Ex(t),Ey(t),Ez(t) for a duroid 9 stack-pair rectangular stack
SIMULATIONS
2
1.5
As a prelude to pushing particles through the
1
computational grid, the electric (E) and magnetic (H)
0.5
0
fields were extracted from the data. In these snapshots
-0.5
(see Figure 7 a-f), the switches were located out of the
-1
-1.5
page and closed at time t=0 ns. The collapsing wavefront
-2
in one Blumlein layer in each stack pair approached the
-2.5
beampipe and the electric fields there added. Note that
-3
-3.5
the integral of E along any off-axis line is the same as the
-4
0
10
20
30
40
5
15
25
35
45
50
integral on axis.  The extra beampipe left/right of the
time (ns)
stack was needed for computational reasons so that
Ez(z,t=tpeak) for a duroid 9 stack-pair rectangular stack
Ez(z=zmid;zb;zc, t) for a duroid 9 stack-pair rectangular stack
evanescent fields were allowed to decay before interacting
2
2
1.5
1.5
with the left/right boundary conditions. Figure 7 shows a
1
1
0.5
0.5
short stack.
0
0
-0.5
-0.5
-1
-1
-1.5
-1.5
-2
-2
ferrite
-2.5
-2.5
-3
-3
-3.5
-3.5
-4
0
20
40
80
100 120 140 160 180 200 220 240 260 280 300
-4
60
0
5
10
15
20
25
30
35
40
45
50
position along the beampipe (millimeters)
vacuum
time (ns)
beam direction
Figure 8: taking cuts across the beampipe at three spatial
oil
locations versus time shows excellent pulse stability and
fidelity across the beampipe
CONCLUSIONS
DWA structures have been successfully simulated using
finite difference time domain tools which also greatly aid
in  designing  the  structures.
Double-ended  feed
configurations allow for balanced bias H fields in the
beampipe region. The presence of the equilibration ring
allows for energy to be distributed around the beampipe
for uniform excitation and avoidance of dipole modes.
Since the structure is modular, assembling large stacks is
straight forward and 7-, 9-, and 20-stack units have been
demonstrated.
ACKNOWLEDGEMENTS
The authors would like to thank Jim Sullivan (LLNL)
for providing the experimental data (Figure 5) and for
supporting the modeling and design efforts with his
insights on "real" data and the kinds of structures that can
actually be fabricated.
Figure 7 a-f: as the collapsing blumlien wavefront reaches
the beampipe, the fields now add thus producing
acceleration in the beampipe.
The bottom figure
REFERENCES
(horizontal) shows the view looking down the beampipe.
[1]
REMCOM Corp., 315 S. Allen St., Suite 222,
Pulse fidelity
State College, PA 16801
For this configuration, it is straight forward to achieve
[2]
J.  S.  Sullivan,  E.  J.  Gower,  "Compact,
rectangular pulses with good fidelity (see Figure 8) as
Repetitively Pulsed, Stacked Blumlein Pulser," Lawrence
long as the switch closure time is not excessively short.
Livermore National Laboratory, UCRL-JC-147159, 2002.
Fast dV/dt switch effects couple directly from adjacent
[3]
W. C. Nunnally, et. al., "Investigation of UV
layers and degrade the baseline of the pulse since energy
Laser  Triggered,  Nanosecond,  Surface  Flashover
is coupled from the switched line to the unswitched line
Switches," Lawrence Livermore National Laboratory,
and thus acts as a baseline offset for the fields in the
UCRL-JC-151646, 2003.
beampipe.