« back to results for ""
Below is a cache of http://server.c-ad.bnl.gov/esfd/bnlsns/102.pdf. It's a snapshot of the page taken as our search engine crawled the Web.
The web site itself may have changed. You can check the current page or check for previous versions at the Internet Archive. Yahoo! is not affiliated with the authors of this page or responsible for its content.
Measurement and analysis of the transverse coupling impedance of the SNS extraction kickers
BNL/SNS TECHNICAL NOTE
NO. 102
D. Davino and H. Hahn
October 22, 2001
COLLIDER-ACCELERATOR DEPARTMENT
BROOKHAVEN NATIONAL LABORATORY
UPTON, NEW YORK 11973
Measurement and analysis of the transverse coupling impedance
of the SNS extraction kickers
- 1 -
Measurements and analysis of the transverse coupling impedance
of the SNS extraction kickers

D. Davino and H. Hahn

Brookhaven National Laboratory, Upton, NY 11973 USA


Abstract

Measurements of the transverse coupling impedance of one full size model representative of the 14
extraction kickers for the SNS accumulator ring are discussed in this technical note. The
measurements were made by means of the standard two-wire method. The dependence on various
terminations of the bus bar is discussed, as well as the behavior of a new ferrite winding, intended
to reduce impedances. Finally, data to enter into the SNS ring impedance budget are quoted and an
analytical expression that fits the impedance measurements is indicated.


I. Introduction

The transverse coupling impedance of the extraction kickers is believed to be the largest
contribution to the impedance budget of the SNS accumulator ring. Previous measurements on the
AGS booster dump kicker had shown values to be of concern for the transverse stability of the SNS
high intensity beam [1,2]. Therefore, a study has been carried out to develop methods and
technical solutions to reduce the transverse impedance.

One of the first solutions proposed was to use a 25 resistive termination in parallel with the
pulse form network (PFN) circuit that feeds the kicker. This solution helps to damp the natural
resonance of the bus bar which represents a resonator with small losses introduced by the
CMD5005 ferrite bricks.

Furthermore, Lee has proposed a new type of ferrite winding which was expected to reduce the
transverse impedance without appreciable degradation of the other extraction kicker performances,
namely the kicker magnetic field strength and its rise time [3]. In the following, we refer to that
winding as the ferrite loop. The measurements on a small kicker model were presented in a
previous technical note and the results for the ferrite loop were encouraging [4]. The advantage of
using a resistive termination was already shown in that report and complemented with a
comparison of the loop versus a resistive termination of the PFN.

However, the impedance values scaled from this model were still unacceptably large predicting
beam instability at SNS design intensity. This justified a series of impedance measurements and
tests on the accessible RHIC kickers [5] and finally the full-size SNS prototype, for which the
results are presented in this report.

- 2 -
The full size SNS extraction kicker in its vacuum vessel is shown in Figure 1. In particular, it is
possible to see the vessel in which the kicker is placed, the feed through to which the PFN circuit is
connected, the eddy current strips and the ceramic assembly to keep the kicker assembled. The
aperture dimensions are about h =25 cm in the vertical plane and w =14.6 cm in the horizontal one,
the bus bar is 40 cm long, and the vacuum vessel has a 62 cm inner diameter and it is 88 cm long.
The bus bar current plates are perpendicular to the vertical plane.



Fig. 1. The SNS extraction kicker: schematic view (left), the kicker in the vessel used for the
measurement (right).


