Piezoelectric techniques for the initiation of TIG arcs

TWI Bulletin, February 1982

by G B Melton, BSc and J A Street

Mr Melton is a Research Engineer and Mr Street is a Senior Research Engineer in the Control Engineering Support Group.

Use of simple piezoelectric spark generators for TIG arc initiation offers a safe and relative interference free alternative to conventional HF systems. This article describes arc initiation tests carried out using a TIG torch modified by the addition of a piezoelectric device.

The feasibility of using a high voltage generated by a piezoelectric device for tungsten inert gas (TIG) arc initiation has recently been investigated. Piezoelectric spark generators are commonly used to ignite combustible gases in domestic appliances, for example gas ovens, fires and cigarette lighters.

These devices offer advantages over conventional high frequency (HF) spark systems in that they produce considerably less electrical interference, and are far less complex. Also their size is such that they can be incorporated into the handle of a conventional TIG torch.

Arc initiation

A TIG welding arc may be initiated by touching the electrode down on to the workpiece and drawing an arc, or by breaking down the electrode to workpiece arc gap by applying a high voltage.

Touch starting is unsatisfactory as contamination of both electrode and workpiece may occur. High voltage breakdown is usually achieved by applying an HF spark. This form of spark injection, shown in Fig.1, is not only inefficient but also creates considerable electrical interference, which can be particularly harmful to electronic control circuitry in its vicinity. High voltage DC arc initiation has been proposed^[1] but such a system requires careful design to ensure a relatively safe output and even then is really suited only to mechanised rather than manual welding. A piezoelectric device can generate voltages of sufficient magnitude to break down typical electrode to workpiece arc gaps. The output is inherently safer than high voltage DC, and electrical interference occurs only at the instant of arc breakdown (compared with an HF system that causes virtually constant interference for the whole time that it is switched on).

Piezoelectric initiation device

A piezoelectric material is one in which a voltage appears across opposite faces as a result of dimensional changes between them caused by the application of a mechanical force. The ceramic materials used are brittle and need to be supported in a suitable housing. In a typical arrangement the ceramic cylinders are squeezed within their housing by a lever and pivot system, as shown schematically in Fig.2a. Pressure on the lever causes the cylinders to be compressed and hence to generate a voltage. The complete piezoelectric device used in these tests measures only 60 x 20 x 10mm and thus can be incorporated into the handle of a TIG torch as illustrated in Fig.2b.

Compressing the piezoelectric ceramic causes a charge to be produced, directly proportional to the applied force. The resultant voltage appearing across the ceramic depends upon internal (and external) leakages. Thus the peak output voltage increases with the rate at which the force is applied, up to a maximum value given by the properties of the ceramic and the load imposed upon it. A typical output voltage trace is shown in Fig.3a for a condition that does not result in external breakdown.

The rate at which the voltage increases depends on the squeeze rate, and both the peak voltage achieved and the voltage decay are caused by loading by the measuring system (a high voltage probe and oscilloscope). The true, no load, voltage has been calculated to be 15.5 ± 0.5kV, which is well in excess of that required to break down typical electrode to workpiece arc gaps in argon. (Allowing for the statistical scatter, a maximum of about 7kV is required for a 2mm gap.) A typical breakdown in argon, which initially required a voltage of around 5.5kV, is shown in Fig.3b. As the ceramic is still being stressed, further breakdowns occur for the next 8msec. This multiple breakdown in the partly ionised gap should prove beneficial for arc initiation. If the first breakdown does not result in an arc, then there is another opportunity. The duration of current flow from each spark can be calculated to be about 5µsec, and this has been confirmed by experiment.

Arc initiation tests

The modified air cooled TIG torch shown in Fig.2b was used for arc initiation tests. One high voltage connection was made to the tungsten electrode, the other to the workpiece. To generate high voltages, leakages must be kept to a minimum. Thus for reliable spark breakdown the workpiece connection was taken to a point on the workpiece as close to the electrode as reasonably possible - as for example via a metal addition to the ceramic, similar to that used for spot TIG welding. Even then, leakages were found to be a severe problem. The extra capacitance of the welding leads alone reduced the output by up to 60% and the welding set provided an extra burden. This problem was overcome by effectively isolating the piezoelectric device from the welding power circuit until the breakdown voltage required had been generated.

