A superconducting wiggler (SCW) with a peak field of 4.3 T has recently been installed at NSLS-II. The wiggler generates a high-flux X-ray beam with a photon energy range from 20 keV to 200 keV for the High Energy Engineering X-ray Scattering beamline. This device induces significant distortions in the electron beam orbit and optics. We have corrected these adverse effects using feed-forward tables. The effects of the SCW on the electron beam emittance, energy spread, and bunch length were calculated and measured. The beam-induced heating of the cryogenic vacuum chamber was also studied. Finally, the superconducting wiggler has been commissioned for user operations.
The goal of the NSLS-II online optimization project is to improve the beam quality for the user experiments. To increase the beam lifetime and injection efficiency, we have developed a model-independent online optimization of nonlinear beam dynamics using advanced algorithms, such as Robust Conjugate-Gradient Algorithm (RCDS). The optimization objective is the injection efficiency and optimization variables are the sextupole magnet strengths. Using the online optimization technique, we increased the NSLS-II dynamic aperture and reduced the amplitude-dependent tune shift. Recently, the sextupole optimization was successfully applied to double the injection efficiency up to above 90% for the high-chromaticity lattice being developed to improve the beam stability and to increase the single-bunch beam intensity. Minimizing the beam perturbation during injection is the second objective in this project, realized by online optimization of the injection kickers. To optimize the full set of kicker parameters, including the trigger timing, amplitude, and pulse width, we upgraded all kicker power supplies with the capability of tunable waveform width. As a result, we have reduced the injection transient by a factor of 29, down to the limit of 60 μm.
We report picosecond bunch length measurements using an interferometric method for a 3 MeV electron beam having bunch charge ranging from 1 to 14 pC. The method senses the single-cycle sub-terahertz (THz) pulse emitted by each electron bunch as coherent transition radiation which, in turn, is analyzed using a Michelson-type interferometer, forming an interferogram that is then processed to yield the nominal electron bunch length. This sub-THz coherent radiation intensity was measured using a quasi-optical detector (QOD) operated at room temperature. This experiment was quite challenging since the divergence angle of the sub-THz pulse emitted by the low-energy electron bunch exceeds ±10°, and its pulse energy at the entrance to the detector was as low as 100 pJ. When compared to a conventional helium-cooled silicon composite bolometer designed for frequencies above 0.5 THz, the QOD provided much better signal-to-noise ratio in the ∼80 GHz frequency range, which was critical for the successful measurement of the bunch length.
The heat transfer analysis has been performed for the stripline kicker, design of which is used at many accelerator facilities as a part of the transverse bunch-by-bunch feed-back system, including NSLS-II. The experimental data of the electrode temperature have been collected from a special diagnostic stripline designed with six infrared view ports. The temperature analysis is based on the power loss calculations and the loss factor computed by the 3-D electromagnetic simulation code GdfidL. To estimate the power loss, we analyze the real part of the longitudinal impedance and the loss factor for two different regimes. We use the ANSYS finite element code to correlate the temperature with the power loss. Lambertson and Shafer formalisms of the longitudinal and transverse beam impedances are discussed for the simplified stripline geometries. The required characteristic impedances and the geometric factors are determined analytically for the simplified geometries and compared with the results for the designed stripline geometry using the 2-D POISSON code by solving the Laplace's equation in two dimensions. Lambertson's equations are compared with numerical results. We introduce the frequency range, within which we can apply the analytical approach.
A diagnostic stripline for experimental study of beam-induced heating has been designed, manufactured, and installed in the NSLS-II storage ring. Two 150-mm long striplines are installed in the round pipe with a 79.25 mm diameter; EIA-type feedthroughs are connected to the ends. Beam-induced power was measured using a broadband oscilloscope connected to the stripline feedthrough by a 28-m long cable. The measurement procedure is described, the measured results are presented.
