No abstract

CERN, the European Laboratory for Particle Physics, has started construction of the Large Hadron Collider (LHC), a superconducting accelerator that will collide protons at a center of mass energy of 14 TeV from the year 2005 onwards. The kicker magnet pulse generators of the LHC beam extraction system require fast high power switches. One possible type is the pseudospark switch (PSS) which has several advantages for this application. A PSS fulfilling most of the requirements has been developed in the past years. Two outstanding problems, prefiring at high operating voltages and sudden current interruptions (quenching) at low voltage could be solved recently. Prefiring can be avoided for this special application by conditioning the switch at two times the nominal voltage after each power pulse. Quenching can be suppressed by choosing an appropriate electrode geometry and by mixing Krypton to the D 2 gas atmosphere. One remaining problem, related to the required large dynamic voltage range (1.7 kV to 30 kV) is under active investigation: steps in forward voltage during conduction, occurring at low operation voltage at irregular time instants and causing a pulse to pulse jitter of the peak current.This paper presents results of electrical measurements concerning prefiring and quenching and explains how these problems have been solved. Furthermore the plans to cure the forward voltage step problem will be discussed. CERN, Geneva, SwitzerlandAbstract CERN, the European Laboratory for Particle Physics, has started construction of the Large Hadron Collider (LHC), a superconducting accelerator that will collide protons at a center of mass energy of 14 TeV from the year 2005 onwards. The kicker magnet pulse generators of the LHC beam extraction system require fast high power switches. One possible type is the pseudospark switch (PSS) which has several advantages for this application. A PSS fulfilling most of the requirements has been developed in the past years. Two outstanding problems, prefiring at high operating voltages and sudden current interruptions (quenching) at low voltage could be solved recently. Prefiring can be avoided for this special application by conditioning the switch at two times the nominal voltage after each power pulse. Quenching can be suppressed by choosing an appropriate electrode geometry and by mixing Krypton to the D 2 gas atmosphere. One remaining problem, related to the required large dynamic voltage range (1.7 kV to 30 kV) is under active investigation: steps in forward voltage during conduction, occurring at low operation voltage at irregular time instants and causing a pulse to pulse jitter of the peak current.This paper presents results of electrical measurements concerning prefiring and quenching and explains how these problems have been solved. Furthermore the plans to cure the forward voltage step problem will be discussed.

be discussed.The incorporation of these results into a 2-gap beam dump switch, soon to be assembled, will differences in the quenching behaviour between pure and thoriated tungsten have been found. materials, pure tungsten, thoriated tungsten and tungsten/rhenium (5% Re). Important deuterium gas reservoir. The erosion behaviour has been investigated for different electrode reported. Two of the prototypes have been mounted in a sealed-off version, equipped with a Experimental results on several 1-gap prototypes of different electrode diameter will be employing a ferroelectric trigger in each of the gaps. operation voltage (~ 15). It is intended to achieve these goals with a 2-gap or 3-gap version dumping system, a very low pretire rate (~ 10*4) and a large ratio of maximum to minimum development with the aim of obtaining a switch tailored to two main requirements of the LHC may guarantee a quasi unlimited operation time. Pseudo-spark assemblies are therefore under for this kind of application, as they do not suffer from cathode-heater lifetime limitations and rate (only a few pulses per day). Cold cathode tubes like pseudo-spark switches look suitable Collider (LHC) require fast high power switches (35 kV, 30 kA, 8 us) of very low repetition

No abstract

Linac4 is an H − ion linear accelerator at CERN, intended to replace Linac2 as injector to the PS Booster (PSB). In order to distribute the 160 MeV beam from Linac4 to the four rings of the PSB, new distributor magnets (BI.DIS) and magnetic septa (BI.SMV) need to be built. The proposed distribution scheme requires the construction of five new pulse generators of Pulse Forming Network (PFN) type, to double the magnitude of the BI.DIS field pulse and to increase the maximum duration of the pulse from 140 μs to 620 μs.A prototype PFN featuring 60 cells assembled in a Guillemin type E network was built. A PFN impedance of 6.25 Ω was chosen in order to achieve a magnetic field rise time of 1 μs (1 % -99 %) and a current amplitude of 1 kA in the lumped inductance magnet with a flat top ripple of ± 1%. The nominal operating voltage of 6.25 kV (maximum of 10 kV) and the repetition rate of 1.1 Hz allow operation in air. A switch comprising two commercial IGBTs connected in series will be used to initiate the transfer of the energy from the PFN to the magnet. This document presents the design of the new PFN and the measurements performed with switches from four different suppliers. IGBT rise time improvements, synchronization problems and ideas for timing drift stabilization are discussed.The general layout of the foreseen control system is shown.