A single phase, transducer controlled welding set was used for the tests. This particular welding set was chosen as it was known to have an initially rapid rate of current rise.^[2] In the 5µsec that the piezoelectric spark ionises the arc gap, the welding set is able to supply 8A, as shown in Fig.4. This should be sufficient to maintain an arc. For welding sets with a slow rate of rise of current (the main current build-up from most commercially available TIG sets is of the order of 10⁴A/sec) this initial current burst may be produced by the controlled discharge of a suitable capacitor across the output terminals of the welding set.

Tungsten 2% thoriated electrodes of 1.6mm diameter axially ground to a 60° included angle were used in a shielding gas of argon. A stainless steel workpiece was chosen and each arc initiation test was made on a fresh part of the plate. These test conditions are summarised in Table 1.

Table 1 Test conditions

Results

The results of the arc initiation tests are summarised in Table 2. Each attempt was made on a fresh part of the plate and the electrode was allowed to cool between successful initiations for a minimum of 20sec. A freshly ground electrode (from the same batch) was used for each gap setting.

Table 2 Test results

*A failure was recorded when a single squeeze did not result in an arc start.

A failure was recorded if a single squeeze did not result in an arc (one squeeze may produce several sparks following the first breakdown). Failure could be a result of a spark not developing into an arc, or to no spark breakdown at all.

The limited tests showed, perhaps surprisingly, that arc initiation was most successful at small and large gaps, but less so in the mid-range. With a gap of 0.5mm spark breakdown was always achieved but an arc did not always develop. At small gaps the corresponding breakdown voltage is low and the discharge energy (which is given by ½CV²) is small. Perhaps this weak spark is unable to maintain ionisation sufficiently long for the welding set to be able to deliver current into the discharge. With large gaps of 2.5mm spark breakdown was not as efficient, but each spark resulted in an arc. The breakdown voltage and, hence, the energy in each spark are considerably greater and ionisation is maintained long enough for an arc to develop. Failure at gaps between these limits is probably caused by a combination of the two effects.

On average the success rate was 80% for a single squeeze. In theory, two squeezes should result in a 96% success and three squeezes should virtually guarantee arc initiation.

Discussion

A piezoelectric device, even manually operated like a gas igniter, is clearly capable of successful arc initiation. Voltages are produced which are high enough to break down typical electrode to workpiece arc gaps in common shielding gases. The discharge energy, although low enough to minimise any shock hazard, maintains ionisation for a sufficient time for the welding set to be capable of maintaining an arc, provided that the set is backed up with a local damped capacitance discharge.

The small size of this device allows it to be incorporated into the torch handle. Thus the welding leads do not radiate electrical interference as with an HF system. The levels of energy involved with piezoelectric initiation are an order of magnitude less than conventional HF and even then interference is produced only by the disruption caused by the initial spark. The welding leads themselves help to attenuate the high voltage impulse and hence prevent damage to the welding set, and any remaining voltage is blocked by conventional HF suppressors.

Electrical leakages are a problem but careful design can overcome this. Initially isolating the device from the power welding circuit and keeping the length of high voltage connections as short as possible reduces leakages to a minimum. The application of a moisture repellent to the device and its leads helps to improve reliability on humid days.

Arc initiation efficiency is high compared with HF starting, as with HF up to one thousand sparks occur for each second that the unit is switched on, and several seconds may elapse before the arc strikes.

Long term reliability has yet to be assessed, although there should be no problem as similar devices are used in domestic gas appliances for many thousands of operations.

A TIG spot welding gun is envisaged as the most obvious application, as this facilitates the workpiece connection. This could be designed with a removable piezoelectric cartridge, stressed by a trigger arrangement. The piezoelectric cartridge would be sufficiently cheap to be replaced if necessary at regular intervals.

An impact device in which the piezoelectric ceramic is rapidly stressed is now under evaluation. This may allow the high voltage connection, particularly to the workpiece, to be extended and allow greater flexibility and more diverse applications of this technique.

Acknowledgements

The authors thank Mr J C Needham for his advice and Miss C T Topping for her help with the experimental work.

References

1. Brown M J: 'Initiation of a gas tungsten arc by high voltage DC'. Welding Institute Report 31/1976/P.
   
   
2. Melton G B: 'Initial current characteristics of welding sets for gas tungsten arcs'. Welding Institute Report 140/1981.