Fast AC Position Feedback Systems

Fast orbit correction systems have been implemented and used during normal operation of the machine since November 1998. They operate in a 0.01 to 200 Hz bandwidth with a position sampling and correction rate of 4.4 kHz.

The first system is a global feedback using 16 beam position monitors (BPM) and 16 correctors to damp the vertical beam motion all around the storage ring. The control of the global feedback has been improved (graphical user interface, diagnostics tools) in order to make its operation more robust and facilitate troubleshooting. The efficiency of the corrections has also been improved by compensating the BPM tilt which was causing a spurious mixing of the large horizontal beam motion to the vertical beam motion measurement. The correction of this effect now enables the beam motion amplitude to remain below 0.80 µm at the feedback BPM locations. The main 7 Hz frequency peak is damped by a factor 10 and the global damping factor is 2.5 when averaging all frequencies up to 100 Hz.

The second system is a local horizontal feedback implemented on the ID24 EXAFS beamline in a high ßx straight section. This system uses two electron BPMs to measure the horizontal electron beam position at both ends of the ID24 straight section and four fast steerer dipole magnets to produce a local correction bump (a closed bump which does not change the beam position in the rest of the machine) the residual amplitude of oscillation is damped from 13 µm down to 1 µm (rms) to be compared with the 400 µm (rms) horizontal beam size.

Following the excellent results obtained on ID24, another local horizontal position feedback was tested on ID3 to check the efficiency of the system in a low ßx straight section. The ID3 local feedback shows the same damping efficiency as achieved on the ID24 high ßx straight section. There is no apparent interaction between the ID3 and ID24 systems, when activated simultaneously. It is planned now to test the common operation of a local feedback with the ID gap opening compensation system (both systems use the same steerer magnets on both side of the straight section). In addition, a local feedback will be also implemented on ID14 after the installation of the missing BPM pick-up during the winter shutdown.

The implementation of these systems has resulted in significant improvements in beam stability. These improvements have been assessed by measurements of relevant machine parameters (position stability) as shown for instance in Figure 137, and by the observation of beneficial effects on beamline operation.

Damping Device for Storage Ring Girders

A damping device, called a damping link, has been developed to attenuate the vibration of the ESRF machine girders. It consists of three parts (see Figure 138):

- a sandwich structure with Aluminium plates and ViscoElastic Material (VEM) : Al + VEM + Al

- a girder mounting fixture (GMF) which links the sandwich structure to the girder.

- a floor mounting fixture (FMF) which links the sandwich structure to the floor.

These devices link the two extremities of the girder to the floor. The mounting fixtures (GMF, FMF) should both accommodate the environment in the tunnel, and be stiff enough to transmit maximal dynamic strain energy of the magnet-girder to the layer of VEM which then dissipates this energy. The mechanical properties of the VEM are key parameters for the successful design of the damping link.

Damping links were first tested on a magnet-girder assembly installed in the experimental hall. Excellent damping results from these first tests encouraged us to continue the tests in the storage ring tunnel. Three Storage Ring girders (G10 and G20 in cell 23, G30 in cell 22) have been equipped with the damping links. Test results show that the fundamental peak (7-9 Hz) in the spectral displacement of quadrupole magnets is reduced by a factor of between 5 and 10 for the vibration in the transverse direction. The fundamental peak has been reduced to such an extent that, in some cases, the vibrations induced by the cooling water flow in the frequency range of 20-100 Hz are now noticeably significant. Compatibility tests with alignment operation were also carried out. It was shown that the damping links are compatible with alignment operation, and that girder displacement, required when re-aligning the machine, does not deteriorate the performance of the damping links.

It is now planned to bond the FMF to the floor with epoxy instead of bolts in order to facilitate installation, to compensate for the irregularity of the floor, and even to obtain a more rigid fixation of the FMF onto the floor. Additional damping devices to further attenuate the water flow-induced vibrations are also being studied by finite element modelling.

