See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/238712986

Part 12. Fundamentals of ultrasonic phased arrays

Article  in  Insight - Non-Destructive Testing and Condition Monitoring · April 2006

DOI: 10.1784/insi.2006.48.4.212

CITATIONS READS

12 1,517

1 author:

Sandy Cochran
University of Glasgow
250 PUBLICATIONS   1,493 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Automatic assessment of super sonic shear wave elastography images of breast tissue View project

FUTURA View project

All content following this page was uploaded by Sandy Cochran on 07 March 2015.

The user has requested enhancement of the downloaded file.

 NDT FUNDAMENTALS

Ultrasonics
Part 12. Fundamentals of ultrasonic phased arrays

S Cochran

U
ltrasonic arrays are now widely used in underwater sonar such as a back-wall or a defect within the beam are detected as
and in more than 25% of medical scans but their use in non- electrical signals.
destructive testing (NDT) has been much less common. As ultrasound occurs as local movement of particles in the test-
However, the cost of the design and manufacture of electronics piece, it is necessary for the particles eventually to return to their
specifically for use with ultrasonic arrays for NDT has recently original position to avoid the test-piece being distorted. This is the
fallen very significantly. In addition, adaptation of the processes to basis of ultrasound as a wave phenomenon.
design and manufacture arrays for medical imaging has taken place A particle in an ultrasound wave moves away from its original
for NDT. Thus, two of the main components for the use of arrays position a very small distance in a particular direction then moves
for NDT are now readily available. The third important component back and away again in the opposite direction. This process is
is the understanding and standardisation of inspection procedures repeated, with the distances becoming smaller, until the wave dies
that will allow common use in routine NDT. Work towards this away at that original position. However, particles nearby take up
has begun. As part of this process, the present article is intended to the motion and this is the basis of wave propagation.
assist the NDT practitioner to extend his or her existing knowledge A very important aspect of wave propagation is the wavelength.
of ultrasound probes and their use to ultrasonic arrays. This is the distance between particles which are moving in the same
Full understanding of arrays is very complicated and previous way because of the ultrasound. Mathematically, the wavelength is
publications often assume a high level of mathematical expertise and related to the speed of the ultrasound and its frequency according
the ability of the reader to invest a lot of time in the topic. Instead, to:
this article seeks to establish understanding of the fundamentals, λ=v/f
allowing the interested practitioner to begin to access other resources
where λ is the wavelength, v is the speed, and f is the frequency at
to extend this understanding according to their choice. The article
which the probe is operating.
begins with a short review of conventional ultrasound probes and
As compression wave ultrasound travels faster than shear wave
the basis of ultrasound from first principles, as a foundation for
ultrasound, it has a longer wavelength, typically approximately
discussion of arrays. This foundation is then developed to take into
twice as long, for the same frequency.
account differences between conventional probes and ultrasonic
To aid the practitioner, a mathematical example is used in this
arrays. The way an array works with electronics is described and
article. To begin with, a probe frequency of 5 MHz is assumed.
the flexibility this provides for the NDT practitioner is emphasised.
In steel, the speed of compression waves can be taken to be
The article concludes with a list of advantages and disadvantages
5900 ms-1 and the speed of shear waves to be 3230 ms-1 (1). Hence the
of ultrasonic arrays and their reasons.
wavelength of compression waves is 1.18 mm and the wavelength
of shear waves is 0.646 mm.
Conventional ultrasound probes Longer wavelengths are affected less by small features in the
Historically, most use of ultrasound in NDT has been based on test-piece such as defects or the microstructure of the material
single- or dual-element probes. An enormous range of such probes itself. This means that, for a given frequency, compression wave
is available. These can generally be divided into two main classes: ultrasound is less sensitive to small defects than shear wave
q Compression wave probes generate ultrasound in which the ultrasound but it has better penetration. Thus, if a practitioner wants
local, almost negligible movement of particles in the test-piece, to achieve the same penetration with a shear wave probe as with a
which is the basis of ultrasound, is in the same direction as compression wave probe, a lower frequency should be used for the
the beam of ultrasound itself. Compression wave ultrasound shear wave probe.
generally travels at high speed and is the only type possible in If the probe is a dual-element type, then one element is used for
liquids such as water. transmitting and the other for receiving. This has some advantages
q Shear wave probes generate ultrasound in which the movement for system design and operation but makes no difference to the
of particles in the test-piece is at right angles to the direction way ultrasound is generated in the test-piece or detected after it has
of the beam of ultrasound. Shear wave ultrasound generally reflected off a defect.
propagates more slowly than compression wave ultrasound and Any given single-element device of either the compression
does not exist in liquids. wave or shear wave type produces a beam in a fixed direction. With
compression wave probes, this is most often at right angles to the
In both cases, a pulse of ultrasound is transmitted into the test- surface of the test-piece with which the probe is in contact, and the
piece and any reflections which return to the probe from features probe is called a 0° probe, as illustrated in Figure 12.1(a). With
shear wave probes, the beam is most often at a specified angle from
the surface of the test-piece, such as 30°, 45° or 70°, as illustrated
Dr Sandy Cochran is with Microscale Sensors at the University of Paisley,
where he presently holds an EPSRC Advanced Fellowship and the position in Figure 12.1(b).
of Reader. His PhD was in ultrasonic arrays for condition monitoring and It is also possible to make a single- or a dual-element probe
since then he has developed his interests in a wide range of topics relating which focuses the beam of ultrasound. In this case, rather than
to NDT. having a beam in a particular direction which gradually spreads out,