being observed. However, because changing BBAs in the BPM readings has direct and significant impact on the machine operation as well as the user experiments, they are not applied instantly. Instead, in NSLS-II, the official BBA values are managed by the snapshot archiving and retrieving system, called MASAR (MAchine Snapshot Archiving and Retrieve), so that they are well organized and can be easily set to any stored set of BBAs values. NSLS-II is implementing MASAR as a part of EPICS control system for the convenient and mistake-free machine operation. Each snapshot is a group of key-value pairs where keys are EPICS Process Variables (PVs). The snapshots are organized by configurations where each configuration has specific PV sets. And each configuration saves snapshots with titles corresponding their purposes such as normal user operations or specific beam studies. Anyone with control system account can save snapshots and they can be restored whenever needed. One of the configurations is “SR_BPM_BBAS" and the configuration includes the set of EPICS PVs for the storage ring BPM BBA values. This configuration also includes the BBAs of the in the straight sections, called ID (insertion device) BPMs playing important roles in user experiments. The ID BPM BBAs are also directly engaged in the active interlock (AI) machine protection system, which prevents the mechanical damage from the ID radiation. However, in estimating ID BPM BBAs, because there is no quadrupoles around them, interpolations from the surrounding sector BPM BBAs are used. This note focuses on the BPM offset variations based on the BBA measurement data and does not discuss about the ID BPM BBAs because they are not the direct measures of the offsets.
wigglers turned off, is strongly affected by IBS. To check for consistency with IBS theory of, for example, the measured vertical beam size, we need to know all dimensions of the beam, including the longitudinal one. But beyond this practical reason for studying IBS, IBS is currently a hot research topic at many accelerators around the world (see e.g. Ref. [1]), and the effect in actual machines is not well understood. Typically, when comparing theory with measurements fudge factors are needed to get agreement (see e.g. Ref. [1]). With its strong IBS effect, the ATF is an ideal machine for studying IBS, and an indispensable ingredient for this study is a knowledge of the longitudinal phase space of the beam. The results of earlier bunch lengthening measurements in the ATF can be found in Refs. [2]-[4]. Measurements of current dependent effects, especially bunch length measurements using a streak camera, can be difficult to perform accurately. For example, space charge in the camera itself can lead to systematic errors in the measurement results. It is important the results be accurate and reproducible. In the measurements of both December 1998[3] and December 1999[4], by using light filters, the authors first checked that space charge in the streak camera was not significant. And then the Dec 99 authors show that their results agree with those Dec 98, i.e. on the dates of the two measurements the results were reproducible. Since IBS is so strong in the ATF, in the Dec 99 measurements an attempt was made to estimate the impedance effect using the following method: First, from the form of the energy spread vs. current measurements it was concluded that the threshold to the microwave instability was beyond 2 mA. Then, by dividing the bunch length vs. current curve by the energy spread vs. current curve the effect of IBS was divided out, and PWBL was approximated. The assumption is that PWBL can be treated as a perturbation on top of IBS. The result was that this component of bunch lengthening was found to grow by 7-15% (depending on the rf voltage) between the currents of .5 mA and 2 mA, about a factor of 3 less than the total bunch length growth. The conclusion was that the inductive component of the impedance was small, in fact much smaller than had been concluded earlier in Ref. [2]. Electron machines generally run in a parameter regime where IBS is an insignificant effect, and impedance measurements and calculations have also normally been performed for machines where IBS is unimportant. To simplify the interpretation of the impedance from bunch length measurements, in April 2000 the energy spread and bunch length measurements of Dec 99 were repeated, but now with the beam on a linear (difference) coupling resonance, where the horizontal and vertical emittances were approximately equal. For this case the effect of IBS was expected to be very small. An energy spread vs. current measurement under such conditions will also allow us to more clearly see whether we reach the threshold to the microwave instability. As part of the April data taking we, in addition, repeated the earlier off-coupling measurements, in order to check the reproducibility of the earlier results. In this report we present and analyze this recent set of data, and compare it with the results of the earlier measurements, particularly those of Dec 99. The measurements and analysis of data in this report follow essentially the same procedure as was used in Ref. [4]. In the present report we will try to be relatively brief. The comparison of our results with IBS theory will be given in a following report. For more details about the measurement and analysis techniques presented in this report, the reader should consult Ref. [4].
Novel features of the longitudinal instability of a single electron bunch circulating in a low-emittance electron storage ring are discussed. Measurements and numerical simulations, performed both in time and frequency domain, show a non-monotonic increase of the electron beam energy spread as a function of single bunch current, characterized by the presence of local minima and maxima, where a local minimum of the energy spread is interpreted as a higher-order microwave instability threshold. It is also shown that thresholds related to the same zero-intensity bunch length depend linearly on the accelerating radio frequency voltage. The observed intensity-dependent features of the energy spread, confirmed by measurements with two independent diagnostics methods, i.e. horizontal beam profile measurements by a synchrotron light monitor and photon energy spectrum measurements of undulator radiation, are given a theoretical interpretation by applying a novel eigenvalue analysis based on the linearized Vlasov equation.
