Introduction

The inspection possibilities with ultrasonic phased array technology are closely linked to the reliable operation of the ultrasonic of the electronic equipment and the soundfield properties of the array probes. Since many of the specific characteristic properties are not comparable to those of standard ultrasonic equipments and -probes and are in addition not common within the medical diagnostic field one has to introduce some procedures specific for phased array systems in NDT.
The reliability and reproduceability of inspections with such complex systems are much more depending on the correct functioning and the correct calibration of the system then any other ultrasonic inspection equipment. Since the calibration procedures are in general not sensitive enough for the detection of all parameter deviations from the their standard values, it is nescessary to consider special steps and tools for the measurement of the most important probe-and equipment characteristics

Fig 1


Figure 1 shows a schematic presentation of an ultrasonic phased array equipment with probes and single elements, cable and cable adapter, the transmitter, the receiver, the A-scan converter and the data acquisition computer. At the different parts of this block diagram are indicated the figure numbers, which contain more detailed information for this specific equipment partition.
All the parameters which will be discussed in the following have to discriminate two different purposes: There are data sheet parameters which should be guaranteed by the equipment manufacturer and there are operational parameters which can be adjusted, calibrated and modified by the user of such a system . Although this basic difference is very common in standards for the characterization of ultrasonic equipments, in the case of phased array equipments we have to reconsider this seperation in view of the different technical concepts of the used equipments.
Data sheet parameters and design criterias are closely linked. E.g. for the probe design (element size, element spacing, distribution of differently dimensioned element sizes, sensitivity apodizing etc.), it is important to know the relationship between grating and main lobes at the minimum and the maximum skewing -or incidence angles. This can be optimized by a suitable modelling of the probe soundfield structure [1,2]. The software may influence the grating lobes by an amplification apodizing at the border elements of an array or by randomly chosen small variations of the delay times. It is therefore recommanded to document the grating lobes for the different wave modes and to regard them also as an important operational parameter (see also chapter 3).

Data Sheet Parameters	Operational Parameters
Proto type, frequency and wave mode size of the effective transducer dimensions range of the skewing or the incidence angle delay path length wedge angle number of elements sensitivity variations of different elements frequency range	Crosstalk between probe element channels sensitivity variations due to: skewing - or incidence angles and focussing software controled parameters (angles, beam divergency, focussing,grating lobes)
number of transmitter and receiver element channels maximum delay time maximum delay time increment number of probe channels (T/R - Mode/Pulse - Echo - Mode) Cable length between probe and equipment	Crosstalk at transmitter/receiver channels amplitude transfer behaviour(linearity, noise, saturation) frequency transfer behaviour(bandwidth) signal to noise ratio delay offset gain compensation of DAC curves (TGC)
Sample rate resolution in bits real time storage rate of A-Scans	dynamic range and amplification(effective number of bits, digital offset, undersampling of echopulses)

Fig 2

Figure 2 gives a list of parameters for both groups. Data sheet parameters are e.g.:

• For the probe:
  probe type,-frequency and wave mode, number of elements, size of the elements, sensitivity variations of different elements, size of the effective transducer dimensions, delay path length, range of variation of the skewing and the incidence angle. Angle of incidence for a zero delay-time distribution or wedge angle.
• For the equipment:
  maximum cable length between the probe and the equipment., number of transmitter and receiver channels, number of probe channels(in T/R-Mode and in Pulse-Echo-Mode), frequency range, maximum delay time, minimal delay time increment.
• For the A-scan converter:
  sample rate, resolution in bits, maximal and minimal dynamic range, maximal pulse repetition rate of the the A-scan conversion and data storage. Due to the software based variability of some of those parameters, many of them have to be regarded as well as operational parameters.

Some typical operational parameters are:
amplitude transfer behaviour (linearity), dynamic range in dependency on the amplification calibration and adjustment, frequency transfer behaviour (bandwidth).
Crosstalk between different element channels at the probes and at the transmitter/receiver channels within the equipment.
Signal to noise ratio for a given target and at a given distance. Software depending parameters like angles, angle beam divergency, focussing.
Sensitivity variations due to changes of the skewing- or incidence angles and of the focussing, sensitivity corrections depending on the element channel and depending on the angle and the focussing, sensitivity corrections related to the DAC curves (TGC).