212 Insight Vol 48 No 4 April 2006

 Figure 12.1(c). Schematic diagram of a 0° focused compression
Figure 12.1(a). Schematic diagram of a 0° compression wave wave probe, showing how the beam narrows at the focal depth, in
probe, showing the configuration of the beam at right angles to comparison with the unfocused probe shown in Figure 12.1(a)
the surface of the test-piece

at all points across the surface of a large test-piece, conventional

use of a compression wave probe will take a long time. Another
example is when the test-piece is a complicated fabrication which
requires different inspection procedures in different places. In this
case, several conventional probes may be needed. A third example
is when it is necessary to have an electronic record of the condition
of a test-piece with full area coverage. In this case, position sensing
and digital signal recording are necessary.
The examples which have been outlined relate to the capabilities
of conventional probes. In the next section, the capabilities of NDT
with an ultrasonic array are described.

Capabilities of an ultrasonic array

An ultrasonic array is a probe which can be used in different ways
to replace many different conventional probes and which can also
behave in ways that are impossible with conventional probes.
(a) An ultrasonic array can be set up to behave as an unfocused
electronically scanned compression probe. This allows much
faster full area coverage than with a conventional compression
wave probe.
(b) An ultrasonic array can be set up to behave as a focused
Figure 12.1(b). Schematic diagram of a 45° shear wave probe, compression probe with the focal depth scanned electronically.
showing the configuration of the beam on the wedge and in the This maintains sensitivity to defects through a continuous range
test-piece
of depths.
(c) Using a wedge, an ultrasonic array can be set up to behave
the beam is deliberately shaped so that the ultrasound converges at as an angle beam shear wave probe with the angle scanned
a particular distance from the probe, as illustrated in Figure 12.1(c). electronically. This maintains sensitivity to defects around a
Focusing increases the sensitivity of the inspection as a defect at the continuous range of angles.
focal point will be subject to a larger ultrasound signal and a larger (d) An ultrasonic array can be set up to focus its beam and receive
electrical signal will be generated when the ultrasound reflection signals from a range of angles and at a range of depths, giving
reaches the probe. However, this effect does not occur unless the an inspection process which maintains sensitivity at all points in
defect is near the focus. the scanning region within the test-piece.
For some inspection procedures, the choice of probe is very (e) By combining the capabilities of the ultrasonic array, electronics
easy. For example, a compression wave probe is ideal to measure and digital data recording, a full inspection of a test-piece can
the thickness of a simple flat test-piece. However, the use of be carried out with a single physical scan. It is possible for the
conventional probes is more difficult in other situations. array then to mimic any conventional probe just by processing
As an example of a difficulty, if thickness measurement is needed the data in different ways on a computer workstation.

