Grudiev 17.06.2009
description
Transcript of Grudiev 17.06.2009
T24_vg1.8_disk 11WNSDVG1.8
CLIC_G un-damped @ 11.424 GHzmeasurements versus simulations
(preliminary)
A. Grudiev17.06.2009
Acknowledgements
CERN:M. GerbauxR. ZennaroA. OlyuninW. Wuensch
SLAC:Z. Li
First cellHs/EaEs/Ea Sc/Ea
2
a [mm] 3.307
d [mm] 1.753
e 1.16
f [GHz] 11.433
Q(Cu) 6814
vg/c [%] 1.83
r’/Q [LinacΩ/m] 15198
Es/Ea 1.95
Hs/Ea [mA/V] 2.6
Sc/Ea2 [mA/V] 0.37
Middle cellHs/EaEs/Ea Sc/Ea
2
a [mm] 2.887
d [mm] 1.402
e 1.15
f [GHz] 11.432
Q(Cu) 6980
vg/c [%] 1.33
r’/Q [LinacΩ/m] 16960
Es/Ea 1.95
Hs/Ea [mA/V] 2.45
Sc/Ea2 [mA/V] 0.34
Last cellHs/EaEs/Ea Sc/Ea
2
a [mm] 2.467
d [mm] 1.05
e 1.13
f [GHz] 11.426
Q(Cu) 7157
vg/c [%] 0.92
r’/Q [LinacΩ/m] 18713
Es/Ea 1.9
Hs/Ea [mA/V] 2.3
Sc/Ea2 [mA/V] 0.28
6
Gradient in 24 regular cells
Number of regular cells: Nc 24Bunch population: N 3.72x109
Number of bunches: Nb 312 Bunch separation: Ncycl 6 rf cycles
Average unloaded of 100 MV/m Average loaded of 100 MV/m
0 4 8 12 16 20 24240
50
100
150
200
250
iris number
P [
MW
] (b
lack
), E
s (
gre
en),
Ea (
red
) [M
V/m
],
T
[K
] (b
lue)
, S
c*50
[MW
/mm
2 ] (m
agen
ta)
7.5 8.4
176
205
3.03.2
90
108
41.1
23.4
Pinload = 41.1 MW, P
outload = 23.4 MW
Eff = 0.0 % tr = 0.0 ns, t
f = 0.0 ns, t
p = 100.0 ns
0 4 8 12 16 20 240
50
100
150
200
250
iris number
P [
MW
] (b
lack
), E
s (
gre
en),
Ea (
red
) [M
V/m
],
T
[K
] (b
lue)
, S
c*50
[MW
/mm
2 ] (m
agen
ta)
15.0 16.8
206
240
4.1
4.5
105
126
56.4
32.2
Pinload = 56.4 MW, P
outload = 14.7 MW
Eff = 28.9 % tr = 20.7 ns, t
f = 54.8 ns, t
p = 238.8 ns
2
1'PI
Q
R
vP
Qvdz
dP
gg
Simulation setup 1 & 2HFSS-quad
HFSS simulation of ¼ of the structure:Surface conductivity (Cu): 58e6 S/mSurface approximation: ds=5 μm, Total number of tetrahedra: Ntetr = 1024991Mesh density: ~ 35000 tetr / ¼ cell
S3P-quadS3P simulation of ¼ of the structure made by Zenghai Li at SLAC-ACDSurface conductivity (Cu): 57e6 S/m different
HFSS v11.1
Simulation setup 3 HFSS-cells + couplers
HFSS simulation of ¼ of input coupler + segment of 5 deg. of the cells + ¼ of output coupler :Surface conductivity (Cu): 58e6 S/mSurface approximation: ds=1 μm, Total number of tetrahedra for cells: Ntetr = 506075Mesh density: ~ 350000 tetr / ¼ cell (10 times higher than in HFSS-quad setup)
HFSS v11.1
Measurement setup 1 (reflections)
Perfectload
Perfectload
VNAport1
VNAport2
1
2 3
4
Reflection = S11+S12
Frequency correction due to air (ε = 1.00059) Frequency in vacuum: f’ = f*sqrt(1.00059)
Reflection: comparison
• There is a shift up in frequency of ~3.5 MHz due to air correction in the measurements• The HFSS simulation results are about 4 MHz higher than air corrected measurements• The S3P simulation results are about 4-5 MHz lower than air corrected measurements• There is very small (~1MHz) or no difference between S3P simulations and the air corrected measurements
