Institut de GSnie Rtomique, Ecole Polytechnique F6derale de Lausanne, CH-1015 Lausanne, Switzerland

Transcription

1 JOURNAL DE PHYSIQUE Colloque C8, supplément au n 12. Tome 48, décembre 1987 C8-335 "CAMELEONDE" CONTINUOUS WAVE AUTOMATIC ULTRASONIC MEASURING SYSTEM A. KULIK and J.E. BIDAUX Institut de GSnie Rtomique, Ecole Polytechnique F6derale de Lausanne, CH-1015 Lausanne, Switzerland Résumé - Un système automatisé de mesure ultrasonore a été construit. Il permet de mesurer l'atténuation et la vitesse d'ondes ultrasonores utilisant la méthode ondes continues. Cette méthode est bien adaptée à la mesure d'échantillons présentant une forte atténuation. La mesure peut se faire quasi-simultanément sur 5 fréquences distinctes, choisies dans la gamme MHz. Ce système a été utilisé pour des mesures de vitesse et d'atténuation dans le cobalt entre 300 K et 850 K. Il est piloté par un ordinateur de table qui contrôle les paramètres expérimetaux. Les résultats sont affichés en temps réel sur écran graphique et stockés sur disque. Abstract - A fully automatized Ultrasonic (US) measuring system was constructed. It measures US attenuation and velocity using the continuous wave method. The chosen method is specially suitable for measuring specimens which exhibit high attenuation. It allows to measure US attenuation and velocity at five distinct frequencies, chosen in the range MHz, in a quasi simultaneous manner. This system was used to measure US attenuation and velocity in cobalt samples in the temperature range K. It is driven by a desktop computer which controls all experimental parameters. Results are displayed in real time on a graphic screen and stored on a magnetic medium. INTRODUCTION Ultrasonic. (US) attenuation and velocity measurements are a convenient method for studying the phase transitions. Specially, velocity measurements for certain acoustic wave propagation directions give important informations on the elastic behaviour of the material at the transformation. From room temperature to 850 K, cobalt exhibits two important peaks of US.attenuation which render the measurement difficult. When attenuation in the sample increases, the signal to noise (S/N) ratio becomes lower. On the other hand, noise of the US receivers is directly proportional to their bandwidth. In the classical pulse echo method, US receiver bandwidth must be kept wide in order to preserve the pulse shape. That means that the noise level is a quite important limiting factor of the classical pulse echo devices. Continuous wave (CW) US method was popular in 1960's [1] and rarely used afterwards due to the lack of suitable equipment on the market. Nowadays, High Frequency Vector Network Analyzers seem to be very convenient for ultrasonic measurements. Such an equipment has been used in the present work to develop a new apparatus which is able to measure US attenuation and velocity as a function of temperature in the range K. Article published online by EDP Sciences and available at

2 JOURNAL DE PHYSIQUE FUNDAMENTALS OF CW US CW transmission spectrometers are well described in [I]. Conceptually system can be divided into three parts : - transmitter (generator) - composite resonator (i.e. transducer - sample - transducer) - receiver and measuring system As a transmitter, receiver and measuring system, a Network Analyzer can be used. CW US techniques means establishment of US standing waves (or resonances) in the resonator. Connecting both parts of the Network Analyzer to the resonator and displaying its transmission response (called usually sl2) as a function of frequency, standing wave frequencies can be measured. Supposing ideal transducers (i.e. very wide bandwidth and negligible mass), series of standing wave frequencies will be observed. Therefore we can calculate US velocity from the frequency distance between two adjacent peaks : where : v = US velocity (m/s) 1 = sample length (m) AF = frequency difference between adjacent standing wave peaks For relative measurements, the value of the center frequency of one peak (Fo) is quite satisfactory. From the -3 db width of each of the peaks (Af), we can calculate US attenuation or Internal Friction (I.F.) using : where : Af is -3 db peak width FO is peaks center frequency Therefore CW US method presents analogies with the classical resonant methad in kilocycle frequency range. In the case of real US transducers, a correction must be made for absolute velocity measurements [I]. Due to limited transducer bandwidth, standing wave frequencies can be measured only inside these bandwidths. Therefore measurements can be done around the central transducer frequency and the odd overtones. Bandwidth of the receivers in the Network Analyzers is typically 100 Hz or less, which is several orders of magnitude lower than in the case of pulsed US method. The Transmitter (generator) of the Network Analyzer usually has a maximum output voltage around lv, which is enough due to above mentioned narrow receiver bandwidth and the good S/N ratio.

