B.C. Brown

The Fermilab Main Injector: high intensity operation and beam loss control

From 2005 through 2012, the Fermilab Main Injector provided intense beams of 120 GeV protons to produce neutrino beams and antiprotons. Hardware improvements in conjunction with improved diagnostics allowed the system to reach sustained operation at ~400 kW beam power. Transmission was very high except for beam lost at or near the 8 GeV injection energy where 95% beam transmission results in about 1.5 kW of beam loss. By minimizing and localizing loss, residual radiation levels fell while beam power was doubled. Lost beam was directed to either the collimation system or to the beam abort. Critical apertures were increased while improved instrumentation allowed optimal use of available apertures. We will summarize the improvements required to achieve high intensity, the impact of various loss control tools and the status and trends in residual radiation in the Main Injector.

Stability tests of permanent magnets built with strontium ferrite

Permanent magnets built using strontium ferrite bricks have been tested for stability against demagnetization. Ten test dipoles were built to monitor ferrite behavior under a variety of stressing conditions, including irradiation, mechanical shock, extreme thermal excursions, and long term magnetization stability. The test magnets were geometrically similar to, but much shorter than, the magnets built for the 8 GeV transfer line at FNAL. No loss of magnetization was observed for bricks exposed to a proton beam, and a magnet exposed to several Gigarads of Co/sup 60/ gamma radiation suffered no measurable demagnetization. The magnet strength was observed to decrease logarithmically with time, consistent with the expected effect of thermal fluctuations. Irreversible demagnetization of /spl sim/0.1% was seen in cooling magnets to 0/spl deg/C, and the loss was /spl sim/0.2% for magnets cooled to /spl sim/20/spl deg/C. No additional demagnetization was seen on subsequent cycling to 0/spl deg/C. Finally, one of the long dipoles built for the 8 GeV line was periodically tested over the course of 3 months, and showed no measurable demagnetization.

Collimation system design for beam loss localization with slipstacking injection in the Fermilab main injector

Results of modeling with the 3-D STRUCT and MARS 15 codes of beam loss localization and related radiation effects are presented for the slipstacking injection to the Fermilab Main Injector. Simulations of proton beam loss are done using multi-turn tracking with realistic accelerator apertures, nonlinear fields in the accelerator magnets and time function of the RF manipulations to explain the results of beam loss measurements. The collimation system consists of one primary and four secondary collimators. It intercepts a beam power of 1.6 kW at a scraping rate of 5% of 5.5E+13 ppp, with a beam loss rate in the ring outside the collimation region of 1 W/m or less. Based on thorough energy deposition and radiation modeling, a corresponding collimator design was developed that satisfies all the radiation and engineering constraints.

Permanent gradient magnets for the 8 GeV transfer line at FNAL

The 8 GeV transfer line feeding protons into the new Fermilab Main Injector has been built using strontium ferrite permanent magnets. This article addresses the design and manufacture of the 67 combined function magnets; permanent horizontal and vertical bend dipoles and quadrupoles were also built. The combined function magnets were built with a mean integrated strength at midaperture of 0.56953 T-m (central field nominally 0.15 T), and a gradient of 3.23% per cm relative to the dipole strength (nominal gradient=0.48 T/m). Thermal compensation of these bricks was effected by use of a nickel-iron alloy. The magnets were thermally cycled from 20/spl deg/C to 0/spl deg/C to condition the ferrite against irreversible thermal losses; the compensation was measured with a flipcoil and verified with a rotating harmonics coil. We present details of the magnet assembly process and also summarize the magnetic measurements.

Hybrid permanent quadrupoles for the 8 GeV transfer line at Fermilab

Hybrid permanent magnet quadrupoles for specialized portions of the 8 GeV transfer line from the Fermilab Booster to the new Main Injector have been built, tested and installed. These magnets use a 0.635 m long iron shell and provide an integrated gradient of 1.48 T-m/m with an iron pole tip radius of 0.0416 m. and pole length of 0.508 m. Bricks of 0.0254 m thick strontium ferrite supply the flux to the back of the pole to produce the desired 2.91 T/m gradient. For temperature compensation, Ni-Fe alloy strips are interspersed between ferrite bricks to subtract flux in a temperature dependent fashion. Adjustments of the permeance of each pole using iron between the pole and the flux return shell permits the matching of pole potentials. Magnetic potentials of the poles are measured with a Rogowski coil and adjusted to the desired value to achieve the prescribed strength and field uniformity. After these adjustments, the magnets are measured using a rotating coil to determine the integral gradient and the harmonics. These measurements are used to calibrate the production Rogowski coil measurements. Similar quadrupoles are included in the design of the Fermilab Recycler.

