Informative booklets for non-destructive testing

NDT
Compact and understandable

Volume 7

THE X-RAY INSPECTION

(RT/RS)

The German Library - CIP-Standard reception

Purschke, Matthias
The X-Ray Inspection (RT / RS)
Castell Publication Inc., Wuppertal 2004
All rights reserved. No part of this publication should
be reproduced, stored in a retrieval system, or
transmitted in any form or by any means, electronic,
mechanical, photocopying, recording or translated.
Printed in Germany
Layout, printing and processing:
Neusser Werbedruck GmbH, Remscheid
ISBN 3 - 934 255 - 22 - 1

Volume 7

The X-Ray Inspection (RT / RS)

by
Matthias Purschke PhD
GE Inspection Techologies (GmbH LLC etc.)

X-ray inspection is used:

to discover inconsistencies and defects of all kinds. This applies to both the
volume and the surface of the workpiece through which the radiation
passes. Not only iron and steel can be subjected to an X-ray inspection, but
almost any type of metal and its respective alloys, as well as ceramics,
synthetic porcelain, concrete, wood and composite materials (carbon-fiber
reinforced and glass-fiber reinforced plastic) can be inspected.
Along with ultrasound and eddy-current methods, X-ray inspection is one of
the most widely used procedures in Non-Destructive Testing (NDT). It is
used to examine welded joints and castings that are used in a diverse array
of applications such as inspection of vehicle components, pressure
vessels, pipelines, chemical systems and many others.

X-ray techniques are divided into 4 types:

Radiography (RT): The X-ray film or digital detector serves as a medium
that makes the X-ray attenuation profile visible. This technique is commonly
applied to welded joints, particularly during field examinations with mobile
x-ray units.
Radioscopy (RS): An electronic detector is used to make the x-ray
attenuation profile visible. This, in turn, enables the inspector to carry out
an on-line inspection through a video monitor. This technique is
commonly used for light alloy castings and assemblies, as well as
stationary inspections (i.e., testing of welded joints during the production
process).
Densitometry: This technique is used to measure the thickness of material
(i.e., steel plates).
Diffractometry: This technique is used to analyze various materials by
examining the diffraction properties of X-rays.
This book covers radiography and radioscopy exclusively. It is important to
note that the X-ray techniques described are only applicable to inspection
objects that can be penetrated by X-rays.
The focus of the book on radioscopic applications is not intended as a
value judgement or a bias towards radiography. It is based solely on the
authors considerable additional experience with radioscopic applications.

Contents:
1.

Fundamentals of X-ray technology

1.1
1.2
1.3
1.4

1.5

2.

7
8
11
12
13
13
16
17
19

Industrial X-ray units

22

2.1
2.2

22
23
23
25
26
27
30
31

2.3
2.4
2.5

3.

Introduction
Generation of X-ray radiation
X-ray adjustments
X-ray attenuation
1.4.1 Radiation attenuation through photo
absorption
1.4.2 Radiation attenuation through scattering
Radiation contrast
1.5.1 Quantitative observations
1.5.2 Noise

Operating/control module
High voltage generation
2.2.1 AC voltage devices
2.2.2 DC voltage devices
2.2.3 High-stability devices
Tube housing
Operation of X-ray units
Verification of device properties
2.5.1 Focal spot measurement according to
IEC and EN 12543
2.5.2 Evaluation of maximum energy according
to EN 12544

31
35

Detectors

37

3.1
3.2

37
39
41
43
45
50
50
53

3.3

Radiographic imaging system

Radioscopic imaging systems
3.2.1 X-ray image intensifier video system
3.2.2 X-ray-sensitive line diode arrays
3.2.3 Semiconductor flat panel detectors
Image quality control
3.3.1 Wire penetrameters
3.3.2 Platinum duplex wire penetrameter

4.

Radiography - Radioscopy

55

4.1

56
57
58
60

4.2

5.

Radioscopic examination systems

65

5.1

67
68
69
69
70
71
72
73
78

5.2
5.3

6.

Radioscopic optical imaging geometry

4.1.1 Ideal imaging geometry
4.1.2 Real imaging geometry
Quantitative description of radioscopy systems

Requirements
5.1.1 Weld examinations
5.1.2 Casting examinations
5.1.3 Recognition of foreign bodies in food
5.1.4 Soft radiation technology
Introduction to digital image processing
5.2.1 Image enhancement
5.2.2 Image evaluation
Computed tomography

Summary and outlook

86

1. Fundamentals of X-ray technology

1.1

Introduction

X-ray radiation belongs to the group of electromagnetic radiation. Figure

1 shows an overview of the entire spectrum of electromagnetic radiation
types according to their respective energy. Because of its higher energy in
comparison to visible light, the X-ray can also penetrate materials that are
opaque to ordinary light. This penetrative ability and the linear
propagation of the X-ray form the basis for all radiographic (with
radiographic film) and radioscopic (without film) examination methods.
These processes provide useful information by generating descriptive
images of the thickness and density properties of the workpiece.

Fig. 1: full spectrum of electromagnetic radiation and energy

specifications, as well as fields of application (according to Becker)
In a technical sense, X-rays are generated through the use of X-ray tubes
(see Fig. 2). Radioactive isotopes are often used in field radiography.
Radiation forms when the atomic nuclei of radioactive substances are
converted. The conversion causes a spontaneous decay; the X-ray is thus
generated. It is not possible to influence the energy and intensity of the
radiation. The type of radioactive substance used is the sole determinant of
these factors.
7

In contrast to the continuous X-ray spectrum (see Fig. 4), radioactive

isotopes have line spectra that vary according to the isotope used. The
isotopes that are most commonly used today are selenium (Se 75), iridium
(Ir 192), and cobalt (Co 60). The isotopes are transported in suitable
containers and used through a mechanical system at the destination. On
one hand, isotopes are easy to handle because no electrical supply is
necessary, but on the other hand, there are substantial legal regulations to
ensure sufficient protection from radiation that can make use and
transportation difficult. Because it is not possible to adapt the energy of the
isotope to the wall thickness of the workpiece, isotopes can only be used
for limited wall thickness ranges (see EN 1435).
1.2

Generation of X-ray radiation

The principle of X-ray generation is illustrated in Fig.2.

Fig. 2: the principle of X-ray generation

A thin wire filament (cathode) inside an evacuated glass or metal-ceramic

tube sends out (emits) electrons. The electrical potential difference (tube
voltage) that exists between the cathode (negative electrical polarity)
and the anode (positive electrical polarity) results in a force that causes
these free electrons to accelerate as they move towards the anode. Once
they strike the anode (target), the electrons decelerate (lose kinetic
energy).
The area of the anode that is struck by the electrons emitted by the
cathode is the thermal focal spot. Its size depends on how focused the
electron beam is, as well as the arrangement and size of the heating wire
filaments. When the electrons decelerate, only approximately 1% of the
energy generated is X-ray radiation. The remaining 99% of the energy
generated is thermal energy. Consequently, the X-ray tube anode must be
cooled intensively. There are several requirements that the anode material
and the degree of focus of the electron beam must fulfill.
The anode material must possess a high melting point. In general,
tungsten (melting point approx. 3000C) is used. The mean temperature of
the focus is about 800C, however, at its peak, it can be far higher.
There must also be a limited degree of focus of the electron beam. The
intensity requirements determine the degree of focal spot focus. High
intensities require large focal spots.
When a thermal focal spot is projected onto an area that is perpendicular to
the central beam, it is referred to as an optically effective focal spot or an
optical focal spot (see Fig. 3). Particularly in radioscopy, the size of the
optical focal spot is fundamentally important - it affects the quality of the
image. Two factors determine the size of the optical focal spot:

the size of the thermal focal spotthe angle of the anode surface in relation to the central beam, called
the anode angle .

This angle is about 20 for usual standard tubes. Anode angles and ray
outlet windows limit the opening angle of the useful bundle of rays to about
40.
9

Fig. 3: thermal and optical focal spot

The X-ray that is emitted by the anode during the deceleration process
leads to a continuous spectrum, from a process called Bremstrahlung
(from the German for braking radiation). A spectrum is described as
radiation intensity I0 over radiation energy E (see Fig. 4). The term
continuous means that within a certain area, all energies are present even if they display different intensities. In contrast, gamma rays display a
specific and discrete line spectrum. The highest point of energy is referred
to as maximum energy Eg.
In radioscopy, depending on the application, the dimensions of the focal
spot vary between 0.3 mm and a maximum of 1.5 mm. In radiography, the
dimensions can measure up to several millimeters. To fulfil the highest
image quality requirements, microfocus tubes with focal spot dimensions
that measure less than 0.1 mm are also used.

10

Fig. 4: energy distribution at a specific tube voltage

The maximum energy Eg can only be used to describe radiation in a limited
manner. This is because it is based on the condition that the kinetic energy
of the electrons is fully converted into radiation. This is a purely theoretical
occurrence. Therefore, we normally use the mean energy or the position
of maximum intensity to describe radiation properties clearly. The mean
energy describes the level of radiation energy at which the shaded areas in
Figure 3 below the curve are equal in size. It is not at the point of maximum
intensity.
The spectrum's total intensity depends on the tube current I and the tube
voltage U values. For the purposes of radiography and radioscopy
applications, the anode material does not need to be considered further.
1.3

X-ray adjustments

The following discusses what impact X-ray operating desk adjustments can
have on the spectrum:

Change of tube current

11

Changes in current I cause a proportional change of radiation intensity I0 for

all energies below the maximum energy. The mean energy, the position of
the maximum intensity and the maximum energy remain unaffected.

Change of the acceleration voltage (tube voltage)

Increasing voltage V causes the point of maximum intensity to shift to a

higher level of energy. It also leads to higher maximum and mean energy
values, as well as a superproportional change of the total intensity.
To affect the radiation quality (penetrative ability and intensity rates),
another option is to insert prefilters into the path of rays between the
radiation source and the workpiece. This is also referred to as hardening of
the spectrum.

Hardening through prefiltration

In this process, the total intensity decreases, but the mean energy and
maximum point of intensity shift to a higher energy value. Materials
commonly used for prefiltration purposes are aluminum (Al) with a
thickness of 1-5 mm and copper (Cu) with a thickness of 0.2 2 mm
upwards.
1.4

X-ray attenuation

The major characteristic of the X-ray is its ability to penetrate. Some of the
radiation passes through a workpiece, but the rest does not. A suitable
detector can record the resulting attenuation and create a radiograph.
In order to understand the interactions that take place when the radiation
passes through the workpiece, it helps to imagine the radiation as
particles that possess a certain energy. In the case of electromagnetic
radiation, these particles are called photons or quanta. You can picture
the attenuation as an intermittent interaction with the electrons of the atom
or the molecules of the workpiece. For the application energy areas
covered in this course (approx. 40-450 KeV), we can describe the
attenuation as the sum of two sub-processes: photo absorption and
scattering.