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.44-45
-40
-35
-30
-25
-20
-15
-10
-5
0
f [GHz]
Ref
lect
ion
[dB
]
input measurementoutput measurement
input measurement - Air corrected
output measurement - Air corrected
input HFSS-cellsoutput HFSS-cells
input HFSS-quad
output HFSS-quad
input S3P-quadoutput S3P-quad
1
2 3
4
Measurement setup 2 (transmission)Perfect load
Perfect load
VNA port1
VNA port2Transmission = S13+S23
1
2 3
4
Perfect load
VNA port2VNA port1
Perfect load
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.44-4
-3.8
-3.6
-3.4
-3.2
-3
-2.8
-2.6
-2.4
-2.2
-2
f [GHz]
Tran
smis
ion
[dB
]
S3P-quad
HFSS-quadHFSS-cells
measurement: S13+S23
measurement: |S13|+|S23|
measurement: 2*S13measurement: 2*S23
measurement: Air corrected
Transmission: comparison • There is a shift up in frequency of ~3.5 MHz due to air correction in the measurements• The HFSS simulation results are about 4 MHz higher than air corrected measurements• The HFSS simulation results shows more transmission by about 0.3 dB at 11.424 GHz• The S3P simulation results shows less transmission than HFSS by about 0.15 dB at 11.424 GHz
Some useful equationsnattenuatio ln 12
2 where SτePP inout
structure lossfreein energy stored -
structurelossy in energy stored -
delay group timefilling - )arg(
factorquality average -
0
120
where
W
W
d
Sd
P
Wt
P
WQePP
inf
loss
Q
t
inout
f
On the other hand
Finally
2
ftQ
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.4450
52
54
56
58
60
62
64
66
68
70
f [GHz]
d /
d
[ns
]
measurement
measurement: Air correctred
HFSS-quadHFSS-cells
S3P-quad
Transmission: comparing group delay
• There is a shift up in frequency of ~3.5 MHz due to air correction in the measurements• The HFSS simulation results are about 4 MHz higher than air corrected measurements• There is no difference between S3P simulations and the air corrected measurements
Transmission: comparing Q-factor • There is no difference in Q-factor (~7000) between HFSS and S3P simulations.• The measured Q-factor of about 6600 is lower than the simulated value by about 6 %.
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.445800
6000
6200
6400
6600
6800
7000
7200
f [GHz]
Q
measurement
measurement: Air correctedHFSS-cells
S3P-quad
Measurement setup 3 (bead pull)
Perfectload
Perfectload
VNAport1
VNAport2
1
2 3
4
E2 ~ S11-<S11>
Bead pull in complex plane
-0.05 0 0.05
-0.05
0
0.05
ImS
11
f=11.415 GHz
-0.05 0 0.05
-0.05
0
0.05
f=11.416 GHz
-0.05 0 0.05
-0.05
0
0.05
f=11.417 GHz
-0.05 0 0.05
-0.05
0
0.05
ImS
11
f=11.418 GHz
-0.05 0 0.05
-0.05
0
0.05
f=11.419 GHz
-0.05 0 0.05
-0.05
0
0.05
f=11.42 GHz
-0.05 0 0.05
-0.05
0
0.05
ImS
11
f=11.421 GHz
-0.05 0 0.05
-0.05
0
0.05
f=11.422 GHz
-0.05 0 0.05
-0.05
0
0.05
f=11.423 GHz
-0.05 0 0.05
