C. Biscari- Phi-Factory Design

8/3/2019 C. Biscari- Phi-Factory Design

1/21

Lectu re given at th e joint US-CERN School on P ar ticle Accelerat ors"Fr ont iers of Pa rt icle Beams, Factories with e+e- Rings",

Benalm aden a, Spa in, 29 Oct - 4 Nov, 1992.

-Factory Des ignC. Biscar i

INFN - Laboratori Nazionali di Frascati. C.P. 13 - 00044 Frascati (RM) - Italia

Introduction

The search for fundamental interaction laws stimulates continuously the

accelerator community to push up the energy of accelerators. Recently the

interest for high accuracy measurements in already explored energy

regions put forwar d th e requirement for a n ew genera tion of colliders, th e

e+ e- factories, work ing at the ener gies of ha dronic r esona nces a n d

producing a very high ra te of event s with a lu min osity increased by at least

two orders of magnitude with respect to the present values. Such high

luminosity poses interesting new problems to both the detector and

accelerator physicists.

The lowest energy factory is expected to work at the resonance(~ 1 GeV) an d is ma inly dedicat ed to the in vestigation of CP violat ion. Th e

construction of a collider in energy range requires a relatively small

investment, in terms of budget and man power, and can therefore be

afforded by medium size laboratories. A strong advantage for this kind of

project is the small size of the groups which allows the participation to all

problems concern ing th e detector an d t he a ccelera tor.

The basic requirements of a -Fa ctory ar e:

- Lum inosity = 1032 1033 cm -2 sec-1 at the c.m. energy of 1.02 GeV. This

is a challenging request: the maximum luminosity in this energy range

has been obtained by VEPP-2M1 and is two orders of magnitude lower

(L ~ 5 1030 cm -2 sec-1). The ma xim u m lum inosity ever obta ined is of th e

same order but has been obtained at an energy one order of magnitude

larger (L ~ 1.8 1032 cm -2sec-1 at th e ener gy of 5.3 GeV in CESR2).

- High average luminosity. The peak cross-section at the resonance is:

peak= 4.4 10-30 cm2


2/21


3/21

3

and equal in the two planes; this is possible if the ratio between the

emittances, the coupling factor ,is equa l to the rat io of the beam r m ssizes at the IP. It follows that is also equal to the ratio of the betatron

functions at the IP:

=xy

=xy

=xy (3)

With =x + y , the beam-beam t une shift can be rewritt en as:

=r e

2N

(4)

By introducing in the luminosity expression, we set:

L = f2

re22(1 + )

y = f2

r e22(1 + )

x(5)

This formula shows the relevant accelerator parameters to play with

in order to maximize the luminosity; a discussion on each of them follows

to illustr at e th e choices tha t can be ma de.

1.1 Linear tune sh ift paramete r

When two beams collide the electromagnetic interaction between particles

acts as a focusing lens producing a tune shift. The interaction force is lin-

ear only up to an extent of about one rm s beam size an d ha s its m a xi m u m

at 1.6 (see Fig. 1). The beam-beam tune shift linear part, given by (2), is

proportional to the beam density (strength of the interaction), to the beta-tron function (sensitivity to interaction) and inversely to (rigidity of the

beam). The focusing strength on a particle and therefore its tune shift

depend on the particle position inside the bunch: due to beam-beam

interaction a tune spread arises whose width is well represented by ,

because the maximum tune shift corresponds to particles which are

inside the core of th e bunch distribut ion a nd t herefore see th e linear par t of

the interaction force.


4/21


5/21

5

Bassett i has considered a larger n um ber of colliders includin g old an d

low energy accelerators, and new and high energy ones like LEP, and

foun d th at th e best fit for y is t he following:

y1

( N i ). 4 (E xx)

.5 1.7

It is interesting to note that the fit does not depend on the energy E,

becau se even if E a ppear s in th e form ula , being x proportional to E2, the Edependence cancels out. The fit favours flat beams against round.

Anyway, when designing a new factory, a more or less conservative

value for mu st be assum ed, so fixing N/.

may change with the tune in a not easily predictable way: the ringlattice should therefore be widely flexible, in order to compensate for un-

expected effects in the beam-beam interaction.

