Transverse impedance measurements of the DARHT-2 accelerator cell

Abstract

A new technique was developed to measure the transverse interaction impedance of the DARHT-2 induction cells, designed to accelerate a 2-4 kA, 2 µ sec electron beam pulse. An extensive campaign to minimize the transverse resistive impedance of these massive metglas-filled structures lead to a design with thin ferrite tiles in the form of a ring placed along one wall of the radial line connecting the pulseline feeds to the accelerator gap. This ferrite ring heavily damped all transverse modes. To measure the broadband m=1 impedance, one cell is excited by a matched twinlead on the axis of the beamtube. The ratio of the radial RF magnetic field in the accelerator gap to the azimuthal RF magnetic field at the beamtube wall (far away from the gap) determines the complex impedance, as explained in the paper. This technique is much simpler than the TSD approach for these very large bore systems. It also covers the full frequency band including the reactive impedance at low frequency related to the "image displacement instability". Results with the 10 inch diameter bore "standard cell" indicate a peak resistive impedance of 280 ohms per meter around 575 MHz (Q = 6).

Long-pulse beam stability experiments on the DARHT-II linear induction accelerator

“…The least-squares fit to the data for this shot corresponds to transverse impedance . This is slightly less than the 193 measured on a single cell at the 168-MHz resonance peak [12], [16], but greater than the 157-average for the 20-MHz band that we analyzed. By using as the independent variable for this analysis, we were able to combine data from several shots having different currents and tunes.…”

“…12 corresponds to a transverse impedance of 184 . Again, this is slightly less than the 193-single-cell 168-MHz resonance peak measurement [12], [16], but greater than the 157-average for the 20-MHz analysis band. The value of corresponding to a nominal tune of the final 17-MeV accelerator configuration is about 200 A/G, and is shown on Fig.…”

“…A sinusoidal excitation at the resonant frequency is predicted to grow to this maximum amplification and then saturate in time due to resistive energy losses in the cells. The transverse impedance measured for a single accelerator cell is a continuous function of frequency above about 100 MHz, and has resonances at 168, 236, and 572 MHz [16]. At the accelerator exit of the final 17-MeV configuration with a nominal tune we predict at 168 MHz.…”

BBU and corkscrew growth predictions for the DARHT second axis accelerator

“…A quite extensive experimental and computational campaign was conducted to lower both the transverse resistive impedance Z ⊥ and the cavity Q in order to limit BBU growth [2]. The resulting measurements were then used as input to the LLNL BREAKUP code, together with the lattice B z and γ(z) profiles to compute the predicted BBU response.…”

“…Let R be the averaged centroid radial displacement over the flattop portion of the current pulse, and A be the averaged corkscrew amplitude. For the first tuning scheme, the tuning-V steering algorithm is used to minimize a figure of merit M, where M = (W*R) 2 + A 2 , and W is a weighting factor, at every BPM with W = 0.1. For the second tuning scheme, only 2 BPMs are used.…”