12

The highly complex secondary processes, which work to undo the

temporary ionization of the atoms, will not be described further at this
instance. In addition, it must be noted that in reality, the reciprocal actions
between X-rays and physical matter are much more complicated. The
simplified description presented here is, however, fully sufficient for a clear
understanding.
1.4.1

Radiation attenuation through photo absorption

Photo absorption is the actual desired and image-creating effect. When the
energy of the X-ray photon hits the test object, it is transferred fully
(absorbed). A portion of the energy is emitted again when a photoelectron is
released, and when X-ray fluorescence radiation is generated (see Fig. 5).

Fig. 5: photo absorption of a photon by an orbital electron

1.4.2

Radiation attenuation through scattering

During the scattering process, the X-ray photon experiences a change of

direction (scattering) with or without a reduction of energy (Compton or
Rayleigh scattering). From energy values of 40-60 keV onwards, Compton
scattering dominates significantly in comparison to Rayleigh scattering.
Therefore, we will concentrate exclusively on the Compton effect. Here, the
intermittent action causes the photon to lose energy while at the same time
a Compton electron is released (see Fig. 6). The scattered radiation thus
generated spreads evenly in all directions and does not generate an
image. This can be compared to illuminating an object from all sides; it
does not produce a silhouette within an optically visible range.
13

Fig. 6: scattering of a photon when it comes into contact with an orbital

electron
The radiation attenuation, which is presented as the sum of both effects,
strongly depends on the radiation energy E. In the case of increasing
energy, the attenuation weakens considerably. In addition, the workpiece
atomic number Z (see Fig. 7) also determines which effect is dominant. For
light alloys and plastics, it is the Compton scattering effect.

Fig. 7: dependence on absorption and scattering of energy and atomic

number
It is not possible to mathematically reconstruct the individual intensity
and/or attenuation profile that results from the radiation attenuation once
the radiation has passed through the workpiece. A quantitative description
14

of radiation attenuation can only be made when simplified significantly. In

addition, two highly theoretical assumptions are made:
narrow bundling of rays and
monoenergetic radiation (only one line in the spectrum).
However, it can be useful for many applications to use the above
assumptions to describe the radiation attenuation behavior mathematically
(see Fig. 8):
I = I0 e - w
I - intensity behind test object
I0 - non-attenuated radiation intensity
- attenuation coefficient
w - irradiated workpiece wall thickness
The attenuation coefficient depends on the:
wall thickness and volume of the workpiece
energy spectrum
material
imaging geometry

Fig. 8: radiation attenuation through absorption and scattering

15

As the radiation energy increases, the attenuation coefficient decreases. In

other words, the workpiece becomes increasingly transparent and thus
absorbs less radiation. This means that as the energy increases,
multithickness workpiece walls can only be imaged with smaller intensity
differences (radiation contrast Cr).
1.5

Radiation contrast

The radiation contrast is of special interest for X-ray examinations. This is

because, in general, there is a significant link between how easy it is to
detect material defects and the contrast quality of the X-ray image.
As radiation energy increases, the attenuation coefficient decreases. This
means that the workpiece becomes more and more transparent (it absorbs
less radiation).
Consequently, as the energy increases, multithickness workpiece walls can
only be imaged with smaller intensity differences I (see Fig. 9).

Fig. 9: dependence of intensity difference on energy

That is to say, as the energy increases, the intensity difference of a
detail (a defect) on the radiograph decreases independently of the
detector used. This must be considered for each respective
examination. "More is better" is the wrong method when adjusting
the X-ray voltage for detecting defects!
16

1.5.1

Quantitative observations

In the following paragraphs, we would like to present a mathematical

definition for radiation contrast. It is based on the intensity conditions during
the X-ray examination. This point of view is intended to explain the major
influence factors on radiation contrast for practical work purposes.
In accordance with the physical processes that take place during
attenuation, the total intensity behind the test object always consists of
the attenuated primary radiation (absorption) and the scattered radiation.
The total intensity Ig behind the test object is the sum of attenuated primary
radiation Ip and scattered radiation Is:
Ig = Ip + Is
Because the scattered radiation spreads evenly in all directions, in cases
where there is only a small difference w in the thickness of the material, the
portion of scattered radiation Is has the same value in all areas behind the
workpiece. The "diffused" character of the scattered radiation appears as a
base intensity or background brightness (radioscopy) / base optical
density (radiography) in the radiograph.
The intensity difference I, which is important for the ability to detect
defects, and which is brought about through the change in wall thickness w
is, in simple terms, the difference between the attenuated primary radiation
values Ip1 and Ip2.
I = Ip1 - Ip2
If we now apply the X-ray attenuation law,
I = I0 e - w
I - intensity behind the test object,
I0 - unattenuated radiation intensity,
- attenuation coefficient, and
w - wall thickness of irradiated specimen,

17

then a connection can be made between I and w as well as the energydependent X-ray absorption coefficient. The result is:
I Ip w
We can then calculate the radiation contrast as the relative difference in
intensity I to the total intensity Ig.
Cr = I/Ig = w / (1 + Is/Ip)
A more general formulation illustrates the difference between two different
attenuation coefficients (test material () and material defect (f)) and
provides a definition of radiation contrast as follows:
Radiation contrast: Cr ( -f) w / (1 + Is/Ip)
This formulation is important for X-ray imaging examinations. It uses the
following definition for the scattering ratio k: k = Is / Ip
Three factors affect the radiation contrast directly:
the difference between the attenuation coefficients
the spread of the defect, parallel to the direction of the beam
the scattering ratio
We can thus infer the following statements for practical examinations:
Air or gas-filled cavities (f 0) such as shrinkholes, etc, generate a
high radiation contrast, which is solely determined by the depth of the
defect w, the workpiece material and the scatter ratio. This also
applies to defects with large surfaces and cracks. This is where the
main advantage of radioscopy comes into play. In terms of their
orientation, defects can be brought into an optimal position in
relation to the direction of the beam.
In the case of metallic inclusions, high radiation contrasts can obtained if
the difference between the attenuation coefficients of the workpiece
material and the inclusion is large. This is often the case with light alloy
castings. As above, an adjustment to the direction of the beam can
enable optimal radiation contrast.

18

 In general, it is possible to achieve a smaller scatter ratio by adjusting

the tube to fade out the primary beam on the specimen section to be
imaged. Such methods should always be used to ensure optimal image
quality or radiation contrast.
1.5.2

Noise

To illustrate the "quantum nature" of radiation, the particle model was

used from section 1.4. It illustrated how helpful it is to imagine radiation as
a stream of particles. Therefore, a mean value is always obtained, even
when we irradiate material with a constant thickness. The intensity
fluctuates around this mean value (see Fig. 10). This variation of the mean
value is called "noise."

19

Intensity
I0

Time

Fig. 10: superposed noise over the mean intensity

The signal-to-noise ratio (SNR) is used to measure the influence of noise
on the actual useful signal. Normally, definitions of this term describe
noise amplitude in relation to the mean intensity value:
SNR large useful signal; mean intensity value I0 is large in relation
to noise good image quality
SNR small useful signal; mean intensity value I0 is small in relation
to noise poor image quality
The SNR depends on intensity. The greater the intensity of the X-ray,
the better/larger the SNR.
In contrast to radiography, the signal-to-noise ratio is very important in
radioscopy. This is because in radioscopy, the images generated can be
viewed live through the use of the appropriate detectors. By necessity,
radiography uses exposure times that are significantly longer in order to
average the intensity. This causes a higher SNR and allows the examiner
to disregard the impact of noise on image quality. Radioscopy either works
with the corresponding higher intensities or uses digital image processing
systems to average the intensity (see Section 3.2).

20

Conclusions regarding radiographic image quality:

Penetration Contrast Intensity SNR
X-ray voltage/power:
increase

more

less

more

more

decrease

less

more

less

less

increase

same

more

more

more

decrease

same

less

less

less

X-ray current:

21

2. Industrial X-ray units

Industrial X-ray units typically consist of the following components:

2.1

operating/control module
high voltage generation
high voltage cable
tube housing
cooling system
Operating/control module

The user can adjust all the necessary parameters through control module
operation:

high voltage
(kV - kilovolts - 1000 V)
tube current
(mA - milliamperes - 0.001 A)
exposure time
(min - minutes)
focal spot
(if the tube allows a choice between two focal spots)
operation modes (constant current, constant power, program)
selection of the warm-up program for tube training
focal spot size and focusing (specific to microfocus systems only)

Fig. 11: operating module of a modern X-ray unit

Modern operation and control modules (Fig. 11) are controlled by
microprocessors. The necessary parameters thus can be configured with
22

ease, and the user benefits from the best protection available against faulty
operation. Images of nominal and actual values are displayed on a screen
simultaneously. The user can operate the device shown in Fig. 11 via soft
function keys. In addition, the system displays messages alerting the user
to all causes of errors that occur through faulty operation or device
malfunction. For a cost-effective integration of the X-ray unit into a
manufacturing process, there is the option to control the entire system
through a serial interface. This can also be done without the control
module, for example with a PC or an industrial control panel
2.2

High voltage generation

X-ray units can be divided into three groups according to the type of high
voltage generation:
AC voltage devices
DC voltage devices
high-stability devices
2.2.1

AC voltage devices

AC voltage devices are used for radiographic weld examinations for ship,
pipe, and steel construction. They are deployed during field examinations
where mobility is essential.
To enable mobility, the high voltage source and the X-ray tube are installed
in a radiation protection housing called the tube housing assembly.
Consequently, high-voltage cables and patch cords are no longer needed.
The tube housing assembly is connected to the control unit through a
fairly long low-voltage cable which allows the configuration of the
parameters (voltage, current, exposure time) (see Fig. 12).

23

Fig. 12: transportable AC device (built around 1960)

These devices are also called single-tank or half cycle systems. The
X-ray tube is supplied with a high voltage of changing polarity. This means
that an X-ray can only be generated during every other half cycle (when the
high voltage adjacent to the anode is positive). The radiation generated is
pulsed at the same rate as the supply frequency. This effect does not
interfere with radiography, because the exposure times are generally much
longer than the net period.
Modern single-tank systems are characterized by the fact that they can
produce a high DC voltage and thus substantially shorten the necessary
exposure times (see Fig. 13).

24

Fig. 13: transportable DC voltage device (built in 2003)

2.2.2

DC voltage devices

DC voltage devices are intended for continuous operation and primarily

used for stationary operation. The high voltage is generated in an area
that is separate from the X-ray tube. Such devices have a rectifier circuit
with filter capacitors that is integrated into the process of generating high
voltage. Therefore, the DC voltage can be generated with a low ripple.
The main advantage when using DC voltage X-ray units, especially for
mobile weld examinations, is that the exposure time is significantly shorter
than for half cycle devices (see Fig. 14). DC voltage devices reduce the
necessary exposure time substantially, particularly when examining larger
wall thicknesses.