1.2 Low beta

From theluminosity formula it is evident that the smaller at the IP thehigher the luminosity: the beam-beam interaction changes the divergence

of particles; where the divergence is large the unwanted deflection is less

troublesome. Anyway cannot be squeezed at will, first of all because ofthe geomet rical limita tion of lum inosity3. The betatron function, starting

from a symmetry point, shows a parabolic dependence on the distance s:

y(s) = y*( 1 +( sy* )2 )

Fig. 2 - Lu m inosity depend ence on vertical betatron function3.


6/21

6

The betatron function within the bunch increases rapidly, so that to

keep the advantage of having a low beta, y* must be larger than the half

bunch length zwhich is determined by the ring rf system, impedances

and single bunch instability thresholds.Moreover low y* implies high chromaticity with a corresponding re-

duction of dynamic aperture and also strong focusing near the IP, where

normally there is lack of free space because of detector requirements.

Reasonable values for both zand oyare in the order of few centime-

te rs .

1.3 Emittance

The equilibrium emittance is determined by the radiation process which

depends mostly on the characteristic radius of the bending field, and on

th e properties of th e latt ice.

To increase the luminosity one would like to have large emittance, but

cannot be made arbitrarily large because of the machine physical aper-tu re n ecessary for a reasonable beam lifetime.

The dynamic aperture is also sensitive to : large emittance meansthat m ost pa rt icles perform betat ron oscillat ions in the ma gnetic field r e-

gion where multipole components are more dangerous.

Furthermore, being N/, increasing means increasing thebunch current, to which beam instabilities are related.

The relatively large value of the emittance (~ 1 mm mrad) can be ob-

ta ined eith er with sm all bending radius dipoles or with wigglers wh ich al-

low also emittance tuning.

1.4 Coupling factor

The coupling factor can be chosen between 0 and 1, corresponding respec-

tively to a very flat or to a r oun d beam.

= 1 implies roun d beam at the IP: x = y , x = y, x = yFor the -Factory, due to the relatively low energy of the beam, it is

possible, at leas t in principle, to focus a r oun d beam to th e low beta valu es

with a solenoidal field, while in higher energy rin gs quad ru poles a re n ec-

essar y to obta in strong focusing an d it is critical to focus in both pla nes at

th e same point in t he latt ice.


7/21


8/21

8

2 Projects and proposal

There are different projects around the world for -factories; two of

them (Novosibirsk and UCLA) follow the compact ring approach. Theoth er t hr ee are t wo rings configura tions. In the following table their m a in

par am eters ar e listed, togeth er with th ose of VEP P-2M for compa rison.

A general description of compa ct r ing char acter istics an d some w ord s

on each proposal follow. The same for the two rings configuration with

special emphasis, of course, on the Frascati project.

T able 1 - Main param eters of-factory projects.

Nov5 UCLA8 DANE 9 KE K10 Mainz6 VEPP2M1

R u ssia US A I taly J apan G er m a n y R u ss ia

C (m) 35 17 98 120 29 18

y * (cm ) 1.0 3.9(.3)* 4.5 1.0 0.12 4.5

z (cm) ~1 3(.3) 3 . 5 1x (m m m r a d) . 47 3.2(1) 1.0 1.1 . 22 . 46

1 0.2 0.01 0.01 1 0.01

N (1010) 20 40 9 6 8 3.7

Itot (A) . 5 1.2(.5) 5 7 .8 . 1

n b 1 1 120 300 6 1

f (MHz) 17 17 368 750 62 17

(m ra d) 0 0 12.5 20 0 0

x .10 .05 .04 .03 .08 .05

y .10 .05 .04 .03 .08 .02

.03 -.06 .11(~0.) .005 .007 .018

(m in ) < 5 40 300 15 60*Lo(10

32) 10 2 (10) 0.04 0.1 1.6 0.07

L (1032) 10 2 (10) 5 30 10 0.07

* Numbers in pa rent hesis correspond t o QIR

** Onlybb


9/21

9

3 Compact ring

The size of a detector is of the order of 3 to 5 m; to fit it inside a ring a

min imu m cir cum ference of ~ 20 m is needed, corresponding to a maxi-mum collision frequency of the order of 15 MHz . This relatively low value

of f implies th at t o obta in t he desir ed lum inosity ne or more of the follow-

ing parameters must be pushed to critical values: low beta; emittance;

bunch density; beam-beam tu ne shift.

The small bending radius can be obtained with superconducting

dipoles, whose high nonlinearities can limit the dynamic aperture, al-

ready critical due t o th e large emitt an ce.