The microwave instability in the NSLS-II has been studied for the current configuration of insertion devices, 9 In-Vacuum Undulators (IVU’s), 3 Elliptically-Polarized Undulators (EPU’s), 3 Damping Wigglers (DWs). The energy spread as a function of single bunch current has been measured based on the frequency spectrum of IVU for X-Ray Spectroscopy (SRX) beamline, for two lattices, the bare lattice with nominal energy spread = 0.0005, and a lattice with one DW magnet gap closed ( = 0.0007). In addition we did measure the energy spread from a Synchrotron Light Monitor (SLM) camera installed in a nonzero dispersive region, for the two aforementioned lattices, and for a third lattice with 3 DWs gaps closed ( = 0.00087). The measurements have been complemented by beam spectra taken from a Spectrum Analyzer, and have been compared with numerical simulations with the particle tracking code SPACE.
Visible Synchrotron Light Monitor (SLM) beamline has been designed and constructed at NSLS2 storage ring, to characterize the electron beam profile at various machine conditions. Due to careful alignment, SLM beamline was able to see the first light even before beam circulating the ring. Besides a normal CCD camera to monitor the beam profile, streak camera and gated camera are used to measure the longitudinal and transverse profile to understand the beam dynamics. Measurement results from these cameras will be present in this paper. INTRODUCTION NSLS2 is a third generation light source at Brookhaven National Laboratory. The 3GeV low emittance storage ring has been commissioned with beam recently. Average current of 50mA beam was able to be stored in the ring with superconducting RF cavity [1]. While electrons pass through the bending magnet, broadband synchrotron radiation will be generated. This can be used to measure the transverse and longitudinal profile. Visible synchrotron light monitor (SLM) diagnostic beamline utilizes the radiation from C30 BM-B, which is the second dipole after injection straight. Nominal source point is ~2.75mrad into the dipole. The beamline has acceptance of +/-1.5mrad horizontal and +/-3.5mrad vertical. Visible light from the dipole synchrotron radiation will be reflected by in-vacuum mirror. The visible light is guided into SLM hutch located on the C30 experimental floor. There are various optics setups on the 4’x10’ optical table, currently there are three branches setups: CCD camera branch; fast gated camera branch and streak camera branch. Visible light can be guided to different cameras. More information on the diagnostic beamline design can be found at [2]. Figure 1 shows the installed beamline and optical table setups. Radiations from the dipole pass through the fixed aperture, which defines the source point and horizontal/vertical apertures. Upstream of the fixed mask are vacuum gate valve (GV) and bending magnet photon shutter (BMPS). These two components are standard for NSLS2 beamlines which can be used to shut down the photon whenever needed. Most high energy photons are blocked by the thin absorber called “cold finger”. The cold finger has vertical aperture of +/-0.5 mrad. Most of the power will be blocked by the cold finger so that downstream first mirror sees less than 1W of power at 500mA. The cold finger is controlled through a linear motor and stage. It can be fully retracted or tracking the electron beam position with 10um resolution. In vacuum first mirror reflect the visible (and near infrared) light 90 degree out of the vacuum window. The mirror was made from Glidcop with Aluminum coating. Both first mirror and vacuum window have flatness better than 50nm, which is about 1/10 of interesting wavelength. Visible light is then reflected by three 6’’ diameter in air mirrors on to the experiment floor, where a dark room and 4’x10’ optical table are located. Synchrotron light first gets focused on the optical table with 6’’ achromatic lens. Focal length of the lens is 2.25 m. Light is then guided to different cameras using beam splitter and reflection mirrors. Figure 1: (top) Installed visible synchrotron light monitor (SLM) beamline components inside the NSLS2 storage ring tunnel; (bottom) Optical table setup in the SLM hutch. ___________________________________________ * This material is based upon work supported by the U.S. Department of Energy, Office of Science, Brookhaven National Laboratory under Contract No. DE-AC02-98CH10886. chengwx@bnl.gov Proceedings of IBIC2014, Monterey, CA, USA TUPF21 Optical Instrumentation ISBN 978-3-95450-141-0 369 Co py rig ht © 20 14 CC -B Y3. 0 an d by th er es pe ct iv ea ut ho rs