25

Fig. 14: comparison of exposure times: iridium, standard single-tank and

DC voltage device (ERESCO 42 MF)
Depending on the high voltage required, one can distinguish unipolar
or bipolar devices. A unipolar system requires one high voltage generator,
while a bipolar system requires two. For unipolar operation the maximum
high voltage is 225 kV, and for bipolar operation it is 450 kV (see Section
2.3).
2.2.3

High-stability devices

Until recently, devices with a converted frequency of 500 Hz instead of a

supply frequency of 50 Hz were referred to as high-stability DC voltage
devices. They enabled voltage fluctuations at a per mille ratio to the
nominal value.
Today, devices that are equipped with the most modern power-electronic
semiconductor components are also considered state-of-the-art for routine
radioscopic examinations. The devices work with a switching frequency of
about 20 kHz and only exhibit high voltage ripples at frequencies of about 40
kHz. These ripples amount to fractions of per mille. Therefore, even at the
intensity of radiation, only the smallest fluctuations occur.
The application of this technology allows a highly compact and durable
power and control electronics set-up (see Fig .15).
26

Fig. 15: high voltage generator cathode with power module (unipolar
system up to 225 kV)
Connecting the power module directly to the high voltage generator
decreases the required usage space. In the case of bipolar systems, which
consist of a cathode generator, power module, and anode generator, it is
possible to stack or place the high voltage generators next to each other.
For the usual radioscopy applications, i.e., routine weld and casting
examinations, no additional cooling is needed for the generators and power
modules.
2.3

Tube housing

X-ray tubes differ according to:

power (maximum 4.5 KW)
nominal voltage (maximum 450 kV)
27

 dimension of the focal spot

design (unipolar, bipolar tubes)
Two different types of design are unipolar and bipolar tubes. One major
unipolar tube characteristic is that only the cathode is supplied with
high voltage. The anode is grounded (0 V), thus only one high voltage
cable and one high voltage generator (cathode generator) are necessary. It
is possible to cool the anode of a unipolar tube of up to 225 kV with water
(see Fig. 16).
water cooling

0 Volt
max. -225 kV
Fig. 16: diagram of a unipolar tube
For isolation reasons, unipolar operation is not useful at voltages greater
than 225 kV. Therefore, bipolar tubes are used for higher tube voltages.
With a -225 kV cathode maximum and a +225 kV anode maximum, the
total voltage amounts to a maximum of 450 kV (see Fig. 17). Bipolar
tubes generally use oil for cooling purposes, and need a heat
exchanger/cooling unit to re-cool the oil.
oil cooling

max. -225 kV
Fig. 17: diagram of a bipolar tube

28

max. +225 kV

Currently, both tube types are called metal-ceramic tubes. They contain
ceramic parts, which are placed between a metal cover and high voltage
parts, providing for electrical isolation. Beryllium is the material used for the
tube window. This material has a small inherent filtration rate and only has
a slight effect on the radiation spectrum.
We can further divide the types of tubes into direct and panoramic
emitting units. Direct emitting units are generally used for standard
applications. The radiation angle of it normally lies at around 40 degrees. In
contrast, a panoramic unit can emit radiation over 360 degrees. These
tubes are often used for weld examinations of round welds centrally located
inside the pipe (the tube is moved into the pipe).
The size of the focal spot, or focus, is an important aspect when selecting
the X-ray tube for a special application. While in radiography, due to special
image conditions, the focal spot sizes are defined in millimeters. In
radioscopy, focal spot sizes are classified as follows:
standard focal spot:
minifocus:
microfocus:

0.4
0.1-0.4
0.1

Geometric magnification of up to 2 and a maximum of 3 can easily be

obtained with the standard focal spot 0.4 IEC (0.85 mm EN 12543).
Minifocus tubes with a 0.2 focal spot (0.3 mm EN 12543) allow for
geometric magnification of a maximum of 6-8. Microfocus tubes are
used for larger geometric magnifications. Depending on the degree of
focus, they can enable magnifications of 50 and more. This is also referred
to as X-ray microscopy.
In contrast to the tubes described so far, microfocus tubes are open
tubes. This means that an additional vacuum pump is necessary to
evacuate the tube body. This type of system allows the user to replace the
wire filament and anode as needed.
Microfocus tubes have focusing systems that can be used to bundle the
electron beam. In contrast to conventional tubes, microfocus tubes contain
a focus coil and a special cathode system in order to bundle radiation. They
also often contain an XY deflection system in order to be able to position
the focal spot on the anode. Typical focal spot dimensions lie within a

29

range of 10 m, for an electrical power of up to approximately 40 W.

Higher wattages result in larger focal spots.
2.4

Operation of X-ray units

The adjustment of the high voltage depends on the workpiece and

determines the penetrative ability of the radiation. The adjustment of the
tube current determines the intensity of the radiation and must also be
adapted to the workpiece.
It is only necessary to pre-select an exposure time in radiography. The
exposure clock is switched off for radioscopic applications.
With double focus tubes, it is possible to switch between a large
"radiographic" and a small "radioscopic" focal spot. Often tubes for
radioscopy are already equipped with two small focal spots so that if one
focus is damaged, the second focus can be used as a substitute without
losing time.
The following modes of operation are available:
Constant current operation: The tube is operated with pre-set operating
values for voltage, current and if necessary, exposure time.
Constant power operation: The tube is operated with pre-set voltage and
if necessary, exposure times. The tube current is automatically adjusted to
the value that the tube power allows for at the specified high voltage.
Program operation: Operating parameters can be entered into programs
to facilitate work on repeated examinations. Entering a program number will
launch the program. This further reduces the opportunity to enter wrong
parameters.
It is necessary to warm-up the tube after each long pause because it is a
vacuum element. The warm-up speed depends on whether the tube has
been out of operation for days, weeks, or months. A microprocessor control
module automatically performs a warm-up according to the length of the
break. Modern X-ray units have a built-in real-time clock. It enables an
optimal adjustment to the new desired operating parameters, while taking
the previous break and operating parameters into account.
30

For X-ray units with sealed tubes, it is only necessary to adjust the
parameters for current, voltage, and exposure time. For microfocus
systems it is also necessary to adjust the focus of the beam. Because the
optimal beam focus changes whenever the high voltage changes, it is
necessary to readjust it for each high voltage value. Since the optimal
focusing current changes whenever the high voltage changes, it must be
adjusted anew for each high voltage. Today, CPU-controlled operator
control panels are available to automate this highly complicated procedure
almost entirely.
2.5

Verification of device properties

The properties of an X-ray unit that are important for the user are the
imaging properties of the tube, and the relation between the adjusted X-ray
voltage and the actual maximum energy attained. The actual maximum
energy attained is a major factor in determining the necessary exposure
time. Because of this fact, the relation to the adjusted X-ray voltage is an
essential factor for quality - especially in radiography. European
standardization authorities took this into account when they drafted EN
12544, which is a binding standard on determining this relation.
It is a fact that particularly in radioscopy, the impact of the focal spot on the
image and image quality is of particular importance. We would therefore
like to introduce the method of measurement according to DIN or IEC,
which has been in use so far, as well as the new reference method of
measurement according to EN 12543. These standards will help the user to
assess the various data provided by tube manufacturers in relation to the
image properties.
2.5.1

Focal spot measurement according to IEC and EN 12543

Focal spots are currently measured in accordance with two standards,

which are identical in content: DIN 6823 and IEC 336. These standards
specify the imaging conditions, the structure of the aperture or pinhole
camera, and the subsequent evaluation of the measurement results.
In order to measure the size of the focal spot, a lead-encased aperture
(pinhole camera) is attached to the front of the tube outlet window. The size
of the aperture is 30 m for nominal focal spot values larger than 0.3, and
10 m for minifocus tubes smaller than 0.3. For nominal focal spot values
31

larger than 1.2 or 2.5, apertures of 75 m or 100 m are permissible. An

image of the focus is projected onto a radiographic film.
For this type of measurement, several problems arise that can considerably
skew the result:
the aperture produces a blurred focal spot image
fixation of the aperture in the central beam
affect of prefilters on the optical density of the film
The most significant disadvantage of the procedure is that measurements
can only be made in a field of energy that is normally limited to a maximum
of 60-70 keV. This is because at higher levels of energy, the radiation
passes through the lead shield of the aperture. Therefore, it is not possible
to measure the focus under nominal conditions.
A magnifying lens is used to evaluate the focus image generated on the
radiographic film (see Fig. 18). The result of this measurement is then
compared to the IEC reference table provided to determine the nominal
focus size.

Fig. 18: two X-ray focal spots on a radiographic film (photographed with a
pinhole camera)
Since it is already possible to weight the measured value towards the
anode-cathode-axis with a factor of 0.7 at the focal spot image evaluation
stage, substantial discrepancies between the data provided by the tube
manufacturers and the true focal spot dimensions will occur in practice.
For example, with a 0.4 focus, the focal spot dimensions that were
determined at that value may be surprisingly high at values of 0.6 mm or
0.85 mm.
32

The nominal focal spot values determined are used in radiography to

compute the minimum distances between the focus and the film (according
to DIN 54111, EN 444 and EN 1435). Naturally, the result of the difference
between the true focal spot dimensions and the specified nominal values
indicates that the image quality in radiography is worse than expected at
the specified nominal values. This does not, however, play a crucial role
because, thanks to the clearly specified standards, all users will obtain the
same distance ratios and the same image quality.
This situation is not as acceptable in radioscopy. Because it uses
geometric magnification in a very specific manner, a physically meaningful
characteristic value that accurately describes the actual focus image
properties is essential.
In the course of the European standardization activities, this was
implemented for both the measuring process and the evaluation of
measurement results. In the future, manufacturers will base the focal spot
data they provide on the actual focal spot measurements. Because this will
force a declaration of larger focal spot measurements, it will have a lasting
impact on the applicable minimum distances in radiography. It will also
have a positive effect on image quality.
The European standard EN 12543 concerning X-ray focal spot
measurement and evaluation also takes into consideration that an accurate
specification of the focal spot size is especially important in radioscopy.
That is why part 1 of the draft standard defined a reference measuring
procedure. All procedures that follow must bear comparison to this
definition. In addition, this standard defines how to accurately specify the
focal spot size based on the measurement results.
This "scanning procedure" is a radiometric method for measuring focal
spots. It scans the focus line by line, computes the size of the focal spot,
and then displays it in a three-dimensional image.
The roughly collimated central beam of the X-ray tube is measured with the
help of a scintillation counter over a double-slit collimator with an effective
opening of 10 m x 10 m. The precollimator, the double-slit collimator, and
the scintillation counter are fitted onto a precision compound slide rest.
Stepper motors are used to move them. A PC that is capable of reading out
the measurements records the output signal. At the same time, a line scan
is carried out in specified increments. The size of the increment depends
33

on the desired resolution. The software facilitates the creation of a highly

descriptive image, and subsequent focal spot processing.
Figure 19 shows the result for a typical focal spot with a clearly pronounced
saddle structure. According to DIN specifications, it is a focal spot with a
nominal value of 1.5. The distribution of intensity displayed in the three
dimensional image has maximum values that correspond to the highest
intensities, and also to the largest optical density that can be measured
when using the pinhole camera method.