The large synchrotron radiation power density on the vacuum cha m -

ber walls must be faced with special design of the vacuum chamber (for

example the antichamber at the UCLA ring).

A fun dam enta l problem of th e compa ct r ing is th e h igh rat e of par ticle

loss due to single beam-beam bremsstrahlung which shortens the lifetime

to few minutes. Continuous filling of the ring is necessary.

3.1 Novosibirsk - Factory

The Novosibirsk proposal is a very innovative idea. The ring is a bone-

shaped machine (see Fig. 3) where the two beams, in the single bunch

mode, circulate in opposite directions and cross at one point thanks to the

intr oduction of negat ive cur vatu re sections in t he a rcs.

Fig. 3 - N ovosibirsk-Factory Layout.


10/21

10

Round beams are used: a sc solenoid (11 T) around the IP rotates the

transverse planes by /2: the betatron oscillation normal modes are verti-

cal in one ar c an d horizont al in the oth er so that t he emitt an ces in the t wo

planes are the same. Beams are focused to very small beta values in both

planes by the same solenoid.

A very h igh valu e of (> 0.1) is claimed to be obta ina ble.

The working point is placed on the main coupling resonance x- y = 0. The tunes are shifted along the coupling resonance and do not

tr espass on t he 2-dimensiona l coupling sidebands.

SC d ipoles (6.5 T) ar e used to obta in th e sma ll bending r adius .

The beam lifetime of few minutes is considered satisfactory and man-

ageable with the injection chain, consisting of a full energy linac and a

positron cooling ring.

A later stage with 3 bunches per beam is envisaged, applying electro-

sta tic beam separa tion at the IP sides to rea ch lum inosities higher t ha n

1033 cm -2 sec-1.

3.2 UCLA - Fac tory

The UCLA -Factory design consists of a small machine of ~ 20 m cir-

cumference, equipped with SC dipoles. A sketch of the proposed layout is

given in Fig. 4.

Three successive phases are envisaged, with increasing luminosity.

The Compact Superconducting Ring (SMC) with L = 2 1032 cm -2 sec-1

working in single bunch mode, with low coupling ( = 0.2) and high emit-

tance.

The Quasi Isochronous Ring (QIR) with L = 1033 cm -2 sec-1 . The idea

is to shorten the bunch to the millimeter range by making the first order

momentum compaction vanish, enabling very small values of the vertical

betat ron fun ction @ IP , so th at , even with a lower cur ren t with r espect to

th e pha se I, a higher luminosity can be achieved.

The Linac against Ring option, finally, should lead to a luminosity

larger th an 1033 cm -2sec-1.


11/21

11

Fig. 4 - UCL A -Factory layout .

4 Double ring-multibunch operation

The choice of increasing the collision frequency by storing many bunches

in two rings leads to a design based on conventional technology, as far as

the lattice is concerned.

The short dist an ce between bun ches obliges the beam s to cross at a n

angle in order to suppress parasitic crossings.

The key problem are multibunch instabilities which must be faced

with feedback systems specially designed, rf cavity design with HOM sup-

pression and low impedance in all vacuum chamber components.

The discussion of the main characteristics of this kind of project will

follow the DANE criteria design. The KEK project will be shortly de-

scribed later, as it follows more or less the same criteria.


12/21

12

4.1 DANE - Frasca ti

DANE, the Frascati -Factory, is up to date the only funded project for

factories in the world.It consist of two rin gs, of about 100 m circum feren ce, int ersectin g a t

an angle in the horizontal plane in two points, to be fitted in the old

ADONE hall. It will accommodate two experiments, KLOE and

FI.NU.DA., the first one specially designed for the CP-violation investiga-

tion, the second one studying nuclear physics. A layout of the rings is

shown in Fig. 5. The ring is not symmetric with respect to the two IPs:

looking at the figure it is clear that because of the crossing angle the arcs

joining the two IPs are of different lengths.

Fig. 5 - DAN E layout

Flat beams ( = 0.01), multibunch operation and crossing angle are the

key feature of the project. The emittance is controlled by means of four

wigglers in each ring, located in the dispersive zone.


13/21

13

Being the two experiment detectors equipped with strong solenoids

(0.6 T x 2.1 m @KLOE a nd 1.5 T x 2 m @ FI .NU.DA), th e per tur bat ion of t he

machine optics obliges a coordinated design of the interaction region and

th e detector.

