Design and Implementation of
Low-Pass Filters for PCT-COSY
Olga Kachko
FH Aachen, Campus Jülich, Nuclear Applications
Supervised by: Prof. Dr. Friedrich Hoyler, Dr. Yury Valdau, and Leonid Eltcov
The purpose of this project was to lter the output of the parametric current transformer (PCT) to obtain a more accurate measurement of DC beam current. Due to
the characteristic features of PCT, excitation frequency 7 kHz of magnetic modulator
is always present in the output signal. In order to attenuate the AC-component of the
PCT signal, it was proposed to design and implement two types of low-pass lters.
The test measurements were conducted in the laboratory and compared with SPICE
simulations.
2. Beam Diagnostics
1. Introduction
To analyze the behavior and the properties
of the accelerator beam, beam diagnostic
devices are of a big importance [3]. These
diagnostic tools can be divided into two main
groups according to the eect they have on the
beam namely destructive and non-destructive.
If the particle beam is lost as a result of the
measurement, a diagnostic device is referred to
as destructive. Destructive methods are based
on detection of energy loss of the charged
particles interacting with matter. A typical
example of a destructive device is a Faraday
cup which is usually used for the beam current
The mini-project was conducted as part of
a TRIC (Test of Time Reversal Invariance at
COSY) experiment [1] at the Institute of Nuclear Physics at Research Center Juelich. The
TRIC experiment is set up as a novel null test
(P-even, T-odd) of time-reversal invariance at
Cooler-Synchrotron COSY-Juelich. Currently,
the main focus of the research group working on TRIC is to develop the high precision
beam current measurement system. In order to study the properties of dierent nondestructive beam diagnostic devices the test
station was designed and built [2].
1
monitoring in the cyclotrons and extraction Fig. 1.
beam lines [4]. Even though it is an accurate
method and widely applied on accelerators it
is not used for routine beam monitoring in
the storage rings since it destroys the beam [3].
If during the operation of a diagnostic device no beam losses are observed, the method
is called non-destructive. Some examples are
beam current transformers, beam position
monitors, wall current monitors, Shottky
pick-ups etc. [4]. The measurement of the
beam current will be discussed in more detail
Figure 1: Simplied schematics of a DC transin the next section.
former consisting of two tori with three types of
windings each. First winding is used as a modulator, second winding as a detector and third
2.1 Beam Current Measurement
winding as a compensator [3].
Non-destructive diagnostic tools are essential at the synchrotron and storage ring
COSY to provide the necessary information
about the beam intensity [5]. Non-destructive
beam current transformers which work on the
principle of magnetic eld detection carried by
the beam are used to give accurate measurement of bunched (AC) and not bunched (DC)
beam current. The goal of this project was to
lter the signal from the parametric current
transformer (PCT) which is used to measure
the DC component of the stored beam at
COZY. The identical device produced by
BERGOZ company is installed at the test
station for non-destructive beam current
measurement [2].
Each toroid consists of three dierent
circuits such as modulating, sensing and
compensating circuits. The square symmetric
AC current is generated within a modulation circuit consisting of two windings with
opposite orientation.
Modulation current
forces the high permeability tori into magnetic
saturation two times for each period. The
sensing winding acts as a detector of the
modulated signal [7].
As shown in Fig. 2 two modes of operation should be considered i.e. no-signal operation and operation with signal. No-signal
operation means that there is no beam going
through the tori. In this case sensor windings
will detect zero average magnetic ux from two
modulation windings thus output voltage will
be equal zero. When beam passes through the
toroidal core, its magnetic eld causes a shift in
B-H hysteresis curve. The change in the magnetic ux will be sensed by secondary windings
and an external current Icomp will be gener-
The PCT consists of three main parts:
toroidal sensor, front-end electronics box and
back-end power supply and control chassis.
Toroidal sensor includes modulator cores and
L/R integrator core [6]. The principal of
operation can be explained with the help of
2
ated via a feedback loop to cancel beam's magnetic eld. The feedback current while passing
through a precision resistor produces the output voltage which is proportional to the beam
current.
Figure 2: PCT modes of operation [3].
Figure 3: Frequency response of rst-order lowpass lter [9].