Fig. 19: result of the scanning procedure

Figure 20 displays the focal spot's distribution of intensity in the form of
contours. This isodose image shows the isodoses between 10% and 90%
of the maximum intensity.

34

Fig. 20: result of the scanning procedure in the form of an isodose image
In the new standard, the focal spot size is defined as the maximum
expansion in length and width that is evident between the 10% isodoses in
the isodose image. In radiography, the larger of these two
measurements is defined as the nominal focal spot. In the near future, two
focal spot sizes will be declared in the specification provided by
manufacturers: the size according to the new EN, and another according to
the old IEC/DIN.
2.5.2

Evaluation of maximum energy according to EN 12544

The relationship between the specified X-ray voltage and the actual
radiation maximum energy attained (the time required to attain the
maximum energy) is predictable when considering exposure times in
radiography. The user should be able to assume that the maximum energy
specified by the manufacturer of the X-ray unit is an accurate measure of
the required exposure time. This, however, is not guaranteed.

35

Figure 21 outlines comparative data for five different transportable X-ray

units. There are remarkable differences in the necessary exposure times
specified for a steel wall thickness of 65 mm. Four of the units provide
maximum energy specifications of 300 keV. However, it is evident from the
table that the necessary exposure times specified vary considerably.
Interestingly enough, one of the units specifies 275 keV as the maximum
energy value. The same unit (ER 60 MF) also specifies the second shortest
exposure time, and not, as one would expect, the longest.

Fig. 21: comparison of required exposure times for various transportable Xray units for 65 mm
EN 12544 will ensure that the data provided by manufacturers will become
more transparent and easier to compare. It is an important step forward in
ensuring that industrial X-ray unit properties are standardized.

36

3. Detectors
It is possible to use radiation examinations to detect differences in density
or thickness in a test specimen. The distribution of intensity behind the test
specimen must be converted by suitable detectors into a visible image for
further evaluation.
It is for this reason that the detector of the X-ray system is of particular
importance. It is also referred to as the actual imaging system because it
converts an invisible radiation image into a visible X-ray image for
evaluation. Different imaging systems are available for various testing
purposes. The most important are:
Radioscopy:
luminescent screen (also with video camera)
X-ray image intensifier video systems
X-ray-sensitive linear diode arrays
semiconductor flat panel detectors
Radiography:
X-ray film
The purpose of the examination determines which specially adapted
imaging system will be used. The most important parameters of the
different detectors are:
spatial resolution or inherent detector unsharpness - this factor
determines how easy it is to detect small details with a certain
geometric magnification.
contrast - difference in intensity when imaging shows slight differences
in wall thickness
dynamic range - maximum wall thickness difference that can still be
shown
examination speed
cost
3.1

Radiographic imaging system

In radiography, the intensity ratios are recorded on an X-ray film.

37

After the exposure process, the intensity distribution exists as a latent

image. The film is then processed into a document for further evaluation.
The relation between the amount of radiation absorbed by the film (dose)
and the resulting optical density produces the characteristic curve. The
optical density difference S resulting from the intensity difference I
depends on the slope (gradient G) of the characteristic curve and is a
property of the film system used (see Fig. 22). Prolonging the time of
exposure leads to a higher contrast because, according to JtB, the
intensity difference I is smoothed in an area in which the characteristic
curve displays a higher gradient.
Radiographs are evaluated through illumination generated by a film
illuminator according EN 25580. This leads to a negative image.
Commonly used film illuminators allow an evaluation of radiographs up to a
film density of S 4, while some more powerful devices allow an evaluation
of up to S = 5. The irradiation standard (EN 444) demands a minimum
density of S = 2.0 for testing class A and S = 2.3 for testing class B.

Fig. 22: characteristic curve

The human eye can still safely perceive density differences of Smin 1%
provided that these changes in absorption occur abruptly. Every imaging
system has a limited spatial resolution or a certain inherent unsharpness ui.
38

This value is energy-dependent; with a voltage of U = 200 kV, the value

reads ui 0.1 mm.
Another property of the film system (EN 584) that determines image quality
is the graininess of the processed X-ray film. A film exposed to a certain
optical density that is viewed locally will display density variations around a
mean value that is analogous to the noise of the X-ray image (see Section
1.5.2). High image graininess hinders the detection of the defect, especially
when displaying small defects optically (very small S). Changing the film
type to one with lower graininess, however, means that the exposure time
must be prolonged.
Radiography is especially advantageous when mobile operation is a
concern. When using a suitable radiation source, the procedure offers high
contrast sensitivity and sharpness in the optical display of defects.
However, radiography is a static examination method, and the detection of
defects is easily complicated by an unfavorable alignment of the defect in
relation to the direction of the beam.
Film, of course, is seldom worth considering as a detector during
production processes because of the required exposure and development
time. In such cases, radioscopic X-ray systems that allow online viewing of
the X-ray image must be used.
3.2

Radioscopic imaging systems

When using radioscopy, an image of the intensity distribution is produced in

its simplest form on a screen. The ability of a certain material to convert Xrays or gamma rays into visible light is called X-ray fluorescence.

39

Fig. 23: radioscopy with luminescent screen

Analogously to the characteristic curve, there is a connection between the
radiation intensity and the brightness (luminescence). The intensity
differences behind the test object are shown as brightness differences on
the luminescent screen. In contrast to radiography, a positive image is
produced. The relation between radiation intensity and brightness is
approximately linear within a very large range. Compared to the
characteristic curve, the small slope of the transfer function must be
considered, and accordingly, the contrast sensitivity of the luminescent
screen is lower. The internal unsharpness is also higher than that of a film
system. Advantages are:

dynamic examination and optimization of beam direction

regarding defect alignment
serial testing of parts
immediate ACCEPT / REJECT decision
low investment and follow-up costs

The X-ray image intensifying screen with a video camera currently has the
largest practical importance as a detector system. Drastic improvements in
contrast and spatial resolution have made the image quality of this system
comparable to the image quality generated in radiography.
40

3.2.1

X-ray image intensifier video system

Inside the vacuum tube of the X-ray image intensifier (see Fig. 24), an Xray sensitive, luminescent coating made of CsI (cesium-iodide) is located
on the front side. In contrast to conventional luminescent screens, this
coating is very thin and composed of specially aligned CsI crystals. This
luminescent coating converts the X-ray image into a visible image. The
resulting light causes electrons to emerge from the photocathode. They are
accelerated through an electric field that exists between the cathode and
the electrodes of electron optics located on the output screen. The output
screen converts the electrons into a visible image.

Fig. 24: design of the X-ray image intensifier

The light intensity is amplified by about 10,000. Despite being an online
image, the image contrast that can be achieved is comparable to that of an
X-ray film. This, however, causes a relatively high noise content in the
image. Nevertheless, in radioscopy it is possible to apply appropriate digital
image processing methods to reduce noise considerably, which is similar to
prolonging exposure time in radiography.
With its inherent unsharpness rate of approx. 0.3 mm, the image
intensifier delivers a distinctly lower spatial resolution compared to the X41

ray film. Geometric magnification (see also section 4) can compensate for
this deficiency. Increased geometric magnification only leads to an
improvement in detail recognition if the focal spot size is sufficiently small.
Another way to improve detail recognition is the zooming of the image
intensifier. This means that the input field of the image intensifier is
decreased, yielding an enlargement effect in the output window.
Consequently, the area of the specimen to be tested at an angle is smaller,
and longer examination times are required. This is especially true in the
case of weld examinations, where a high spatial resolution is required to
detect defects. Zooming helps improve defect recognition in this situation.
The output image of the X-ray image intensifier is scanned through a video
camera and transmitted to a monitor or a digital image processing system.
The transmission is done according to the CCIR standard (or NTSC in
North America)
The light intensity on the output screen of the image intensifier is scanned
line by line and converted into an electrical voltage. In the video signal (see
Fig. 25), a high light intensity (white) corresponds to a high voltage level.
The individual video lines are displayed on the monitor in the line feed
process. The whole image comprises 625 (640) lines. The full image
frequency is 25 (30) Hz and the half image frequency is 50 (60) Hz. The
attainable resolution is 5 MHz-bandwidth horizontally, and approximately
400 (480) pixels vertically.

Fig. 25: CCIR video signal

42

Depending on the application, different camera types are used. Two

different types of cameras are available for a large range of applications:
CCD Cameras and the Conventional Tube Camera.
Tube cameras are used only if at all for test positions that require a high
spatial resolution, i.e., resolution of small details. Tube camera properties
generally depend on the camera tube used. However, certain properties,
such as inertia, are the same across all tube types. This means that when
examining specimen that are in motion (i.e., casting examinations),
dragging effects or blurs appear, which can considerably reduce the
recognition of small details. In the past, this has increasingly lead to the
primary usage of CCD Cameras, which, along with other positive
properties, have very low inertia.
The CCD camera has a fundamentally different design compared to that of
a tube camera. It is composed of discrete light sensitive diode elements.
The typical number of pixels (picture elements) is 756 x 581. Compared to
the tube camera, it has a somewhat lower threshold of resolution, but a
considerably better contrast transmission in the middle spatial frequency
range (see section 4). It is characterized by a very low inertia and a low
tendency towards glare. It is presently used as the standard camera for
weld and casting examinations.
The properties and main configuration of the video monitor used are of
particular importance for the ability to detect defects in X-ray images.
Interestingly enough, they are not taken into consideration in most
discussions of radioscopic image quality. This is probably because monitor
properties are highly intangible and an examination of these properties
would occur under considerable subjective influences. It should be noted,
however, that no changes should be made to a configuration after
optimizing the monitor adjustments (i.e., with a test image). However, this
does not correspond to common practice. Usually, the testing personnel
adjust the monitor according to their own subjective criteria.
3.2.2

X-ray-sensitive line diode arrays

X-ray sensitive line diode arrays are composed of coated photodiodes that
convert the radiation into the electrical signal. The photodiodes are
arranged along a line so that, in contrast to the X-ray image-intensifying
43

screen, only one-dimensional information about the test object is obtained.

This means that the test object must be guided past the line diode array at
a constant speed. This leads to the test set-up shown in Figure 26.