To get accurate beam position measurements along the accelerator complex, beam position monitor (BPM) sensitivity and electric offset need to be calibrated. A good calibration is essential for day one commissioning and large beam offset measurements. For various types of BPMs used in NSLS2, the sensitivity curves were calculated and fitting by high order polynomial fit. Fitting errors are typically less than 50um. BPM electric offsets are measured using four port network analyzer. These offset values will supply a good reference for beam based alignment. INTRODUCTION NSLS2 is a third generation light source under construction at Brookhaven National Laboratory. With sub-nm horizontal emittance, the beam orbit needs to be measured and controlled with very high precision. To achieve that, BPM detectors need to be calibrated for their sensitivity and electrical offset. Figure 1: Types of the NSLS2 BPM. Top left: transfer line round type. Top right: transfer line elliptical type. Bottom left: booster elliptical type I. Bottom right: storage ring large aperture BPM. There are various types of BPMs used in the NSLS2 complex. Table 1 lists the BPMs and their geometric sizes. The linear sensitivities of the BPMs are calculated and fitted with beam offset +/2.5mm. Storage ring BPMs may be rotated to have better vertical sensitivity. These rotated BPMs will be installed at the ends of insertion devices. Figure 1 shows some of the BPM assemblies. With larger beam offset, nonlinearity of the BPM’s response needs to be included. Figure 2 shows a storage ring large aperture BPM sensitivity curve, it’s clear that nonlinear terms need to be included when beam deviations are large. With a 3 or 5 1-dimension polynomial fit, one can get good agreement of the raw value and fitted value. The maximum fitting error is ~100 μm with +/-10mm range. However, this 1-dimension fitting is good only for large beam offset in x plane with small y position offset (or y plane with small x position offset). Due to coupling, beam may have large x and y plane offsets. Huge systematic errors exist when only considering 1-dimensional fitting, as shown in Figure 3. Figure 2: BPM sensitivity with 1-dimensional polynomial fit. Fitting errors are <100 μm with 3 order fit and <50 μm with 5 order fit. Table 1: NSLS-II BPM geometry and 1 Order Sensitivity -10 -8 -6 -4 -2 0 2 4 6 8 10 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Vertical Sensitivity SR LA-BPM Y, [mm] Δ y / Σ Sy=0.084 0.085 0.085 Ky=11.923 11.826 11.813 calculation point 1st order fit 3rd order fit 5th order fit Proceedings of BIW2012, Newport News, VA USA MOPG022 Beam Position Monitoring ISBN 978-3-95450-121-2 77 C op yr ig ht c ○ 20 12 C C -B Y3. 0 an d by th e re sp ec tiv e au th or s Figure 3: Huge errors at large x and y offsets. Red dots show the 2D beam offsets; blue dots, the fitted value using the coefficient from 5 order 1-D polynomial fitting. The 1-D fit is good for the x=0 or y=0 line. BPM NONLINEARITY 2D nonlinearity fitting of BPM sensitivity is required to get rid of large systematic fitting error. Due to symmetry of the BPM geometry, six coefficients are needed, as shown in the following equations.
Modes with transverse electric field (TE-modes) in the NSLS-II multipole vacuum chamber can be generated at frequencies above 450MHz due to its geometric dimensions. Since the NSLS-II BPM system monitors signals within 10 MHz band at RF frequency of 500 MHz, frequencies of higher-order modes (HOM) can be generated within the transmission band of the band pass filter. In order to avoid systematic errors in the NSLS-II BPM system, we introduced frequency shift of HOMs by using RF metal shielding located in the antechamber slot. We demonstrated numerical modeling and experimental studies of the spurious TE modes in the NSLS-II vacuum chambers with antechamber slot. Calculated frequencies of TE-modes in considered chambers with and without RF shielding were verified experimentally. Flexible BeCu RF shielding inside each chamber at proper location shifts frequencies of H{sub 10p}-modes above {approx}900MHz, except chambers S6 odd and even. These chambers need special attention because of synchrotron radiation from downstream magnets. S6 odd multipole vacuum chamber needs to be measured and the RF shielding length has to be optimized. RF shielding looks adequate for baseline design. Fifty percent of open space provides adequate pumping speed.