RFQ simulation results

1
RFQ simulation results

• Gennady Romanov
• September 7, 2006

2
We committed to perform 3D simulations of RFQ
with radial matchers.
Solid model from MWS
Original solid model
Input/output radial matcher are introduced
3
Original vane ends shape without matchers
Original vane end shape. 319 MHz
Original vane end shape with fingers. 318 MHz
This low frequency is OK with and without tuners,
because plunger tuners will increase frequency.
4
Input matcher
The frequency of the end area is higher then
nominal, because the matcher reduces capacitive
loading. The fingers dont help much they lower
frequency only by 1 MHz (327.8 -gt 326.8) for
given model length of 150 mm.
5
Mesh density enhancement
In MWS mesh density may be defined separately for
separate bodies. Introducing fake metal tips
(rods) one gets an opportunity to dense mesh
in critical area without limitations of
automeshing . This trick makes the simulations
of full or ½ length models more reliable.
J.Staples et al presented at Linac06 the detail
simulation of SNS RFQ, and they think that MWS
is already accurate enough to simulate RFQs.
6
Model of ½ of full length with input matcher
0.25 MHz for 30 mm stroke
Nominal
Nominal
Field distribution and frequency of Q-mode vs
finger length. Field is evaluated along line with
xy2.4 mm
At nominal length of tuners 24.5 mm Fq
324.1 MHz Tilt 5.5
½ of full length, magnetic boundary at the open
end
7
Model of ½ of full length with output matcher
18
Field is evaluated along line with xy2.4 mm
At nominal length of tuners 24.5 mm Fq
324.6 MHz Tilt 18
½ of full length, magnetic boundary at the open
end
8
Model of ½ of full length with output matcher
0.18 MHz for 30 mm stroke
Field distribution and frequency of Q-mode vs
finger length
Field is evaluated along line with xy2.4 mm
Output matcher has higher frequency and it is
less sensitive to end-wall tuners.
½ of full length, magnetic boundary at the open
end
9
Field distribution and frequency of 2nd dipole
mode vs finger length
1MHz for 30 mm stroke
Dipole mode spectrum. Finger length is one inch.
Fq 325.4 MHz Fq-Fd2 ? 3 MHz
10
Conclusion. What could be done with the matchers
The bigger rounding off decreases frequency too
and makes surface current distribution more
uniform
Cut-backs can be increased slightly to bring
frequency down. The cut-backs can be done of
rectangular shape, or deeper, or both.
These additional tuners are not very effective.
But can help. Diameter 20 mm, h 5.26 mm