In the following specific problems with some of the listed parameters are described in detail.

Probe related parameters

Beside the classical soundfield and spectral characterization of conventional probes there are three typical sets of parameters which must be recorded and measured at a phased array probe:

• element sensitivity
• element separation (acoustic and electric insulation)
• directivity pattern and grating lobe suppression

Fig 3

Fig 4

Figure 3 presents two directivity pattern measurement arrangements with electrodynamic sensors as contactless microphons on test blocks with a cylindrical scanning surface. Recommanded is a fine grain ferritic or austenitic steel block, similar to that used for the IIW block (ISO 2400). With the help of such a block the individual maximum amplitudes of the elements can be measured as well as their pulse shape and spectrum. The measurement equipment of figure 3 should use as a standard pulse e.g. a needle pulse at a 50 ohm resistor with 5 nsec rise time and 250 V exitation voltage. If squarewave pulses are used the pulse duration should be in the range of 0.8 times half the period of the center frequency. The quality of the directivity measurement with the arrangement of fig. 3 can be characterized by the comparison between a theoretical calculation based on the model of [1] and [2] and the measurement as shown in Figure 4
Since the element sensitivity may deviate from each other it is nescessary to measure it e.g. with semicylindrical test blocks like those shown in fig. 3 for probes with a skewing or an incidence angle variations.

Fig 5


Figure 5 shows an example for the variation of the element sensitivities, which can be compensated by a corresponding software based correction curve individually applied to the different element receiver channels. The variability of the different elements is strongly depending on the production of the probe, especially the cutting between the elements. For the modern composite piezoelectric materials, there are different procedures to cut a transducer surface into different elements resulting in a more or less strong sensitivity variation. A strong variation of the sensitivity reduces the dynamic range of the whole system.

Fig 6

Fig 7

Fig. 6 shows a typical pulse and spectrum of a 2 MHz array element.
The possible presence of secondary vibration modes is strongly depending from the design of the probe concerning the element thickness, the cutting, the element width and the kind of the piezoelectric material (standard ceramic material or piezo composite material). The pulse shape and the spectrum are a good indicator for those modes. Therefore it is recommanded to document pulse shape and spectrum of the elements in order to check for possible changes due to aging or other phenomena.
Deviating spectrum and pulse shapes of the individual elements are also giving first hints on irregularities e. g. due to a bad element separation.

The element separation or the crosstalk between different channels is the second important influence being typical for phased arrays. The separation can be reduced by acoustic or electric cross talk. Whereas the acoustic crosstalk is mainly defined by the transducer design and is not changing to much due to aging influences, the electric crosstalk can be influenced by many other parameters like cable, cable connectors, aging of some electronic components within the probe, the receiver and the transmitter and others. At the probe it is interesting to observe the element directivity patterns in order to document possible crosstalk deficiencies. As shown in Figure 7, elements with a too strong crosstalk can easely be detected due to a strong reduction of their directivity divergency. The theoretical curve in fig. 7 , calculated with the model described in [1] and [2], defines the ideal separation status between the elements. A whide divergency is needed in order to guarantee a large range of steerable angles for a phased array probe. If the directivity pattern is too strongly reduced by crosstalk influences, it is nescessary to compensate the reduced steerability by sensitivity corrections which is resulting in a decreased dynamic range of the whole system.
The grating lobes of a probe system can also be calculated according to [1] and [2] and be compared with the practical measurements. (see fig. 10 in section 4).
It is recommanded to store all the probe parameters into a data bank for the purpose of fast comparison and the early detection of aging phenomena.

