Constant Current AC Source using Improved Howland Pump for Exciting Eddy Current Testing (ECT) Probe

Constant Current AC Source using Improved Howland Pump for Exciting Eddy Current Testing
(ECT) Probe
Sumayya Abbas*, Uzair Gillani†, Shayan Ahmed‡, Salar B. Javed§, Iqra Sajidς
Electronic and Power Engineering Department
Pakistan Navy Engineering College, National University of Sciences and Technology*†‡§ς
Karachi, Pakistan
Email: *, †, ‡,
§
, ς
Abstract—an endeavor has been done to produce a cost
effective Eddy Current Testing System. For this
purpose, study was carried out on using the Improved
Howland Pump as a constant current AC source for
exciting inductive load transducer i.e. the Eddy Current
Testing (ECT) probe. Experimental validation was
performed by fabricating the constant current AC
pump and using it for excitation purposes in the eddy
current testing system. As a result, through-hole surface
cracks were successfully detected which proved the
working of the Improved Howland Current pump as a
constant current AC source. It is a valuable technique
for exciting Eddy Current Probes.
I. INTRODUCTION
Nondestructive testing (NDT) is a wide group of
analysis techniques in science and industry to evaluate the
properties of a material, component or system without
causing damage [1].
In this study, a constant current AC source is designed
to be used as an excitation source in an Eddy Current
Testing system. Eddy Current Testing is a nondestructive
testing technique that is based on the principle of Faraday’s

law of Electromagnetic Induction to inspect cracks/flaws in
metallic non- magnetic materials [2]. If a crack in the
material disturbs the eddy current circulation, the magnetic
coupling with the probe is changed and a defect signal can
be read by measuring the coil impedance variation [2] [3].
The magnetic field generated in the coil is directly
proportional to current. A constant current AC source is
needed so that a constant magnetic field is generated in the
coil and remains the same throughout the inspection. Only
that can offer uniform and precise testing of the metal
specimen.
The aim of the study on Eddy Current Testing system is
to devise a system that would detect surface cracks on
aerospace structures. The basic objective is to fabricate an
affordable system.
These systems are very expensive. An effort has been
made to produce a cost effective solution at laboratory level
so that further improvements can be easily done. The
waveforms acquired at every step of instrumentation section
prove that harmonics are not an issue. They are very stable

508

and are not affected throughout the running of the system.
That is why no harmonic filter has been used.
Eddy Current Testing system has been designed by
using the constant current AC source designed and
developed through the Improved Howland Current Source
with modifications for the particular study.
In this paper, we describe the need for constant current

AC source in Eddy Current Testing. This is a must for Eddy
Current Testing and is the key behind the whole concept. It
has been further explained in II. The hardware design and
its implementation using a sinusoidal oscillator is then
explained.
Moreover, protection circuitry has been
designed for driving inductive load (the ECT probe/coil)
which is the main purpose of the constant current AC
source. Finally, results are shown using the designed
Constant Current AC source. These results prove the
feasibility of the design for the said purpose and have been
experimentally confirmed as well.
II. CONSTANT CURRENT AC SOURCE
The concept of eddy current testing is based on the
principles of Electromagnetic Induction. The Maxwell’s
Ampere Law states that an alternating current sets up a time
varying magnetic field [4], which is known as a primary
magnetic field. When a coil excited with a constant peak
alternating current source is placed in close proximity to a
conducting surface, the alternating magnetic field will
induce a secondary time varying magnetic field of fixed
amplitude in the nearby surface as a result of which an EMF
is produced [5]. This induced voltage produces eddy
currents in the material which by Lenz’s Law oppose the
primary magnetic field [6]. This mutual inductance causes a
change in the impedance of the coil. The path of these eddy
currents is circular in nature and is disturbed due to the
presence of a crack in its way. As a result the impedance of
the coil changes [7]. In the case of Eddy Current Testing, if
the current source providing the excitation to the eddy

current coil is not constant, the presence of cracks will not
be detected. It is necessary that the primary magnetic field
remains constant and uniform throughout the course of
inspection in order to detect surface cracks on metal
specimen. This is so because the phase shift that is detected
by the demodulator compares the two input and output

_____________________________________________

ISBN: 978-1-4799-5754-5/14/$26.00 ©2014 IEEE


waveforms of the Howland circuit. That is why the primary
magnetic field (input waveform) should remain constant at
all times so that the output signal can be measured in
comparison with it.
The current source using Improved Howland Pump has
been designed for the same purpose; Eddy Current Testing.
The working and results are discussed ahead.
III. METHODOLOGY

A. Sinusoidal Oscillator
An oscillator is required to provide input to the
excitation source. This is a sinusoidal oscillator with
frequency of 30 kHz and 4 Vpp amplitude. The design
chosen is the phase shift oscillator [9]. It offers a stable
frequency and absolutely no noise to drive the source. The
schematic of the circuit can be seen below in Fig. 2
developed on MultiSim software v13.0.