II. Scheme of measurement

Transverse impedance measurements of the SNS extraction kicker were made using the standard
method [6] in which a double-wire "Lecher" line is inserted into the kicker, and the forward
transmission coefficients
21
S of the "Device Under Test" and of a reference line of equal length are
interpreted according to

21
21
2
ln(
/
)
DUT
L
DUT
REF
Z
Z
S
S

,






(1)

from which one obtains the transverse impedance as

(
)
21
21
2
2
2
ln
/
DUT
L
DUT
REF
cZ
cZ
Z
S
S =
=
,





(2)

with being the two wire spacing, and Z
L
the characteristic impedance of the line. In order to
obtain sufficiently strong signals and sufficient rigidity to assure alignment and repeatability,
measurements were performed with a cable made at this laboratory, as shown in Figure 2.
Matching of the 165 characteristic impedance of that cable to the 50 cables of the network
analyzer was achieved by means of transformers with a center-tapped secondary winding, which
- 3 -
serves as 180 hybrid. The transformers have a 50 /150 ratio and resistive matching by
resistors was applied thereby covering frequencies up to 100 MHz. All measurements were made
with the network analyzer, Agilent 8753ES, set for a logarithmic frequency scale from 30 kHz to
100 MHz, 1601 points, a 300 Hz bandwidth, and averaged over 5 sweeps.
a
b
d

Fig. 2. The transverse section of the cable used for the measurements. The dimensions are
a=b= 6.35 mm and d=10.8 mm (or d=35.6 mm).

Due to some electromagnetic background noise and due to the transformer properties, rated only
for frequencies above 100 kHz, measurement results at lower frequencies become suspect and must
be used with care. The typical ripple for the transmission coefficient, due to the instrument and
connectors reproducibility, was 3·10
-4
for the amplitude and 20·10
-3
degree for the phase units.
Those values are represented, through the formula (2), in the Figure 3. Those curves can be used as
amplitude error bars for all the coupling impedance measurements presented in this report.

2
10
20
30
40
50
60
70
80
90
100
0
1
2
3
f [MHz]
Real Z
tr
[k /m]
0.48
0.10
|S
21
| 0.3mU
2
10
20
30
40
50
60
70
80
90
100
0
1
2
3
f [MHz]
Imag Z
tr
[k /m]
0.56
0.12 S
21 20m
°

Fig. 3. The error bars amplitude of the measurement setup.

The calibration of the cable

The characteristic impedance,
L
Z = 165 , was measured by using an internal functionality of the
communication signal analyzer, Tektronix CSA 803. The effective spacing, , of the "home-made"
cable was determined by means of a comparison with a known cable, the commercial line CQ 551
with a characteristic impedance
L
Z = 450 and a spacing =13/16" ( =2.0637 cm). The
calibration procedure consists in measuring the transmission coefficient in the vertical plane
between the PFN port and the two cables, used without transformers and terminated, on one side,
- 4 -
into a short-circuit. In the low frequency range, 30 kHz 1 MHz, a simple transformer model can
represent that configuration, and the following expansion applies

(
)
2
1
2
21
2
2
2
C
C
L
L M
j M
S
Z
Z
+
=
+
+

where M is the mutual inductance, L
1
and L
2
are respectively the inductances of the kicker,
measured at the bus-bar port, and of the cable.
C
Z is the characteristic impedance of the
instrument. The mutual inductance depends linearly on the effective cable spacing and is given by,

0
l
M
w
µ
=


where
l is the magnet length and w is the horizontal width of the magnet. Comparing the forward
transmission coefficients allows directly to determine the spacing of the cable used for the
measurements. The cable in Figure 2 has an effective spacing of 1.43 cm, roughly
corresponding to the gap plus rod thickness.

In order to increase the signal and reduce the error bars, measurements were also made with the
same cable but with a bigger gap, d=35.6 mm, and
L
Z =275 . The effective spacing was here 40.6
mm. It is to be noted that these measurements gave cleaner signals, but did not result in noteworthy
changes.


III. Experimental results

Dependence on the bus-bar termination

The transverse coupling impedance of the SNS extraction kicker, without the ferrite loop, was
measured in the vertical plane with terminations directly connected at the bus-bar port. Figure 4
shows the comparison of impedances for different terminations, namely open, 200 , 50 , 25 ,
and short-circuit.

The transverse coupling impedance is significantly lower than the one measured with the AGS
booster dump kicker, the difference being a factor 3 between the two kickers. This is not surprising
and it is mainly due to the different apertures of the two kickers, since a larger aperture lowers the
impedance. The AGS booster dump kicker h