Equipment related parameters

Amplitude Transfer Function Frequency Transfer Function Fig 8

The cable and the equipment have to be checked for transmitter, receiver and A-scan conversion parameters.
At the transmitter channel the constancy of the pulse exitation voltage and the pulse shape as well as the independency from the different delay times are essential parameters which should be documented. This can be measured qualitatively by a suitable control software or more quantitatively with a special hardware support. (see section 5)
The receiver channel can be characterized by two transfer functions (Fig. 8):

• the amplitude transfer function to check the linearity of the amplifier behaviour
• the frequency transfer function to check possible limitations of the frequency band width.

From the amplitude transfer behaviour one can derive the maximum dynamic range between noise and saturation level. The noise level depends on the amplification settings and for digital equipments also on the handling of the lower bits after the A /D conversion. It is recommanded to check the amplitude transfer behaviour for the whole equipment and for the single channels at different amplification setting. It must be mentioned that especially during the sensitivity calibration procedure using large reference reflectors like a back wall or a larger side drilled hole it may happen that one element channel delivers a signal amplitude above the saturation level of that individual receiver channel. This will not be detected observing only the total receiver output signal which adds up all individual element channels. It is therefore an essential for phased array equipments to provide a kind of monitoring for the saturation at individual element channels.

Fig 9

Fig. 9 shows the dynamic range problem at the different stages of a receiver channel. It is of course depending on the kind of amplification settings in the preamplifyer stages and on the kind of digitalization used. The different corrections to be applied at the element channel amplification - e.g. to compensate for the element sensitivity variations, the angular sensitivity dependency or the DAC curves with the TGC - are complicating the problem and are often requiring an individual optimization. It is interesting to remark that the relationship between the sampling frequency of the digitization in relationship to the carrier frequency of an ultrasonic pulse plays an important role for the maximum transferable dynamic range.
Depending on the kind of sensitivity needed for a specific job one has to choose an optimal amplification setting during the calibration procedure in order to guarantee that the inspection is not limited by problems with the signal to noise ratio or with the saturation. This problem exists also at automatic inspections with conventional ultrasonic equipments using linear amplifiers but it is much more dramatic for phased array equipments, where the limitations of the signal to noise ratio cannot be detected observing the total output signal at the receiver.
For the frequency transfer function it is recommanded to use a continuous wave exitation at the input of all element channels in order to avoid a too strong influence from slightly deviating delay times. The use of a sinus burst pulse results in a more or less strong dependency of the measured transfer behaviour on the burst length. Fig. 8 (bottom) shows examples for the frequency transfer function of an ultrasonic equipment with the total transfer behaviour and the behaviour of a single element. The typical oscillations of that transfer function are indicating some problems with the homogeneity of the delay times of the different channels. Those oscillations should not be greater than +-0.5 dB.
The A-scan conversion can be characterized by the transfer of the stepped pulse sequence of Fig. 12 (bottom), which should produce at the output an equally stepped signal. The linearity or an intended logarithmic or other behaviour can be checked by that curves.

Specialized methods for the probe parameters

Fig 10

Fig 11

Most important to characterize the probes are directivity patterns at the different angles and focussing parameters. (Fig. 10). With the arrangement presented in Fig. 3 (right) an example of the patterns for a 0° - L-wave probe, covering several angles of incidence has been measured. The directivity patterns in Fig. 10 have grating lobes, which are as more important as the main lobe is decreasing. Those grating lobes appear especially at the limiting edges of the usable ranges for the incidence or the skewing angles.
From the dependency of the maximum amplitude of the main lobe from the angle of incidence or the angle of skewing one may derive sensitivity correction curves in order to guarantee for each angle of incidence a constant defect detectability.
The angles choosen by the controling software have to be compared with the ones determined by the measurements at a block according to fig. 3. The curves in fig. 11 are giving an example for the resulting skewing (Ø=phi) and incidence (=alpha) angles and for the sensitivity changes at the different angles for the case of a skewing angle beam probe.

