Title Cold Model Test of a DTL with Transition from 47π to 2π-Mode Acceleration

Author(s)Ikeda, Kazumi; Kan, Taro; Yoshizawa, Kunio; Yokobori,Hitoshi; Hirota, Jitsuya; Iwashita, Yoshihisa; Okamoto,Hiromi; Noda, Akira; Inoue, Makoto

Citation Bulletin of the Institute for Chemical Research, KyotoUniversity (1993), 71(1): 26-36

Issue Date 1993-03-31

URL http://hdl.handle.net/2433/77493

Right

Type Departmental Bulletin Paper

Textversion publisher

Kyoto University

Bull. Inst. Chem. Res., Kyoto Univ., Vol. 71, No. 1, 1993

Cold Model Test of a DTL with Transition

from 47r to 27r-Mode Acceleration

Kazumi IKEDA*, Taro KAN*, Kunio YOSHIZAWA*,

Hitoshi YOKOBORI*, Jitsuya HIROTA*, Yoshihisa IWASHITA**,

Hiromi OKAMOTO**, Akira NoDA** and Makoto INouE**

Received February 9, 1993

A cold model test has been carried out using a constant-velocity model for the investigation of a DTL with transition from 4,r to 2,r-mode acceleration. The model cavity consists of sixteen cells which are resonant at 425 MHz and correspond to a proton energy of about 2 MeV. On-axis electric field distributions were measured with the Bead Perturbation Method. The measurements reveal that a difference of about 5% exists between the average fields in the 4,r and 27r-mode cell. Field stabilzation by post coupler is improved by using slightly different spacings between the drift tube and the post coupler for the 4,r and 2n-- mode cells. The extra stem on each drift tube improves the field stability when the stems are installed in line each other. The stability, however, is not improved when the stems are installed cross-wise each other.

KEY WORDS : Proton Linac/ Alvarez DTL/ 47r-mode Acceleration/ Post Coupler/ Field

Stabilization/ Single Stem/ Double Stem

1. INTRODUCTION

A design study of a 10 MeV proton linac for multi-purpose use has been carried on since

April, 1990, in cooperation of the Institute for Chemical Research of Kyoto Univ. and Mitsubishi

Atomic Power Industries, Inc.. It is intended in this study to investigate an optimum

combination of a four vane RFQ linac and DTLs of 47r and 2 n--mode acceleration whose

operating frequncy is 425 MHz. The transition energy from the RFQ to the DTL of 47r-mode

acceleration is tentatively fixed at 1 Mev so that the fabrication and adjustment of the RFQ may

be facilitated. A cold model test of the RFQ indicated that good field stabilization would be

achievable by magnetic coupling between the quadrants.'

The transition energy from the 47r to 2a-mode acceleration in the DTL is tentatively fixed at

2 MeV. The DTL is designed to be a single-tank structure because of the simplicity. A cold

model test was planned for the investigation of the DTL with transition from 4n- to 27r-mode acceleration.

Drift tubes supported by two stems were considered for cryogenic DTLs to reduce vibration

and to minimize drift tube deflection. It was reported that a DTL with double stem 180° apart

should be inherently more stable against tuning errors than a similar structure with single stem.21 It is interesting to examine how the double stem improves the field stability of the DTL with

* 7 ®t811 , A 7 n , atIM41, 5LE i : Mitsubishi Atomic Power Industries, Inc. ** W * , #1. 'f : Nuclear Science Research Facility, Institute for Chemical

Research, Kyoto University.

( 26 )

Cold Model Test of a DTL with Transition from 47r to 27r-Mode Acceleration

transition from 47r to 27-mode acceleration.

It is also intended to examine the field stabilization by post coupler. The post coupler is in

principle a quarter-wave resonator. Empirical observations identify a "safe" operating range for field stabilization3>. The model cavity is fabricated so as to satisfy this operating range.

2. STRUCTURE OF MODEL CAVITY

The structure of the constant-velocity model cavity is shown in Fig. 1. The cavity consists

of 16 cells whose lengths correspond to the proton energy of about 2 MeV ; 5 cells (each 90.0 mm

long) for 47r-mode acceleration, 1 cell (67.5 mm long) for transition from 47r to 27r-mode

acceleration and 10 cells (each 45.0 mm long) for 27r-mode acceleration. The equivalent inner

diameter and length of the cavity are 430.0 and 967.5 mm respectively. The diameter of the drift

tube (DT) and bore hole are 80.0 and 10.0 mm respectively. The gap length between the DTs is

determined by the SUPERFISH calculation to bring the resonant freqency at 425 MHz.