Insight Vol 48 No 4 April 2006 213

(f) With careful control, a single ultrasonic array can act as a
transmitting compression wave probe and a receiving shear
wave probe in the same position. This allows mode conversion
between compression waves and shear waves within the test-
piece to be recorded.
Configurations (a) to (e) could be achieved with conventional
probes, but only much more slowly and with results which might
be significantly poorer. Configuration (f) is impossible with a
conventional probe.
To explain how the capabilities of the array are realised, both
the physical structure of the array and the electronics that are used
with it must be considered. The physical structure is described in
the next section.

Physical structure of an ultrasonic array

The physical structure of an ultrasonic array is shown schematically Figure 12.2(a). Schematic diagram of the internal structure of
a phased array, shown from the side. The wiring layout in an
in Figure 12.2(a). From this diagram, it can be seen that the array actual array would be different and the details of the structure
is actually just a set of small probes arranged side-by-side in a might also differ
single integrated package. This is the physical reality, but it is not
much help in understanding how the array can work so flexibly.
To explain this, a single one of the small probes, called an array
element, is considered first.
The type of element considered here is shown in Figure 12.2(b).
Like a conventional probe, it has a backing material to control
the length of the pulse of ultrasound it produces, and a protective
faceplate to prevent damage from contact with the test-piece. From
Figure 12.2(b), it can be seen that the array element is much longer
than its width. This means that when it is placed side-by-side with
other elements a long line of elements is formed. This leads to the
description of this type of array as a one-dimensional (1D) array.
Because of the shape of the element, the beam of ultrasound
it produces has a broad cylindrical shape. This is also shown in
Figure 12.2(b). The cylindrical beam of the array element means
that ultrasound is transmitted away from it with relatively uniform
amplitude at all angles in the test-piece. As a receiver, the array
element is also sensitive to reflected ultrasound arriving back from
any angle. This is different from a conventional probe, which usually
produces a relatively narrow circular beam with an almost uniform,
large amplitude within the beam and a very small amplitude outside
it, as shown in Figure 12.1.
The beam width can be estimated mathematically using two
different equations. For a conventional circular probe, the beam
width defined by the points at which it has dropped by -6 dB from
its maximum on both sides of the beam is given by(2):
Figure 12.2(b). Schematic diagram of one element of a phased
θc = 2 sin-1 (0.70 λ / φ) array, showing the element detail and indicating the compression
where φ is the diameter of the probe and for a rectangular probe wave beam pattern
such as a single element of an array, the beam width is given by:
ultrasound beam from each element merges with the others to form
θr = 2 sin-1 (0.44 λ / l)
a single beam, as illustrated in Figure 12.3. Thus, an ultrasonic
where l is the length of the side of the rectangle parallel to the plane array with all its elements interconnected will behave similarly to a
in which the beam width is measured. 0° compression wave probe of the same size as the whole array.
Putting in some figures, a 0° circular compression wave probe The aperture of an array can be defined as the outer dimensions
10 mm in diameter has a -6 dB beam width in steel of approximately of the area covered by its elements, as shown in Figure 12.4. The
10°. In the direction parallel to the length of an array element 10 mm length of the array is calculated as:
long, the -6 dB beam width in steel is 6°. In the other direction,
L = (N - 1) δec-c + we
parallel to the width of the array element, assuming it is 0.5 mm
wide, the estimated -6 dB beam width extends to ±90° on each where N is the number of elements in the array, δec-c is the centre-
side of the element. This shows that the beam perpendicular to the to-centre spacing of the elements, and w is the width of an element.
width of the array element is much wider than it is perpendicular The width of the array, WA, is equal to the length of an element. As
to its length, corresponding to the description of the beam as an example, an array with eight elements each 0.5 mm wide, 10 mm
cylindrical. long and with a centre-to-centre spacing of 0.6 mm, is 4.7 mm long
A single array element behaves just like a long narrow and 10 mm wide.
compression wave probe. However, if a number of elements are Using the length and width of the array, it is possible to calculate
used side-by-side, all transmitting the same pulse at once, then the the beam width for the array operating at 0°.

