Home  |  Past Issues  |  Contact Us  | RHIC
 

 

About the Author

Christoph Montag is a physicist in Brookhaven's Collider-Accelerator Department, and serves as the run coordinator of Run 8's polarized proton run.

Bad Vibes in RHIC

By Christoph Montag

Horizontal beam orbit jitter at frequencies near 10 Hz has been observed in both RHIC rings for several years. These oscillations are observed at all beam position monitors around the rings. The rms amplitude of this beam orbit jitter scales rather well with the local horizontal β-function of the RHIC lattice. This indicates that the oscillation is caused by multiple sources.

Since orbit jitter spectra (Fig. 1) in the two rings are almost identical, the source of this vibration is likely a common magnetic element.

Figure 1: Simultaneously measured horizontal orbit vibration spectra in the "blue" (top) and "yellow" (bottom) RHIC rings.

The only potential sources common to both RHIC rings are located in the interaction regions (Figure 2), which consists of superconducting magnets. Horizontal beam jitter could be caused either by magnetic field ripple in the DX separator dipoles, or by transverse horizontal vibrations of the triplet quadrupoles.

Figure 2: Schematic view of the RHIC "6 o'clock" interaction region. Both beams collide head-on at the interaction point (IP) and are separated by the DX dipoles. Both orbits are bent back by the D0 dipoles. Strong low-β focusing is provided by superconducting triplets consisting of the quadrupoles Q1, Q2, and Q3.

Though the latter do not share common cold masses between the two rings, the mechanical support inside common cryostats causes mechanical coupling of the two rings. Vibration measurements at all RHIC triplet cryostats revealed that each individual triplet vibrates at a distinct individual frequency, all of which can also be identified in the orbit jitter spectra. As an example, Figure 3 shows simultaneously measured spectra of horizontal orbit jitter an in "yellow" RHIC ring and horizontal vibration of the 4 o'clock triplet cryostat. Two peaks at 10.14 Hz and 16.133 Hz can be clearly identified in the beam jitter spectrum.

Figure 3: Simultaneously measured spectra of beam orbit vibration vibration in the "yellow" RHIC ring and of mechanical vibration of the 4 o'clock triplet.

Frequencies of all major peaks within the beam orbit jitter spectrum are summarized in Table 1, together with the triplet at which these lines were identified in the vibration spectrum of the respective cryostat, and the rms amplitude of that cryostat vibration.
 

Frequency/Hz Triplet Amplitude/nm
7.75 12 42
8.825 7, 8, 9 28, 33, 83
10.14 3, 4, 5, 11, 12 172, 211, 24, 90, 16
10.625 9 57
10.825 1, 2, 10 33, 141, 11.5
11.00 11 33
11.325 5 250
12.700 (10) (23)
13.000 1 15
13.275 unknown  
13.55 9, (2) 6, (63)
14.325 1, 2, 3 30, 18, 39
15.950 2, 4, 5, 6, 7, 9 18, 14, 9, 155, 14, 9
16.133 3, 4 13, 22
16.500 8 2

Table 1: Dominant frequency lines of beam orbit vibration as shown in Figure 1, and triplets where these frequencies have been detected in the horizontal vibration spectrum.The third column contains the corresponding rms triplet vibration amplitudes. Numbers in brackets indicate locations where the corresponding frequency is present in the vertical triplet vibration spectrum only.

The individual cold masses comprising the RHIC triplets are supported by two posts each, thus essentially being inverted pendulums. Figure 4 shows a cross section of a RHIC triplet, with the cold masses for the two rings side-by-side inside the common cryostat.

Figure 4: Cross section of a RHIC quadrupole triplet.

Using a simple mechanical model, the eigenfrequencies of these pendulums can be analytically estimated at about 15 Hz. Using a detailed finite-element analysis, eigenfrequencies were calculated to be around 10 Hz.
The superconducting RHIC magnets are cooled to 4.5 K by liquid helium, which circulates around the two rings at a mass flow rate of 100 - 150 g/sec. When pressure transducers were installed at one of the RHIC triplets, they showed a helium pressure oscillation at 10.7 Hz (Fig. 5) that vanishes when the helium circulator is turned off and the flow stops.

Figure 5: Measured helium pressure P in the RHIC helium system vs. time. The dominant frequency of the oscillation is 10.7 Hz.

To study the effect of the helium flow on horizontal beam orbit jitter, dedicated experiments were performed. Both helium circulators (for the "blue" and "yellow" RHIC ring) were turned off for a brief period of several minutes, while beams were circulating in the machine. Beam orbit positions were recorded by a million-turn BPM in the "blue" ring, which was triggered roughly every 30 seconds. The rms orbit jitter amplitude in the frequency band from 8.5 to 14.5 Hz was computed from the power density spectrum of each million-turn BPM data set. As Figure 6 shows, there is a clear correlation of the rms orbit jitter amplitude in the "blue" ring and the "blue" circulator RPMs.

Figure 6: Measured horizontal rms beam jitter in the "blue" ring (solid line), when the helium circulators in both rings were turned off, in the frequency range from 8.5 Hz to 14.5 Hz. The dashed line shows the helium circulator RPMs for the "blue" ring.

While this 10 Hz RHIC orbit oscillation has been known for many years, it has never been an obviously limiting factor in RHIC operation, though it may probably lead to emittance growth due to modulated beam-beam offsets at the interaction points. However, when RHIC was operated at a new, near-integer working point, unacceptably high background levels were observed at the detectors STAR and PHENIX. At the same time, measured 10 Hz orbit oscillation amplitudes in the triplets increase from just over one millimeter to seven millimeters. Since for a given dipole kick anywhere in the machine the resulting orbit change at any observation point scales as 1/sin(π·Q), where Q is the "tune" of the machine, the latter was expected. However, the effect of this increased jitter amplitude on detector backgrounds was previously unknown, and could only be tested by actually running RHIC in this configuration. Since the 10 Hz orbit jitter cannot be eliminated, we had to back off and operate the machine at the regular working point around Q=0.69 instead.