Some%20Timing%20Aspects%20for%20ILC

1
Some Timing Aspects for ILC

• Heiko Ehrlichmann
• DESY
• GDE, Frascati, December 2005

2
central componentthe damping rings

• are defining the ILC timing
• are defining some global ILC parameters
• could provide some flexibility
• together with an undulator based positron
  generation also the ILC geometry is influenced
• gt damping ring parameters should be well chosen!

3
some general points

• the damping ring circumference C is given by the
  HF wavelength lHF and the harmonic number h (
  number of HF-buckets)
• C h lHF (assuming that all particles have
  the speed of light)
• usually not all HF buckets are used, but only a
  fraction, giving the number of equally spaced
  buckets NB (not mandatory)
• NB h / i
• the bucket distance in the damping ring t(DR) has
  to be a multiple of one HF bucket distance
  tHF(DR)
• t(DR) i tHF(DR) i / fHF(DR)
• bunch distance in the damping ring t(DR) is much
  smaller than bunch distance in the LINAC t(L)
• gt compressed storage of the desired LINAC
  pulse train
• to produce the pulse train bunch structure the
  damping ring ejection will run with a certain
  feed k (-gt decompression)
• t(L) k t(DR)
• the ejected bunch pattern has to fit to the LINAC
  HF buckets
• t(L) j tHF(L)

4
damping ring ejection

• as an easy example
• damping ring for NB100 filled buckets
• compression factor ejection feed k10
• gt after one revolution an already emptied
  bucket will be reached
• step solution
• allow a step in k after each revolution
• -gt in our example 9x k10, 1x k11, 9x k10, 1x
  k11 etc.
• gt different bunch distances in the LINAC pulse
  train, k not constant
• gt neglected (but still a solution)
• filled solution
• no common divider for k and NB
• always fulfilled for NB prime number or k
  prime number and not divider of NB
• -gt in our example e.g. NB 101 or k 11,
  but also k 9
• gt restrictions for NB and k NB p k e
• special case NB p k /- 1
• gt constant bucket feed per damping ring
  revolution of exact one bucket d k - e -/1

5
once the circumference is fixed

• filled solution (e.g. TESLA TDR)
• both rise time and fall time of the ejection
• (and injection) kicker pulses must be shorter
• than the bucket distance t(DR)
• k can be changed, as long as the restrictions are
  satisfied
• ( simply the bucket feed d will vary)
• -gt flexibility in LINAC bunch distance and HF
  pulse length
• -gt in our example NB 100, k...,7 ,9, 11
  ,13 ,17 ,....
• NB can be changed, as long as h NB i stays
  constant
• -gt flexibility in DR bunch distance and number
  of bunches
• gt h should contain many dividers i
• a desired gap in the LINAC pulse train can be
  produced with single missing bunches in the
  damping ring
• an artificial single gap of empty buckets in the
  damping ring would transform into missing single
  bunches in the LINAC pulse train
• special case p equidistant gaps, fixed
  ejection feed k and NB p k /- 1
• -gt gap solution the gaps are transformed to a
  shorter overall bunch train

6
once the circumference is fixed

• gap solution (fixed bucket feed per revolution d
  empty buckets)
• k can be changed, as long as p k stays constant
• -gt in our example NB 101, k p10 10, 5 20,
  20 5, 50 2, 2 50
• -gt some flexibility
• bucket number NB is nearly fixed h i NB i
  (p k 1) const
• p artificial gaps of empty buckets can be
  implemented without creation of missing single
  bunches in the LINAC pulse train
• ejection of always the bunch before the gap
• -gt more freedom for kicker pulse needs
• -gt only the rise time of the ejection kicker
• pulses must be shorter than the
• bucket distance t(DR) (with
  re-injection bucket feed g)
• as before
• a desired gap in the LINAC pulse train can be
  produced with single missing bunches in the
  damping ring

7
damping ring HF frequency

• as mentioned already above
• the ejected bunch pattern has to fit to the
  LINAC HF buckets
• t(L) j tHF(L) k t(DR) k i tHF(DR)
• gt for flexibility in NB, especially for the
  filled solution, a good choice of the HF
  frequency of the damping ring is important
• fHF(L) / fHF(DR) j / (k i)
• gt flexibility in k and i is determined
• fHF(L) given 1.3GHz examples
• j / (k i) 2 fHF(DR) 650MHz
• 3 433MHz
• 4 325MHz
• 5/2 520MHz
• 13/5 500MHz
• (not necessarily equal in both damping rings)

