E. C. Raka

Increased Intensity Performance of the Brookhaven AGS

With the advent of H- injection into the Brookhaven AGS, circulating beams of up to 3 × 1013 protons at 200 MeV have been obtained. Rf capture of 2.2 × 1013 and acceleration of 1.73 × 1013 up to the transition energy (¿ 8 GeV) and 1.64 × 1013 to full energy (¿ 29 GeV) has been achieved. This represents a 50% increase over the best performance obtained with H+ injection. The increase in circulation beam current is obtained without filling the horizontal aperture. This allows the rf capture process to utilize a larger longitudinal phase space area (¿ 1 eV sec/bunch vs ¿ 0.6 eV sec with H+ operation). The resulting reduction in relative longitudinal density partially offsets the increase in space charge effects at higher currents. In order to make the capture process independent of injected beam current, a dynamic beam loading compensation loop was installed on the AGS rf system. This is the only addition to the synchrotron itself that was required to reach the new intensity records. A discussion of injection, the rf capture process, and space charge effects is presented.

High Intensity Beam Instabilities in the Brookhaven AGS

Transverse resistive wall instabilities and longitudinal reactive wall instabilities have both been observed in the Brookhaven AGS. Recently an instability has been observed characterized by individual coherent phase oscillations of the twelve beam bunches. This instability is believed to be caused by high frequency resonance modes in the rf cavities introduced by recent modifications. A simple analysis shows that voltages induced in these parasitic resonances by the beam can couple the phase oscillations of the individual bunches. A normal mode analysis of the system predicts unstable normal modes with properties similar to the observations. Modifications to the rf cavities to suppress these undesirable effects have been devised.

Damping Bunch Shape Oscillations in the Brookhaven AGS

Bunch shape oscillations are present in the AGS both before and after the transition energy (?8 BeV) is reached. At transition, they can be excited by timing, phase and radial misadjustments, and beyond by rapid large radial position changes used for targeting or by rapid large (> 20%) targeting losses. They can also be excited before transition by misadjusting the low level RF control system. In addition, excitation before transition can occur when 2fs = 720 Hz, the magnet ripple frequency, or when 2fs = 360 Hz, the ripple frequency on the RF cavity tuning servo systems. An analysis of the RF phase control systems including second order effectsl indicates that it should damp these oscillations but at a rate~fs/s? except in the vicinity of transition. At 200 ms, E ? 6.5 BeV, ? ? 6.9, fs 200 Hz, and the damping rate is ? 0.46 s-1. Thus adequate inherent damping is present only early in the acceleration cycle. This, plus the fact that the sensitivity of the phase control electronics to bunch shape variations can at times produce antidamping, has led to the design and testing of various feedback damping systems to control these oscillations. An analysis of RF phase and amplitude feedback, using a peak detected signal, as applied to the RF control system, will be given. Operational results using both types of feedback will be described.

AGS Lattice Corrections and Tuning Using Backleg Windings

The excitation of a novel backleg winding system has significantly contributed to increasing the intensity of the AGS. The benefits of the improved magnetic lattice properties at fields in the vicinity of injection were demonstrated, both for future operations and for quantitative study of high intensity phenomena. A record intensity of 3.2 × 1012 protons per pulse was attained when some portions of these windings were activated. This occured approximately two years ago, too late for the previous Conference. The underlying ideas and computations preceded this date by about two years. While many turn backleg windings already existed on the AGS magnets, the appropriate systems connections had to be constructed and installed during a shutdown. It was hoped by now to have experimental studies using all the systems, but time commitments have prevented further work to date. However the ideas used and the particularly simple approach to orbit analysis developed warrants being described.