The project Eddy Current Testing system consists of
two main parts; hardware and the software. The basic
hardware flowchart is shown below in Fig. 1 [8].

Fig. 2. Phase Shift Oscillator Schematic.

The working formula of the above circuit is shown
below in (2) where R= resistance, C= capacitance. [9].
Fig. 1. Hardware Flowchart.

√

The oscillator drives the current pump that excites the
eddy current probe. The probe is used for detecting surface
and through-hole cracks. Whenever a crack is present, the
ECT probe undergoes a change in impedance, which in turn
causes a phase shift between the input and output signals of
the probe. The output from the probe is prepared for
demodulation in order to recover the message signal, which
is basically the crack signal. Once the signal is demodulated
or extracted from the carrier, it is sent to the PC through a
Data Acquisition (DAQ) card for analysis and real-time
plotting.
The excitation portion of an eddy current system
consists of signal generation amplifiers to drive the
transducer (inductive load in this case). The signal generator
is basically an oscillator that provides the Sine Wave
excitation for the eddy current coil. The design emphasis at
this stage is to produce a signal of adequate frequency
accuracy, stability and low distortion levels for reliable

testing [8]. If frequency is inaccurate, the depth of
penetration given in (1), is not as expected to get the desired
results. In (1), ς= skin depth, µ= permeability and σ =
conductivity.

(2)

Using for this circuit V=5V, R=2.2 kΩ and C=101µF,
the frequency output and voltage were approximately 30
kHz and 6 Vpp. These frequency and voltage levels are not
large enough as to produce significant heat in the circuit. As
a result, the output signal is not affected by this reason. All
of the designs have been experimentally tested and no phase
difference occurs due to heat in the system. The only heat
losses produced are by the operational amplifier OPA548
but the output remains stable due to high input impedance
of the amplifier. A PCB was manufactured and can be seen
below in Fig. 3.

(1)
Since excitation is the beginning of the signal train of
the instrument, it is necessary that this signal is accurate and
stable.
In simpler words, we require an alternating current
source that has a constant maximum/peak value. This is a
difficult task as there is a need of a purely sinusoidal current
output that produces efficient and reliable magnetic field for
eddy current testing purposes. The excitation source is the
fundamental requirement of the system.

Fig. 3. Phase Shift Oscillator PCB.

Testing was done by using a digital oscilloscope for
amplitude and frequency. Pure sinusoidal waveform can be
seen in Fig. 4.

509


Amplitude (V)

4
2
0
-2
-4

0

20

40

60

80
Time (us)

100

120

140

160

Fig. 4. Phase Shift Oscillator Output Waveform.

B. Constant Peak Alternating Current Source
A constant peak alternating current (AC) source is
needed to excite the eddy current probe of above 100µH
inductance. The design shown in Fig. 5 is used for this
purpose [10]. It basically uses the concept of voltage
controlled current source [11] [12]. It changes voltage
signal into current signal.

Fig. 6. Improved Howland Current Pump Schematic.

For the Howland current pump, a power amplifier
OPA548 is used due to the following advantages.
•
•
•
•

Fig. 5. Improved Howland Current Pump.

For the circuit shown in Fig. 5, the load current is given
by (3) where Vl = load voltage and Vi = input voltage.
(3)

The circuit works on the principle that all four resistors
should be perfectly matched and the operational amplifier
should have a large gain in order to keep the current directly
proportional to input voltage and the current sense resistor
(R2B).
We have used 0.1% precision resistors in order to keep
the current to remain constant throughout. The voltage at
the two inputs of the operational amplifier remains the same
due to presence of these resistors. The circuit also
introduces a separate resistor R6 for current flow control
and negligible current flows through the four precision
resistors. As a result power dissipation due to them is also
least and only one power resistor is required in the circuit
which is the ‘current sense’ resistor. It is used of only 1Ω
and so I2R losses are not an issue.
The design was simulated on MultiSim software v13.0
along with the protection for driving inductive load of eddy
current coil. Given below in Fig. 6 is the schematic of the
circuit.