8
different circumferences

• the collisions should always take place at the IP
• gt the bunch train structure must be equal
• t(L)e t(L)e-
• ke ie tHF(DR)e ke- ie- tHF(DR)e-
• C h c tHF(DR) i NB c tHF(DR)
• for the step solution
• impossible, since the steps in k would appear at
  different bunch train positions
• for the filled solution
• possible if both NBe and NBe- are prime
  numbers
• -gt by definition without flexibility in NB
• -gt missing bunches in the larger ring
• -gt in our example NBe- 101, NBe 199 for
  a nearly doubled circumference
• impossible, if k should be a prime number

9
different circumferences

• ke ie tHF(DR)e ke- ie- tHF(DR)e-
• for the gap solution
• with unchanged bunch distance
• tHF(DR)e tHF(DR)e- , ie ie- gt ke
  ke-
• pe z pe-
• (z circumference factor, not necessarily an
  integer)
• gt Ce z Ce- (1 - z) i lHF gt possible
• (e.g. Ce 2 Ce- impossible, but Ce 2 Ce- -
  i lHF )
• with changed bunch distance
• t(DR)e z t(DR)e- gt ke ke- / z
• gt pe z pe-
• gt Ce z Ce- gt possible
• (e.g. Ce 2 Ce- possible )
• restriction k must be dividable by z

10
example for flexibility
long damping ring C 17.434km (h 25200,
fHF(DR ) 433MHz) some of the possible operation
parameters
k 17 19 23
THF(ms) 0.988 1.105 1.338
i NB t(DR) (ns) t(L) (ns) t(L) (ns) t(L) (ns)
3 8400 6.92 117.7 131.5 159.2
4 6300 9.23 156.9 175.4 212.3
5 5040 11.54 196.1 219.2 265.4
6 4200 13.85 235.4 263.1 318.5
7 3600 16.15 274.6 306.9 371.5
8 3150 18.46 313.8 350.7 424.6
9 2800 20.77 353.1 394.6 477.7
10 2520 23.08 392.3 438.5 530.8

• high flexibility in
• number of bunches
• bunch distance in the DR
• bunch distance in the LINAC
• overall bunch train length
• -gt just by changing the DR timing between two
  cycles

11
examples for flexibility
C 6.477km (h 10802, fHF(DR ) 500MHz) NB
5400 , t(DR) 4ns some of the possible operation
parameters

• some flexibility in
• bunch distance in the LINAC
• resulting changes in bunch number
• -gt also just by changing the DR timing between
  two cycles
• by (trivial) omitting of bunches the number of
  bunches is changed, but not the number of buckets
• the bunch distance in the damping ring is fixed

p k t(L) (ns) THF(ms) N (e20)
60 90 360 1.512 4200
72 75 300 1.188 3960
90 60 240 0.864 3600
108 50 200 0.648 3240
C 6.643km (h 14403, fHF(DR ) 650MHz) NB
4800 , t(DR) 4.61ns some of the possible
operation parameters
p k t(L) (ns) THF(ms) N (e20)
60 80 369.2 1.329 3600
64 75 346.1 1.218 3520
75 64 295.4 0.974 3300
80 60 276.9 0.886 3200
96 50 230.8 0.664 2880
12
consequences up to now

• if some flexibility is required the damping ring
  parameters
• HF frequency
• harmonic number and thus circumference
• should be well chosen
• filled solution
• more changes in global parameters allowed during
  operation
• the circumference is given by the kicker pulse
  needs
• both rings should have the same circumference
• probably two long rings are better
• gap solution
• easier adjustment for kicker pulse needs
• (especially for asymmetric pulse shapes)
• requires fixed number of buckets
• different circumferences are possible

13
re-injection

• with independent particle sources the
  re-injection of ejected bunches can be done every
  time (between immediately or after one complete
  ejection cycle)
• in the gap solution one might refill with
• a deliberate bunch feed d for the ejection
• kicker pulse needs
• with the undulator based positron source the
  particle generation time is given by the
  electron LINAC timing
• always an already ejected bucket has to be
  refilled
• gt the path length of the positron transport line
  must fit to the damping ring timing
• most flexible solution self reproduction
• gt an ejected positron bunch is refilled by its
  own electron partner
• essential for single bunch ejection
• (commissioning scenario, pilot bunches, machine
  protection system.....)