Radio Frequency Systems

Upgrading of RF Units

Since the construction of the new SRRF3 unit, an upgrade programme is under way for the three other RF units. The upgrade consists of:
- complete reorganisation of the control system, both hardware and software
- rewriting of the control software that uses all the ESRF standards and tools
- re-design of the low power electronics (RF low level, hardware interlocks)
- re-building of klystron auxiliaries: klystron lead garage, klystron cooling system,

The goal is to obtain complete redundancy in case of a major hardware failure so that it will be possible to:
- run the Storage Ring even if one transmitter is down
- feed the booster cavities from the SRRF2 transmitter (target for year 2000).

Following the cavity control upgrade (cavities 1 to 4), which took place in 1998, the SRRF2 transmitter was upgraded and switched to these four cavities.

The SRRF1 transmitter is currently being dismantled for its upgrade in 2000.

Improved Cavity Temperature Regulation System

An improved temperature regulation system that allows the adjustment of the individual cavity temperatures within a range of 23°C to 60°C with ± 0.1°C was put into operation in January 1999. It is used to tune dangerous Higher Order Modes (HOM) of synchrotron beam resonances, thereby substantially increasing the current thresholds for Longitudinal Coupled Bunch Instabilities (LCBI). This was applied in the course of 1999 to develop new operation modes for the ESRF storage ring, for instance:

- 100 mA at 4 GeV: The natural synchrotron damping being much lower at reduced electron beam energy, the thresholds for HOM driven LCBI are about 5 to 20 mA at 4 GeV, depending on the RF voltage and the level of Landau damping that results from the partial filling of the ring. Thanks to the improved cavity temperature control, a stable working point with 4.5 MV of RF voltage was established allowing the delivery of 100 mA in 2/3 filling with a lifetime of 10h.

- 2x1/3 and uniform filling at 200 mA with a reduced RF voltage of 9 MV: the current thresholds for HOM driven LCBI are reduced thanks to the combination of an RF voltage decrease from 12 to 9 MV for higher Touschek lifetime and a symmetric filling pattern that does not provide any Landau damping. Stable operation was obtained by means of a careful temperature tuning of the cavities. As a result, the symmetric 2 x 1/3 filling pattern at 9 MV was implemented as the standard setting for 200 mA operation. Uniform filling of the storage ring at 200 mA with similar tuning has also been tested successfully and will be delivered to the users early in the year 2000.

Cavity R&D Studies and Outlook

Partial filling is one of the standard methods applied at the ESRF to avoid HOM driven LCBI. The mechanism of Landau damping, which results from a spread in synchrotron frequencies when only a portion of the storage ring is filled, has been extensively investigated and quantified in a PhD thesis within the RF group. Within this thesis, published in 1999, there was also a contribution from the ESRF to the development of a HOM free superconducting cavity for SOLEIL. The possible use of such a cavity at the ESRF is presently being evaluated. Finally, this thesis also gives the first theoretical calculations on the impact of a harmonic cavity at the ESRF. Designed to provide bunch lengthening in order to maximise the Touschek lifetime, especially for few bunch fillings with high currents per bunch, such a harmonic cavity actually reduces the current thresholds for HOM driven LCBI. In order to take these preliminary results further, a new PhD thesis started within the RF group in autumn 1999 to explore the means of controlling the bunch length at the ESRF.

Insertion Devices

More than 60 segments of Insertion Devices are now in operation around the ring. In addition to the installation of new conventional segments, 1999 has seen the successful operation of the second quasi-periodic undulator segment (on ID27), the fast switching planar/helical undulator (on ID12) and the prototype in-vacuum undulator (on ID11). This in-vacuum undulator is now operating in USM with a minimum gap of 6 mm without any perturbation to the beam lifetime in both 16 bunch and 2 x 1/3 filling modes of operation. The in-vacuum undulator technology is more expensive and less flexible than conventional technology with magnet blocks in air. However, an in-vacuum undulator produces a spectrum shifted by 1.57 (1.3) towards higher energies compared to that of a 16 mm (11 mm) gap undulator with magnet blocks in air. This makes this technology attractive for beamlines operating at high X-ray energies.

The recent operation of the ring with a 10 pm vertical emittance has resulted in a further enhancement of the brilliance which is now higher than 4. 1020 on the 5m long U34 undulator installed on ID23 (see Figure 139).