510

high current output
current limit feature controllable through RC
capable of driving transducers
TO-220 transistor packaging allows for efficient
heat sinking and easy use

The PCB was designed and maximum output current
limit has been set at 1A by using the relevant resistance.

Other resistance values used in the Howland Current Pump
are given below in TABLE I.
TABLE I. IMPROVED HOWLAND PUMP RL VALUES

S. No.
1
2
3
4
5
6
7

Label
R1
R2A
R3
R4
R2B
Rl
Xl

Value
10 kΩ
1 kΩ
10 kΩ
1 kΩ
1Ω
1Ω
115.7 µH

All precision resistances used are of 0.1% tolerance
level. Fig. 7 has been taken from the datasheet [13] that
shows the safe operating region of OPA548.


Fig. 9. Improved Howland Current Pump with Snubber.

The main component of the snubber is the capacitor
that is comparative to the inductance of the eddy current
probe. During turn off and turn on periods, the capacitor
does not allow sudden change of voltage across it. And
hence the inductor discharges slowly. The time constant is
set by using a resistance in series with the capacitor. The
hardware circuit for the Howland Current Pump can be seen
below in Fig. 10.

Fig. 7. Safe Operating Area Graph (OPA548).

Values of voltage and current were set accordingly
shown in Table II.
TABLE II. VOLTAGE AND CURRENT PARAMETERS

Parameter
Frequency (f)
Vin
Io
Vs-Vo

Value

30 kHz
+/-15V
1A max limit
22V

1) Improved Howland Pump Protection for Inductive
Load
Since the transducer is inductive in nature, protection
against inductive loads is also necessary. A simple RC
Snubber Circuit [14] [15] has been used in this case that
incorporates a resistor and capacitor in series with each
other and parallel to the load. As seen in Fig. 8, the voltage
spikes at turning on/off of supply is greater than 4 times the
output voltage.

Fig. 10. Improved Howland Current Pump PCB.

The PCB in Fig. 10 shows an external heat sink with 11
fins at both left and right sides. It allows for efficient heat
removal from the amplifier as air passes across the heat sink
at all times. The only two power resistors in the circuit are
that of 1Ω and hence other heat losses is not a problem
except that generated by the IC.
The output of the circuit can be seen below in Fig. 11.
The yellow waveform is that of input and the blue
waveform is that of output which is being sent to the eddy
current probe to excite it for the detection of cracks.
20
Amplitude (V)

S. No.
1
2
3
4

10
0
-10
-20

0

20

40

60

80
Time (us)

100

120

140

160

Fig. 11. Improved Howland Current Pump Output Waveform.

Fig. 8. Improved Howland Current Pump without Snubber.

Hence a snubber circuit has been used for this purpose.
This means that any kickback of the inductor would
circulate through the snubber circuit and the IC would
remain safe. The result can be seen below in Fig. 9.

Since the input is pure sinusoidal, the output follows
showing that the input impedance of OPA548 operational
amplifier is very high.
The Howland current pump has been made using the
inverting amplifier configuration. And hence the input and
output signals are 180 degrees out of phase from one
another as clear from Fig. 4 and Fig. 11.
IV. RESULTS AND FUTURE DIRECTION
This constant current AC source has been used for the
purpose of eddy current testing. The Eddy Current Probe [8]
[16] is shown in Fig. 12 and with its specifications in
TABLE III.

511


The probe then moved/dragged over the specimen for its
inspection.
In terms of future work, efforts are being made to
design a very portable Eddy Current System using the same
Improved Howland current pump as a constant current AC

source. And to improve the sensitivity of the system so that
it can detect minor surface and sub-surface cracks as well.
V.

A constant current AC source has been designed by
using the Improved Howland Pump and results prove their
usefulness in Eddy Current Testing System. The constant
current AC source gives a very stable output even on
different values of the inductive load. This study allowed
detection of through-hole cracks and by improving the
sensitivity of the ECT Probe, smaller cracks can also be
detected. The overall system is cost effective as compared
to the internationally available systems being imported to
Pakistan and also acts as a precursor to further NDT
endeavors within the country.