-0.05
0
0.05
ReS11
ImS
11
f=11.424 GHz
-0.05 0 0.05
-0.05
0
0.05
ReS11
f=11.425 GHz
-0.05 0 0.05
-0.05
0
0.05
ReS11
Measurements: S11-<S11>
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.421 GHz
-50 0 50-50
0
50f=11.422 GHz
-50 0 50-50
0
50f=11.423 GHz
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.424 GHz
-50 0 50-50
0
50f=11.425 GHz
-50 0 50-50
0
50f=11.426 GHz
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.427 GHz
-50 0 50-50
0
50f=11.428 GHz
-50 0 50-50
0
50f=11.429 GHz
-50 0 50-50
0
50
ReEz [kV/m]
ImE
z [k
V/m
]
f=11.43 GHz
-50 0 50-50
0
50
ReEz [kV/m]
f=11.431 GHz
-50 0 50-50
0
50
ReEz [kV/m]
f=11.432 GHz
HFSS-cells: E
0 100 2000
20
40
|Ez|
[kV
/m]
f=11.421 GHz
0 100 2000
20
40
f=11.422 GHz
0 100 2000
20
40
f=11.423 GHz
0 100 2000
20
40
|Ez|
[kV
/m]
f=11.424 GHz
0 100 2000
20
40
f=11.425 GHz
0 100 2000
20
40
f=11.426 GHz
0 100 2000
20
40
|Ez|
[kV
/m]
f=11.427 GHz
0 100 2000
20
40
f=11.428 GHz
0 100 2000
20
40
f=11.429 GHz
0 100 2000
20
40
z [mm]
|Ez|
[kV
/m]
f=11.43 GHz
0 100 2000
20
40
z [mm]
f=11.431 GHz
0 100 2000
20
40
z [mm]
f=11.432 GHz
200 400 6000
0.1
0.2
|Ez|
[a.
u.]
f=11.415 GHz
200 400 6000
0.1
0.2
f=11.416 GHz
200 400 6000
0.1
0.2
f=11.417 GHz
200 400 6000
0.1
0.2
|Ez|
[a.
u.]
f=11.418 GHz
200 400 6000
0.1
0.2
f=11.419 GHz
200 400 6000
0.1
0.2
f=11.42 GHz
200 400 6000
0.1
0.2
|Ez|
[a.
u.]
f=11.421 GHz
200 400 6000
0.1
0.2
f=11.422 GHz
200 400 6000
0.1
0.2
f=11.423 GHz
200 400 6000
0.1
0.2
z [a.u.]
|Ez|
[a.
u.]
f=11.424 GHz
200 400 6000
0.1
0.2
z [a.u.]
f=11.425 GHz
200 400 6000
0.1
0.2
z [a.u.]
Bead pull: field magnitudeMeasurements: |sqrt(S11-<S11>)| HFSS-cells: |E|
Field distribution at 120o phase advance per cell
0 50 100 150 200 250 300 3500
0.2
0.4
0.6
0.8
1
z [mm]
|Ez|
[a.
u.]
HFSS-cells: f=11.428 GHz
measurement: f=11.421 GHz
0 5 10 15 20 25-126
-124
-122
-120
-118
-116
-114
iris number
d [
deg]
measurement: f=11.421 GHz, per cell
measurement: f=11.421 GHz, averageHFSS-cells: f=11.428 GHz, per cell
HFSS-cells: f=11.428 GHz, average
2 5 8 11 14 17 20 23-5-4-3-2-1012345T24_vg1.8 @ 11.424 GHz
cell number
[Pa
dv-
12
0]
(de
g)
S3P simulation resultscourtesy of Zenghai Li
The best frequency fit for 120 deg average phase advance per cell is the same for S3P simulations and air corrected measurements: 11.424 GHz and is different for HFSS: 11.428 GHz
Field distribution at the best match• The best match frequency is 3-4 MHz lower than the 120 deg phase advance frequency both in HFSS simulation and measurement• The average phase advance per cell is about 3 deg/cell different at the best match frequency both for HFSS simulation and measurement• There is still residual standing wave even in the case of HFSS simulations where the match is about -40 dB
0 50 100 150 200 250 300 3500
0.2
0.4
0.6
0.8
1
z [mm]
|Ez| [a
.u.]