Specialized methods for the equipment parameters

There are several different possibilities to check the equipment parameters, either at the laboratory or at on site conditions. Some checks can be based on a special control software without the use of additional hardware components. By this the following parameters can be checked:

• crosstalk at the transmitter- and the receiver channels (switching some channels on or off)
• functionality of probe-, cable- and equipment channels (e.g. using internal probe echos)
Most of the software based checks are part of an integral equipment surveillance and give a good first impression on the functionality of all single channels. But they in general are not able to indicate quantitatively degraded parameters like the crosstalk behaviour or the sensitivity variations at the different channels.
The digital A-scan converter and the data aquisition system can also be checked by a suitable control software, e.g. testing the linearity of the A-scan converter by a stepwise amplification variation using a reference echo from a testblock, which can easily be compared with a stored reference curve.

Fig 12


If a more quantitative figure about the different parameters is required e.g. in order to trace aging influences it is nescessary to use a special hardware support. The Fig. 12 top shows a hardware adaptation box, which has been conceived at BAM. This box acts as a probe replacement and produces with the aid of a function generator different pulse shapes and amplitudes. It allows to measure quantitatively the different transmitter-pulses, the receiver channel transfer functions and the delay time influence. Fig. 12 (middle and bottom) are showing the transmitter pulses produced by that box and an artificial echo sequence of sinus bursts at different amplitudes with a 6 dB step in between. The amplitude accuracy of this test hardware can be better than 0.1 dB allowing its use as a reference source for a quantitative equipment evaluation.
Most of the specialized methods for the equipment parameters can be carried out as a quick surveillance, that means that they can be a build-in part of an inspection procedure. Even the hardware supported checks can be carried out very easily because all of the different checks with the parameter box are software based and do not need more than 5 or 10 minutes.
Of course most available equipment are provided with some build-in self-check procedures, e.g. carried out during the switch on procedure of an equipment.

Surveillance of the control software

Fig 13

A very important factor for the reliable inspection by ultrasonic phased array equipment is the quality assurance of the used software.
A first basic check of a control software is of course the verification of the intended skewing or inclination angles and of the intended focussing and also of the relationship between grating lobes and main lobe. This can be done either by the directivity pattern measurement device in fig. 3 or with specialized test blocks containing suitable side drilled holes. An interesting check could also be based on an auto-focussing procedure searching automatically the optimal angle and focussing at side drilled holes in order to compare it with the values calculated by the controling software. Another problem for the software check is the surveillance of the administration purposes of a control software. The control software of a phased array equipment has not only to guarantee at the different probe positions the required angles and focussing values, but also the amplification, the individual TGC curves, the sensitivity correction values among others have to be selected from look-up tables and the data of the A-scan converter have to be stored on files allowing the off line evaluation. The correctness of this administration task of the control software can only be verified with measurements at suitably tailored test blocks together with an automatic scanning. Fig. 13 gives an example of TD-scans measured at a block with side drilled holes at three different angles. The different positions of the reflector indications at the TD-Scans are allowing an immediate check of the correct operation of the software structure.
In order to detect possible influences of the A-scan conversion and data storage rate, the scanning should be executed at different speeds.
A carefully designed software quality assurance system has to guarantee that a software once checked is not changed during the whole inspection procedure.

Summary

The laboratory and on-site parameter checks of ultrasonic phased array equipment based on software and on hardware tools are demonstrating the necessity to prepare a generally accepted document for the characterization of those equipment. It is hoped that the different manufacturers of probes, equipment and the users of those systems are cooperating in future to define the basic rules for this purpose.

References

1. H. Wüstenberg, A. Erhard, G. Schenk, W. Möhrle: Construction des palpeurs-multiéléments avec ultrasons; Revue Pratique de Contrôle Qualité Industriel, No. 133 bis 24. Année, Septembre 1985, p. 43-53
2. A. Erhard, H. Wüstenberg, G. Schenk, W. Möhrle: Calculation and Construction of Phased Array-UT Probes; Proceedings 3rd German-Japanese joint Seminar on Research of Structural Strength and NDE Problems in Nuclear Engineering.
3. Non-destructive Examination III; August 29 and 30, 1985, Stuttgart/West-Germany; Editor Prof. Kussmaul, MPA-Stuttgart III.2.2