Since PCT modulation frequency is around
7 kHz it was proposed to use a low-pass lter
The rst-order LP lter gives a slope of
to reduce the modulation ripple in the output -20 dB/decade attenuation of frequencies
voltage.
above the cut-o frequency. However, if this
-20 dB/decade is not enough to damp the
3.
Low-Pass Filters: Principle and signals at unwanted frequencies, then two
Schematics
rst-order lters can be cascaded together
resulting in a second-order lter. The more
As the name implies, low-pass lters (LP) RC stages are added the higher becomes the
are specially designed electric circuits that pass lter order.
the lower frequency signals and attenuate signals of higher frequencies. The main character4. Passive Low-Pass Filters
istic of LP lters is a specied frequency, called
cut-o frequency (corner frequency in Fig. 3),
Passive lters are made up of passive circuit
above which a high attenuation of the signal elements such as resistors, capacitors and
is produced. Below cut-o frequency very lit- inductors. The simplest rst-order passive
tle or no attenuation of the signal is observed. lter consists of a resistor connected in series
The gain/attenuation of the lter, generally with a capacitor. A capacitor is a reactive
expressed in Decibels, can be calculated as fol- device which exhibits very high resistance
lows:
to low-frequencies or DC signals and low
Vout
[8]
(1) resistance to high-frequency signals [8].
K = 20 · log
Vin
The cut-o frequency or -3 dB point, can be
found using the standard formula:
fc =
1
2π · RC
[8]
Fourth-order passive LP lter was soldered
according to the circuit diagram shown in
(2) Fig. 4. It was equipped with the switch to
change between lters of dierent orders,
from rst to fourth, to choose the optimal
attenuation at 7 kHz.
3
Figure 4: Fourth Order Low-Pass Filter
Schematic Circuit Diagram.
Figure 6: Phase Shift of Passive LP Filter simuFrequency response curves of the passive
lated in SPICE.
LP lters presented in Fig. 5 were obtained
from SPICE simulation program and plotted
The comparison between the SPICE simuby using OriginPro software. Phase shift as lation and experimental data is presented in
a function of frequency is represented in Fig. 6. Table 1. The experimental values are in good
agreement with theoretical calculations.
Table 1: Passive LP Filter, Attenuation [dB]
Order
Attenuation
SPICE Experiment
First
-3.82
-3.7
Second -7.30
-6.9
Third -9.31
-8.8
Fourth -9.75
-9.3
Table 2: Passive LP Filter, Phase Shift [◦ ]
Order
Phase Shift
SPICE Experiment
First
-24
-18
Second
-53
-48
Third
-82
-72
Fourth -100
-96
Figure 5: Frequency Response of Passive LP Filter obtained from SPICE simulation.
Figures 7 and 8 represent the results obtained by using HMO 3004 series mixed-signal
oscilloscope. It can be seen from the gures
that attenuation of the PCT signal increases
with the increased order of the lter. However,
there is a very small dierence between the
third and fourth order lter. Increasing the
order even further will not result in signicant
attenuation of the signal because each lter order will load its neighboring network [10].
4
(a) First-Order LP Filter
(a) Third-Order LP Filter
(b) Second-Order LP Filter
(b) Fourth-Order LP Filter
Figure 7: In Fig. 7 (a) the PCT output signal Figure 8: In Fig. 8 (a) the PCT output signal
(ch. 1) and output signal of rst-order LP lter
(ch. 1) and output signal of the third-order LP
(ch. 2) are presented. Attenuation of -3.7 dB
lter (ch. 2) are shown. Attenuation of -8.8 dB
is reached. In Fig. 7 (b) the PCT output signal
is obtained. In Fig. 8 (b) the PCT output signal
(ch. 1) and output signal of second-order LP
(ch. 1) and output signal of the fourth-order LP
lter (ch. 2) are shown. Attenuation of -6.9
lter are presented (ch. 2). Attenuation of -9.3
dB is obtained.
dB is reached.
for lter 1 and 3 kHz for lter 2 were obtained. Dierent combinations of resistors and
capacitors were considered and simulated in
Active low-pass lters contain active comSPICE to reach the desired cut-o frequencies.
ponents such as transistors, operational
ampliers or eld-eect transistors. The most
A schematic circuit diagram for each lter
common active lter is an active low-pass lter
is depicted in Fig. 9. Frequency response
which principle of operation and frequency
curves of the active LP lters are presented in
response is similar to passive low-pass lter [8].
