4. Method of testing and instrument technology

4.1 The ultrasonic flaw detector
Before we concern ourselves with further test tasks and their solutions, we must firstly acquire more detailed knowledge about the most frequently applied ultrasonic technique, including test instruments and probes. Based on what has already been stated concerning the location of discontinuities, we must transmit short sound pulses into the test object in order to measure the sound pulse's time of flight from the probe to the reflector and back. This is only possible when there is a clearly defined start time and target time. As long as the test object's sound velocity is known it is then possible to determine, using simple calculation, the distance of the reflector and thus its exact position in the test object, Fig. 12.

Sound reflections in the audio range are called echoes (think of the yodeler in the mountains). Therefore why should we not use this short appropriate term for the reflection of an ultrasonic pulse? Thus the name of the method came into being which is applied in most areas of application for material testing with ultrasonics: the Pulse Echo Method, Fig. 13.

Fig. 12 The priciple of time of flight measurement

Fig. 13 Block diagram: Pulse Echo Method

The time measurement starts with the electrical transmission pulse, the initial pulse. This is an extremely short electrical discharge which triggers a sound pulse at the probe crystal. This pulse travels through the material and is reflected by a discontinuity or the opposing wall and returns back to the probe. The received oscillations are converted into an electrical pulse which stops the time measurement. The distance to the reflector can now be instantly determined by the following formula:

s = sound path [mm] c = sound velocity [km/s] t = time of flight [ms)

Fig. 15 The Display scale Fig. 14 Ultrasonic Testing in practice

If the time of flight is graphically displayed then we are not far from the universal Ultrasonic Flaw Detector, Fig. 14. In order to evaluate the visual signals (echoes) on the screen there is a grid on the inside of the CRT. The exchangeable attachment scale, which has a horizontal scale with 10 graduations is called the display scale, Fig. 15. Using this scale, the ultrasonic operator is able to measure echoes on the display.

How is this done? As already stated, the electrical transmission pulse triggers the sound pulse at the probe crystal. At the same time this voltage pulse is feed to the input of the amplifier so that the high voltage causes a vertical deflection of the display sweep, this is called the initial pulse, Fig. 16a. With this initial pulse, the sweep starts in the lower left corner of the display synchronous to the start of the sound pulse in the test object and moves along the base line at a constant speed to the right, Fig. 16b.

Fig. 16a Initial pulse = Start

Fig. 16b after 10 ms

The speed of the pulse is dependent on the material of the test object (sound velocity = material constant). The sweep speed of the instrument's display can be varied within wide limits. Thus the speed of the display sweep can be exactly matched to the sound velocity. In our example the electron beam reaches scale division 4 while the pulse is at the opposing side of the test object, Fig. 17 a, then it will of course need the same time to return, i.e. the beam spot will be at the 8th scale graduation, Fig. 17 b.

The part of the sound pulse, which is transmitted through the couplant and into the probe, generates a small electrical reception signal at the crystal which, via the amplifier, causes vertical deflection of the beam spot, this is the backwall echo Fig. 18.

Fig. 17a Beam spot at the 4th scale graduation

Fig. 17b Beam spot at the 8th scale graduation

Fig. 18 Backwall echo at the 8th scale graduation

The deflection takes place quickly because the sound pulse is short, therefore can only trigger a short voltage pulse at the probe crystal. The electron beam returns quickly back to the base line and continues to the right, whilst the largest part of the sound pulse is reflected at the coupling surface and travels through the test object a second time. The display indications can now be allocated into two measurement values:
1. Horizontal position.
   left flank of the echo at the 8th scale graduation
2. Vertical amplitude:
   70% screen height

At the moment this does not tell us very much, however, later we will see that nearly all usable results which we obtain from ultrasonic testing are based on these two readings. Let us take a look more closely at the current result: The high initial pulse starts at the left in front of the scale zero point. The rising flank corresponds to the time at which the electrical signal is on the crystal and starts the sound pulse. However, before it is fed to the surface of the test object it must travel through the protection layer of the probe (probe delay). Although it is relatively thin, a short period of time is required. The initial pulse is exactly shifted to the left by this period of time, Fig. 19a.

Fig. 19a Straight-beam probe: initial pulse delay

Fig. 19b Angle-beam probe: initial pulse delay

With angle-beam probes the sound pulse in the probe must travel through a much longer delay path made of perspex before it is transmitted into the test object. Depending on the type of probe, the initial pulse delay can be so large that it no longer appears on the display, Fig. 19 b. We already explained the echo at the 8th scale graduation before: It is the pulse reflected at the opposite wall of the test object, the backwall echo. Now it is not too difficult to guess how the display changes when there is another reflector within the sound beam, e.g. a material separation: between the initial pulse and the backwall echo another echo will appear, caused by partial reflection of the sound wave on a discontinuity, Fig. 20.

Such an echo is called an intermediate echo. It is easy to foresee the position changes of the intermediate echo on the display if the reflector is at different depths. Fig. 21 a+b: the position of the intermediate echo on the display in relation to the position of the backwall echo behaves the same as the distance of the discontinuity related to the total thickness of the test object. We already know a method of determining the distance of an internal flaw; the ultrasonic tester speaks of location of the discontinuity.