Fig. 26: schematic examination set-up with linear diode array

An image processor processes the electrical signal and enables its display
on the video monitor. The read out speed of the line diode array is variable,
which means that the signal-to-noise ratio can be improved. However, the
read out speed must be synchronized with the movement speed of the test
specimen, otherwise the image will show geometric distortions. This means
that a fast-moving object requires a higher read out speed, and only a slight
improvement in the signal-to-noise ratio is possible.
The read out speed also depends on the structural design of the linear
diode array. Linear diode arrays with few photodiodes allow for higher read
out rates and object speeds than those with more photodiodes.
The usual width of a photodiode is around 0.45 mm. For special
examination tasks that require a higher spatial resolution, line diode arrays
with a photodiode width of 0.225 mm are also available. The total width of
44

an X-ray-sensitive line diode array depends on the number of photodiodes

built into the sensor (i.e., 1024 photodiodes x 0.45 mm yields a sensor
width of 460 mm).
In practice, X-ray-sensitive line diode arrays exhibit several advantages.
The tube-sided collimation of the ray cone on a fan beam that illuminates
the linear diode array slightly, exposes only a narrow area of the test
specimen to radiation. Consequently, the resulting scattered radiation in the
test object is minimized, and the radiation contrast of the image is
increased considerably.
Moreover, because the individual layers of the test object are only exposed
to radiation for the duration of one image line recording, the exposure of the
test object to radiation is low compared to other imaging systems. For this
reason, such detectors are installed wherever it is necessary that the
exposure to radiation be minimal, such as baggage checks in airports (no
film damage allowed in the test) and inspection for foreign bodies in food.
3.2.3

Semiconductor flat panel detectors

The desire to combine the advantages of an X-ray image intensifier (high

luminance on output, good spatial resolution) with the advantages of a
fluorescent coating coupled directly to an image recorder led to the
development of flat panel detectors or FPDs (see Fig. 27).

Fig. 27: flat panel detector

45

In flat panel detectors, a fluorescent screen is applied directly to a large

surface made of light-sensitive and radiation-resistant semiconductor
material. Usually, the size of the input screens is within the same range as
the size of the X-ray intensifier input screens. However, the distortions
typical of X-ray intensifiers do not appear.
The working principle of flat panel detectors is based on the conversion of
incident X-ray radiation into an electrical charge, which can then be read
out electronically. Amorphous silicon is used as an electrical semiconductive material for this process.
The two principles of conversion are:
scintillator method (indirect conversion)
photoconductor method (direct conversion)
Each method has special advantages and disadvantages, as well as limits
for practical use in imaging systems. For all methods of conversion, charge
carriers are integrated over a certain period before they are read out
electronically. The measurement data is then recorded using appropriate
method, and displayed as an image.
With the scintillator method, a fluorescent screen is applied to the
amorphous silicon to convert the X-ray radiation into visible light. Various
materials are available for this method. The basic design of a flat panel
detector with a conversion layer of phosphor compounds (Gadolinium OxySulfide (Lanex) or similar) is shown in Figure 28.
The incident X-ray radiation produces flashes of light in the phosphor
compound that are transformed into an electrical charge by a photodiode.
Here again, the charge carriers are read out with the help of thin film
transistors, which are approached line-by-line or column-by-column.
By using various chemical compounds and particle sizes in the scintillator
material, different properties concerning spatial resolution, light output or
noise can be enhanced or optimized. A drawback of all phosphor
compounds is the effect of light scattering in the scintillator. A compromise
between image sharpness and signal-to-noise ratio must be made
depending on the material used. In spite of this, it is currently the most
widely used technology for flat panel detectors.

46

Varian comp.
Fig. 28: scintillator method (phosphor)
A flat panel detector consists of millions of light-sensitive pixels that are
arranged in a grid on a rectangular surface. (see Fig. 29).

Varian comp.
Fig. 29: electrical design
47

It is possible to also view this structure on the flat panel detector through a
microscope. (see Fig. 30).

Varian comp.
Fig. 30: actual design
Flat panel detectors with varying pixel sizes and dimensions are available
(see Tab. 3.1). A flat panel detector weighs about 8 - 10 kg. Therefore, it is
well below the weight of an X-ray image intensifier.
FhG-IZFP

Detector 1

Detector 2

0.4 mm

0.127 mm

1.2 LP/mm

4 LP/mm

< 1%

3%

Dynamic range

14,000 : 1

1,000 :1

SNR *

2,500 : 1

1,100 : 1

Pixel width
Spatial resolution
Contrast resolution @200kV

* calculated for: 200 kV, 1mm Cu pre-filter, 1 mA, 1 sec. integration time, 1 mm detector
surface

Tab. 3.1: overview

48

There are distinct differences in spatial resolution between the two flat
panel detectors, depending on the width of the pixels. With its distinctly low
spatial resolution, detector 1 has considerable advantages in dynamic
range and signal-to-noise ratio. Accordingly, a compromise must always be
made between spatial resolution and dynamic range, depending on the
usage, and the appropriate detector must be implemented.
One major difference between a flat panel detector and an X-ray image
intensifier video system is that some flat panel detectors do not have a
real-time function. In addition, there can be problems when it comes to
displaying grey values on a normal VGA monitor, because many detectors
have a very high dynamic range of 14-16 bits (Tab. 3.1, Detector 1). This
has considerable consequences for visual radioscopic examinations. For
example, the test object can only be moved in a stop-go mode;
consequently, the object examination requires a manual drive through the
individual grey value ranges that correspond to the various wall
thicknesses.
A flat panel detector (detector 2) is shown in Figure 31.

Fig. 31: flat panel detector

49

The entire system consists of the actual X-ray-sensitive detector, a

connection cable and an electronic read out system with data acquisition to
display the X-ray image on a monitor.
When flat panel detectors are manufactured, the individual pixels exhibit
varying sensitivities. Furthermore, temperature changes influence the
behavior of the detectors. For this reason, the detector must be calibrated
periodically (gain and offset calibration).
Different manufacturer-specific data acquisition solutions are available for
the flat panel detector read out and image display on a monitor. The data is
immediately processed digitally, and the image is transmitted to a computer
monitor directly (see Fig. 27).
Flat panel detectors are higher in price than X-ray image intensifiers. The
fact that prices are acceptable despite the expensive mechanics and
manufacturing process is largely because electronics and a-Si structures
are mass produced. The manufacturing process is also the basis for
notebook screens (albeit with a reversed working principle), which helps
control pricing in the market.
3.3

Image quality control

The Image Quality Indicator (IQI) generally specifies how good or poor
the image quality is. It is determined using an Image Quality test piece.
Important Note:

Image Quality Indicators serve only to determine

a metric for the image quality or the unsharpness
of the system.
BUT:
There is not a perfect correlation between the
detectability of Image Quality Indicators and the
detectability of natural defects!

3.3.1

Wire penetrameters

The question of how good or poor an X-ray image is can be answered with
the specification of an Image Quality Indicator (IQI). To determine the
50

quality index, for instance, wire penetrameters according to EN 462-1

(see Fig. 32) are laid out on the workpiece and imaged. The diameter of the
wires decreases according to a geometrical progression from 3.5 mm (wire
W1) to 0.05 mm (wire W19). The indicator of the thinnest detectable wire
penetrometer yields the Image Quality Indicator. For example, if wire W11
is still visible, the image has an Image Quality Indicator of IQI 11. The wire
should be of the same material as the irradiated workpiece.
Frequently, the quality of an X-ray image is also specified as a relative
wire detectability, or IQI sensitivity. This is especially true in radioscopy.
This means that a ratio is composed of a wire diameter and a irradiated
wall thickness and then specified as a percentage. For example, a wire
diameter of 1 mm on a 100 mm wall thickness yields a wire detectability of
1%.

Fig. 32: wire penetrameter

Figure 33 shows the wire detectability of an X-ray image intensifier video
system (in percent) taken from the irradiated wall thickness for aluminum
(low percent values mean high wire detectability).

51

Fig. 33: wire detectability with aluminium

Figure 34 shows the wire detectability of an X-ray image intensifier video
system (in percent) taken from the irradiated wall thickness for iron (low
percent values mean high wire detectability).

Fig. 34: wire detectability with iron

52

Fig. 35: comparison of an image intensifier and flat panel detectors

(aluminum)
A comparison of the wire detectability (aluminum) between an image
intensifier and flat panel detectors (see Fig. 35) clearly shows that both flat
panel detectors 1 and 2 deliver a better wire detectability than an image
intensifier system without digital image processing. Except for detector 2 in
real-time mode, wire detectabilities are attained that are nearly equal in
quality to the image intensifier with digital image processing.
3.3.2

Platinum duplex wire penetrameter

The double wire penetrameter (acc. to EN 462-5) (see Fig. 36) is

comprised of a row of 13 elements. Each element is comprised of a pair of
wires with a circular cross-section. The distance between the individual
wire penetrometers in a pair corresponds to the wire diameter.
The largest pair of wires whose separate wire images become blurred into
a single shape is considered the threshold of differentiability. The total
unsharpness is then obtained by using the size of the double wire
diameter. The duplex wire IQI serves only to determine unsharpness
and should always be used in conjunction with wire penetrameters.

53

Fig. 36: platinum double-wire penetrometer

54

4. Radiography - Radioscopy
The essential differences between radiography and radioscopy lie in the
different detector properties and the resulting inspection conditions (see
Tab.4.1).
While noise is of little importance in radiography, it plays a considerable
role in X-ray image intensifier video systems. This is because in radioscopy
it is possible to carry out X-ray image examinations online. Because in
radiography exposure times sometimes reach several minutes, the
quantum noise of x-ray radiation is reduced. In contrast, when using
image intensifiers the exposure time is limited to the 40 ms (0.04 seconds)
specified in the video standard. This leads to a distinct graininess in the
live image. Using digital image processing, it is possible to considerably
reduce the noise in a static image by summing up a number of video
images, similar to prolonging the exposure time in radiography.
Furthermore, the spatial resolution of common detectors in radioscopy is
distinctly lower than the spatial resolution of film (film approx. 0.1 mm,
image intensifier camera approx. 0.3 mm). This means that in
radioscopy an optical imaging geometry is required to permit geometric
magnification. Common factors of magnification are around approx. 1.5 2.5 in casting examinations. The geometric magnification compensates for
the low spatial resolution or higher inherent unsharpness in comparison to
X-ray film.
There are also considerable differences in the flexibility of examination
perspectives. Because of the arrangement of the film, only a fixed, rigid
examination perspective directly behind the test subject is possible in
radiography. However, very few restrictions exist in radioscopy, depending
on the handling system and the geometry of the casting.

