RF Cavity Development for FFAG Application on ERLP at Daresbury
E. Wooldridge, C, Beard, P. McIntosh, B. Todd, B. Fell, ASTeC, STFC Daresbury Laboratory, Warrington, UK
B. Spencer and R. Jones, Cockcroft Institute & University of Manchester, Manchester, UK
Abstract Cavity Optimisation Power Distribution
Funding for a non-scaling, Fixed Field Alternating Gr adient (FFAG) facility has been To minimise the power architecture for EMMA the shunt impedance, Rsh , must be as high as There are multiple sources available to power these cavities in the form of Inductive Output
approved for inst allation on the Energy Recovery Linac Prototype (ERLP) at Daresbury. The possible. The relationship between the shunt impedance, the cavity volt age, V, and the Tubes (IOTs) or klystrons.
RF syst em specification for this project requires the development of a high efficiency, 1.3 power, P, is given below .
GHz, norm al conducting accelerating structure, capable of delivering the required
accelerating volt age, whilst adhering to stringent space limit ations imposed by the extremely
compact nature of the FFAG ring. We have optimised a cavity design, providing the Rsh =
necessary acceleration and minimising the RF power requirement s to m atch with
commercially available power sources.
The ELBE and PEP II t ype cavities shapes were altered to m aximise the shunt impedance
below are before and after im ages are given below.
EMMA ELBE Type Cavity
30 kW (Pulsed)
20 kW (Pulsed)
30 kW (CW)
160 kW (CW)
These are peak power values; it is common for these supplies to be run at lower than peak
EMMA (Electron Machine for Many Applications) is a proof-of-principle non-scaling FFAG power where the effect s are linear. This is a drop of 1 to 2 dB for klystrons and 1 dB for
designed to accelerate electrons from ERLP. The electrons are extracted from ERLP after IOTs.
PEP II Type Cavity
the first pass through the m ain linac and injected into EMMA at an energy of 10 MeV. A
schem atic of EMMA is shown below , this im age also show s the first arc of ERLP and the The torus design with a gradient of 120 kV requires, from (1), 2.09 kW to be delivered to
extraction line from ERLP to EMMA. each cavity. When calculating the power requirement s for the ring losses in transmission
lines will need to be included. Assuming 2.09 kW per cavity the entire ring will require
around 40 kW of power; this is before losses in the distribution to the cavities which will
increase the power requirement s to 65 kW. For high gradient s such as the 180 kV
mentioned earlier the power requirement becomes 90 kW.
To distribute the power to the cavities there are two methods. The First is to cascade the
power from cavity to cavity, this is the preferred method if the cavities are to be by a klystron.
Using the trends from these two cavities the toroid cavity w as created. It is a simple torus
intersected by the beam pipe. The shape provides vast ly increased shunt impedance. To
confirm the result s the cavity w as modelled in both Microw ave Studio and HFSS.
An EMMA Cell is shown below . The ring consist s of 21 of these cells, identical except for the
injection and extraction cells where the cavity has been removed.
The second method is to split the power using w aveguide tees, this split s the power equally
into two further pieces of w aveguide. This is repeat ed until the correct power is reached.
This is the preferred solution if the cavities are to be powered by w aveguide.
The figures of merit for the cavity can be seen in the t able below . The values for final toroid
design given in the t able are for a cavity that is 10 mm longer than the other designs. Extra
space has been m ade available for the cavity due to a change of flange.
ELBE ELBE PEP II Toroid Final
Cavity Like Like Desig Toroid
(Baseline Design Desig n Design
Design) n (Cavit y
Longer) Summ ary
Shunt Impedance / M1 1.4 2.52 2.4 3.41 4.3 The RF requirement s for EMMA are challenging, and it s cavity has been optimised for high
Practical Shunt Impedance / M1 1.12 2.016 1.92 2.728 3.44 shunt impedance and hence minimising the cavity losses. The cavities will be powered by
commercially available klystrons or IOTs. There are two options to distribute the power to
Power Requirement s @ 120 kV / 6.43 3.57 3.75 2.64 2.09
the cavity by cascading the power through splitters from cavity to cavity or by dividing the
power through splitters until the required number of output s is achieved. The method used
Power Requirement s @ 180 kV / 14.46 8.04 8.44 5.94 4.71 will be dependant on the power source chosen.
Accelerator Science and Technology Centre www.astec.ac.uk