214 Insight Vol 48 No 4 April 2006

 To make the beams from all the array elements merge together at
this angle, the element farthest away from the weld must produce
a pulse of ultrasound first, then the element next farthest away, and
so on until the nearest element is pulsed.
Mathematically, the difference in time between each element,
δt, can be calculated using the formula:
δt = δec-c sin θs / f λ = δec-c sin θs / v
so that the element farthest from the weld is pulsed at time t = 0 s,
the next element at t = 60 ns, and the element nearest the weld at
time t = 420 ns. In this way, the array behaves like a compression
wave probe producing a beam at 45°.
It is important to notice that the previous example is different
Figure 12.3. Diagram showing how the broad beams with from a conventional angle probe operating at 45° on a wedge as the
cylindrical wavefronts emerging from the individual elements of conventional probe produces a shear wave beam whereas the array
a phased array merge together to form a single 0° compression
wave beam produces a compression wave beam. An array can also be used to
produce a shear wave beam by operating it on a wedge. In this case,
the calculation of the time delays is more complicated and software
supplied with the array controller should be used to generate what
is called a focal law.
As an example of focal law calculation, the arrangement in
Figure 12.5(b) can be considered. To produce the focused beam
shown there, in which the pulses of ultrasound produced by the
array converge at a focal point 10 mm below the array, the same
kind of calculation process as before can be used.

Figure 12.4. The aperture of a phased array operating at 0°.

The dark areas are the elements and the overall shaded area
indicates the total effective aperture of the array

Because the structure of the array is much more complicated

than the structure of a conventional probe and because some of its
dimensions are much smaller, it is much more expensive to make
an array. The need for a cable containing many individual co-axial
connections also increases the cost of the array. As a guide, an array
may cost around ten times the price of a conventional probe of
the same engineering quality. However, if the array can be made
to behave like many different probes or if it can be used for an
inspection which is otherwise impossible, then this price difference
Figure 12.5(a). Diagram showing how a phased array can be
may not matter. made to produce a compression wave beam propagating at
Now that the physical structure of the array has been described, an angle such as 45°. Excitation of the farthest left element
the operation of the electronics can be understood. This is covered must take place first, then the next element, and so on until the
in the next section. farthest right element has been excited last

Electronics used with an ultrasonic array

Almost all ultrasonic arrays have a number of elements which is a
power of two, such as 8, 16, 32 or 64. This is because it is usually
easiest to design electronics, called an array controller, with a
number of channels which is a power of two, with each channel
behaving like a separate digital pulser-receiver.
The single most important difference between an array and
a conventional probe is that, by connecting each element in the
array to a separate channel in the array controller, it can be made
to transmit or to act as a receiver at a slightly different time instant
from the other elements. By setting up the array controller so
that each channel operates at a different time instant, the NDT
practitioner can control the way the transmitted pulses or received
signals merge together. Thus, the array can be made to mimic
different conventional probes.
As an example, consider an array with eight elements each Figure 12.5(b). Diagram showing how a phased array can be
used to produce a focused compression wave beam. As the
0.5 mm wide and with a centre-to-centre spacing of 0.6 mm which position of the focal point is controlled electronically, it can
is to be used to produce a beam steered at an angle θs = 45° for a be moved around without altering the physical structure or
procedure such as weld inspection, as illustrated in Figure 12.5(a). position of the array