The interaction regions in between the two splitting magnets which

separa te the two beam tra jectories, are sh ar ed by the two rin gs. The m a g-

net ic axis of th eir elemen ts is th e bisectr ice of th e nomin al closed orbits of

the two rings. The two interaction regions are equivalent from the optics

point of view, i.e. with th e same first order tr an sfer ma tr ix, so that the op-

tical functions and the beam separation are the same at the Interaction

Region ends. This solution simplifies injection, tuning and chromaticity

corr ection in th e differen t modes of opera tion.

In the I nt era ction Regions the tr an sverse coupling due to the detector

solenoid is corrected by two compensating solenoids at the detector ends

whose total field integral is opposite to the detector one. In both experi-

ments all or some of the low beta quadrupoles are immersed in the

solenoidal field. To elimina te t he coupling effect qua dr up oles should be r o-

ta ted following th e a ngle given by:

sol =

Bz dz

2B

This continuos rotation is not feasible with a real quadrupole, so each

qua dru pole will be r ota ted by th e avera ge valu e ofsol along its length. To

eliminate the residual coupling four parameters are needed, and the rota-

tion angle of the three quadrupoles plus the field of the compensator

solenoid can be used to block-diagonalize the transfer matrix from the IPto the split ma gnet.

4.1 .1 Crossing angle

In multibunch configuration the reduced bunch to bunch distance forbids

head-on collisions and beams must cross with an angle. Crossing can be

in the horizont al or in t he vert ical plan e. When beams collide with a n a n -

gle synchro-betatron coupling arises: this has been proved in Doris11 a n d

by beam-beam simulations12 . This effect can be avoided with a crab


14/21

14

crossing scheme. The original idea was given by R. Palmer 13 for linear

colliders and it has been later applied to ring colliders 14. The principle is

the following: a transverse deflecting rf cavity, located /2 in -phase ad-

vance from the IP and with the zero-crossing of the electrical field syn-

chronous with the bunch center, gives to a particle an angular kick pro-

port iona l to its distan ce from t he bun ch center. This kick t ra nsform s int o a

transverse displacement at the IP, thereby making the two bunches tilt

and collide head-on. Another cavity, away from the first, restores the

original bunch orientation. This geometry allows a bunch spacing closer

than a normal head-on collision, and consequently a higher collision fre-

quency.

The required cavity voltage, assum ing th at t he r ms bun ch length z issma ll compa red t o r f/4, where r f is th e rf wavelength of the cavity, is:

Vr f=E

4er f

* c

being ha lf th e crossing an gle, * th e beta tr on function a t t he IP, c at the

cavity locat ion, both in t he crossing plan e.

This scheme r equires rf phase and amplitude stability:


15/21

15

It is a measur e of th e coupling between th e radial a nd the longitu dina l

phase spaces generated by the crossing angle. This coupling, experimen-

tally observed in DORIS I11 , limits the ma ximu m achievable tune shift

an d, consequent ly the luminosity.

The simplest hypothesis one can make on the relation between a and

is :

a = const

Suggestions on this value come both from observations at DORIS I and

computer simulations. At DORIS I the maximum value of the tune shift

ever obtained a nd t he value ofawere:

= 0.01 , a = 0.5

In DANE a ~ 14 , so th at a n angle of ~ 10 mrad seems safe from th is

point of view. Simu lat ions 15 of particle tracking including beam-beam in-

teraction with crossing angle confirm that no synchro-betatron dangerous

coupling occur s u p t o an a ngle of ~ 100 mra d.

The effect of parasitic crossings must also be considered: when two

beams collide at an angle 2 th e first par asit ic crossing (p.c.) occur s at dis-

tance l/2 (l is the spacing between bun ches) from th e IP, being the bu n ches

cent ers a t a dista nce given by d = l.

The tu ne sh ift due t o the par asitic crossing is:

x =ro Nx

2d2 y =

ro Ny

2d2

where x, y are the betatron functions at the p.c. From the consideration

th at one beam could act as a scra per for the opposite beam par ticles pa ss-

ing inside its core, a criteria ha d ar ised16 : the full separation should be at

least 7so that particles one wishes to keep in the beam stay out of the

opposite beam center.

To check this criteria simulations have been done with DANE pa-

rameters17

, using t he a pproximat ion of weak-str ong int eractions.


16/21

16

Two differen t effects ha ve shown:

- Blow-up in th e vert ical emitta nce for par ticles inside the core wh ich af-

fect t he luminosity (y increase).