HFSS-cells: f=11.424 GHz
measurement: f=11.418 GHz
0 5 10 15 20 25-124
-122
-120
-118
-116
-114
-112
-110
-108
iris number
d [
deg]
measurement: f=11.418 GHz, per cell
measurement: f=11.418 GHz, average
HFSS-cells: f=11.424 GHz, per cell
HFSS-cells: f=11.424 GHz, average
Power for <Eacc> = 100 MV/m
11.42 11.422 11.424 11.426 11.428 11.43 11.43240
45
50
55
60
65
70
75
f [GHz]
P[M
W]
for
<E
acc>
=10
0MV
/m
S3P
HFSS
Summary table for T24_vg1.8_diskparameters at 11.424 GHz if not stated otherwise
S12 Pout/Pin
τ=0.5ln (Pout/Pin)
tf=dφ/dω [ns]
tf=W0/Pin [ns] Q=ωtf/2τ Q=ωW
/PlossPin100 24 cells[MW]
Measurements
0.7077 0.5008 0.3457 63.11 - 6600 -
HFSS ¼ 0.741 0.549 0.2998 58.67 58.65 7025 7006
HFSS InCoupOutCoupCells 24+2
0.74120.998750.998890.743
0.54940.99750.99780.552
0.29940.00130.00110.297
58.590.4600.36257.77
702012700118006980
700513200123007001 42.1 @
11.428 GHz
3-cells model
0.569-0.533
0.282-0.315
- 54.8-59.37
6974-6764
6980 41.1
S3P (SLAC)
0.730 0.533 0.315 60.8 6927(σ=57e6)6988(σ=58e6)
42.4
Conclusions• There is no difference between HFSS simulations of ¼ of the structure and HFSS simulation of couplers + cells despite a factor 10 difference in the mesh density. This demonstrates that numerical convergence has been reached• All 3 different ways of calculating Q-factor: S3P, HFSS-S-parameter solver and HFSS-eigenmode solver give very close values of 6990, 7020 and 6980, respectively• The measurements of the T24_vg1.8_disk structure made at CERN show 6 % lower Q-factor of 6600• The difference in the simulated transmission between HFSS-S-parameter and S3P comes from the “geometrical” type of error which is equivalent to a difference of 4 MHz in the frequency for the 120 deg phase advance per cell• S3P simulations show 120 deg phase advance frequency of 11.424 GHZ. This is the frequency used in design of the cells using HFSS-eigenmode solver meaning that HFSS-eigenmode solver and S3P agrees and that the systematic error of 4 MHz in the case of T24_vg1.8_disk geometry comes from the HFSS-S-parameter solver• The 120 deg phase advance frequency in the air corrected measurement is 11.424 GHz which is the same as design frequency with the accuracy of ± 0.5 MHz. This demonstrates extremely high (sub-micron) precision of machining. For example, it is equivalent to ± 0.6 μm tolerance on the outer wall radius
Simulation setup 4 HFSS-cells + couplers
HFSS simulation of ¼ of input coupler + segment of the cells made in HFSS by 5 deg. Sweep (3D geometry created in HFSS)+ ¼ of output coupler :Surface conductivity (Cu): 58e6 S/mSurface approximation: ds=1 μm,
HFSS v11.1 HFSS v10.1versus
Reflection
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.44-45
-40
-35
-30
-25
-20
-15
-10
-5
0
f [GHz]
Ref
lect
ion
[dB
]
input measurementoutput measurement
input measurement - Air corrected
output measurement - Air corrected
input HFSS-cellsoutput HFSS-cells
input HFSS-quad
output HFSS-quad
input S3P-quadoutput S3P-quad
HFSS v11.1
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.44-45
-40
-35
-30
-25
-20
-15
-10
-5
0
f [GHz]
Ref
lect
ion
[dB
]
input measurementoutput measurement
input measurement - Air corrected
output measurement - Air corrected
input HFSS-cellsoutput HFSS-cells