The DT can be supported by either single or double stem 180° apart. The diameter of the

stem is 15 mm. The half DT on each end plate can be moved to detune the end cell. A post

coupler (PC) can be installed opposite each DT of the 47-mode cell and transition cell, and

opposite every other DT of the 27-mode cell. Successive PCs are located at the alternate side of

the tank. The diameter of the PC is also 15 mm. The model cavity, including the PC, are made of aluminum alloy (5052). Photo. 1 and 2 shows the inside of the cavity under assembling

and the outside of the cavity after completion.

Fig. 2 shows measured deviations of the bore center of each DT from the beam axis in

horizontal and vertical directions. The target accuracy of the DT alignment is i- 0.05 mm in

both directions which are achieved as follows : Horizontal x : —0.017±0.023 mm

Vertical y : —0.015±0.029 mm

Radial r : 0.023±0.025 mm

A !mull

-- © Askimon en Lirewmi...

IT IF3Bliniiiii.....M.11111.11.1.1111111111111111111.1 o°Aii

A

A—A

Fig. 1. Structure of the model cavity. 1 Drift tube 4 End plate 7 Post coupler

2 Stem5 Cavity wall 8 End cell detuning device 3 Garter 6 Channel 9 Measurement hole

( 2 7 )

K. IKEDA, T. KAN, K. YOSHIZAWA, H. YOKOBORI, J. HIROTA, Y. IWASHITA, H. OKAMOTO, A. NODA and M. INouR

- --...i,iFyti„

---—itlf: „irf

'-

Cold Model Test of a DTL with Transition from 4r to 2r-Mode Acceleration

10-------------------------------------------------------------------- O Horizontal

8• Ve rtical

E 6_ E •

00 4_ • 0•

2-

o•----------------0o-----------------------------------------------•• • OOO

0•• s • • O OO 0

o•0 0 0 d _ 4._•

•00 •

m -6-0

-8- •

_101 1 1 1 1 I I I II I I I I I I I In 1 2 3 4 5 68 10 12 14 Out

NO. of DT

Fig. 2. Deviation of the bore center from the beam axis in the model with

single stem.

20--------------------------------------------------------------------

16-0 0

12-O O0 0 •

-

N00 4—0 O

O •0-000 0O--o- 0 -4-

(0 4 -12- A

-16-

_201 I I I I I I I I I I I I I I-------------------------------------------------------I 1 2 3 4 5 68 10 12 14 16

No. of Cell

Fig. 3. Difference of the gap length from the design value in the model with single stem.

Fig. 3 shows differences of the measured gap lengths from the design values which are 21.82,

15. 36 and 8.54 mm for the 47r-mode, the transition and the 2nr-mode cell respectively. The

averaged value of these differences is 0.04 ± 0.06 mm. It is suspected that the present

measurement method of the gap length tends to give a larger value than the real one. The Q value of the cavity is measured to be 2.5 X 104 and estimated with SUFERFISH to be

4.57 X 104 including the effects of the stems and end plates. Considering the difference of the

electric conductivities between copper and aluminum alloy, the calulated value is corrected to

2.70 X 104.

3. RESONANT FREQUENCY AND FIELD DISTRIBUTION

Resonant frequencies for various modes were measured on the model cavity with single

(29)

K. IKEDA, T. KAN, K. YOSHIZAWA, H. YOKOBORI, J. HIROTA, Y. IWASHITA, H. OKAMOTO, A. NODA and M. INOUE

700------------------------------------------

600-----------------------T110 In--------------

>, 500----------------------------------- TElln-----------

.- 5 400

0 300

•0

ai 200----- _

---

Stem Mode

100 ------------------------------------------------ 01 23

Mode Number

Fig. 4. Dispersion curves in the cavity with single stem.

stem. Dispersion curves for these modes are shown in Fig. 4. The measured resonant

freauency is 425.613 MHz for the TM010 mode which agrees quite well with the calculated one.