Fig. 20 Test object with discontinuity, display with flaw echo

Fig. 21a Discontinuity in front of the backwall

Fig. 21b Discontinuity near the surface

4.2 Near resolution
So, what can we do when a small discontinuity is just below the surface of the test object, i.e. directly in front of the probe? Can this discontinuity still be detected? The answer is no, because the intermediate echo is now within the initial pulse, it is therefore covered by it. Probably there are also no further indications that there is a near-to-surface discontinuity here, Fig. 22.

Fig. 22 A non-detectable near-to-surface discontinuity

Fig. 23 Shadowing of the backwall echo by a larger near-to-surface reflector

Fig. 24 Echo sequence of a near-to-surface discontinuity

Or do we perhaps have a clue which will lead us to the unseen intermediate echo (a near-to-surface discontinuity)? The answer is yes, when the discontinuity is large enough and shadows a noticeable part of the sound beam so that the backwall echo becomes smaller, Fig. 23. If the near-to-surface discontinuity is also smooth and parallel to the surface, then there is an echo sequence which is more or less well formed because the pulses are reflected many times between the surface and the discontinuity, Fig. 24.

Fig. 25 Dead zone: display, test object

In this case, the amplitudes of the echoes become smaller as the distance increases. The more dense the flat echoes advance to the surface, the more the echoes of the echo sequence disappear into the initial pulse, this causes the echoes to become even more dense. In such cases there is a limit to detection.

From everything, we see that the initial pulse is not welcome on the display, however it is a technical necessity: it limits the detectability of near-to-surface discontinuities. Reflectors in the dead zone, the non-testable area immediately beneath the surface, can no longer be detected, Fig. 25. The dead zone is dependent on the test setup, this means from the probe and the test instrument. However, it can be minimized by suitable selection of the testing device.

4.3 The probe
Probes whose beams are normal to the surface are called straight-beam probes, Figs. 1a and 26. Most standard straight-beam probes transmit and receive longitudinal waves (pressure waves). The oscillations of such a wave can be described by compression and decompression of the atoms propagating through the material (gas, liquid and solid), Fig 27.

Fig. 26 Straight beam probe


Fig. 27 Longitudinal wave

There is a large selection of straight-beam probes in various sizes and range from frequencies of approximately 0.5 MHz to 25 MHz. Distances of over 10†m can be obtained thus enabling large test objects to be tested. The wide range enables individual matching of probe characteristics to every test task, even under difficult testing conditions. We have already mentioned a disadvantage of straight-beam probes which, under certain conditions, can be decisive: the poor recognition of near-to-surface discontinuities due to the width of the initial pulse.

Probes whose beams enter at an angle are called angle-beam probes because they transmit and receive the sound waves at an angle to the surface of the test object, Figs. 1b and 28. Most standard angle-beam probes transmit and receive, due to technical reasons, transverse waves or shear waves. With a transverse wave the atoms or molecules oscillate vertical to the wave's direction of propagation, Fig. 29, due to the fact that excitation is made by shear force (transverse to the propagation's directive forces).

Fig. 28 Angel-beam probes








Fig. 29 Transverse wave

Transverse waves only occur in solid materials never in liquids or gases because these do not have a shear modulus and therefore do not effect any shear forces. In addition to this, they propagate much slower than longitudinal waves in the same material. There is no quick reply to the question about why angle-beam probes do not transmit longitudinal waves. In this case a detailed examination is required.

4.4 Refraction and mode conversion
Inclined sound waves are almost exclusively generated so that they occur at an angle to the probe/test object interface, Fig. 1b. This is simply achieved by cementing the element onto a wedge shaped delay path which is normally made of perspex. If a longitudinal wave, at a fixed angle of incidence (the wedge angle), hits the perspex/steel interface then this wave is firstly split-up into a reflected and a transmitted wave, Fig. 30a. Reflected waves obey the reflection law (angle of incidence = angle of reflection) and transmitted waves the refraction law (Snell's law):

a = angle of incidence b = angle of refraction c 1 = sound velocity in medium 1 c 2 = sound velocity in medium 2

Fig. 30a Refraction and reflection without transverse waves

Fig. 30b Refraction and reflection with transverse waves

Fig. 31 Evaluation: one echo - two possible reflector locations

Moreover something strange happens: In addition transverse waves are created at the sound beam's point of impact, Fig. 30b. This happens with reflection as well as with refraction! Due to the fact that the transverse waves propagate at around only half the sound velocity of longitudinal waves, other propagation directions are automatically produced due to the refraction law, i.e. reflection and refraction angles.

If, with inclined scanning, this wave conversion is not taken into consideration, then location and evaluation of discontinuities is not possible in many cases, even detection becomes questionable because one echo on the display leads to two different reflector locations depending on whether one takes longitudinal waves or transverse waves as a basis, Fig.31.

But where is the discontinuity? A clear answer can only be given by the operator when one of the wave modes does not occur. That is undoubtedly the precondition for the universal application of angle-beam probes. This precondition can be derived from the refraction law: firstly we recognize that the refraction angle of longitudinal waves is for steel approximately twice as large as that of the transverse waves, Fig. 30b.

