franck - Indico
Download
Report
Transcript franck - Indico
Modulator and overmoded RF components
status for the 12GHz Test Stand at CERN
RF meeting – 17/03/2010
F. Peauger – 17/03/10
The High Voltage Modulator System
The modulator is a high voltage pulsed power supply fabricated by Scandinova
(Sweden). It delivers high voltage pulses to the XL5 klystron with maximum amplitude of
450 kV – 335 A and a pulse length of 1.5 µs (2.3 µs FWHM) at 50 Hz
Klystron
Switch units
HV-DCPS
Oil Tank
The modulator includes the main HV-DCPS, the oil tank with a special pulse transformer made by
Stangenes, some solid state switches, the three power supplies for the magnet, one DC power supply
for klystron filament heating, five ion pump power supplies and a command/control system. It includes
also a hydraulic network, a 380 V - 50 Hz 3-phase line electrical distribution network and some racks
containing all the elements
F. Peauger – 17/03/10
2
Two key elements in the Scandinova technology
•Pulse transformer
The particularity is the use of multiple magnetic cores on
the primary side. The transformer is connected to an energy
storage capacitor. To avoid the droop when discharging the
capacitor, RL tuning circuits are mounted near the
transformer. There is also a core bias transformer. The ion
pump at the klystron gun is connected to an external DC power
supply via a separate winding.
•Solid state switch
Switch Unit including 6 IGBT’s
F. Peauger – 17/03/10
Pulse transformer
Each transformer core is driven by low voltage solid
states switches called IGBT operating at 1kV-15A
and connected in parallel . The IGBTs come Infineon –
U-Pack. They are specified to 1.7kV-32A but operate
with margin to get more reliability. There are 72
connections on the primary side of the transformer. A
Swith Unit (SU) includes 6 IGBT. In case of a default
of one IGBT, the system can continue to work at a
lower HV. Each switch can dissipate up to 25 kW at
200 Hz and are cooled by a water circuit.
3
Modulator procurement follow-up
Three reviews in 2009
Visit at PSI Villigen in Dec 09:
Review of problems met on similar S band Scandinova modulator
Meeting at Scandinova in Jan 10:
X band PSI modulator shown
Status of the mechanical feature of elements presented
interfaces discussed (SLAC klystron, water cooling, command control…)
Technical report in Feb 10
mechanical simulation of klystron lifting device
Schedule:
-
Factory Acceptance Test
Delivery at CERN
Installation
Start up/System testing
Final Acceptance Test at CERN
=>
=>
=>
=>
=>
W12
W13
W14-W15
W16-W17
W18
=> FAT cancelled
“ Assembly
yet finished due to late
lastnotweek
shipping of the DC Power Supply
(should be delivered next week)”
“ Arcing occurred in the tank during
the testing of the PSI modulator… oil
insulation under investigation”
Weekly conference call with Scandinova proposed
F. Peauger – 17/03/10
4
RF components development strategy
RF design by CEA
HFSS
Scaling of existing SLAC components or new design based on
SLAC experience and publications
Mechanical design and drawings by CERN
Tolerance of 20 µm
Internal surface roughness of Ra=0.1
Precise machining of elementary pieces by CEA
Done in french firms
Cleaning, coating, brazing and re-machining by CERN
Vacuum brazing at high temp (>780°C)
Hydrogen treatment to be investigated
F. Peauger – 17/03/10
90° mode converter
The mode converter is made in two parts. The first part
is a rectangular waveguide bend on H plane. It converts
the TE10□ mode into a TE20□ mode. The second part is a
circular waveguide with two posts positioned at 180 ° at
a certain distance of the rectangular section. This design
is based on an original idea of S. Kazakov. It is compact
and relatively easy to fabricate
Reflection
F=36.27mm
Bandwidth of 150 MHz @ -20dB reflection
and -0.0618 dB transmission at 11.994 GHz
giving 98.6% conversion efficiency in power
Esurfmax=52.5MV/m
with 150MW input
power at 11.9942GHz
F. Peauger – 17/03/10
Hsurfmax=162 kA/m
with 150MW input
power at
11.9942GHz giving
DT=5.8° with 250 ns
pulse length
6
New « bayonet » mode converter
This second kind of mode converter is also composed
of different independent sectors.
The first part is a rectangular waveguide on H plane
with an enlarged section and shifted transversally
compared to the WR90 flange axis.
The small side of the waveguide is increased with a
smooth taper and the TE10□ mode is transformed into
a TE20□ mode.
The second part is a regular transition between the
rectangular section and an elliptical section first, and
then from the elliptical profile to a circular one to
obtain a TE01o mode in waveguide diameter of 36.27
mm.
The reflection coefficients S11 (mode n°3) and S22
(mode n°1) are respectively -39.5 dB and -40.8 dB
and the transmission coefficient S12 from the TE01o
mode to the TE10□ mode is -0.00708 dB (99.8%
transmission in power) which is very satisfactory.
Esurfmax=37.2 MV/m
with 150MW input
power at 11.9942GHz
TE10□
WR90
TE20□
Elliptical profile
TE01O
F=36.27mm
Half of the structure is represented
Total length = 175 mm
F. Peauger – 17/03/10
Mechanical design under progress
7
RF Valve
The RF valve has been introduced in the CTF3 30 GHz test stand. It works on the circular mode
TE01o mainly to avoid surface electric field and have steps in diameter to “focus” the wave in the
center of the guide. Based on the same principle, RF valves working at 11.4 GHz have also been
developed at SLAC for accelerating structure testing. The 12 GHz RF valve is a scaled version of
the SLAC one.
19 dB
Magnetic field on the valve
surface at 11.9942 GHz for 1 W
injected
F. Peauger – 17/03/10
8
36 mm/50 mm Circular Taper
A circular taper is needed to connect the existing circular waveguides (50 mm diameter) with the RF
valve and the mode converter calculated previously (with a diameter of 36.27 mm). It is composed
of different steps of diameters with various lengths optimized for matching on a large bandwidth. It
has been scaled from part of the 3 GHz power phase shifter.
Electric field at 11.9942 GHz for
1 W injected
The reflection is less than -30 dB between 11.82 and
12.18 GHz, which gives a bandwidth of 360 MHz
F. Peauger – 17/03/10
9
Future work proposed…
Double height 3dB - 90° hybrid power splitter:
- it equally divides the power into two arms while keeping a low RF
reflection to the source
- the pulse compressor contains a 3 dB hybrid splitter with constant
height of 10.16 mm
-> a double height splitter may reduce the risk of RF breakdown
- Also a second splitter would be needed in the CTF2 test structure area
in order to power two structures at the same time
90° circular waveguide bend:
- it bends efficiently the circular TE01o mode at 90°, which simplify the
RF layout
F. Peauger – 17/03/10
10
RF components status
Components
Status
90° mode converter
2 ordered + 2 to be launched
RF Valve
2 ordered
36/50 circular taper
4 ordered
Bayonet mode converter
RF design ready
Mechanical design in progress
90°circular waveguide bend
Under RF design
Double height hybrid
Study to be started
+ Collaboration with TRIESTE
F. Peauger – 17/03/10
11