The measured resonant freqency is 449.375 MHz for the TM011 mode and the frequency

difference between the TM010 and TM011 mode is 23.760 MHz which is considerably large.

The larger the frequency difference, the harder it is to achieve an appreciable admixture of the

+'10 _0

9.0

ro 0 0

10 0 0 O

8.0

00

b 7.0

I14 6.0 ro

ro

s_o1IILI I I 1 I I I I I I I

1 2 3 A5 6 8 10 12 14 16.

No. of Cell

Fig. 5. Distribution of the average electric field in the cavity with single stem.

( 30 )

Cold Model Test of a DTL with Transition from 47r to 27r-Mode Acceleration 4 ----------------------------------------------------------------------------------------------------------1 1 I I I 1. I I

A °P

~1-

3 _.. N 11 -

o

" 2 _....

U

1 C./ N

-

ICVYit~~) I ti )1i..14~11A ./1it 1I I1/A0~~ 0 10 20 30 4050 60 70 80 90 100

On—Axis Distance Z(cm) Fig. 6. Distribution of the electric field calculated with SUPERFISH.

TM011 with the TM010 mode, and therefore, the more stable the field distribution. This large

frequency difference, however, is only owing to the shortness of the model cavity. The stem

modes distribute in a range from about 200 MHz to 140 MHz on the cavity with single stem.

The on-axis electric field distribution was measured with the Bead Perturbation Method.

An aluminum cylindrical bead of 3.0 mm diameter and 1.5 mm length was used. It was

confirmed that the length of the bead was short enough to measure the fine field distribution in

the gap. Fig. 5 shows the distribution of the average electric field (integrated electric field

divided by the cell length) of each cell in the cavity with single stem. The average fields in the

47r and 2ir-mode cells are uniform within ± 1.6% and ± 1.3% respectively, except for the

transition cell.

The averaged electric field in the 4n--mode cells is 5% larger than that in the 27r-mode cells.

Fig. 6 shows the distribution of the electric field calculated with SUPERFISH. This distribution

gives the difference of 4% which agrees satisfactorily with the measured one.

4. FIELD STABILZATION BY PC

Fig. 7 shows the shifts of resonant frequencies for the TMOln modes by 10 PCs which are

inserted with the same spacing between DT and PC each other. These modes were identified by

the on-axis field distributions. This figure suggests that PC may be effective in a range less than

20 mm of the spacing between DT and PC.

The field stabilization was investigated as follows : Two field distrubutions for different detuning of the end cell are carried out : For the first measurement, decreasing the first gap

(cell-1) lowers the TM010 mode frequency by ,6f and increasing the last gap (cell-16) restores the frequency. This type of the detuning tends to "tilt" the field towards the cell-1, resulting in the

higher field in the cell-1 and lower field in the cell-16. The second measurement corresponds to

the opposite sign perturbation to both the end cells, which tends to tilt the field opposite direction. A distortion parameter, Ds, which indicates the effectiveness of field stabilization by PC is

defined with similarity to Dx4) as follows :

( 31 )

K. IKEDA, T. KAN, K. YOSHIZAWA, H. YOKOBORI, J. HIROTA, Y. IWASHITA, H. OKAMOTO, A. NODA and M. INOUE

460----------------------------------------------

450-•

• NTM013 - TM011

x4401-/•/ U/ TH012

~~ /

N/i

Q

v 430-X 0/TM010 -O -'O-0,O-O—O—O------0-----0

A

/ 0 420-I •

X / N

N a• 410 -

400 I I 1 I 1 I I I I

0 10 20 30 40 50 60 70 80 90. 100

Spacing between DT and PC(mm)

Fig. 7. Shifts of the resonant frequencies for TMOIn modes by PC.

50.0------------------------------------------------------

N •

t N ix'40 _ 0 -

N Q

•

N\ •/ 4 30 _ 0 - N

\• G• (0\•~•

P,

20-0 -4 (1 5, 1 4)

0

y,X(14,16)

0O1s> 1, 10.0 -

1 N

AI

0.0 I I I 1 1 I I I I

10 11 12 13 14 15 16 17 18 19 20

Spacing between DT and PC(mm)

Fig. 8. Distortion parameter vs. spacing between DT and PC. (Si, S2) : S1 and S2 are the spacing in 4,r and 27r-mode cell respectively.