Permanent dipole magnets for the 8 GeV transfer line at FNAL

The transfer line that will serve to transport 8 GeV protons from the Booster to the new Fermilab Main Injector has been built using permanent magnets. A total of 46 horizontal bend dipoles and 5 vertical bend dipoles were built for this beamline; 67 gradient magnets were also built. The magnets were built using magnetized strontium ferrite bricks. Thermal compensation of these bricks was effected by use of a nickel-iron alloy. The dipole magnets were built with a mean integrated strength of 0.56954 T-m, and an rms spread of 0.06%. The magnets were thermally cycled from 20/spl deg/C to 0/spl deg/C to condition the ferrite against irreversible thermal losses, and the compensation was measured with a flipcoil. The magnet strength was adjusted by varying the number of bricks installed at the magnet ends. Details of the assembly process and a summary of magnetic measurements are presented here.

Design for Fermilab Main Injector magnet ramps which account for hysteresis

Although the dominant fields in accelerator electromagnets are proportional to the excitation current, precise control of accelerator parameters requires a detailed understanding of the fields in Main Injector magnets including contribution from eddy currents, magnet saturation, and hysteresis. Operation for decelerating beam makes such considerations particularly significant. Analysis of magnet measurements and design of control system software is presented. Field saturation and its effects on low field hysteresis are accounted for in specifying the field ramps for dipole, quadrupole and sextupole magnets. Some simplifying assumptions are made which are accepted as limitations on the required ramp sequences. Specifications are provided for relating desired field ramps to required current ramps for the momentum, tune, and chromaticity control.

Experience with the procurement of ferrite and temperature compensator for permanent magnets for accelerators

The use of permanent magnets for transporting the 8 GeV proton beam from the Fermilab Booster to the new Fermilab Main Injector accelerator has been implemented and the magnets for a new 8 GeV ring to be installed in the Main Injector tunnel for increasing the luminosity of pbar/p collisions in the Tevatron are about to start being produced. Strontium oxide ferrite was selected for the magnets due to its low cost and satisfactory magnetic properties for the 1.5 kilogauss fields required in the 2-inch gap magnets. Fermilab has received 96,000 pounds of ferrite and by working with the Vendor (HITACHI, Edmore, MI) improved uniformity of Residual Induction (Br) has reached 3905 gauss /spl plusmn/0.65%. Further details are given in the paper. Overcoming the magnetic field variation when the temperature of the magnets changes is accomplished by incorporation of approximately 30% nickel steel alloy. The ferrite changes approximately-0.2% per degree C, which is compensated for by the 13% by volume of compensator alloy incorporated in the magnet. Fourteen thousand (14,000) pounds of this material has been received and in order to obtain sufficient uniformity we mixed equal amounts from each batch into each magnet. Results of this process are given in the paper.

A wide aperture quadrupole for the Fermilab main injector synchrotron

During the design of the Fermilab Main Injector synchrotron it was recognized that the aperture was limited at the beam transfer and extraction points by the combination of the Lambertson magnets and the reused Main Ring quadrupoles located between the Lambertsons. Increased intensity demands on the Main Injector from antiproton production for the collider program, slow spill to the meson fixed target program, and high intensity beam to the high energy neutrino program have led us to replace the aperture-limiting quadrupoles with newly built magnets that have the same physical length but a larger aperture. The magnets run on the main quadrupole bus, and must therefore have the same excitation profile as the magnets they replaced. We present here the design of the magnets, their magnetic performance, and the accelerator performance.

Tune control in the Fermilab Main Injector

We describe methods used to measure and control tunes in the Fermilab Main Injector (FMI). Emphasis is given to software implementation of the operator interface, to the front-end embedded computer system, and handling of hysteresis of main dipole and quadrupole magnets. Techniques are developed to permit control of tune of the Main Injector through several acceleration cycles: from 8.9 GeV/c to 120 GeV/c, from 8.9 GeV/c to 150 GeV/c, and from 150 GeV/c to 8.9 GeV/c. Systems which automate the complex interactions between tune measurement and the variety of ramping options are described. Some results of tune measurements and their comparison with the design model are presented.