Transparent lattice characterization with gated turn-by-turn data of diagnostic bunch train

Methods of characterization of a storage ring’s lattice have traditionally been intrusive to routine operations. More importantly, the lattice seen by particles can drift with the beam current due to collective effects. To circumvent this, we have developed a novel approach for dynamically characterizing a storage ring’s lattice that is transparent to operations. Our approach adopts a dedicated filling pattern which has a short, separate diagnostic bunch train (DBT). Through the use of a bunch-by-bunch feedback system, the DBT can be selectively excited on demand. Gated functionality of a beam position monitor system is capable of collecting turn-by-turn data of the DBT, from which the lattice can then be characterized after excitation. As the DBT comprises only about one percent of the total operational bunches, the effects of its excitation are negligible to users. This approach allows us to localize the distributed quadrupolar wakefields generated in the storage ring vacuum chamber during beam accumulation. While effectively transparent to operations, our approach enables us to dynamically control the beta beat and phase beat, and unobtrusively optimize performance of the National Synchrotron Light Source-II accelerator during routine operations.

Figure 1

Bucket filling pattern with an extra DBT for transparent lattice characterization.

Figure 2

Analog-to-digital converter (ADC) signals (counts) from BPM buttons of an excited bunch (in blue) without BBFB suppression and an unexcited bunch (in red) with BBFB suppression. Excitation is performed through resonance driving for about 700 turns ($\le 2\mathrm{ms}$). Free betatron oscillation then decays through radiation damping.

Figure 3

In-phase excitation of the DBT through resonance driving. Here the vertical ADC counts of three out of ten bunches are shown. The signals are measured by the dedicated BPM used as the BBFB’s pickup, and its digitizer can distinguish different bunches in the train.

Figure 4

The schematic diagram of the gated signal processing. Two signal-processing channels with separated gates are provided. One of them (Gate 2) is dedicated to processing the DBT.

Figure 5

ADC signal of BPM button “A” with and without gated functionality. Ungated BPM signals are the sum of contributions from all buckets (top). After applying a gated waveform as illustrated by the red lines, the DBT signals can be filtered out and processed by the newly added channel (Gate 2) as seen in Fig. 4.

Figure 6

Comparison of the BPM resolutions for gated and ungated data. Gated BPM data resolution is measured $\sim 3$ times better than the ungated data in both the horizontal and vertical planes.

Figure 7

Comparison of the gated BPM TbT data of the DBT (top) and the ungated TbT data averaged over all filled buckets (bottom). Although the DBT amplitude reaches 1 mm for a few ms, the disturbance on the global beam stability is negligible. The ungated data of the combined bunch trains sees noise from either the power supplies or the rf cavities. Note that the two subplots’ vertical scales are different.

Figure 8

Tune shifts with stored beam current.

Figure 9

$\beta$-beat (top) and phase beat (bottom) in relation to stored beam currents.

Figure 10

Variation of the vertical phase advance as a function of beam current around the location of a 7 m long damping wiggler (section 166), which has a flat chamber.

Figure 11

Quadrupoles’ virtual variations at different beam currents obtained with Eq. (11) for a NSLS-II supercell. They can be interpreted as horizontal focusing quadrupolar wakefields that gradually increase with the beam current, which provides a guide for localizing the impedance distribution around the ring.

Figure 12

Integrated quadrupole strength which is needed to compensate for the quadrupolar wakefields at different beam currents.

Figure 13

Vertical emittance is drastically increased when the tune approaches a linear resonance during beam current accumulation.

Figure 14

Comparison of two coupling $\beta$-functions at a low beam current (2 mA) and at a higher beam current (276 mA).

Figure 15

Dynamic aperture reduction at different stored beam currents without linear lattice correction at NSLS-II. In addition to the quadrupole nominal settings, localized quadrupolar wakefields are introduced in the lattice. The systematic and random multipole errors, closed orbit distortion and linear coupling are included in the simulations. Each dynamic aperture is obtained, accounting for random multipole error distributions in specific magnets, and averaged over 50 random seeds.