3 REALIZATION OF THE MEASURING SYSTEM Mechanical The sample has a cylindrical form about 10 mm diameter and 5 mm length. Two Lithium niobate or quartz US transducers (overtone polished, Valpey Fisher Co.) are bonded on the opposite sides of the sample by means of a coupling high temperature glue (ZGM Krautkramer). Such a composed resonator is placed TEMPERATURE inside the sample holder (fig. 1) made from OFHC copper, tightened with thin gold foil and upper cover. COAXIAL LINE Electrical contact with the transducers is assured using the elasticity of the central conductors of the coaxial lines, through spheri- BNC CONNECTOR cally shaped contacts. Around the sample holder, a thermocoax heats; is wound and PtlOO temperature sensor is mounted in close proximity of the sample. During the measure- Fig 1. Cross section of the sample ments the complete system is placed holder. in a vacuum chamber (- Torr). Electrical Network Analyzer with S-parameter set (fig. 2) is connected to the sample holder and used in transmission measurement mode (SI2). Therefore -- manual measurements are done at five frequencies and all experimental parameters are saved in the instrument using the "SAVE SETUP" facility. All subsequent operations are done using "MARKER" functions build into the Network Analyzer. Using the computer and the fact that all the functions of the Network Analyzer can be computer controlled, one measuring cycle takes roughly 1 min for five frequencies. The only time consuming task is the Analyzer's sweep time which must be chosen sufficiently long in order to avoid distorsions of the resonant curve. r Network Analyzer SAMPLE HOLDER S-Parameter set Fig. 2. Block diagram of the electronic part of Ultrasonic System.

4 C8-338 JOURNAL DE PHYSIQUE The problem of impedance mismatch [I] between the Analyzer and transducers is not critical in practice. Sensitivity of the Analyzer is sufficiently high, and quite important impedance mismatch is even necessary in order to avoid electrical attenuation background in the system. A typical experimental result is shown in fig x U >. V : 3 (1: 14.6 L IZ: m m m m a 6 3 m m m c r ~ c s m ~ m m ~ r n m m v v ) m p c r n m TEtlPERRTURE CK3 TEMPERATURE CKI Fig 3. Raw results obtained in 99.98% polycrystalline cobalt on heating. US propagation direction parallel to c-hexagonal axis. TEMPERATURE CONTROL SYSTEM Anelastic measurements, specially of the phase transition can be very sensitive to the heating or cooling rate. Work in vacuum leads to long thermal time constants. Temperature controller was constructed using a commercially available equipment (fig 2). A low cost Data Acquisition Unit is used for measuring the resistance of the temperature sensor. The value is sent to the HP71B Pocket Computer through HPIL Interface and used to calculate the temperature [2]. A linear setpoint ramp can be created if needed. Using PID algorithm [3], an appropriate value of the heater power is calculated and sent back to the Data Acquisition Unit. A Laboratory made digital to analog converter Card is addressed and the output voltage is amplified using Power Supply, and finally used for heating. Dialogue with the regulator is assured using the redefined pocket computer keyboard, or from the main computer side, using the HPIL-HPIB Interface. The typical deviation between programmed and actual temperature, during lk/min ramps, does not exceed 0.05 K, and, in the steady state, 0.01 K. CONCLUSION The described automatic measuring system enables US velocity and attenuation measurements to be performed with a good accuracy in a wide temperature range. Measurements are done quasi simultaneously at five different.frequencies, and continuous wave ultrasonic method allows to study highly attenuating samples.

5 REFERENCES 1. W.P. Mason, Phys. Acoustics VIII, 96 (1971). 2. Hewlett Packard, System Software 3054A, 44453A opt Hewlett Packard, Application Note AN Using the HP3497A to Control Industrial Wastewater Treatment. ACKNOWLEDGEMENTS The authors are grateful to G. Analog converter Card. Gremaud for conception of Digital to This work Foundation. was partially supported by the. Swiss National Science