Insight Vol 48 No 4 April 2006 215

 Using Pythagoras’s theorem, the distance from each array An array system can produce A-, B-, C- and D-scans. However,
element to the focal point can be calculated as shown in the second less probe movement is required. An A-scan is produced by an
column of Table 1. The time it takes for a compression wave array in a single position, but a B-scan can also be produced in this
beam to travel from the array element to the focal point can then way. C- and D-scans can be produced by moving the array only in
be calculated, as shown in the third column, finally allowing the a straight line, whereas a conventional probe must be moved in two
appropriate time differences to be calculated, as shown in the fourth dimensions.
column. A different type of scan produced by an array is the S-scan (sector
scan). This usually has a characteristic triangular shape as shown
Table 1. Calculation of the focal law to focus a compression in Figure12.6. The array is taken to be at the top of the triangle,
wave beam 10 mm below the example array and the signal amplitude is shown with colour or brightness. An
Path length to focal point Beam travel time one Relative time for array S-scan can be obtained from an array in almost any configuration
(mm) way (µs) controller (ns)
including with beam focusing, beam steering, used with a wedge,
Array
element
Lp = √(Dhorizontal2+Dvertical2) t = Lp / v δt = t - tminimum in direct contact with the test-piece or with immersion testing.
1 10.218 = √( 2.1 + 10 )
2 2
1.732 = 10.218 / 5.9 36 = 1732 – 1696 An uncorrected S-scan is like a B-scan but with the A-scans laid
2 10.111 = √( 1.52 + 102 ) 1.714 = 10.111 / 5.9 18 = 1714 – 1696 out in a sector. It therefore presents the data in the form in which
3 10.040 = √( 0.92 + 102 ) 1.702 = 10.040 / 5.9 6 = 1702 – 1696 they are recorded digitally. A corrected S-scan takes into account
4 10.004 = √( 0.32 + 102 ) 1.696 = 10.004 / 5.9 0 = 1696 – 1696 phenomena such as ultrasonic refraction which changes the angle
5 10.004 = √( 0.32 + 102 ) 1.696 = 10.004 / 5.9 0 = 1696 – 1696 of the beam as it propagates. It therefore shows actual positions of
6 10.040 = √( 0.92 + 102 ) 1.702 = 10.040 / 5.9 6 = 1702 – 1696 features in the test-piece such as flaws and surfaces.
7 10.111 = √( 1.52 + 102 ) 1.714 = 10.111 / 5.9 18 = 1714 – 1696 The most important aspects of an ultrasonic array system for
8 10.218 = √( 2.12 + 102 ) 1.732 = 10.218 / 5.9 36 = 1732 – 1696 NDT have now been described. The next section outlines the main
advantages and disadvantages of such a system.

By combining the type of calculation for beam steering with the

calculation for focusing, it is possible to produce a focal law for any
combination of focusing and steering, allowing the array to focus
anywhere within the test-piece consistent with various limitations.
These limitations include the length of the array, L, which affects
the maximum distance at which focusing can be achieved; the
centre-to-centre spacing of the elements, δec-c, which affects the
generation of artefacts in the beam; and the frequency, f, which
affects the penetration depth in the same way as for a conventional
probe.
As an array controller must contain all the electronics for a
separate digital pulser-receiver for each channel, it is much more
expensive than a conventional single-channel pulser-receiver. In
addition, it is vital to have software to calculate focal laws and to
drive the array controller and this means that a PC must be built
into the controller or a separate PC must be connected to it. This
also adds to the cost. However, there is no need for a separate
display or separate manual controls for each channel and as the
electronics for all the channels are the same, the cost is not as high
as might be expected. As a guide, a 32-channel array controller
may cost approximately ten times as much as a single conventional
pulser-receiver. Figure 12.6. Schematic diagram of the S-scan, uniquely used
The two main physical components of an NDT system based to display results from a phased array. Other possible displays
include the conventional A-scan, B-scan and C-scan
on an ultrasonic array have now been described. The next section
of this article briefly describes the way that results from an array
system can be presented, comparing them with results from use of Advantages of an array system
conventional probes.
An array system has several advantages. These include:
q At the simplest level, an array system can mimic many different
Results from an array system
conventional probes, reducing the need to stock multiple probes
The results from a conventional probe in a single position are for different purposes.
usually displayed as an A-scan (amplitude scan) which allows the q For simple configurations, the S-scan should provide a much
NDT practitioner to see the amplitude of the signal as a function more immediate and easily interpreted presentation of results
of time. As ultrasound usually travels at a fixed speed, this is than is possible with conventional probes.
equivalent to the amplitude as a function of distance in the test- q Sensitivity to defects can be maintained throughout the depth of
piece. If the probe is moved in a straight line and A-scan data a test-piece by using dynamic depth focusing in which a set of
are recorded regularly, a B-scan (brightness scan) display can be focal laws is used for different depths in a single inspection.
obtained. This comprises multiple A-scans laid side-by-side usually q The inherent accuracy of the locations of the elements in an
with amplitude represented as colour or brightness. If the probe is array may allow better defect position measurement accuracy if
moved in two dimensions then a top view of the test-piece can be the ‘dB drop’ method can be used without moving the array.
displayed, with the maximum amplitude shown for each point over q Provided ultrasonic coupling can be maintained between the
which the probe has been moved. This is called a C-scan. The final full surface of an array and the test-piece, full area coverage can
conventional way to present results is a D-scan, which is similar to be completed much more quickly with a suitable array than with
a C-scan, but with the data taken at right angles to the C-scan. a conventional probe.

