Non-intrusive Diagnosis of Individual Cell Frequencies in a Coupled Cavity Chain

Yi Ni

1997-01-01

Abstract:To research and manufacture a linear accelerator, it is necessary to tune the cavity carefully. In order to determine the tuning direction of each cell, the ordinary means is to insert a pair of perturbing probes into the cavity to measure each cell frequency cell by cell; therefore, it takes much time and work. It would be inconvenient to tune a long cavity or a non-uniform one. Moreover, if a sealed cavity becomes out of shape in the course of sealing or installing, its detuned state can not be diagnosed with perturbing objects. On the other hand, for a superconducting cavity, field flatness is strictly demanded, thus more careful tuning is required. But when the cavity is under operating conditions and in a bath of liquid helium, it is much more complicated to measure the cell frequencies with perturbing probes. With the development of computer science, numerical computation and intelligent instruments, an important task is to develop a method that can estimate each cell’s frequency without perturbing probes. In the early part of 1980’s, while testing a superconducting cavity in CERN, E.Hable and J.Tuckmantel developed a method based on a coupled resonant model that can calculate each cell’s frequency through measured dispersion frequencies of the cavity . Their cavity was a uniform single period one consisting of 5 cells. Because of superconducting, loss of each cell could be ignored. The varying range of each cell’s frequency was assumed small, so it was treated as a first order perturbing problem. J.Sekutowicz developed a method using measured dispersion frequency and its relevant field distribution to calculate each cell’s frequency. This method can reduce tuning work, but can not be used in a sealed cavity. In this paper, a general objective is discussed, which concerns a cavity that has a long or non-uniform chain, and loss of each cell is considered. The fundamental method is by means of the passband information obtained from an RF network analyzer and computer data acquisition system to calculate each cell’s frequency. A special calculating program is designed and has been well checked by theoretical values. Some experimental results are also given.