A Measurement of the Longitudinal Coupling Impedance in the Brookhaven Ags

The imaginary part of the longitudinal coupling impedance has been measured as a function of energy from 5 to ~ 28 GeV. This impedance is proportional to ¿f = (fq - 2fd) where fd is the coherent dipole frequency and fq the coherent quadrupole frequency. These frequencies are obtained by stimulating coupled bunch oscillations. If the dominant impedance is due to inductive wall plus space charge effects, then one has (Z/n) = j[¿oL - goZo/2ß¿2]1 where L is the inductance per turn and ¿o = 2¿fo the particle rotation frequency. The expression (Z/n) = 4j¿f¿2hVocos¿sB3/ 3Iofd can be used to find the impedance if the synchrotrop phase space distribution is proportional to (1 - r2)¿. Io is the current per bunch, B = fo × Tl the bunch length and Vo is the external voltage. For a distribution given by (1 - r2) the right hand side should be multiplied by 27/4¿2. If the latter is assumed, an inductive impedance of 20.4 ¿ is obtained with a null at ¿ GeV (¿tr = 8.5) for a transverse emittance of 22 ¿¿ rad-m. At 5 GeV the reactance is negative but larger than the simple relation assumed for (Z/n) would predict. If the bunches are parabolic, then the inductive impedance would be 29.7 ¿ with a null again at 6.6 GeV but only for an emittance of 2.5 ¿rad-m. Again the .5 GeV reactance is much too large. The significance of these results is discussed.

Beam Transfer from the AGS to ISABELLE

The ISABELLE design current is built up by repetitive transfer of charge from the AGS. To do this, a momentum stacking method similar to the one used at the ISR has been chosen. The AGS beam bunches are synchronously transferred from the AGS into waiting rf buckets on the injection orbit at the ISA, and then slowly accelerated into the previously established debunched beam stack. This process is repeated until all the available momentum aperture in the ISA is filled up, whereupon the beam stack is rebunched and accelerated to the desired operating energy.

Measurement of the Linear Coupling in the Brookhaven AGS

The magnitude and sign of the zeroth harmonic skew quadrupole component of the magnetic field at 28.5 BeV are determined by exciting the normal mode frequencies in part of the debunched beam present during a flat top extraction cycle. Simple rf excitation of the (9-Q) mode is employed. Filtered difference signals from pick-up electrodes are used to measure the frequencies and relative phases of the H and V oscillations. During acceleration when the beam is bunched it is kicked horizontally and the radial position adjusted until the coupled vertical motion in the (9-Q) mode reaches a maximum. Correction quadrupoles are then powered to minimize the observed amplitude. The magnitude of the coupling roughly tracks with the beam momentum. Saturation effects at high fields plus the powering of backleg bumps and tuning quadrupoles on the SEB flat top are possible sources of the somewhat larger coupling observed under these conditions.

Damping Coherent Oscillations in the AGS

In order to further study the vertical instability in the Brookhaven AGS a narrow band feedback damping system has been developed. The damping force is obtained from a pair of 6 ft long coils located in one of the straight sections. The damping signal is provided by a pair of pickup electrodes at an angle ? = 10.7° upstream from the coils. Since ?V ? 8.8, ?v ? ? ?/2. The damping is proportional to sin [n? - (n-?) 2? foT] where n is the mode number, fo the rotation frequency and T the time delay between the pickup and damping elements. The system is designed to work over a range in fo from 200 kc to 370 kc. In this range, with T = .14 ?sec, and for n = 7-11, the angle in brackets varies ± 19° from ?/2 if ? changes about ± 1%. With a simple filter in the feedback loop it is possible to provide damping for the lower order modes n = 9,8,10, without exciting those for n = 7,11 etc. under normal operating conditions.

Calculations of Capture Efficiency of the Debunched Stack in ISABELLE

A computer program which simulates particle behavior in the longitudinal phase space has been developed to investigate the rebunching process and capture efficiency of the debunched stack in ISABELLE. A simple expression for a capture efficiency for an arbitrary initial distribution has been derived and used to estimate capture efficiencies for several assumed distributions.

Injection into the ISA

Three modes of injection into the ISA storage accelerators are discussed. The three are: 1) Energy stacking in the ISA. 2) Acceleration of 12 bunches in the AGS followed by single bunch transfer to the ISA. Application of a moving bucket technique then allows the transferred bunch to be brought closer to the bucket train circulating in the ISA. 3) Acceleration on the first harmonic in the AGS. The single bunch is then transferred directly to the ISA into a matched bucket.