- Long vert ical tails for par ticles which passes n ear the core of the oppo-site bunch which can affect the beam lifetim e.

Both t he effects depend on th e cur rent , on th e beam separa tion, an d on

the betatron functions at crossing, but we got no evidence of the scraper

effect. In fact a particle passing inside the opposite beam gets a tune shift

which affects its betatron motion, but not necessarily cuts it out, also

because a particle which sees the other bunch at one turn will probably not

pass inside it again in the successive turns during a betatron damping

t ime .

With DANE design parameters 7 at the p.c. is a safe number, butfor example with a higher current a separation of 7 would affect beam

lifetime.

4.1.2 Rf and feedback system

The rf system mu st face one of th e most ha rm ful problems in high cur r en t

storage rings working in multibunch operation: the high order modes

(HOMs) of the rf cavities induce coupling of the relative motion of the n b

bunches through the wake fields left behind in the cavity by the bunches

themselves giving rise to multibunch instabilities. The instability growth

rate must be kept below the radiation damping rate by a proper cavity de-

sign a nd a n a ppropriate feedback system.

The rf voltage necessary to compensate the energy losses due to syn-

chrotron radiation and to HOMs is not very high (


17/21

17

Fig. 6 - DAN E cavity sha pe.

Even with HOM damping or frequency shift the rise time of coupled

bunch (cb) modes of oscillat ion can be much faster th an the n at ur al r a dia -

tion and Landau damping time; moreover the probability for a damped

HOM to cross a cb mode frequency is larger, due to the wider bandwidth.

An all-mode feedback system (f.s.) capable to damp all the cb modes and

the injection transients is then necessary.

In most operating rings the usual approach consists in detecting the

bunch synchrotron phase error, rotating the signal in the longitudinal

pha se plane with a filter a nd finally amp lifying an d giving an energy kick

with a kicker. This m ethod requir es a n um ber of filters equal to the n u m -

ber of bun ches, which of cour se becomes very dema nd ing in m u lt ibu n ch

operation.

On the cont ra ry a bunch to bunch, time doma in system19 is now possi-

ble with the available electronic technology: in a mixed analog/digital f.s.

the Digital Signal Processors (DSP) filters can implement several different

channels, reducing the overall complexity. Moreover the digital system

can be programmed in such a way as to maintain the correction signal

just below the saturation limit of the power amplifier even in presence of

large ph ase excur sions. A modular appr oach allows to implement the sys-

tem for a redu ced number of bunches an d to updat e it when a h igher nb is

needed for th e luminosity upgrade.


18/21

18

4.2 KEK-Factory

The two beams are stored in two rings super posed an d h orizont ally cross-

ing with an angle. The lattice provides flat beams with high currents and

large emittance.

The rf frequency is 1.5 GHz. Every two buckets are filled to avoid very

large values ofx.

Fig 7 - KEK -Factory layout.

4.3 -Factory Design Studies at Mainz

Studies for designing a -factory are being developed at Mainz. The pro-

posal is a compromise between the compact ring and the double ring oper-

at ion. The lat tice consist s of two rin gs inter sectin g at one point (see F ig. 8).

Beams are round and following the Novosibirsk idea they are both focused

an d rotat ed by a su per cond uct ing solenoid; few bunches per bea m (nb = 6)

give a collision frequency of 50 MHz. The shor t bunch len gth (z = 1 cm )

should be obtained with a special insertion made of dielectric material

which should focus the beam longitudinally. A high design value is

proposed: = 0.08, checked with beam-beam simulations.


19/21


20/21

20

b. S ingle bunch ef fect s

- Touschek effect: when scattered part icles ins ide the bunch get a

longitudinal momentum variation they can exceed the momen-

tu m acceptan ce of th e ring, or if the scatt ering occur s in a disper -

sive zone the induced horizontal oscillation can exceed the trans-

verse acceptance of the ring. This effect is very strong for low en-

ergy rings, especially for very flat beam s becau se of its depen den ce

on the bunch density and is in fact the main limitation for

multibunch operation rings working with flat beams.

- Multiple Touschek effect: scatt ering between part icles ins ide the

bunch tr an sfers momen tu m between the planes to which can be

associated an increase in the transverse emittance which could

exceed the tr an sverse a ccepta nce, eith er physical or dyna mic. I n

-factories this effect is not of much concern because the emit-

tance is already large and its increment is only of few percent,

well cont ained in t he r ing acceptan ce.

c . Beam-gas in teract ion

The scattering between beam particles and residual gas molecules

can be elastic, producing betatron oscillations that can exceed the

transverse acceptance, or inelastic (bremsstrahlung), producing

energy oscillations exceeding the longitudinal acceptance.