55

Radiography
Film detector:
high spatial resolution
high contrast resolution
low noise (with reduction
through long exposure time)

Radioscopy
X-ray image intensifier detector:
medium spatial resolution
high contrast resolution
strong noise
digital image processing: integration

Imaging geometry 1 : 1
geometric magnification approx. 1
influence of optical focal spot:
low if minimum distance is kept

Imaging geometry 1 : 1.5 and higher:

geometric magnification larger than 1
influence of optical focal spot:
large with strong magnification

Test perspective: fixed

Test perspective: flexible

Tab. 4.1: comparison of radiography and radioscopy

Despite the lower spatial resolution of radioscopic detectors compared to
X-ray film, it is possible to attain an imaging quality at least equivalent to
that of film. The prerequisite for this is the optimization of the radiation
image using the flexible options that radioscopy systems offer, especially
geometric magnification and noise reduction through the use of digital
image processing methods. This is of special importance for the field of
radioscopic weld examinations, because it has been clearly defined
through standards.
Conclusion: Assuming that sufficient geometric magnification is used and
the focal spot is correspondingly small, it is possible to achieve an imaging
quality and detail detectability equivalent to that of film.
4.1

Radioscopic optical imaging geometry

The schematic design of a radioscopy system is shown in Figure 37. The

development of the radiation image can be divided in two components: the
generation of a visible X-ray image and the subsequent image transfer. All
imaging systems used in radioscopy produce a positive image of the
attenuation profile (low attenuation equals high intensity in radiation image).
The X-ray source produces the radiation image of the test object through
central projection. An X-ray image intensifier converts the radiation image
into a visible X-ray image. The transfer system that follows consists of a
56

lens system, camera, and monitor. Its only task is to make the output image
of the X-ray image intensifier visible to the operator on a monitor. Through
digitization it is possible to send the image to a digital image processing
system. Digital image processing is used both to enhance the image
quality, and to obtain an automatic image evaluation.

Fig. 37: radioscopic system set-up and imaging system

A consequence of the geometric magnification of test objects is that the
properties of the X-ray focal spot have a decisive influence on the image
quality. To better understand the influence of the focal spot, the ideal
imaging geometry is discussed here, even though it is irrelevant in practice.
4.1.1

Ideal imaging geometry

Figure 38 shows the ideal imaging geometry of a point-shaped X-ray focal

spot. The optical display of the test object is dependent only on its position
between X-ray focus and detector. Thus theoretically, any small object
detail could be displayed or visualized as a radiation image through the
appropriate increase in geometric or direct magnification. The
magnification factor M is consequently calculated as the ratio of the
focus-detector distance (FDD) to the focus-object distance (FOD):
geometric or direct magnification factor M = FDD / FOD

57

 Therefore, the magnification factor M tends towards 1 when the object is

positioned directly on the detector (corresponds to the radiographic
imaging situation)
The closer the object moves to the X-ray source, the greater the
magnification factor M
The magnification factor is exactly 2 when the object is located in the
middle between focus and detector

Fig. 38: ideal imaging geometry

4.1.2

Real imaging geometry

The real imaging situation however, is somewhat different because of the

X-ray focus size (see Fig. 39). The focal spot now causes a geometric
unsharpness ug, which is dependent on focal point size d and on the
selected geometric magnification M. The geometric unsharpness is
calculated as:
geometric unsharpness ug = d (FDD - FOD) / FOD
ug = d (FDD/FOD - 1)
Employing geometric magnification yields: ug= d ( M - 1)

58

or

Fig. 39: real imaging geometry

Geometric unsharpness has such a pronounced impact on the intensity
profile that an edge or sharp structure will have a blurred appearance in an
image, and will not appear as a jump in intensity. For a small object detail,
this means that the loss of contrast due to blurring can lead to the detail
being drowned out by noise. It is then no longer detectable on the image.
The geometric unsharpness caused by the focal spot leads to a limited
resolution of the whole radioscopy system, regardless of the detector
used. The larger the focal point, and the larger the geometrical
magnification M, the more distinct this effect becomes.
Since a geometric magnification must generally be used in radioscopy to
compensate for the low detector resolution compared to film, there are
higher requirements as to the size of the focal spot. Focal spots in
radioscopy are approximately 0.4 to
1.5 mm maximum, while in
radiography focal spots are up to several millimeters in size. Thus in
radiography, a minimum distance between focal point and film is specified
in order to achieve a certain optical imaging quality (EN 1435 / EN 444).
Increasing the distance between tube and detector is not sensible in
radioscopy for several reasons. First of all, a magnification of the FDD
would lead to a decrease in geometric magnification. A further significant
argument against increasing the distance lies in the observation of the
intensity conditions.
The inverse square law states that doubling the FDD means only a
quarter of the surface element intensity will reach the detector. This loss of
intensity would have to be compensated for through a longer integration
59

time with digital image processing in order to achieve a reasonable signalto-noise ratio in the radiation image.
Conclusion: Since, due to the thermal conditions of the focal spot, a
compromise must always be found between required intensity and focal
spot size, it is always necessary to set up a usage-specific system that has
a fixed minimum detectability value.
4.2

Quantitative description of radioscopy systems

For several years, scientists have been working on procedures that will
enable the quantitative capture and assessment of radioscopy system
properties. The definition of the modulation transfer function has proven to
be especially suitable for this technology. With the modulation transfer
function (MTF), it is possible to establish a relationship between
unsharpness and the threshold resolution of a system, as well as describe
the transfer of different image details.
MTF describes the relationship between contrast and spatial frequency.
The spatial frequency is defined as the number of line pairs per millimeter:

1 mm
1 Lp/mm

60

1 mm
4 Lp/mm

How is MTF generated (see Fig.40)?

Fig. 40: generation of the modulation transfer function

In diagram a), a sinusoidal intensity distribution of increasing spatial
frequency is displayed as it would appear on the input screen of an X-ray
image intensifier.
Diagram b) displays the optical image of the intensity distribution from
diagram a), just as it would appear in the video signal. As expected, the
more the spatial frequency increases, the more the contrast decreases until
the threshold resolution is reached.
In diagram c), the contrast found in diagram b) for each spatial frequency is
displayed. The diagram is applied to 1 or 100%. This diagram represents
61

the modulation transfer function. The threshold resolution is, by definition,

reached when the contrast has faded to 10%. The MTF describes both the
threshold resolution of the whole system or the individual components, and
the contrast transmission of small details of the whole system or of the
individual components.
All components of the imaging system are described with an MTF.
However, for the practical user, only the MTF of the entire system is
important in the end. Not only is the threshold resolution important in order
to describe the radioscopy system (spatial frequency, with the MTF faded
out to 10%), but also the whole process - even when spatial frequencies lie
below the threshold resolution. X-ray systems with identical threshold
frequencies may exhibit differing MTF courses, and thereby, differ
considerably in their transmission properties.

Fig. 41: schematic diagram of the optical imaging process

Even if this description of a radioscopy system seems abstract, the MTF is
the quantitative description of what the user may observe in the daily use of
the radioscopy system. To illustrate, the imaging procedure for an ideal
detail (i.e., a pore) is schematically displayed over the whole image chain
up to the monitor in Figure 41. Through every component of the system,
the first ideal orpin-shaped detail, which may quantitatively be described
with the appropriate MTF of the respective components, is blurred.
62

The diagram shows how, due to its specific transfer properties, each new
component widens the detail that was originally ideal with a simultaneous
reduction of contrast. The result of the diagram, i.e., the MTF of the whole
system, may be described as the product of the MTFs of the individual
components. Figure 42 demonstrates this through an example with two
MTFs.

Fig. 42: total MTF as product of individual MTFs

The curves a) and b) in Figure 42 represent the MTFs of the image
intensifier and the camera. The resulting total MTF curve c) is generated
from the product of curve a) and curve b) and is clearly below curve b).
Curve d) is the result of what happens when the output signal of the
camera is reproduced by the monitor.
The diagram makes clear that the total MTF is always worse than the
worst MTF of an individual component of the image chain. This means
that for an optimized system, the best transfer properties possible
must be used for EACH COMPONENT.
To get the complete picture, an approach should be introduced which gives
an extensive description of imaging systems.

63

Image quality is typically thought of in terms of resolution, signal contrast

and noise. There is a contribution to these effects that are determined by
the X-ray spectrum, the part to be inspected and the inspection time,which
puts a limit on the best possible image quality to be expected. In addition,
there is a loss of image quality during the detection process. The ability of
the detector to accurately measure the information present in the X-rays
during the imaging process is the Detective Quantum Efficiency (DQE).
DQE can best be thought of as the square of the signal-to-noise ratio of the
x-rays incident on the detector, divided by the square of the signal-to-noise
measured by the detector for a given frequency:

DQE ( f ) =

( SNROut ) 2
( SNR In ) 2

DQE is the ability of the detector to accurately represent all of the

information present in the X-ray field as a function of frequency. A perfect
detector will give a DQE of one over all frequencies, while a poor detector
will have a DQE that approaches zero. The frequency dependence ensures
that spatial resolution and signal-to-noise are contained in DQE leading to
image quality and probability of detection (POD) estimates.
With DQE, the final image quality can be estimated through this formula:

SNROut = DQE ( f ) SNR In .

For example, assume there are two detectors with different DQE. For the
same X-ray exposure conditions, the detector with the higher DQE will give
the larger SNROut and better image quality. The same image quality can be
achieved in both detectors, but the detector with the higher DQE would
require fewer X-rays, which translates to shortened exposure times and/or
lower tube current.
In general, DQE consolidates the detector performance covering many
individual parameters (resolution, efficiency, noise, etc.) into a single
parameter. Therefore, specifying the DQE for a detector will also help
determine the final image quality and inspection times required for a given
application.

64

5. Radioscopic examination systems

Apart from the technical properties of imaging systems, economic
properties, such as cost, reliability of examination, capacity for automation,
and the speed of inspection, are also important when it comes to practical
use. This realization led to the development of single-function customized
systems in the start of the 1980s.
The radioscopic workstation DP 351 for the automatic examination of light
alloy vehicle components in the production line (see Fig. 43) is used as an
example to illustrate the design and operation mode of an examination
system of this kind.

Fig. 43: radioscopic examination workstation for automatic inspection of

castings
The manipulator principle enables radiation without shadow effects for all
necessary angles of inspection. Additional features (type recognition, sort

65

function, etc.) facilitate integration into the production process, up to the

complete automation of the Feed, Examine, and Sort operations.
In order to achieve a maximal output of parts, high-performance
examination systems of this kind are equipped with a twin manipulator. This
means that two part manipulators are installed. While one loaded
manipulator is inside the radiation protection cabinet and the examination is
in process, the loading, unloading, and marking of the examined parts can
take place simultaneously outside of the cabinet with the second
manipulator. The time required to change parts and the idle time (where no
examination takes place), are solely determined by the rotation of the twin
manipulators. This takes only a few seconds.
Apart from such single-function systems optimized for maximum output,
standard or universal systems have been widespread for years. Flexible
concepts, easy handling and an advantageous price-performance ratio
have led these systems to use in the most diverse areas. Universal
systems are operated either manually or using programmed sequences.