Front-Ends

Status in 1999

42 front-ends have been installed and are now connected to the storage ring : 27 are insertion device front-ends (23 for users and 4 for machine diagnostic and development) and 15 are bending magnet front-ends (14 for users and 1 for machine development).

Design, Construction and Test of a High-Power Front-End Configuration

A first prototype of a high power X-ray absorber was built in 1998 but proved to be very difficult to manufacture. Therefore, at the beginning of 1999 a second design was developed, alongside with a complete re-design of the front-end (Figure 140).

Before the summer shutdown the first prototype of the newly-designed absorber was installed and tested successfully on ID23 front end with the maximum power available (200 mA, three U34 undulator segments closed at 11 mm gap, giving a power density of 270 KW/mrad2).

After the summer shutdown a completely new front-end configuration was installed on ID23 in order to complete the tests on the absorber, to validate the new module 2 with its diamond window, and also to finalise its design. This new front-end configuration was successfully operated during the second half of the year without any problems, and it is expected to be able to operate with a larger power density (up to 400 KW/mrad2).

The new front-end configuration requires more diagnostics and different equipment. A new generation of PLC has to be used for the high-power front-ends. Preliminary tests have been carried out with the computing service group to improve the control of the existing front-ends and to make it compatible with future evolutions. The first step of this change will occur at the end of 1999.

Machine Diagnostics

The 'Thousand Turn' Extension to the Closed-Orbit BPM System

The closed-orbit BPM system consists of 214 active BPM stations distributed around the Storage Ring. It measures the closed orbit of the electron beam with high resolution and good reproducibility at a frequency of about 1 Hz. It is this measurement that is used for the so-called global orbit correction scheme.

An extension to this slow BPM system has now been conceived, installed and commissioned and it permits the measurement of the beam position, in both planes, at each BPM station, at each orbit turn, for a large number of consecutive turns.

This new functionality makes it possible, in a very short time, to extract a large number of beam and machine parameters such as local beta values, phase advance, dispersion, damping times, tune shifts and widths, phase-space diagrams etc., that are otherwise not accessible or only accessible after long and tedious manipulations and measurements.

The system can measure up to 2048 turns and average over a maximum of 4096 cycles. Typical results have shown a resolution of 3 µm rms for an averaging of 128 cycles (with a response time below 1 minute) on a 30 mA partially filled (200 ns) SR beam. In addition to the measurement of certain lattice parameters this system has been used to measure and optimise the closure of the injection bump.

Development and Results with the Jitter-Free Streak Camera

The development at the ESRF of a jitter-free, laser-triggered streak camera has now yielded time resolution results as short as 460 fs whilst operating in accumulating mode. The so-called jitter-free synchronisation between the laser light and the streak camera is performed through a GaAs photo-switch in a simple HV circuit that connects directly to the Streak tube's deflection plates.

The novelty of this technique is that excellent dynamic range measurements in a shot-to-shot accumulation of ultra fast (laser stimulated) events at up to 1 Khz can be obtained without degrading the time resolution.

Important insight was obtained regarding the quality of this optical synchronisation and its dependence on the laser characteristics, the switch circuit, and the structure of the GaAs switch itself. This permitted the suppression of the causes of the jitter and today the 500 fs limitation is imposed by the streak tube's intrinsic time resolution. This work was done by measuring the 3rd harmonic (i.e. 267 nm) of a 100 fs Ti:Saph laser (with Au or Pd photo-cathodes).

Important progress has also been made concerning the reliability of the photo-switch. Furthermore the problems of HV breakdown and structural degradation have been completely resolved.

Since the principal use of this system at the ESRF is in ultra-fast X-ray diffraction experiments, the exchangeable photo-cathode structure of this tube covers the entire UV-to-X-ray spectrum. The QE of various photo-cathode materials was measured in the 8-30 KeV range.

The system was used for the first time in July 1999 by external users on ID9 to measure ultra-fast surface disordering of GaAs induced by a 100 fs laser pulse at 800 nm (see Figure 141). The rapid modulation of the structure was observed by orienting a single crystal sample of GaAs to the positive side of the rocking curve of the 111 reflection. A 10 ps timescale drop in reflectivity was observed. We do not believe this to represent the ultimate time resolution of the camera but rather the timescale imposed by the experimental conditions.