Ultrasonic welding - the next chapter

TWI Bulletin, July/August 1991

Recent commercial access to a rare earth alloy known as Terfenol has opened the door to development of a new type of transducer for ultrasonic welding equipment. Roger Wise takes up the story.

During the exhibition polystyrene testpieces were welded together using 400W of input electrical power. This is thought to be only a fraction of the maximum ultrasonic power generation capability of Terfenol transducers and machines with a multi-kilowatt power output characteristic are expected to be commercially available within the next five years.

Magnetostriction

The property of magnetostriction was first discovered in nickel which is able to produce a maximum strain of 40 parts per million (ppm).

Electromechanical devices using nickel and alloys of nickel, cobalt and iron found many uses in the 1940s and 50s, for example in ultrasonic cleaning, welding and machining. However in the 1960s transducers working under a different physical principle, the piezoelectric effect, began to emerge. The effect (which is the electrical anologue of magnetostriction) is the name given to a property of some electrically insulating materials which experience a slight change in dimension on the application of an electric field. One example of this is the ceramic lead zirconate titanate (PZT) which is capable of a maximum strain of 100ppm and is also significantly more efficient than nickel in converting electrical energy to mechanical energy.

Most currently available ultrasonic welding machines use piezoelectric transducers as their source of ultrasonic energy, but the maximum power output from these machines is limited to approximately 3kW because of limitations in the properties of the ceramic materials.

Terfenol

Although this material has been studied for over 15 years it has only recently been seriously considered for commercial devices and applications. ^[2] Johnson Matthey-Rare Earth Products is currently producing the material and potential industrial applications include active noise and vibration control, sonar, sonochemistry, robotics and ultrasonic processing. ^[3]

Ultrasonic transducer

To obtain maximum strain from a rod of Terfenol it is necessary to clamp it mechanically to a prescribed pre-stress. ^[4] This can be achieved by tightening clamping bolts to the correct strain which can be measured on strain gauges attached to the Terfenol core. Since a magnetic field applied in either direction along the axis of the rod produces an increase in its length, it is necessary to apply a DC field while superimposing a high frequency AC field to produce ultrasonic vibration at the frequency of the AC excitation. If no DC field were applied, the rod would oscillate at twice the frequency of the AC signal because an expansion would occur during both the positive and negative parts of the electrical cycle ( Fig.3). The DC excitation, which could be supplied electrically or with permanent magnets, has the effect of offsetting the equilibrium position of the AC displacement. This means that over one half of the AC cycle, the magnetic fields combine to produce a large expansion of the rod, while over the other half of the cycle the AC and DC fields oppose one another to produce a net contraction of the rod from the equilibrium position.

Welding equipment

The prototype welding equipment commissioned for Interplas '90 is illustrated in Fig.4, and includes the press for the welding horn, the transducer and the power supply.

A typical magnetostrictive ultrasonic welding machine may have three sections down which the ultrasonic energy is transmitted ( Fig.5). First, the transducer converts the electrical energy delivered by the power supply from a magnetic field into mechanical vibrations. In this case the transducer has a Terfenol rod clamped to a prescribed mechanical pre-stress and under the influence of a DC magnetic field.

Secondly, the transducer is bolted to a booster which is a rod of titanium or aluminium alloy of a certain length which may have a variable cross section as a means to amplify the amplitude of ultrasonic vibration.

Finally, the welding horn is bolted to the booster and provides the medium for introducing the ultrasonic energy to the workpiece. Welding horns may have variable cross section to alter the amplitude of vibration of the ultrasound and are usually shaped at one end to direct the ultrasonic energy to the area where the weld takes place.

In production welding, the transducer and booster usually remain unchanged while the welding horn is redesigned to suit each different welding application.

Each of the three sections of the ultrasonic system usually corresponds to one half wavelength of the resonant ultrasound and this length can be calculated for rods of uniform cross section with knowledge of the speed of sound in the material involved. Figure 5 shows the node and antinode distribution for the system.

The weld pressure is usually applied via a clamp attached to the node on the booster component since this is a position of minimal ultrasonic vibration.

The working frequency of the majority of ultrasonic welding machines is 20kHz, which is the lowest round figure above the audible limit for most human ears.

Welds in the polystyrene testpieces using the Terfenol machine were made at a working frequency of approximately 19.5kHz although the resonant frequency of the system varied over the weld cycle. This variation was an expected effect of a variable load, i.e. as the weld was made under pressure the tip of the welding horn experienced variable resistance to movement during the weld cycle.

A potential problem with magnetostrictive transducers is the generation of eddy currents in the magnetic core material. Nickel and other ferromagnetic core materials are usually laminated with electrically insulating layers so that electric currents cannot circulate and cause harmful heating. For this reason a composite Terfenol core was manufactured from six rods with hexagonal cross section and each rod was electrically insulated. This core configuration was designed as a compromise between a truly laminated structure and the range of material geometries which could be manufactured at that time.

A gap was left in the centre of this core to allow compressed air to be forced through the core to remove any heat generated by eddy currents or hysteresis losses in the material. The design of the AC and DC coils was formulated with the aid of a finite element computer model which was also able to predict the distribution of eddy currents within the core. ^[5]

During welding trials it was found that the temperature of the core material increased from 40 to 60°C over the three seconds it took to weld the polystyrene testpieces, even with compressed air cooling and the provision of fans on the outer casing of the transducer. This heating effect could be attributed mainly to eddy current generation in the Terfenol cores.

The next stage of the development will incorporate a redesign of the Terfenol cores to include laminates (which are now available) in an attempt to reduce eddy current generation further. A target of 3kW output power has been set for the next phase of work which brings the Terfenol transducer into direct competition with its piezoelectric counterparts.

Member companies interested in this development are invited to contact Roger Wise at Abington.

References