( 32 )

Cold Model Test of a DTL with Transition from 47r to 27r-Mode Acceleration

Ds= s—En s I /(En s+ En, s). Ai X 100 (%)(1)

where En s and En, , are the on-axis electric field of the cell-n in the first and second measurement for the spacing, s, between DT and PC. Fig. 8 shows a plot as a function of s. As seen in this

figure, a smaller Ds is obtainable if the slightly different spacing is used for the 47r and 27r-mode

cell.

5. DOUBLE STEM AND TILT SENSITIVITY

The dispersion curves shown in Fig. 9 are those for the cavity with double stem. The

measured resonant frequencies are 426.432 MHz for the TM010 and 477.617 MHz for the

TM011 mode in the cavity with double stem (in line). The extra stem on each DT increases the

TM010 mode frequency by only 0.819 MHz, but increases the TM011 mode frequency by 28.244

MHz. The frequency difference between the TM010 and TM011 mode is more than two times

larger with double stem (in line) in comparison with single stem.

800--------------------------------------------

TMOIn

700TElln

N •1 (cross)

7

0600------------------------------------------------ TTMOM

(in line) N/ u

/ u 500-----------------------------------------------

di • N d~ 0 No

a 400----------- Stem Mode

Q(cross)

Q

300I

0 123

Mode Number-

Fig. 9. Dispersion curves in the cavity with double stem.

Table 1. Fitting results of the TMOln frequencies to second order polynomials.

Single Stem Double Stem (in line)

Co425.613 MHz426.430 MHz CI1.44832.685 C222.31218.505

TMO1n frequency f(MHz)=Co+Cn C2n2 (n=0,1,2)

( 33 )

K. IKEDA, T. KAN, K. YOSHIZAWA, H. YoKOBORI, J. HIROTA, Y. IWASHITA, H. OKAMOTO, A. NODA and M. INouE

Table 1 gives the fitting results of the resonant frequency (f MHz) for the TMOln mode (n=

0, I, 2) to second order polynomials of n for the cavity with single and double stem (in line). The

first order coefficient C1 for double stem (in line) is much larger than that for single stem.

Therefore, a DTL with double stem (in line) should be more advantageous than that with single

stem for a longer cavity at this energy region.

The results calculated with MAFIA5) are shown in Fig. 10 and a typical one-cell geometry

used in these calculations is shown in Fig. 11. Only the zero modes and STEM or mode are

calculated with this geometry. The zero modes for TE passband are not seen in a real finite

1,000 --------------------------------------------------1,000 ----------------------------------------------------------1,000. -------------------------------------------------- TM1101381I TM/10 BillTM110 - 900 - •—^ -900 -t -900 -

TM110E/11I TM110 Bid 11E210 E8116210 E//l°

800 - °---1 TE210 on I---b-800 --800 --

N 700 - .-----ITE210 ELI—.700 'I TE210 Ed _700-I- TE210 E1

TE110 Eil_ .....eTE110 v600 --600 -600 -

0, U° I TE110 Ed—, UUIBM C 500 --q500 --C 500 -- NTMO70 ••1TM010I•TM010 • •400 -~• - O, 400 -8_1TE110E1I~,-a400 - O.I TE 110 E1 r° _ cuI STEMO l 1" 300 -- - 1" 300 -s" 300 --

200 — 0ISTEMO1• —200 - 0 STEMn0 -200 - 0 I SIak I 0

100 --100 --100 -

0 ---------------------------------------0 --------------------------------------------0 121212

ARXRx

(a)(b)(c) Fig. 10. Results calculated with MAFIA. (a) Single stem, (b) Double stem (in line) and

(c) Double stem (cross).

?M IA •" ' o.~ „u ..,.. :,c~ovze,s3 za,~,ss---------------------------,

II'Tq q~r.,~_

W DW

Fig. 11. Typical mesh geometry used in the MAFIA calculation.

( 34 )

Cold Model Test of a DTL with Transition from 4,r to 2,r-Mode Acceleration

length cavity. Fig. 10-a shows the results for the cavity with single stem and Fig. 10-b and c

show those with double stem(in line and cross). Because the two radial directions x and y are identical in the calculation for the cavity with double stem (cross), the dipole modes perturbed by the stem are degenerated.

