Mark A. Kemp

ILC Marx modulator development program status

A Marx-topology klystron modulator is under development for the International Linear Collider (ILC) project [1]. It is envisioned as a lower cost, smaller footprint, and higher reliability alternative to the present, bouncer-topology, baseline design. The application requires 120 kV (+/−0.5%), 140 A, 1.6 ms pulses at a rate of 5 Hz. The Marx constructs the high voltage pulse by combining, in series, a number of lower voltage cells. The Marx employs solid state elements; IGBTs and diodes, to control the charge, discharge and isolation of the cells. Active compensation of the output is used to achieve the voltage regulation while minimizing the stored energy. The developmental testing of a first generation prototype, P1, has been completed. This modulator has been integrated into a test stand with a 10 MW L-band klystron, where each is undergoing life testing. Development of a second generation prototype, P2, is underway. The P2 is based the P1 topology but incorporates an alternative cell configuration to increase redundancy and improve availability. Status updates for both prototypes are presented.

IGBT PEBB technology for future high energy physics machine operation applications

Terascale physics is driving the demand for innovative pulsed power modulators having greater compactness and better manufacturability with increasingly superior performance. A particularly promising route for such modulators is Marx-architecture based. Moreover, there is opportunity for improvement and gain of greater benefits through further development of topology and architecture, gate driver method, and control schemes. Prior work discussed a new concept of droop correction, which was the result of topology hybridisation using a nesting approach, and illustrated its great potential. This is further investigated here. This paper details various design aspects of a hybrid Marx cell Power Electronic Building Block (PEBB) and includes specifics about estimated losses and efficiency, thermal management issues, protection strategies, gate driver development, and control implementation. In addition, figures-of-merit of the cell design are given for comparison and evaluation purposes. Experimental results, based on both single-cell and three-cell hardware prototypes, are presented demonstrating the functionality and performance of the new topology. This is a significant milestone in the progression toward constructing a full 32-cell PEBB-based Marx klystron modulator with nested droop correction. Lessons learned during various stages of the prototype development and future directions are commented on.

P1-Marx modulator for the ILC

A first generation prototype, P1, Marx-topology klystron modulator has been developed at the SLAC National Accelerator Laboratory for the International Linear Collider (ILC) project[1]. It is envisioned as a lower cost, smaller footprint, and higher reliability alternative to the present, bouncer-topology, baseline design. The application requires 120 kV (+/−0.5%), 140 A, 1.6 ms pulses at a rate of 5 Hz. The Marx constructs the high voltage pulse by combining, in series, a number of lower voltage cells. The Marx employs solid state elements; IGBTs and diodes, to control the charge, discharge and isolation of the cells. Active compensation of the output is used to achieve the voltage regulation while minimizing the stored energy. The P1-Marx has been integrated into a test stand with a 10 MW L-band klystron, where each is undergoing life testing. A review of the P1-Marx design and its operational history in the L-band test stand are presented.

Final design of the SLAC P2 Marx klystron modulator

The SLAC P2 Marx has been under development for two years, and follows on the P1 Marx as an alternative to the baseline klystron modulator for the International Linear Collider. The P2 Marx utilizes a redundant architecture, air-insulation, a control system with abundant diagnostic access, and a novel nested droop correction scheme. This paper is an overview of the design of this modulator.

A prognostic method for scheduling maintenance on the P2-Marx modulator

The SLAC National Accelerator Laboratory is developing a second generation Marx-type modulator for the ILC, the P2-Marx. The modulator is expected to operate reliably in excess of 105 hours with minimum downtime. A prognostic system is being implemented with the development of the P2-Marx to monitor and track the health of key high voltage components. This paper discusses the way in which the prognostic system will be implemented and used to monitor the health of the P2-Marx modulator.

Lifetime tests on a high ohms/square metalized high crystalline polypropylene film capacitor with application to a Marx modulator

This paper presents accelerated lifetime tests on a polypropylene film capacitor. Experimental parameters (20% droop, 5 Hz repetition rate) simulate anticipated operating conditions encountered in the SLAC P2 Marx. Elevated film electric field stress is utilized as the acceleration parameter. Results indicate that, for the particular film of interest, a film stress of ∼290 V/μm corresponds to a 105 hour lifetime. In addition, the voltage scaling exponent for this film is 13.1.

Design considerations for a PEBB-based Marx-topology ILC klystron modulator

The concept of Power Electronic Building Blocks (PEBBs) has its origin in the U.S. Navy during the last decade of the past century. As compared to a more conventional or classical design approach, a PEBB-oriented design approach combines various potential advantages such as increased modularity, high availability, and simplified serviceability. This relatively new design paradigm for power conversion has progressively matured since then and its underlying philosophy has been clearly and successfully demonstrated in a number of real-world applications. Therefore, this approach has been adopted here to design a Marx-topology modulator for an International Linear Collider (ILC) environment where easy serviceability and high availability are crucial. This paper describes various aspects relating to the design of a 32-cell Marx-topology ILC klystron modulator. The concept of nested droop correction is introduced and illustrated. Several design considerations including cosmic ray withstand, power cycling capability, fault tolerance, etc., are discussed. Details of the design of a Marx cell PEBB are included.

A Self-Biasing Pulsed Depressed Collector

Depressed collectors have been utilized successfully for many years to improve the electrical efficiency of vacuum electron devices. Increasingly, pulsed, high-peak power accelerator applications are placing a premium on electrical efficiency. As RF systems are responsible for a large percentage of the overall energy usage at accelerator laboratories, methods to improve upon the state-of-the-art in pulsed high-power sources are desired. This paper presents a technique for self-biasing the stages in a multistage depressed collector. With this technique, the energy lost during the rise and fall times of the pulse can be recovered, separate power supplies are not needed, and existing modulators can be retrofitted. Calculations show that significant cost savings can be realized with the implementation of this device in high-power systems. In this paper, the technique is described along with experimental demonstration.

The SLAC P2 Marx

A proposed high energy physics accelerator, the International Linear Collider, will require greater than five hundred rf stations. Each station is composed of a klystron driven by a modulator. Recently, the SLAC P2 Marx was designated the baseline modulator for the ILC. This paper describes some key features of this modulator and presents recent experimental results.

Status update on the second-generation ILC Marx modulator prototype

This paper is a status update of the SLAC P2 Marx. This Marx-topology klystron modulator is a second-generation modulator which builds upon experience gained from the SLAC P1 Marx. There are several fundamental differences between these modulators including the correction scheme, bus voltages, and the control system architecture. These differences, along with preliminary experimental results and the schedule for further development, are detailed in this paper.