Analysis of meander line deflecting structure for a traveling wave chopper

1
Analysis of meander line deflecting structure
for a traveling wave chopper

• G.Romanov/R.Madrak
• March 16, 2006

2
Chopper for the MEBT

• Point of the chopper
• Linac Bunches are spaced by 325 MHz (3.1ns)
• In MI, bunch spacing is 53 MHz (19ns)
• Dont want bunches in the 53 MHz separatrix
• gtChop out 1 of every 6 bunches
• - Additional complication 325?n53
• gtSometimes chop 1, sometimes 2
• Only 3ns between bunches gt use a traveling wave
  deflector which follows the beam
• Since b0.073 _at_2.5 MeV, need to slow down pulse
  gt use meander wave structure
• Giorgio obtained a prototype at CERN (from Fritz
  Caspers)

3
Basic layout of a double meander type deflector
3 MeV, b 0.08
Metallization Molybden/mangan silver
Stolen from F.Caspers
4
How does it work
We need to redesign the meander structure for
b 0.073 and check a coverage factor.
5
Pulse velocity along single strip
Simulated impedance is 90 W for W 0.45 mm and
45 W for W3 mm Simulated b 0.425. For strip
line of length L42.5 mm resonant frequency is
GHz,
which is the lowest measured resonance of
prototype structure.
6
Pulse velocity along beam path I.
Pulse velocity along strip is still too high. To
synchronize beam and pulse the meander line has
been proposed. But it turns out that the simple
formulas to define parameters of meander do not
work
L42.5 mm, d 6 mm (period of the meander)
SNS chopper
7
Pulse velocity along beam path II Double meander
as a periodic structure
In periodic structures a propagation of a signal
in terms of space harmonics is described by
dispersion equation or Brillouin diagram
Energy (pulse) propagates along structure with
velocity
We want bgroup 0.073 in region close to 0-mode
of propagation.
Only ¼ of a cell is needed to calculate b
for whole structure. Its good for optimization.
8
Parameters to play
In general we have two parameters to get desired
group velocity strip resonance frequency and
coupling between strips, which is standard for
any periodic structures. In practice both depend
on a number of geometrical and material
parameters dielectric constant, shape of
dielectric substrate, strip dimensions, distance
between strips, gap between platesThis
dependence may be a problem. For example lets
consider gap
Depending on the gap we need strip length
43.7-44.4 mm According to the simulations Fritzs
structure is not of beta0.08, but 0.075-0.076.
Whatever the reason is for such behavior,
most geometrical parameters should be fixed
before optimization.
9
Direct measurement of pulse velocity in a long
structure (final check).
Direct measurement is clear, but it has two
difficulties 1) Velocity of center, back and
front of pulse are different. For the given plot
b_front0.073, but b_center0.075. Usage of
gaussian pulse may be a solution. 2) In real
structure with imperfect ends pulse velocity
depends on structure length. We have to simulate
pretty long structures to get the right numbers.
Gaussian pulse
10
Coverage factor.
Coverage Factor Em/E0 where Em Electric
field seen by beam due to meanders E0
Electric field seen by beam due to perfect
conducting planes, with same V and gap
Field along axis in median plane
Gap 6 mm
Gap 12 mm
E0 V/gap ?
11
Conclusion

• Since the pulser is the most difficult (and
  expensive) part, this will be the driving factor
  in choice of parameters gap, length, meander
  period, meander width
• With Petr Ostroumovs parameters (deflection
  EL2/2bp)
• gap 16mm, meander length 50cm, deflection
  6mm
• gt Need V 1.9kV (perfect coverage)
• gtReally need V 2.5 kV (coverage factor 0.75)
• Kentech 500V prototype pulser in progress 50k
  (one pulser)
• They would like to make a higher voltage one,
  but so far price unknown
• (said I am reasonably confident that 2kV is a
  possibility)

12
Conclusion, contd

• Considerations
• Fritzs pcb meander
• More simple mechanically
• We have a prototype on hand, and have performed
  measurements
• Can be manufactured commercially
• But, has a lower coverage factor than the SNS
  one shown earlier (gt higher voltage pulser)
• Should we work this or continue with the pcb?
• To proceed with simulation
• Choose gap, other structure parameters this
  determines coverage factor and required voltage
• Fine tune to get 50 Ohm impedance
• Tune to get correct b, width 42.5 -gt 43.7-44.3
  mm
• minimizing dispersion gt shorter structure, more
  deflection gt longer structure