Fig. 12. Eddy Current Probe.
TABLE III. EDDY CURRENT PROBE PARAMETERS

S. No.
1
2
3

Parameter
Coil Radius
Coil Inductance
No. of turns

Value

0.4in
115.75 µH
100

The eddy current probe was used to test through-hole
surface cracks on an Aluminum sheet of thickness 2-3mm
and dimensions of 12’*12’. The signal from the probe was
demodulated and filtered to get the required signal of the
crack. The demodulated signal can be seen below in Fig. 13.

ACKNOWLEDGMENTS
The authors pay special thanks to Dr. Tariq Mairaj, our
Advisor; Dr. Faisal Amir, the Co-Advisor; Mr. Taha Ali,
Research Assistant at NDT Center, PNEC; and the rest of
the NDT team for their continuous support and technical
advices.

Amplitude (V)

20

10

0

-10

REFERENCES
0

20

40

60

80
Time (us)

100

120

140

160

Fig. 13. Demodulator Output Waveform.

The output of the demodulator is a square wave whose
pulse width changes when a crack is detected. But the
change was very minute and could not be picked up
visually. So, this signal was filtered out and the result can
be seen below.

Amplitude (V)

2

Voltage level in Air

1

Voltage Impulse on crack
0

-1

0

2

4

6

8
Time (s)

10

12

14

16

Fig. 14. Low Pass Filter Output Waveform.

In this step, the square wave has been filtered out to get
pure dc voltage. It was observed that the voltage level

dropped when the probe was moved from air to the metal
specimen and then a rising impulse occurred when the
probe passed over a crack. This is clearly depicted in the
waveform of filter output i.e. Fig. 14. In this section, it is
important to note that the position of the probe is changed
uniformly over the test specimen so that all cracks can be
detected. Moreover, the probe is touched with the specimen
to avoid liftoff errors while testing. Otherwise, cracks may
be missed or wrong indications may be received. Hence, it
is very important to keep the probe very close to the metal
specimen in order to concentrate the magnetic field on it.
512

CONCLUSION

[1] E.P. Mix (1987), ‘Introduction to nondestructive
testing: a training guide’. New York, (2nd Ed) Wiley.
[2] Hagemaier, D. J. (1990), ‘Fundamentals of eddy
current testing’, Columbus, OH: American Society for
Nondestructive Testing.
[3] H. L. Libby (1979), ‘Introduction to Electromagnetic
Nondestructive lest Methods’, Wiley.
[4] R. E. Beissner P. HÃller (1988), ‘Theory of eddy
current characterization of magnetic conductors’, Proc.
3rd. Int. Symp. Nondestructive Characterization of
Materials, Saarbrück, pp.541 -548.
[5] R.C. McMaster. (1958), ‘Nondestructive testing
handbook edited for the Society for Non-destructive
Testing’, New York: Ronald Press Co.
[6] J. Garcia-Martin, J. Gomez-Gil, E. Vasquez-Sanchez.

(2011), ‘Non-destructive techniques based on Eddy
current testing sensors’.
[7] Kitchin, C., ‘Analog devices: Amplifier circuits’,
Analog Dialog, p41.
[8] E. James, J. David and J. Eric, ‘Electromagnetic
testing’, chapter 6, pp 172-187.
[9] V. P. Ramon, ‘RC phase shift oscillators’, Universidad
Nacional De Ingenieria.
[10] R.A. Pease (2008), ‘A comprehensive study of the
Howland current pump’.
[11] D. X. Chen, X. Deng and W. Q. Yang (2010),
‘Comparison of three current sources for singleElectrode Capacitance Measurement’, Review of
Scientific Instruments, Vol. 81, No. 3, pp. 1-3.
[12] Apex Micro technology Inc. (2012), ‘Voltage to
Current Conversion’, Apex.


[13] Texas Instruments Inc. (2012), ‘High-voltage, highcurrent Operational Amplifier’.
[14] William McMurray (1972), ‘Optimum Snubbers for
power semiconductors’, IEEE IAS transactions, Vol.
IA-8, No. 5, pp. 593-600.
[15] S. Rudy, ‘Design of Snubbers for power circuits’.
[16] A. Ayad, A. Semmah, A. Lahcen and Y. Ramdani,
‘Numerical and analytical analysis of eddy current
Non-destructive testing in JSAEM benchmark’, JEE
Journal of Electrical Engineering
[17] D. X. Chen, X. Deng and W. Q. Yang (2010),
‘Comparison of Three Current Sources for SingleElectrode Capacitance Measurement’, Review of
Scientific Instruments, Vol. 81, No. 3.

513