Fig. 10. Phase shift as a function of frequency
is presented in Fig. 11.
Two active low-pass lters were constructed
and tested during this project. The components of the active LP lters were chosen in
such a way that cut-o frequencies of 1 kHz
5. Active Low-Pass Filters
5
Figure 10: Frequency Response of Active LP Filter, SPICE.
Figure 9: Schematic Circuit Diagram for Active
Low-Pass Filters.
In order to test the cut-o frequency of each
lter, the wave generator was used to generate
the sinusoidal wave with peak-to-peak voltage
of 2V and frequencies of 1 and 3 kHz respectively. Signal from the generator (ch. 1), output of LP lter 1 (ch. 2), output of LP lter 2
(ch. 3) obtained with the help of oscilloscope
are presented in Fig. 12. In Fig. 12 (a) a sine
wave of 1 kHz frequency and attenuation of
the signal corresponding to -3 dB are shown.
Fig. 12 (b) represents a generated sine wave of
3 kHz frequency and attenuation of the incoming signal by -3 dB. By analyzing the peak-topeak voltage of input and output signals, one
can estimate lter's cut-o frequency, or -3 dB
point i.e.
Figure 11: Phase Shift of Active LP Filter,
SPICE.
lter 2 (ch. 2) are shown in Fig. 13 (b). The
signal attenuation was calculated by using
equation (1) and presented in Table 3. The
measured values are in perfect agreement with
the SPICE simulation at 7 kH.
Vout = 0.707 ∗ Vin
In Fig. 13 the attenuation of PCT signal
by using two active LP lters is presented.
The PCT signal (ch. 1) and output signal
of lter 1 (ch. 3) are shown in Fig. 13 (a).
The PCT signal (ch. 1) and output signal of
6
(a) Filter 1, test.
(a) Filter 1
(b) Filter 2, test.
(b) Filter 2
Figure 12: Cut-o frequency test. Sine wave Figure 13: Active Low-Pass Filter. The PCT
(ch. 1) of 1 kHz and output of active LP loutput signal (ch. 1) and output of active LP
ter 1 (ch. 3) is depicted in Fig. 12 (a). Sine
lter 1 (ch. 3) are depicted in Fig. 13 (a). The
wave (ch. 1) of 3 kHz and output of active LP
PCT output signal (ch. 1) and output of active
lter 2 (ch. 2) are shown in Fig. 12 (b). The
LP lter 2 (ch. 2) are shown in Fig. 13 (b).
attenuation of - 3 dB is reached with both lters.
The attenuation of -26 dB and -10 dB respectively was obtained.
References
6. Conclusions
[1] D. Eversheim and Y. Valdau. Test of timereversal invariance at cosy (tric). 2012.
Two types of low-pass lters were constructed and the attenuation of the ACcomponent of PCT output signal was reached.
The test measurements were in good agreement with SPICE simulation. However, the
alternative techniques and methods should be
developed and tested in order to suppress the
modulation ripple in the output signal even
further.
[2] S. Kirfel. Design and implementation of
a test station for beam current measurement devices. Master's thesis, FH Aachen,
2014.
[3] J.C. Denard.
CERN
Beam current monitors.
Accelerator School, 2008.
[4] D. Brandt. Beam diagnostics for accel7
Table 3: Active LP Filters
Filter 1
Filter 2
Filter 1
Filter 2
Attenuation [dB]
SPICE Experiment
-25
-26
-9.6
-10
Phase Shift [◦ ]
SPICE Experiment
-160
-168
-111
-106
erators, proc. cern accelerator school cas.
2009.
[5] R. Maier and et al. Non-beam disturbing diagnostics at cosy-jülich. Conf.Proc.
C900612, 1990.
[6] BERGOZ. Parametric current transformer. Instruction Manual, July 1989.
[7] E. J. Kletsky. Design criteria for lowlevel second-harmonic magnetic modulators. Technical report, Electronic Nuclear
Instrumentation Group, 1977.
[8] P. Horowitz and W. Hill. The Art of
Electronics. Cambridge University Press,
1980.
[9] Electronics tutorials.
http://http:
//www.electronics-tutorials.ws/
filter/filter_2.html.
Accessed:
2016-02-03.
[10] E. Hughes. Electrical Technology. Addison
Wesley Longman, 1995.
8