FLASH X-RAY (FXR) LINEAR INDUCTION ACCELERATOR (LIA) OPTIMIZATION Sensor Delay Correction
The radiographic goal of the FXR Optimization Project is to generate an x-ray pulse with peak energy of 19 MeV, spot-size of 1.5 mm, a dose of 500 rad, and duration of 60 ns. The electrical objectives are to generate a 3 kA electron-beam and refine our 16 MV accelerator so that the voltage does not vary more than 1%-rms. In a multi-cell linear induction accelerator, like FXR, the timing of the acceleration pulses relative to the beam is critical. The pulses must be timed optimally so that a cell is at full voltage before the beam arrives and does not drop until the beam passes. In order to stay within the energy-variation budget, the synchronization between the cells and beam arrival must be controlled to a couple of nanoseconds. Therefore, temporal measurements must be accurate to a fraction of a nanosecond. FXR Optimization Project developed a one-giga-sample per second (gs/s) data acquisition system to record beam sensor data. Signal processing algorithms were written to determine cell timing with an uncertainty of a fraction of a nanosecond. However, the uncertainty in the sensor delay was still a few nanoseconds. This error had to be reduced if we are to improve the quality of the electron beam. Two types of sensors are used to align the cell voltage pulse against the beam current. The beam current is measured with resistive-wall sensors. The cell voltages are read with capacitive voltage monitors. Sensor delays can be traced to two mechanisms: (1) the sensors are not co-located at the beam and cell interaction points, and (2) the sensors have different length jumper cables and other components that connect them to the standard-length coaxial cables of the data acquisition system. Using the physical locations and dimensions of the sensor components, and the dielectric constant of the materials, delay times were computed. Relative to the cell voltage, the beam current was theoretically reporting late by 7.7 ns. Two experiments were performed to verify and refine the sensor delay correction. In the first experiment, the beam was allowed to drift through a cell that was not pulsed. The beam induces a potential into the cell that is read by the voltage monitor. Analysis of the data indicated that the beam sensor signal was likely 7.1 ns late. In the second experiment, the beam current is calculated from the injector diode voltage that is the sum of the cell voltages. A 7 ns correction produced a very good match between the signals from the two types of sensors. For simplicity, we selected a correction factor that advanced the current signals by 7 ns. This should reduce the uncertainty in the temporal measurements to less than 1 ns.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 899421
- Report Number(s):
- UCRL-TR-223183; TRN: US0702011
- Country of Publication:
- United States
- Language:
- English
Similar Records
"Real-Life" Pulse Flattening on the LLNL Flash X-ray (FXR) Machine
FXR LIA Optimization - Time-resolved OTR Emittance Measurement