1989 Particle Accelerator Conference, Chicago, IL, Harch 20-23, 1989.

lIN.-413 32

A FAST CHOPPER FOR PROGRAMMED POPULATION OF TIlE LONGITUDINAL PHASE SPACE OF THE AGS"

J.M. Brennan, L. Ahrens, J. Alessi, J. Brodowski, J. Kats, W. van Asselt
AGS epirtunt , lrooklvuien Nut tunal Iaburiat.ry [  (   \1  l!
Aouuuic ted UniveruLLe, tUe., Uptul, NY  I. ll/J U41L

K11(JS

BNL--41832

DE89 012539

Sumina ry
A fast beam chopper hae been built that enn
produce an arbitrary pulse program of the 200 HeV H-
beam for synchronous injection into moving rf buck-
ets in the AGS. The chopper will eliminate rC cap-
ture losses and can be used to tailor the initial
distribution in longitudinal phase space by varying
the pulse parameters, width and phase, on a bunch-
by-bunch time scale, during multi-turn injection.
The chopper also serves as a studies tool since It
can provide controllable beam intensity with fixed
longitudinal emlttance (and converi.ly) and/or mIHs-
ing bunches.    it is an Llectrunlulte dellectLour
device with 15 pairs of plates located above and
below the 35 keV ti- beam between the ion source and
the ItFQ preInjector.  The pate a ore apa ed 20 mm
apart in the beam direction and connected at a Nlow-
wave structure by coaxial cables.   They are driven
to = 760 V by dc-coupled pulse generators.     Beam
current rise and fall tImes are Less thnUM 10 us.
Introduction
Injection, rf capture, and early acceleration
are crucial steps to successful operation of the
synchrotron, in particular, the vast majority of
particle losses in an acceleration cycle occur dur-
ing these steps. Rf capture losses could, in prin-
ciple, be eliminated by an adiabatic turn on of the
rf voltage.   But for a high intensity machine such
as the AGS, the long time required for adiabatic
capture would be prohibitive, because other loss
mechanisms, such   ae transverse stopband losses,
driven by large betatron tune spreads due to space
charge at low energy, would defeat the advantage of
effective rf capture.  OptinIzed non-adlahcn cap-
ture schemes have been used with some success,1 but
efficiencies are limited. In order to eliminate the
rf capture procepn altogether, a fnat chopper hall
been built that can prepare the rf bunch structure
of the beam before injection into phase stable buck.
ets in the AGS. The elimination of rf capture loss-
es may relax some of the constraints on optimizing
other facets of the early acceleration process.
Moreover, the presence of time structure on the
injected beam allows investigation of the causes of
other early losses by enabling ac-coupled instru-
mentation devices to see the beam at injection.
To be effective, the chopper must have complete
controllability of the injected pulse program (see
below). This controllability also opens the poisi-
bility for using the chopper as a studies tool.   A
specified longitudinal emittance can be obtained by
populating only the appropriate area of phase space
while the beam intensity can be independently con-
trolled by the number of turns injuectd. ConvureU-
ly, at fixed intensity, the initial distribution in
*Work  1p rrnrmn.d  mugilr  i.he  aunpl,* .g n  of  l.i   [. .- .
Department of Energy.

i olM' 'ijit*  Cll .'~  1..'. d
as the dependence Wf tet'atFn tune depression on
I ng.LLd nul hunching fatctr and the Htab itIty of
hollow distributons in the presence of beam control
feedback loops.2   Zero is an allowed value of the
wtdLh, which gives bunch-to-bunch Litensity control,
useful for studies of transient beam loading on the
acceleration cavities.
Applications Program
Injection into the AGS is done while the mag-
netic field is rising at a rate of 0.5 T/s.1 This
Lta  dulle  Lu  wii itm(.e  Lhe  Ll.e  Lhat.  the  bcui  muUt
spend at the lowest energy where space charge ef-
fects are greatest.   Also, the power supply ripple
W   I be eX1:uHnLnve at the tow vult tge  cel'ed fot
ver rLhe rate at the Injection field level (250
Gauss). The synchronization energy (E - the energy
of a partLcle that does not change piuse with re-
spect to the rf), therefore, depends on time during
the  injerc il)  proriln.  The t imp dcpendencp (F
governed by the two controlled variables Bi and E
(- df/dt) through3
2
is  - (f/f- 4-B/B)
Ytr
where 9 is the normalized particle velocity, and n -
Yr - Y      One normally chooses f/f _ Y 2/B so
that the radius of the orbit of the synchronous
particle is constant.      In that case, E9 - 7.0
HeV/ms. At the ocher extreme, a capture strategy
eoUld choose f/    "  -2 B/B so that the synchronous
energy is constant.   The radius of the synchronous
particle will change then at the rate of -3.5 cm/ms.
Since the linnc (energy is constant, the former
choice implies that the ph-se coordinates of the
separatrix at the injected beam energy will change
continuously during injection.. While for the latter
case one is injecting into a stationary bucket at
the synchronous energy and the boundaries are always
1.
The chopper accommodates this variety of choice
by using pre-calculated phase coordinates for the
edges of each pulse that is injected. The coordi-
nates are fetched out in real time from a fast
memory and used to trigger the high voltage pulsers
that switch the beam on and off.    The memory con-
tents are calculated by an interactive applications
program in the AGS Apollo-based computer control
system.   An example of the graphic output of the
program is shown in Figure 1.   The input variables
are:   rf voltago/turn, IL, t, injected synchronous
energy, beam current, and number of turns to inject.
The output is a graphic display of the separatrix
, D  WIm  I1..  1.11  w11r1 OQCM 1E."n  WT  i RT   Injnervl.
DiSTRIBUTION OF YVA   QYi11              TD