With further enlargement of the angle of incidence the angle of refraction balso increases until finally, at an angle of incidence of a = 27.5° (1st critical angle) , the longitudinal wave, with an angle b of 90°, is refracted. This means that it runs along the interface whilst the transverse wave is still transmitted into the test object, Fig 32a.

Fig. 32a Refraction: 1st critical angle

Fig. 32b Refraction: transverse wave under 45°

Our precondition for clear reflector evaluation is fulfilled: now only one sound wave occurs in the test object, this is the transverse wave with a refraction angle of 33.3° (for perspex/steel). With further enlargement of the angle of incidence various refraction angles of the transverse wave (= beam angle) can be set, e.g. exactly 45°, Fig. 32 b. Finally, with an angle of incidence of about 57° (2nd critical angle) , the transverse wave, with an angle of 90°, is refracted and propagates along the surface of the test object, it then becomes a surface wave, Fig. 32 c.

That is the limit over which no more sound waves are transmitted into the test object. Total reflectionstarts from here, Fig. 32d. The area in which an angle of incidence is present between the 1st and 2nd critical angle (27.5° - 57°) gives us a clear evaluable sound wave in the test object (made of steel), namely the transverse wave between 33.3° and 90°, Fig. 33.

Fig. 32c Refraction: 2nd critical angle, surface wave

Fig. 32d Total reflection

Fig. 33 Usable range for angle-beam probes in steel

4.5 Characteristics of angle-beam probes
Due to the fact that steel is tested in most applications, the angle-beam probes are designed so that suitable angles of incidence are produced in steel. To achieve clear evaluation there are angle-beam probes with angles of 35°, 45°, 60°, 70°, 80° and 90° (surface waves), Fig. 33.

Angles of 45°, 60° and 70° are mostly used. With regard to frequency, angle-beam probes do not have such a wide selection as straight-beam probes. This is primarily due to the fact that high frequency transverse waves in non-alloyed fine grain steels are subjected to high attenuation. As the sound energy of the waves travels through the material it is so strongly absorbed and scattered that only relatively small test objects can be tested with sufficient sensitivity.

If discontinuities have to be detected over larger distances (in thicker test objects) then angle-beam probes with larger crystals and lower frequencies are to be used; e.g. a reflector with a size of 2 mm in low alloyed fine grain steel with a 2 MHz angle-beam probe with a large crystal can be detected up to a distance of 700 mm. The near resolution of angle-beam probes is often better than with straight-beam probes because the initial pulse is shifted far to the left due to the relatively large perspex delay path. The falling flank of the initial pulse could sometimes still cover near-to-surface discontinuities. Figs. 34a+b show, when using an angle-beam probe, how a near-to-surface drilled hole (1 mm deep) can be reliably detected.

Fig. 34a Scanning a 1 mm transverse hole at a depth of 1 mm

Fig. 34b Detection of a hole with a MWB70-4E

4.6 The TR probe


Fig. 35 TR probe: section

If you want to obtain a similarly good near resolution with straight-beam scanning you should use a TR probe, Fig. 35. This technique uses two crystal elements which are acoustically and electrically separated from each other in the same housing. In addition to this, both elements are stuck to a relatively long delay path (made of perspex) and are slightly inclined towards each other. Connection of the TR probe on the instrument is made in the TR or dual mode, i.e. one element is connected to the transmitter and the other with the input of the receiver amplifier. The initial pulse is positioned far left of the display due to the long delay path, Fig. 36.

Multi-reflections within the delay path of the transmitter do not interfer because the transmitter element does not have any reception function. Only when the sound pulses come out of the test object and into the receiver element of the TR probe do evaluatable echoes appear on the display.

Fig. 36 TR probe on the test object: CRT with backwall echo

Fig. 37 TR probe on the test object: discontinuity echo in the cross-talk echo

The electrical and acoustic separation is, due to technical reasons, not completely possible. Especially high gain adjustments and rough test object surfaces cause portions of sound to be directly transferred from the transmitter to the receiver. This generates an interference echo on the display which is called the cross-talk echo. The cross-talk echo can exactly cover the near-surface area of the test object and once again there is a loss in detection sensitivity, especially of small discontinuities. However, most cross-talk echoes are so small, or even negligible, that they can be clearly distinguished from possible discontinuity echoes, Fig. 37.

Fig. 38 Wall thickness measurement with a digital thickness gauge in practice

TR probes are therefore ideally suited for the detection of near-to-surface discontinuities and for thickness measurements on thin test objects. The TR probe reacts considerably less sensitive to coupling variations which may be caused by rough or curved material surfaces. This characteristic explains why TR probes play a valuable part in the chemical and energy generating industries: they are ideal for testing all types of tubes and containers, for the detection of discontinuities in tube walls, and for measurements of inside corrosion and remaining wall thicknesses. Special high temperature probes are even able to measure the wall thickness on test object surfaces up to about 550°C so that installations can be tested during operation.

© NDT.net - info@ndt.net

|Top|