets
initial power requirements
Second straight reserved for
third , fourth 500 MHz and
second 1500MHz cavities as
additional user insertion devices
increase RF power requirement
Klystrons located in
adjacent RF building to minimize loop delays in feedback systems

11. RF Power Sources for Ring System
Manufacturer/model
Frequency
MHz
Power kW
Efficiency %
CPI VPK-7957A
500
800 60
Thales TH2161B/2178
310/800 61
Toshiba E3774
180 53
Klystron can be sourced by multiple vendors

12. Low Level RF System Effect of RF jitter on the beam:
Amplitude modulation of the RF fields leads to momentum deviations of the beam, Likewise phase modulations translate into amplitude modulations again leading to momentum deviations. The momentum errors affect beam size and orbit jitter as follows:
The beam size in the center of the 5 m straight is given by
Since the dispersion is near zero (~1mm) and the natural energy spread is <0.001 the second term is negligible and the beam size becomes
σ x,y= 40μm, 2.4 μm
Orbit jitter is given as σx = σδ∙ηx , σy = σδ∙ηy
σx′ = σδ∙ηx′ , σy′ = σδ∙ηy′
This work is in progress, and preliminary tolerances of +/- 0.5% amplitude
and 0.5 degree phase have been adopted. These values have been achieved at
other 3rd generation light sources.

13. Modeling: Future work NSLS-II bunch length of
4.5mm excites modes up to
elliptical beam-pipe cutoff frequency
C48 Ferrite loss factor decreasing rapidly ε′
Courtesy M. deJong, CLS
μ′′
Losses declining μ′
ε′′
By combining different ferrite
tiles broad range of frequencies
can be damped
V. Shemelin
Impedance limit for 8ms damping
time, 4.5mm bunch length 40

14. Booster Ring Requirements
Full energy booster in same
tunnel 780m circumference
Energy loss /turn =493 keV
Vacc = 1MeV for 1% RF
acceptance
7nC/macropulse, 1 per min.
Beam current = ~3mA
Beam power 3kW
Delta-E/E
Phase in 500 MHz (radians)
1 “Petra” type cavity and (1) kW IOT

15. Conclusions
RF systems for third generation light sources are mature technology;
RF cavities and power sources available from industry
The short bunch length of 4.5mm in NSLS-II means operating the
CESR-B cavities in a new regime, C-48 ferrite may not meet
requirements at higher frequencies: This is being studied further, new
ferrite combinations appear to be direct substitutes and can be
incorporated in the preliminary design phase.
Work continues to define RF tolerances, however modern digital
RF control systems can achieve an order of magnitude improvement
over existing (APS, ESRF) systems mitigating any concerns