input HFSS-quad
output HFSS-quad
input S3P-quadoutput S3P-quad
HFSS v10.1
Transmission
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.44-4
-3.8
-3.6
-3.4
-3.2
-3
-2.8
-2.6
-2.4
-2.2
-2
f [GHz]
Tra
nsm
isio
n [d
B]
S3P-quad
HFSS-quadHFSS-cells
measurement: S13+S23
measurement: |S13|+|S23|
measurement: 2*S13measurement: 2*S23
measurement: Air corrected
HFSS v11.1
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.44-4
-3.8
-3.6
-3.4
-3.2
-3
-2.8
-2.6
-2.4
-2.2
-2
f [GHz]
Tran
smis
ion
[dB
]
S3P-quad
HFSS-quadHFSS-cells
measurement: S13+S23
measurement: |S13|+|S23|
measurement: 2*S13measurement: 2*S23
measurement: Air corrected
HFSS v10.1
Group delay
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.4450
52
54
56
58
60
62
64
66
68
70
f [GHz]
dphi
/dw
[ns]
measurement
measurement: Air correctred
HFSS-quadHFSS-cells
S3P-quad
HFSS v11.1
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.4450
52
54
56
58
60
62
64
66
68
70
f [GHz]
dphi
/dw
[ns]
measurement
measurement: Air correctred
HFSS-quadHFSS-cells
S3P-quad
HFSS v10.1
Q-factor
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.445800
6000
6200
6400
6600
6800
7000
7200
f [GHz]
Q
measurement
measurement: Air correctedHFSS-cells
S3P-quad
HFSS v11.1
11.4 11.405 11.41 11.415 11.42 11.425 11.43 11.435 11.445800
6000
6200
6400
6600
6800
7000
7200
f [GHz]
Q
measurement
measurement: Air correctedHFSS-cells
S3P-quad
HFSS v10.1
Phase advance per cell
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.421 GHz
-50 0 50-50
0
50f=11.422 GHz
-50 0 50-50
0
50f=11.423 GHz
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.424 GHz
-50 0 50-50
0
50f=11.425 GHz
-50 0 50-50
0
50f=11.426 GHz
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.427 GHz
-50 0 50-50
0
50f=11.428 GHz
-50 0 50-50
0
50f=11.429 GHz
-50 0 50-50
0
50
ReEz [kV/m]
ImE
z [k
V/m
]
f=11.43 GHz
-50 0 50-50
0
50
ReEz [kV/m]
f=11.431 GHz
-50 0 50-50
0
50
ReEz [kV/m]
f=11.432 GHz
HFSS v11.1
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.421 GHz
-50 0 50-50
0
50f=11.422 GHz
-50 0 50-50
0
50f=11.423 GHz
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.424 GHz
-50 0 50-50
0
50f=11.425 GHz
-50 0 50-50
0
50f=11.426 GHz
-50 0 50-50
0
50
ImE
z [k
V/m
]
f=11.427 GHz
-50 0 50-50
0
50f=11.428 GHz
-50 0 50-50
0
50f=11.429 GHz
-50 0 50-50
0
50
ReEz [kV/m]
ImE
z [k
V/m
]
f=11.43 GHz
-50 0 50-50
0
50
ReEz [kV/m]
f=11.431 GHz
-50 0 50-50
0
50
ReEz [kV/m]
f=11.432 GHz
HFSS v10.1
Phase advance per cell
5 10 15 20 25-124
-122
-120
-118
-116
-114
-112
niris
dphi
[de
g]
measurement: f=11.421 GHz, per cell
measurement: f=11.421 GHz, averageHFSS-cells: f=11.424 GHz, per cell
HFSS-cells: f=11.424 GHz, average
HFSS v10.1
2 5 8 11 14 17 20 23-5-4-3-2-1012345T24_vg1.8 @ 11.424 GHz
cell number
[Pa
dv-
12
0]
(de
g)
S3P simulation resultscourtesy of Zenghai Li
Outlook• To cross-check above conclusions S-parameter frequency sweep calculated with S3P would be very useful• Similar study of the other structures (T18, TD18, TD24, …) would also help to understand the limitations of structure manufacturing as well as the structure design using HFSS-S-parameter solver
Recipes for structure matching• Use HFSS-S-parameter solver VERSION 10• Use S3P• …