Depends on vacuum and apertures: it is more demanding in small

rings because of the higher synchrotron radiation pressure.

d . B ea m -b ea m b rem s st ra h lu n g

The beam-beam decay rate is proportional toL/(n b N). In single bunch

modes and at high luminosity the beam lifetime is reduced to few

minu te s .

6 Injection

The total number of positrons required to obtain the quoted luminosity is

very high: 1011 1012 for compact rings and ~ 1013 for double rings.

Furthermore reliable operation asks for a short injection time, typically

few minutes for the double ring configuration, 1 minute for the compact

ring, to store the full current of both beams. It is clearly convenient to

inject at full energy.


21/21

21

An accumulator, at the same energy of the main ring, is usually pro-

posed to accumulate at lower fr f to accept LINAC pulse. It is used also to

damp the e+ emitta nce and ener gy spread t o values acceptable by the m a in

ring system a nd t o equalize th e two beam cha ra cter istics.

References

1. P.M. Ivanov et al, Luminosity an d the Beam-Beam Effects on the Electr on-Positr on

Storage Ring VEPP -2M with Su perconductin g Wiggler Magnet , Pr oceedin gs of the

3rd Advanced ICFA Beam Dynamics Workshop, Novosibirsk, 1989, p. 26.

2. D.L. Rubin and L.A. Schick, Single In ter action Point of Opera tion of CESR, 1991

Particle Accelerator Conference, San Francisco, USA, 1991, p.144.3. J .T. Seeman , Observations of the Beam-beam In teraction, Nonlinear Dyn a m ics

Aspects of Pa rt icle Accelerat ors, S. Mar gher ita di Pu la, I ta ly, 1985, Lectu re Notes i n

Ph ysics 247, p.121.

4. M. Bassett i , What a re the Beam-beam limits in DANE?, 1992, Unpublished.

5. L. Bark ov et al., Novosibirsk P roject of-Meson F actory, Pr oceedings of Work sh op

on Ph ysics an d Detectors for DANE - Fra scati, April 9-12 1991, p.67.

6 . A. S tr eu n , -Factory Stu dies at Main z, P roceedings of Work shop on Ph ysics a n d

Detectors for DANE - Fr ascati, April 9-12 1991, p.99.

7. A.N. Filippov et al., P roposal of the Round Beam Lat tice for VEPP -2M Collid er ,

Present ed at the XVth International Conference on H.E.A., Hamburg, 1992.

8 . C. Pel legr in i , The UCLA -Factory Project, Proceedings of Workshop on Physics

an d Detectors for DANE - Fr ascati, April 9-12 1991, p.83.

9 . DANE Project Team, DANE Status Report, Proceedings of 3 r d EP AC Conferen ce -

Berlin, 1992.

10. K. Hirata and K. Ohmi, Feasibili ty of a -Factory in KEK, 1991 Particle Accelerator

Conference, San Francisco, USA, 1991, p. 2847.

11. A. Piwinsk i, IE EE Tra ns. on N ucl. Scie. NS-24 , 1408 (1977).

12. A. Piwinski - "Synchro-betatr on resonan ces" in Nonlinear Dynam ics Aspects of

Particle Accelerators, S. Margherita di Pula, Italy, 1985, Lecture Notes in Physics

247, p.104.

13. R.B. Pa lmer - SLAC - PUB - 4707 (1988).

14. K. Oide and Y. Yokoya, Ph ys. Rev. D40,315 (1989).

15. M. Bassett i an d M.E. Biagini, private commu nication.

16. D.H. Rice, Beam-Beam In ter action: Experiment al, AIP Conf. Proceedings 214,

Berk eley 219 (1990).

17. M. Bassetti and C. Biscari, Beam-beam simulat ions with pc, un published.

18. P. Baldini et al . - DANE Accelera tin g Cavit y: R&D - Pr oceedin gs of 3rd EPAC

Conference - Berlin, 24-28 March, 1992.

19. M. Bassetti et al - DANE Longitudinal Feedback - Proceedings of 3rd

EPACConference - Berlin, 24-28 March, 1992.

Documents

C. Biscari- Phi-Factory Design