rf design options for quenching breakdowns

CLIC Breakdown Workshop 5/20/2008

1

RF Design Options for Quenching Breakdowns

Igor Syratchev, Riccardo Zennaro


2

HDS 60 L

PINC

HDS 60 S

PINC

Very often we do observe, that after accelerating structure processing the most of the surface modifications take place in a few first cells. Also the number of cells involved is correlated with the group velocity, the less the Vg the fewer cells modified.

As one of the conventional explanation one could expect the statistical distribution of the events in a chain model. However with adopted processing strategy (trip rate ~10-3) the event probability and normalized to that damage distribution is calculated to be very flat.

0 20 40 600

0.5

1

1.5

Cell no

Nor

mal

ized

tota

l dam

age

HDS 60

HDS 11 titanium

So why?


3

What do we certainly know, the breakdown ignition is a very fast

process: 0.1 -10 ns. If so, one can propose the main difference

between the “first” and “second” cell is accessible bandwidth.

And the lower group velocity the more the difference.

The first cell, if breakdown occurs is loaded by the input

coupler/waveguide and is very specific in terms of bandwidth.

Other words, the first cell can accept “more” energy during

breakdown initiation then consequent ones.

Worse to mention that we do not know the exact transient

behavior of the breakdown and the structure bandwidth could play

important role.


4

RF current source Iejwt

Structure: 2pi/3 aperture 3.5 mm (Vg=4.5%)

Ib, kA/mm2

Pout

Pin = 50MW

28.5 29 29.5 30 30.5 310.01

0.1

1

F, GHz

Breakdown ‘naive’ modeling in HFSS

Radiation spectra(breakdown in cell#1 )

To the output coupler

To the input coupler

Missing energy plot


5

29 29.5 30 30.5 3140

30

20

10

0

F, GHz

S-p

ara

mete

rs,

dB

Traveling wave

Configuration #1 (breakdown resonant fuse) :Resonant cavity with reduced electric surface field (HO1) is located between structure and waveguide.

Standing wave


6

NLC inline taper

V. Dolgashev (May 2002)

0 2 4 6 80

1

2

3

4

Vg=PLc/W

0 2 4 6 80.2

0.4

0.6

0.8

1

1.2

Stored Energy/cell

Cell #Cell #

Inline taper with a “speed” bump

1-st cell


7

Why speed bump?

Can we reduce the bandwidth and provide a tapering without increasing the overall length?

We can tray by reducing vg in the matching cell


8

Speed bump (TM03)

R=14.398 mm

R_iris= 2.428 mm

Iris_thickness= 1mm

Bandwidth

-10

-9

-8

-7

-6

-5

-4

-3

-2

-1

0

29.5 29.6 29.7 29.8 29.9 30 30.1 30.2 30.3 30.4 30.5

f (GHz)

S12

(dB

) Speed bump TM03

Speed bump TM02

Original

0

0.01

0.02

0.03

0.04

0.05

0.06

0 1 2 3 4 5

cell #

vg

/c

2nd mode speed bump

regular cell nominal value

3rd mode speed bump

3rd mode,final version

To be compared with the C30_vg4.7


9

Same phase advanceSame P/cSame aperture and iris shapeSame field configuration in the iris region

but

Different group velocity: 4.7% & 2%

Different R/Q: 29 kΩ/m & 12 kΩ/m

TM02 structureIs it possible to change some global parameter without changing local field distribution?

Only by changing the propagating mode

TM02 regular cell

TM01 regular cell “reference”

To be compared with the C30_vg4.7


10

TM02 structure

Predicted gradients for the test structures calculated for different parameters (normalized to the experimental results of T53 for b.d.r.=10-6 and pulse length=100 ns)

0

20

40

60

80

100

120

140

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

vg (%)

E (

MV

/m)

P/c

Sc

E on surface

1/sqrt(vg)

Measured

2p/3TM02


11

Conclusions

Experimental results shows a correlation between

RF damage and group velocity

This correlation could be explained in terms of

accessible bandwidth

The speed bump structure should inhibit the large

breakdown damage in the first cells

The TM02 structure should provide indications on

the correlation between the gradient at a given

b.d. rate and the group velocity


12

Extra slides….just in case


13

d [mm]a [mm]

2.79 2.13 2.00 1.66 1.37 1.25

2.53 Vg: 0.7% CLIC _vg1output 1.0%

2.85 T53 output1.0%

3.0 CERN-X1.1%

Vg: 1.35%

3.873.89*

Vg: 2.25%(*)

30 GHz 2π/3 ≈2.6%

T53 inputVg: 3.3%

4.38 30 GHz 2π/3 4.7%

5.00 30 GHz π/2 7.4%

30 GHz2π/3 8.2%

The test matrix (all structures in disks)

In red: 11.4 GHz new structures

In blue: 30 GHz new structures (scaled values for a and d)

(*) not very different from input vg1 (d=2.79; a=4.06)

Direct comparison of variation of a and P/c

Direct comparison of variation of d

Direct comparison of variation of P/c

Test for a relatively large group velocity


14

0

20

40

60

80

100

120

140

160

180

200

220

240

0 1 2 3 4 5 6 7 8 9 10

vg (%)

E (

MV

/m)

(P/c)

(Sc)

E on surface

1/sqrt(vg)

Reference structure:

C10_vg3.3 (T53 input)

C10_vg0.7 C10_vg1.35

C10_vg2.25

C30_vg2.6

C30_vg8.2

Predicted gradients for the test structures calculated for different parameters (normalized to the experimental results of T53 for b.d.r.=10-6 and pulse length=100

ns)

The test matrix

rf design options for quenching breakdowns

Documents