Horizontal Focusing Optics

The virtual focusing of the electron beam downstream of the straight section of a beamline gives the possibility of further reducing the horizontal size of the photon beam at the sample. This can be achieved by operating the lattice with modified optical functions in the corresponding straight section (x # 0, ßx = 55 m). As already reported in the 1998 issue of the Highlights, very promising results were obtained on ID6, the machine diagnostics beamline, with a gain in beam sizes by a factor of two and a significant increase in spectral flux per unit surface.

These modified optics will be implemented on ID20 at the beginning of 2000, once the 6 individual power supplies for the magnets of the straight section have been procured. In order to prepare for the changeover, the characteristics of the new optics have been fully assessed. With such a breaking of the 16 fold periodicity of the lattice optics, the challenging issues come from the requirement to maintain the performances of the standard optics (electron beam emittances, ß values at the other source points, ) without experiencing severe penalties from a one-fold symmetry lattice (energy acceptance, dynamic aperture and lifetime reduction).

All of these aspects have been carefully checked. No distortion of the optics or degradation of its characteristics could be seen. As shown in Figure 135, which gives a comparison of measured and theoretical ß-values in the HFO straight section, the properties of the optics are close to those predicted. The momentum acceptance and transverse acceptance are equal to the usual values. The lifetime in multibunch mode was over 50 hours at 200 mA. This looks reasonable for operation, even if this figure is 5 to 10 hours shorter than on the periodic lattice. Further optimisation of the sextupole distribution is likely to minimise this slight drawback.

Energy Acceptance

Even if the lifetime of the machine is dominated by the Touschek lifetime in all operating modes, the request for a large momentum acceptance is crucial only for the few bunch modes of operation. With the present acceptance of 3%, the lifetime stands at 11 h in 16 bunch mode (90 mA, 1% coupling). This is mainly due to the numerous resonances on the path of Touschek scattered particles resulting from the high chromaticities (x = 4, z = 8) required to raise the transverse instability thresholds.

The first step to better understand the impact of sextupole distributions on machine performance was to refine the sextupole calibration based on magnetic measurement. The additional calibration with beam is based on the processing of measured chromaticities. By running the lattice with only chromaticity sextupoles and with one harmonic sextupole family at a time, a large amount of data could be obtained and a model was built from the fitting of the data. This reliable calibration now enables deeper comparisons of theoretical and experimental non-linear effects of sextupoles. The objective is to refine the sextupole distribution in order to optimise the tune path of scattered particles.

Low Coupling

Within a Storage Ring, the finite vertical emittance is generated by a transfer of electron motion from the horizontal to the vertical plane. For this reason it is expressed through the coupling which is the ratio between the vertical and horizontal emittance.

A record vertical emittance of 12 pm.rad was measured on ID8 and D9 X-ray pinholes at the end of 1998, with the developed correction method consisting of a response matrix modelling and minimisation of vertical beam sizes measured with the two pinholes. Following this achievement, activities in the year 1999 started with the pursuit of the limit of correction by investigating various surrounding effects.

Having seen that the optimal skew corrector strengths shifted systematically from predictions, their strengths were re-calibrated utilising the measured dispersion, as carried out previously. By taking into account local variations of the dispersion, the calibration coefficient turned out to be ~20% larger than previously estimated for most of the correctors, explaining the aforementioned discrepancy. Re-calibration was also carried out on the pinhole camera's set up in the large coupling mode, where it was known that measured couplings give large overestimations. The new calibration managed to reproduce the expected values correctly affecting the measurement in the low coupling regime.

Since the lowest measured vertical emittance of 12 pm.rad is still notably higher than the ~5 pm.rad predicted by the model, and as the correction often saturates in a non-continuous manner, it was suspected that the correction could be limited by the influence of vertical beam motions. It was seen that the global orbit feedback that corrects the vertical beam oscillations up to 100 Hz could bring the measured vertical emittance a few picometres below the level without feedback. Similar results were obtained by reducing the integration time of the pinholes from 20 to 2 ms.

The effect of fast beam oscillations arising from resistive wall instabilities were also investigated. In uniform filling mode, the magnitude of the instability, which varies with the chromaticity, was found to be accurately reflected in the vertical emittances of the 12 pm.rad level, giving a long tail off of the effect up to the chromaticity of ~10. However, the observation appears to be in contradiction with the fact that, so far, 8 pm.rad could be achieved with 200 mA beam in 2 x 1/3 filling, despite the presence of instabilities. This puzzling issue must be investigated in more detail.

With non-uniformity revealed in the skew error distribution obtained from the response matrix analysis, an attempt was made to eliminate the two outstanding peaks in the distribution (cell 5 and 15) by moving the magnets. Figure 136 shows the initial distribution of errors.

As the method of localisation of error did not have a resolution high enough to identify a faulty magnet, between the two equivalent actions of rotating a quadrupole or vertically displacing a sextupole, the latter was chosen for reasons of feasibility. By displacing S22 and S24 magnets by ~0.5 mm in the proper directions, both outlying peaks were successfully removed, which enabled the standard correction to reach a slightly lower vertical emittance (~20 pm.rad).

Further efforts are underway to distinguish the sextupole contributions to the skew errors from that of the quadrupoles. Settings were found to enable a response matrix measurement with and without a (harmonic) sextupole family turned on. Analysis of a pair of such measurements should solely identify the contribution of the selected family.

Low coupling was tentatively introduced during USM on two time-limited occasions (27 Oct ~ 1 Nov and 19 ~ 22 Nov). The correction procedure was automated and kept running in order to start the correction as soon as the measured vertical beam sizes exceeded a certain threshold following ID gap motions. Vertical emittances were mostly kept to 9 ~ 11 pm.rad with a 200 mA beam in 2 x 1/3 filling. The lifetime reduction due to the lower coupling was in the ~15 hours range (60 to 45 hours). Actual ID gap variations during USM did not bring vertical emittances beyond the ~15 pm.rad level.

Studies of Transverse Beam Instabilities

Single bunch and associated time structure modes are limited by strong vertical transverse instabilities. For operation, the vertical chromaticity is increased in order to raise the current threshold and to maintain a low vertical beam size, but the drawback of this remedy is lifetime reduction. A study ­ including theory, simulations and experiments ­ has been initiated to come to a better understanding of the instability mechanism and the associated machine impedance.

At zero chromaticity, the single bunch current is limited below 1 mA by the merging of mode 0 and mode ­1. From this result, the peak of the BBR model of the impedance can be identified at high frequency (22 GHz). With the increase of the chromaticity and the large variation of bunch length with the current, the observation of successive head tail mode interacting with the impedance is also in accordance with this model. Frequency and time domain tools have been developed for the simulation of these observations.

Damping times and growth rates have also been studied using simulation and measurement data. The strong discrepancy between the two approaches needed a complementary study in order to explain the lower threshold observed. Hence, the simulation helped by showing that adjacent modes ­ widened by the increased current ­ could couple close to the threshold. This assumption was verified during machine studies.

The inefficiency of the transverse feedback is also linked to high order modes that have low components at low frequency.

Multibunch modes are affected by smooth transverse instabilities that are produced by the low-frequency resistive wall impedance. For operation, the vertical chromaticity is also pushed to increase the current threshold and to maintain a low vertical beam size.

A systematic study has been initiated to characterise the mechanism and to prepare the installation of a transverse feedback. For the first experiments, which were made in uniform filling, only the first modes (mainly the first two) were observed. In this case, we were able to measure the relation between the beam size and the level of the excitation and to successfully apply a feedback on the first line.