14
ILC geometry (positron part)

• C circumference of the damping ring
• L distance between the IP and the beginning of
  the linear tunnel (BDS, LINAC, BC)
• T1 distance between the IP and the damping ring
• T2 distance between the damping ring and the
  beginning of the 180 return arc
• B pass length of the 180 return arc
• A linear tunnel length between both 180 return
  arc ends
• b additional path length for the IP bypass
  line, artificial detours somewhere in the
  positron transport line or other reasons (e.g.
  also for particle velocity differing from c)
• D damping ring bucket feed length for the
  re-injected bunch
• (D0 for self reproduction in the filled
  solution)
• gt n C D 2 L (B-A) b

15
path length restriction

• n C D 2 L (B-A) b
• independent of the damping ring shape or position
  along the LINAC
• valid for all ILC stages 500GeV, 1TeV upgrade
  ...
• (B-A) b -gt detour path lengths, e.g. due to
  180 return arc
• -gt small in comparison to U or L
• -gt geometry can be used for path length
  adjustment
• for self reproducing fills (filled solution) D
  0
• gt strong geometry restriction, but high
  operation flexibility
• for non-reproducing filled solution D multiple
  of k i lHF
• gt geometry restriction reduced, but operation
  flexibility also (k fixed)
• for the gap solution D multiple of (g k) i
  lHF
• gt operation flexibility is reduced anyhow (k,
  NB), some geometry restrictions, also the bucket
  feed g will be fixed
• in general all geometry conditions can be
  satisfied with artificial detours, maybe
  adjustable during operation -gt costs?

16
some additional comments

• second IP at a different longitudinal position
• the path length equation is valid for both IPs
• 1. IP distance c t(L) c k t(DR)
• gt k fixed by geometry
• 2. switch able additional path length in the
  positron transport, just compensating the IP
  distance effect
• HF frequency changes during the damping times
  could be used for shifting bunch patters between
  two LINAC pulses, but are not able to relax the
  flexibility or geometry restrictions
• the exact IP position can be adjusted by the
  LINAC HF phases
• since the damping ring HF phases must fit to the
  corresponding LINAC HF phases and the positron
  generation phase is determined by the electron
  LINAC HF phase, an adjustment of the damping ring
  injection phase is only possible with positron
  path length adjustment

17
conclusions

• if the overall kicker pulse length can be small
• and a high flexibility in operation parameter
  choice is required
• gt long damping rings with equal circumference,
  every bucket filled
• accept strong design parameter restrictions
• if the kicker pulse fall time is expected to be
  long
• gt gaps for the kicker pulse needs
• accept the reduced flexibility

18
only one kicker system

• in case of circular damping rings (small tunnel
  overlap with the LINAC) one kicker system could
  be used for ejection and re-injection (two
  independent septa)
• the timing shift between the ejection pulse
  sequence and the re-injection pulse sequence is
  constant for the gap solution and depends on the
  kicker feed k for the filled solution
• a double kicker system with two parts and 180
  phase advance and the septum in between could be
  used for compensation of long kicker pulses (180
  bump)
• a special case would be the synchronous ejection
  and re-injection within one kicker pulse, using
  e.g. neighbored HF buckets (by definition a bunch
  pattern cannot be self reproducing)
• gt the path length has to be well adjusted to
  the corresponding special set of parameters (no
  flexibility)

19
damping ring position

• long damping rings, filled solution
• due to cost reasons maybe a dogbone shape is
  preferable, using the LINAC tunnel
• gt one of each bending sections can be used for
  the required (and independent) 180 return arcs
• gt damping rings at both ends of the ILC
  preferable
• for damping rings in a separate tunnel the
  position is a free parameter
• maybe cost reduction by putting both rings in
  the same tunnel
• then coupling possibilities
• -gt use both rings together for electrons or
  positrons
• -gt use the electron ring for positrons in case
  of technical problems with the positron
  ring
• -gt operate the positron ring with electrons
• during commissioning
• as keep alive solution, when positrons are not
  available

20
switch yard

• for using the e- ring with e
• for using the e ring with e-
• bypass lines
• mirrored view for better visibility, both rings
  of course in tunnel

• all switches can be slow (DC)
• only for using both rings for positrons a fast
  switch is needed