Fig. 44: the universal radioscopic inspection system X-cube compact

66

Internal illumination, an Emergency stop combination, and a safety switch

with isolated contacts are all part of the electrical equipment of an X-ray
cabinet. The protection cabinet, which is made of layers of steel, lead and
another layer of steel, has the dimensions and equipment of a full
protection apparatus based on the current version of official X-ray
regulations (RV, in Germany).
Figure 45 shows an example of an ergonomic operating desk. The display
level contains a monitor with a live image, as well as a monitor for
displaying digitally processed images.

Fig. 45: operating console of an X-cube compact

The control level contains sets of buttons for choosing operation modes,
additional functions, parameter specifications, manual control, buttons for
the examination decision as well as the necessary signal lamps.
5.1

Requirements

To fulfill specific inspection needs, it is always necessary to optimize the

radioscopy system for a particular application. The individual components
67

need to be selected so that an optimal image quality can be attained as a

prerequisite for a reliable examination result. The requirements for different
inspection tasks, such as weld and casting examinations etc., will be
discussed further on.
5.1.1

Weld examinations

The radioscopy standard EN 13068 (Radioscopic Testing) was published in

the year 2001. In recent years, radioscopic weld examinations have
considerably increased in importance. Consequently, radioscopic
examinations have become indispensable in the serial production of
longitudinally or helically welded pipes. Various supervisory institutions
(TV in Germany etc.) have released this examination technique for a
number of users as well.
In weld examination it is often important to be able to discover the
smallest defects (i.e., pores), as well as cracks. Except for the weld
elevation, these components are even. For radioscopic systems, these
constraints mean that high demands are placed on the spatial resolution.
Generally, geometric magnifications of 2.5 or more are used depending on
the wall thickness to be penetrated. This requires a suitable X-ray tube,
often with focus dimensions of 0.3 mm. The imaging system must be able
to transmit even weak contrasts. This requires a high contrast sensitivity,
which is normally tested using image quality indicators (IQI). In addition to
the good contrast sensitivity, the signal-to-noise ratio must be high in order
to be able to assess even the smallest of details. Therefore, the
examination of an item in motion is generally not possible because the
signal-to-noise ratio often must be improved using digital image processing
to integrate the stationary image.
In economic terms, radioscopic examinations generally perform very well
compared to radiography, because there are no film, film processing, or
development costs. However, there is a possibility that it will become a
requirement to fully document all examination results, which could lead to
problems. Normal documentation methods such as video recorders or hard
copies generally have an insufficient image quality. Storing the results
without loss of quality can only be performed digitally. This, however,
requires considerable volumes of storage media, such as magnetic-optical
storage drives, CDs or the like.

68

5.1.2

Casting examinations

Castings generally have distinctively large differences in wall thickness.

Moreover, the size of permitted defects is relatively high in comparison to
weld examinations.. Defects may however, be relatively flat, in regards to
wall thickness. In other words, the depth range may be relatively low. As a
result, the radioscopy system is required to have a large dynamic range
and high contrast sensitivity. The spatial resolution requirements are more
moderate. In any case, the contrast transmission of the system at midrange spatial frequencies, (for mid-range defect sizes) should be excellent.
Over a sufficiently large area the spatial resolution should only decrease
slightly with relation to the modulation transmission function. It should
not decrease to the threshold resolution until higher spatial frequencies
have been reached. The signal-to-noise ratio requirements are also high, in
addition to the high contrast resolution.
Examination speed is of particular importance for the serial inspection of
castings. For this reason, examinations are generally performed on moving
castings. The examination speed, and thus the casting output, is further
increased by the use of specially designed single-purpose systems which
have special casting manipulators, casting feeds and outlets.
An alternative to a programmed examination procedure is a fullyautomatic examination, which involves an image evaluation for casting
assessments. This unmanned fully-automatic examination is now the
industrial standard for various castings made of light alloys (wheels,
suspension components, etc.). This objective examination guarantees high
and constant examination reliability, since it lacks the possibility of normal
human error. When considering the possible cost increases resulting from
the delivery of defective workpieces, computer-controlled inspection has
great potential for lowering costs by eliminating this problem.
5.1.3

Recognition of foreign bodies in food

Due to the product liability law enforced in the European internal market,
new questions have risen relating to quality guarantees for food production.
Alongside the criteria of taste, appearance and adherence to the legal
requirements for provisions, packaged foods must also be kept free of
foreign bodies. Radioscopic examinations offer a solution for both metallic
and non-metallic foreign bodies.
69

The largest technical challenges for this type of examination system are the
vast quantities in which foodstuffs are produced, as well as the high
production rate at which they are produced and packaged. In order to
obtain a 100% examination in a production line, the inspection must be in
motion at a very high speed. A line diode array is usually used instead of
an X-ray image intensifier. Apart from the speeds at which the test items
travel, the need for a low radiation exposure of the test items also plays a
significant role. One of the most important requirements for a radioscopy
system is a good contrast resolution, since foodstuffs contain
predominantly non-metallic foreign bodies, which are only visible via weak
contrasts. Due to the factors mentioned above, visual evaluation by an
inspector is generally impossible. Production line rates also place
considerable demands on the computer-based fully automatic evaluation,
since this also has to be performed on a test piece in motion.
5.1.4

Soft radiation technology

The term soft radiation technology indicates that, for such applications,
low accelerating voltages are used (starting from 20kV). These techniques
are generally used in the examination of plastics or electronic components.
An area of application which has gained importance significantly in the
recent years is compound materials (e.g., carbon-fiber, glass-fiber
reinforced plastics), as well as honeycomb structures (used for the
construction of airplanes). The latter are made of plastic and are built to
carry honeycomb-like structures of aluminum. A typical example of
application is the testing of bonds in critical areas.
The major demand on the radioscopy system is the requirement to produce
images with high contrast and low noise, even at low tube voltages.
Conditions may be entirely different for applications in electronics. The
inspection of integrated circuits (ICs) under certain circumstances makes
the use of microfocus tubes necessary. The reason being, to inspect the
connection wires, which have a diameter of just 2 m, extreme
magnifications of 50 - 80 are required. This is often termed X-ray
microscopy. The tube voltages required for this may also be considerably
greater than for normal soft radiation applications.

70

5.2

Introduction to digital image processing

Digitization provides the opportunity to make the output image of an X-raysensitive detector accessible to an image processing system. Digital
image processing is used to enhance image quality or evaluate images
automatically.
The first step involves converting or transforming the image obtained from
radioscopy to make it computer-compatible, so that it can be understood
by a computer. This conversion involves the digitization and quantization
of the image.
Digitization refers to the analysis or scanning of the image into individual
points called Pixels (Picture Elements). The image is divided into mosaiclike individual values. Common formats are 512 x 512, 640 x 512, and
1024 x 1024 pixels.
Moreover, each pixel has to be assigned a value, called a grey value,
according to the intensity of the spot in question. This process is known as
quantization. X-ray image intensifier video systems divide the intensity
scale from black to white into approximately 256 grey values. The grey
value zero corresponds to black in general, and the grey value 255 to
white.
The image that has been digitized and quantified in this way is also called
the image matrix or digitized image. This process is carried out on the
video signal through the use of an analogue-digital converter. The actual
image processing is, put simply, performed between these two converters
in the image processing system. In image data processing, it is possible
to distinguish between two areas of the application:
image enhancement
image evaluation or image analysis

71

5.2.1

Image enhancement

Various image enhancement techniques have become standard in

radioscopic inspection:
arithmetical averaging or integration
contrast stretching
filtering
The output images from radioscopy systems have a high proportion of
noise. This is mainly due to the quantum nature of radiation. The
integration or arithmetical averaging of several individual video images
during recording leads to a considerable reduction in noise, and thus to an
improvement in the signal-to-noise ratio in the image. This improvement is
highly desirable since it allows details to be recognized more easily.
However, this operation is only possible on a frozen image, because any
movement of the test piece would lead to long streaking and blurs in the
image.
Contrast stretching allows certain grey value areas in the image to be
emphasized and thereby brings low-contrast details of the image into the
foreground. However, in radioscopic weld examinations, this application is
often only useful after a shading correction.
In digital image processing, filtering has a particular importance. Its
possible applications range from noise reduction to the smoothing of small
image details, as well as the emphasis of structures and small details.
Another area of application with this filter operation is the detection of
simple structures, such as edges, borders of objects, corners and points.
Furthermore, it is possible to change the image into a vivid, threedimensional graph. Attempts to achieve automatic defect detection for
welded joints and castings are also based on special filter structures.
The main problem with filtering image data relates to the deliberate
manipulation of the image content. Just as it is possible to make material
defects more visible in the radioscopic image, using a different filter can
produce the opposite effect. This means that users of the digital filters must
have a fundamental understanding of the relations and effects of the
various filters. At the very least, however, the use of filters should be
practiced thoroughly on model parts before actual examination begins.
72

5.2.2

Image evaluation

Image evaluation or image analysis is understood both as the measuring of

structures or intensities in the image, and as automatic defect recognition in
weld and casting examinations. The current performance capacity of such
systems will be illustrated here by taking a closer look at, for example, the
fast automatic image evaluation system (SABA) for the fully automatic
inspection of castings.
Castings which are used in the automotive industry such as wheels, gear
boxes, and suspension components are often important for security and
have to satisfy specific quality regulations. In the corresponding
specifications, details are given on critical values concerning the type, size
and number of defects. Generally a 100% X-ray examination is required to
guarantee the functionality of the part.
The number of parts to be checked of the same type is very high, because
of their mass character. Visual examinations by qualified operators, which
are still common today, are a tiring and monotonous activity that can be
hindered by various distracting subjective influences. For this reason, there
has been a desire for years to automate the X-ray examination of light alloy
castings. These castings are important for security, and therefore make
the examination more objective. The fully automatic serial examination of
light alloy castings has now become the state-of-the-art process for
industrial practice.
The set-up of the fully automatic inspection system only differs from a
traditional radioscopy system in that it has an image evaluation computer.
The design of the examination machine is largely determined by the kind of
castings to be examined. However, the components X-ray source, imaging
system and manipulator for the automatic handling of the workpiece are
always part of the examination system. The image evaluation computer
decides, without any human participation, whether or not the workpiece
satisfies the quality regulations.
In practice, a very diverse set of high demands are made on a fullyautomatic examination system for defect detection in radioscopy images:

73

 A fully-automatic system must guarantee that all defects which

would lead to exclusion according to the quality regulations are
recognized and classified as rejects by the system.
The false reject rate, or the number of parts which the examination
system mistakenly classifies as reject parts, must be negligible, if
only from an economic point of view. This also includes the
requirement for higher system reliability.
The unhindered functionality of a fully-automatic examination
system must be periodically ascertained through an independently
performed test of the system.
The fully-automatic examination should not last longer than a
visual inspection of the part. On the contrary, it should be faster
than the visual inspection, if possible, to gain further economic
advantage.
For practical use, these requirements led to a solution that satisfied the
economic and reliability requirements. While considering the prior-known
structures and shapes of a casting, an optimum defect detection filter
(see Fig. 46) can be used for each area of the workpiece. For a practical
inspection, it is necessary to create a type-specific set of parameters
containing information on the geometry of the workpiece, the inspection
positions and the limits permissible according to the quality specification.

