Liu Lin Liu@lnls.br
LNLS, Laboratorio Nacional de Luz
Sincrotrón
Cx. P. 6192,
Campinas 13083 Sãn Paulo,
BRASIL
3mm
The Brazilian National Synchrotron Light Laboratory, LNLS, funded
by the National Research Council of the Ministry of Science and
Technology (CNPq), consists of a 1.37 GeV electron storage ring
with a 120 MeV injector Linac. Commissioning of the Linac started
in December 1995 and after almost simultaneous completion of all
sub-systems in April 1996, commissioning of the storage ring at low
energy started on May 1st. The first stored beam was observed a month
later, in May 30. Some weeks later we could capture a current of 3 mA
(after one damping time) with the one-shot on-axis injection. The
off-axis accumulation process (with 3 kickers producing a closed
bump) was still very difficult and the accumulated current saturated
at about 0.3 mA. Our efforts at this time split roughly into 2 main
tasks with the machine being operated 24 hours a day, 7 days a week:
energy ramping and accumulation at 120 MeV. While energy ramping
evolved quickly with a rather systematic approach, accumulation
turned out to be much more `intriguing' and subtle.
The first stored beam at 1.15 GeV was observed
late in July 96 but only in October 19, after many measurements (with
some results which are understood and others which are not), empirical
adjustments and Ttrial and frustrationsU have we succeeded in
accumulating 20 mA at 120 MeV. Presently (November 96) we can
store 48 mA at 120 MeV with about 90 seconds lifetime for an
average pressure of about 
Torr. The lifetime is
limited by elastic scattering on the residual gas. At 1.37 GeV the
lifetime depends on the beam current and varies from about 15 minutes
to an hour for currents from about 25 to 3 mA.
The limiting effect in this case is Touschek scattering for higher
currents. This limitation will become less restrictive with the
upgrade (in progress) in the 476 MHz RF cavity cooling system,
which presently allows maximum gap voltages of only 240 kV.
In order to optimize accumulation at 120 MeV the machine working
point has to be set surprisingly close to the integer resonance,
and 
. With the present beam lifetime (90 s)
the Linac pulses are injected every 2 seconds, about a fifth of
the radiation damping time. The repeatability of the
injection conditions is very sensitive to the magnets cycling procedure.
Due to the proximity of the magnets in the storage ring even a change
in the magnet cycling sequence introduce large variations in the initial
conditions of the magnets at low excitations. Following a standard
cycling procedure the injection conditions can be reasonably repeated.
Sometimes fine adjustments are needed, remarkably in the Linac energy.
(especially if it is late afternoon!)
No clear evidence of ion trapping has been observed so far, although small (200 V) voltages in the clearing electrodes have been used to avoid slow (few seconds) oscillations in the accumulated current.
The machine has been intensively characterized at injection energy. Betatron and dispersion functions were measured and show good agreement with theoretical values. The chromaticity can be corrected by setting the sextupoles to calculated values. The beam horizontal size, as seen and measured on a synchrotron light monitor, is compatible with the expected from intrabeam scattering effects whereas the vertical size is larger than predicted from measured coupling effects. Several measurements were made to determine the ring horizontal and vertical apertures: with localized beam bumps, with single corrector excitation, with scrapers and with kickers. The experiments gave different results and we learned that clear experiments with clear results are very scarce. We could only conclude that there was no indication for a localized physical obstruction in the ring since no large asymmetries in the results showed up. The aperture measured with scrapers seem to give results which would not allow for accumulation.
Energy ramping from 120 MeV to 1.37 GeV can be optimized
for minimum loss during ramp by setting up intermediate
configurations with corrected orbits and tunes. In our present
best ramping path there are configurations every 10 MeV from
120 to 200 MeV, and then at 300, 400, 1150 (nominal energy) and
1370 MeV. The tunes are kept at and
up to 400 MeV; after this energy the working point tend to the
design value of 
and 
at 1.15 GeV.
The ramping time is 44 seconds and the ramping efficiency is
about 75 % with beam losses occurring mainly at the very
beginning of the ramp. The RF gap voltage is increased during the
ramp from the initial 50 kV to 240 kV. This is done `manually'
and improvements in the RF system may still increase the
ramping efficiency.
Any one of the three implemented orbit correction algorithms (matrix, best correctors and harmonic correction) can be applied successfully both at low and high energies, with the difference that at high energy the ring model can be taken from the measured quadrupole strengths, whereas at low energy the tune has to be fitted to the measured values to produce an adequate model. At high energy, orbit stability is better than 20 mm. Localized beam bumps using 3 correctors can also be produced and have already been used to move the photon beam in an experimental station. The beam bumps were also used during the hard commissioning days (when nothing seemed to work) to scan the physical aperture looking for obstructions in the ring and to probe non-linear fields with tune measurements as a function of bump amplitude.
The commissioning results obtained up to this date show that there
is no basic limitation in the low energy injection and accumulation
preventing us from achieving the design specifications of the ring.
The accumulated current at high energy and lifetime, although still
small, allowed already the start-up of the commissioning of some
beam-lines. The first EXAFS spectra were obtained in November
3rd; and the first 
VUV reflectivity spectrum in November
7th. Some hardware upgrades are in progress, including the
construction of a new thin septum with thickness less than the
present 6 mm; construction of a new thick septum with improved
magnetic shielding to decrease its effect on the stored beam;
improvement in the ripple of orbit corrector power supplies;
improvement in the RF cavity cooling system; and upgrading
the injection energy to 170 MeV by adding two klystrons to the
Linac. Also vacuum will be improved by baking procedures as well
as washing with high energy photons. (Concerning washing, we have
accumulated up to November 13, 160 mA.h of high energy beam).