216 Insight Vol 48 No 4 April 2006

q The flexibility of electronic configuration of an array means that Conclusions
it can be optimised for a particular inspection by the practitioner
It has long been a question why ultrasonic arrays have been much
at point of use, potentially avoiding the need for a custom-
less used in NDT than in biomedical imaging but the reasons are
manufactured conventional probe.
becoming increasingly clear. First, real-time imaging is less of an
q It may be possible to complete a complex inspection with a
issue in NDT as the test-piece is static or so quickly moving in
single array configured in several different ways by the array
production that human monitoring is impossible. Second, there
controller rather than with several conventional probes.
is much greater variation between NDT test-pieces than between
q An array will often allow simplification of the mechanics of an
human subjects. Together, these have restricted the market for arrays
automated scanning system, replacing physical movement with
and instrumentation, leading to high prices, low volumes, and lack
electronic control and hence enhancing the reliability of the
of development. Now, with the costs of array controller design and
system.
production falling and the ease of development of software with an
accessible user interface, widespread use of arrays in NDT looks
Disadvantages of an array system imminent. Hence, an understanding of the technology is becoming
Although an array system has several advantages, it also has increasingly important for the practitioner.
several disadvantages, some fundamental and some which will be The most important aspect of the ultrasonic array for the NDT
overcome as the technology and its use mature. These include: practitioner to understand is that it may be viewed simply as a
q The complexity of an array leads to an inevitably higher cost than highly flexible probe which can be made to behave in the same way
for a conventional probe. This may be offset by its flexibility in as many different conventional probes using the control capabilities
replacing several conventional probes, but physical damage to of the electronic array controller and its software. Taking this
the array will lead to a greater one-off financial penalty. approach, three main developments are needed. The first is to
q The complexity of an array means that reliability may be reduce the price of an array further so that it is similar to that of
reduced. In relation to this, there is not yet a recognised and the probes it will be used to replace. The second development is
routine way to check that all array elements are working. to reduce the price and physical size of the array controller so that
q The complexity of an array controller leads to an inevitably they match those of a flaw detector used with conventional probes.
higher cost than for a conventional flaw detector. However, The third development is to improve the software so that an NDT
the price may be expected to reduce as sales accumulate, practitioner requires as little additional training as possible to use
development costs are written off, and competition in supply the array.
increases. Although an array and electronics can replace many different
q Conventional probes and their use are the subject of many years probes, this ignores the possibilities they open up to improve NDT
experience and development of standard procedures. Array beyond the capabilities of conventional systems. In the next article in
systems lack this strong foundation in NDT practice. this series, a number of variations on the theme of ultrasonic arrays
q Exploiting the advantages of an array system demands relatively for NDT will be considered to help the NDT practitioner and the
complex system setup and careful choice of many different NDT development engineer to appreciate what may be possible in
possible configurations. This raises the skill level required by future, so that new developments can be explored most effectively.
the NDT practitioner and increases the risk that a defect may be These variations include curved arrays, arrays with more than one
missed because of an incorrect choice. line of elements, and arrays operated in different ways, not just
q The beam generated by an array is likely to be more complex mimicking conventional probes, for example through the synthetic
than that generated by the equivalent conventional probe. aperture focusing technique and the total focusing method.
This can lead to problems where, for example, the beam from
an array is steered away from the back wall, but the precise References
configuration of the beam leads to unexpected detection of the 1. J Krautkramer and H Krautkramer, Ultrasonic testing of
back wall. materials, 4th ed, Springer Verlag, chapter 31, pp 495-506,
q An array is likely to have a larger contact area than a conventional 1990.
probe, leading to difficulties with acoustic coupling on rigid 2. R/D Tech, Introduction to phased array ultrasonic technology
surfaces. applications, R/D Tech, pp 44-48, 2004.

AEROSPACE NDT Symposium

26-27 April 2006
BAWA Conference Centre, Filton, Bristol

Tickets:
Single day – £70/Both days – £110

For information contact

BINDT Conference Department
1 Spencer Parade, Northampton NN1 5AA, UK.
Tel: +44 (0)1604 630124; Fax: +44 (0)1604 231489; E-mail: conf@bindt.org

Insight Vol 48 No 4 April 2006 217

View publication stats