The frequency dependece on cell length is not large in these geometries except for the stem and TE modes in the cavity with double stem (cross). The STEMO mode in the cavity with single stem has the frequency much lower than the TM010 and that with double stem (in line) has the frequency closer to the TM010 mode. The STEMO mode in the cavity with double stem

(cross) has the frequency higher than the TM010 and the stem-mode passband overlaps the TM010 mode in both cell lengths. In this stem geometry, it is supposed that the STEMO mode has a frequency lower than the TM010 in longer cell geometries.

The effectiveness of the extra stem on the field stabilization was investigated by using tilt sensitivity (TS)2). TS is defined as follows:

TS=(E\ En )l (En+En)•,Af X 100 ((Y0)(2)

where En and En are the on-axis electric field of the cell-n in the first and second measurement as defined for Ds. Fig. 12 shows three TS measurements ; one with single stern and two with double stem 180° apart. The TS slope for the cavity with single stem and double stern (in line) are —38 and —12%/MHz/m respectively. The extra stem on each DT results in an almost threefold redution in the slope when the stem are installed in line each other. This higher

stability is consistent with the observed large increase in the TM011 mode frequency. The TS behaves quite differently when the stems are installed cross-wise as seen in Fig. 12 : The TS slope for the 47-mode cells is smaller than that for the 2n-mode cells, and almost equal to the TS slope for the cavity with double stem (in line) although opposite in the sign. This fact indicates

20-•/ A

N•o Q/ 10- \/.

0a v • A

• G' r4 0 00

Double StemO 070/0 (in 1ine) ,0 _0 _0

N.__o--o~•~ F+-10 Double Stem -•

^-1(crOSSj •

Single Stem •\

•

-'LO- I I 1 I I I I I 1 1 1 1 1 1 1 1

1 2 3 4 5 6 7 8 910 1 2 14 16

No_ of Cell

Fig. 12. Tilt sensitivities for the cavities with single and double stem.

( 3 5 )

K. IKEDA, T. KAN, K. YOSHIZAWA, H. Yoxosom, J. HIROTA, Y. IWASHITA, H. OKAMOTO, A. NODA and M. INOUE

that a DTL with double stem (cross) may be advantageous in an energy region higher than 8

MeV.

6. CONCLUSION

The resonant frequency measured for the TM010 mode agrees quite well with the value

calculated with SUPERFISH. The average field difference of about 5% is observed between the

47r and 27r-mode cell, which agrees satisfactorily with the SUPERFISH calculation . As for the field stabilization by PC, the use of slightly different spacings between DT and PC

in the 47r and 27r-mode cells results in an about twofold reduction in the distortion parameter . The extra stem on each DT improves the field stability resulting in an almost threefold

reduction in the TS slope when the stems are installed in line each other . This higher stability is consistent with the almost twofold increase in the frequency difference between the TM010 and

TM011 mode. The stability is not improved when the stems are installed cross-wise each other.

A DTL with double stem (cross) may be advantageous in an energy region higher than 8 MeV.

ACKNOWLEDGEMENT

Computation time for the MAFIA calculation was provided by the Supercomputer

Laboratory, Institute for Chemical Research, Kyoto University.

REFERENCES

(1) K. Ikeda et al., "Cold Model Test of A Four Vane RFQ for Multi-Purpose Use Linac", Bull. Inst. Chem. Res., Kyoto Univ., 70, 99 (1992).

(2) J.H. Billen, "Field Stability in Two-Stem Drift-Tube Linacs", Mo3-35, 1988 Linac Conf. at CEBAF, CEBAF-Report 89-001, 125 (1989).

(3) J.H. Billen, "Survey of Drift Tube Linacs with Post Couplers", AT-1 : 84-74 LANL memorandom (1984). (4) M. Sawamura and H,Takekoshi, "Field Stabilization of the Alvarez Linac by Post coupler", Bull. Inst.

Chem. Res., Kyoto Univ., 63, 9 (1985). (5) The Mafia Collaboration, "MAFIA USER GUIDE", March 16, 1988.

( 36 )