74

Fig. 46: automatic image evaluation approach

In principle, this approach offers several opportunities to develop fast
examination programs. During the practical implementation, attention was
paid to optimized software structures and a cycle-displaced evaluation was
created. This means that a large portion of the operations were necessary
for the image evaluation take place at the same time as the positioning of
the workpiece. The results concluded that, for Accept parts only, the total
examination time was determined by the time needed to actually capture
the image, as well as the time needed to position the workpiece.
To make certain that the unit is in working order, the system periodically
checks the most important components of the unit. These include the
image evaluation computer, the imaging system (X-ray tube, image
intensifier) and the manipulator. Clearly, these system solutions require
cooperation of the image evaluation and the control of the inspection
system.
Since many of the inspection machines used today were created with the
requirements of visual inspection in mind, there are additional requirements
for the examination machines themselves. The differences between
visual and fully automatic inspection must be considered here.

75

If the examination procedure and the inspection requirements of the visual

and fully-automatic examination are compared, clear differences emerge:
Unlike a visual examination, in a fully-automatic system the workpiece is
not examined while in motion, but rather in pre-specified examination
positions. It must be possible to enter these positions and reproduce
them with sufficient accuracy.
Both computers and humans rely on high image quality for reliable
defect detection. However, the respective requirements on image quality
differ considerably. Strong contrasts are required for automatic
examinations. This makes a spectral adaptation of the X-ray radiation
necessary through the use of various prefilters to harden the radiation.
Therefore, swivel mechanisms should be included on the examination
machines. The same basic need of strong contrasts for better defect
recognition exists for visual examinations. However, the operator prefers
a lower-contrast image to avoid their eyes tiring too quickly. Another
difference is in the tolerance of noise in the radioscopic image. A human
compensates for and accepts considerably higher levels of noise if this
allows them to gain additional information for the inspection. However,
for the computer a low-noise image is a prerequisite for reliable defect
detection. Therefore, for fully-automatic examinations, noise reduction
must be provided through digital pre-processing of the image.
The examination specifications and quality standards currently in use
were formulated with regard to visual examinations. They are of limited
use for fully-automatic examinations, since occasionally, algorithmic
conversion of the rejection criteria required is not possible.
Currently, casting designers, when designing a part, do not often take
into consideration the ease with which that particular part can be
automatically examined. This sometimes causes problems for human
inspectors as well. In an automatic examination system, this can be
problematic if labels or stamps are added to areas that require rigorous
examination.
Examining various types of castings (for instance, cast wheels of
various styles, etc.) poses a few problems for human inspectors. A fullyautomatic examination must first be run to recognize which of the
various types of examinations is necessary before the actual
radioscopic examination is performed. This can be affected, for
76

instance, by using a casting code that is removed again in the final

processing of the casting.
The image processing computer controls almost all functions of the
examination system via a digital interface. The harmonious coordination
of the examination system and the computer must be guaranteed for the
most diverse examination units (e.g., for wheels, gear boxes, engine
blocks etc.). This places high demands on the examination and control
software of the fully-automatic examination system.
The first fully-automatic examination systems were tested in practice in
various installations and for several years. Since then, they have become
the industrial standard in the serial examination of a diverse range of
castings. In the test phases, interesting results were also obtained
compared to the reliability of visual examinations.
During a comparative practical test, three operators performed a spot
check on 100 wheels. This spot check contained 18 parts that were
obviously defective and had been certified unanimously as such
beforehand by several operators.
The inspection occurred under similar conditions to those in practice, over
a period of two and a half hours. The result was impressive proof of the
weaknesses of human operators compared to a fully-automatic system
(see Tab. 5.1).

77

Comparative practical test

100 wheels with 18 obvious rejects
Operator 1
Overall acceptance

96

Rejects missed

94

Agreement on rejection

77.8%

Incorrect defect detection

Time taken

Operator 2

6
66.7%

Operator 3
94
6
66.7%

SABA
100
0
100%

2.4 h

2.3 h

2.5 h

2.4 h

Tab. 5.1: comparative practical test (1989)

No operator was able to find all of the reject parts in the spot check. On the
contrary, all three inspectors let 4-6 of the 18 parts pass. It became
apparent that towards the end of the examination, the number of parts
missed rose strongly as a result of increasing exhaustion. The inspection
times of the fully automatic system were of the same order as those of the
inspectors.
It is also worth noting that, in all of these tests, the operators knew the
nature and the goal of the examinations. We may therefore conclude,
without a doubt, that the operators were highly motivated and concentrated
during the tests.
Thanks to an optimal coordination of the examination system and highperformance image processing, the systems used in practice today for
fully-automatic examinations allow for a considerably higher output of parts
than visual examinations, with significantly higher examination reliability.
5.3

Computed tomography

High demands on the quality of materials and components make it

necessary to apply non-destructive inspection mechanisms to optimize the
production process. The need to determine the existence of defects in
materials and components, and to measure and characterize them requires
78

examination procedures with a high spatial resolution. Computed

tomography is particularly suitable for this, and has become increasingly
significant due to the development of computers in the last decade.
Computer tomography (CT) enables a reconstruction of object profiles from
X-ray inspections. This involves irradiating the object from at least 180, but
normally 360, and capturing the individual attenuation profiles (the
projections from various angles). A corresponding object profile can then
be reconstructed via mathematical reconstruction procedures. The
procedure most often used for this is a filtered back projection. The
mathematical basis for this was described in 1917 by Radon. Figure 47
shows the principle of data capture with a CT with parallel ray geometry.

Fig. 47: measuring principle of CT and the use of parallel ray geometry

79

The measuring set-up of CT is displayed in Figure 48. The object to be

examined is rotated step-wise in a fan-shaped X-ray beam and the
individual projections are recorded. By using the fan-shaped beam, the
object has an enlarged image on the detector, which improves the spatial
resolution.

Fig. 48: working principle of the 2D CT using fan beam geometry and direct
magnification.
The complete examination of a component with two-dimensional computed
tomography (2D CT) takes a great deal of time, since many sections of the
object have to be captured. Because the object has to be turned through
360o for each section, several hours or days may be required to measure
its entire volume.
Unlike two-dimensional CT, three-dimensional computed tomography only
requires one 360o rotation of the object. An array detector is used instead
of a detector line to take the radioscopic images (projections). This results
in short inspection times in comparison to 2D CT. The three-dimensional
distribution of the X-ray attenuation coefficients in the object volume is
calculated from the projections using a mathematical algorithm.
The 3D-CT requires a high level of calculation. To achieve short calculation
times, despite the efficient reconstruction algorithms used, multi-processor
systems are necessary in order to calculate the volume while measuring.
The algorithms were implemented on dual Pentium PCs in such a manner

80

that adding further PCs could easily raise the calculation efficiency. The
PCs are connected to each other over a fast network (see Fig. 49).

Fig. 49: schematic structure of an industrial 3D CT system

Various 3D CT systems were developed with this concept. Depending on the
resolution and X-ray absorption required, 450 kV tubes with a focal spot of 1.5
mm, minifocus tubes with a 0.3 mm focal spot, and microfocus tubes are used.
Voxel sizes vary according to the source from (0.35 mm)3 to (0.01 mm)3.

1
2

Reconstruction
volume
[voxel]

Projection
number

Measurement data
[Mbyte]

Volume
Typical
Recondata
measuring struction
[Mbyte] time [min] time2 [min]

255x255x255

200

100

16

2.5

2.5

511x511x511

400

200

128

1023x1023x511

800

800

512

201

60

With extension of measuring area, using a detector array of 512 x 512 Pixels
Using 4 dual Pentium processors with 500 MHz

Tab. 5.2: measuring and calculation time according to reconstruction

volume
Table 5.2 provides data volumes and calculation times according to the
size of the reconstructed volume. The calculation times refer to the use of a
450 kV or a 225 kV minifocus tube and a 16-bit detector array with (512)
detector elements. Shorter calculation times can be achieved by using
81

several computers for the reconstruction. More precisely, the calculation

time is inversely proportional to the number of reconstruction computers.
The volumes are evaluated on a separate visualization station and placed
in an archive.
Figure 50 shows the volume visualization of an aluminum casting, which
originates from a series of experiments for optimizing the manufacturing
process.

Fig. 50: cast aluminum part, 3D visualization

The diameter of this component is about 15 cm. At a (0.3 mm) voxel size
and 5 minutes measuring time, pores measuring 0.4 mm are detected in
82

this object. The visual examination of the volume indicated the presence of
several pore groups, which is attributable to the disadvantageous way in
which the casting was carried out. Based on the distribution of the pores,
experts can draw conclusions about the casting process and optimize it.
Figures 51 and 52 provide us with two xy-planes calculated from the
reconstructed volume through points 1 and 2. Figures 53 and 54 display
two xz-sections through points 3 and 4.
In figure 51, single pores and chains of pores can be seen arranged around
the central borehole. The size of the pores varies between 0.5 mm and
approximately 3 mm. Figure 52 shows a second xy-plane at a distance of
10 cm from the first level through point 2. This image displays a pore group,
as well as a single pore (a). The diameter of the single pore is around
0.5 mm.

Fig. 51: section through the casting Fig. 52: section through the casting
xy-plane through point 1
xy-plane through point 2
Figures 53 and 54 show the depth range of the pore groups. The pores
extend up to 1.5 cm in the z-direction in Figure 53 and 1 cm in Figure 54.

83

Fig. 53: section through the casting Fig. 54: section through the casting
xz-level through point 3
xz- plane through point 4
A 17 cm-long area of an aluminum cylinder head is volumetrically
examined through the use of 3D CT. Because of the high level of X-ray
absorption, the examination is done using 450 kV, with the voxel size
equalling (0.35 mm). Figure 55 shows a cross-section of the cylinder head,
and figure 56 shows a 3D display of a 2 cm high area.

84

Fig. 55: CT-section of cylinder head

Fig. 56: 3D display

85

6. Summary and outlook

For decades, X-ray examination has been one of the most important
procedures in non-destructive testing technology. The major advantages of
this procedure are the direct creation of an image of the test object and, in
the case of radiography, the ability to store the results of the examination
as an image on an exposed X-ray film.
Radioscopy cannot and does not intend to replace classic radiography in
the future. Because of high-performance system technology and the
technology of new X-ray sensitive detector systems, entirely new
application areas will emerge for X-ray examination in the future. Threedimensional computed tomography, for instance, will be used increasingly
for product optimization, reverse engineering, and to measure structures
within the test object.
Even with a history already spanning 100 years, X-ray examination will
continue to play an important part in NDT and other fields in the future.

86