It is estimated that approximately 20,000 commercial laser sheet metal cutting systems have been installed worldwide since 1980, when sales figures became generally available. The value of these systems is probably about $7.5 billion. Over 65% of these machines are installed in Japan, with Europe accounting for 24% and the US (perhaps surprisingly) only about 9%.

　　Perhaps the ultimate application of laser cutting is via 5 axis manipulation for 3-dimensional work. Such gantry beam delivery systems did not appear on the market until the 80s and the first of these systems was produced by the Italian company Prima Industrie. In the UK, the first such system was installed at the Swindon plant of Austin Rover in 1983 and was used very successfully for the trimming of pre-production car body panels during press tool development. This application is still the major use of multiaxis laser cutting throughout the world, with all the major automotive suppliers utilising the technique.

　　In the course of preparation of this article the author had the opportunity to speak to Peter Houldcroft (now retired) and one of the questions asked was: what had given him the original idea for gas assisted laser cutting? The answer was very surprising. In 1965 Peter Houldcroft had visited BMC (British Motor Company), where he was told some preliminary cutting trials had been undertaken using a plasma torch, for the application of body panel trimming during press tool development. The problem was that the system was not accurate enough and produced burning. Peter was asked if he could think of any other suitable cutting process. On the drive back to Cambridge, the idea of combining an oxygen-jet with a focused laser beam began to form. The necessary catalyst for this idea was provided by the availability of 300W of CO laser power at SERL.

　　In the introduction to his book 'Laser Materials Processing', Professor Bill Steen presents the argument that since the invention of the laser in 1960, we have entered into a new industrial revolution, based on the use of coherent optical energy. If you are prepared to subscribe to this idea, it is difficult to think of a better example of how this industrial revolution has progressed, than that of gas assisted laser cutting.

　　Early attempts to weld metals with the slow flow lasers described earlier, found that although thin steels could be melted on the surface, the fusion zones were intermittent. These trials suggested that a laser beam with a low order transverse mode structure and a power of about 2kW was needed for practical welding of thin materials, say up to 3mm and at speeds above 0.5m/min. In principle, it would be possible to simply increase the length of a slow gas flow laser until the required power was reached. The power output is about 60 watts per metre of discharge from such a laser, and on this basis a 2kW slow flow laser would need more than 30m of discharge tube and would have to be optically folded many times. Cumulative power losses and optical distortion from so many reflections became a serious consideration, and the search was on for a more promising design for a 2kW CO laser for welding, one which produced a relatively high power output per unit length of discharge, in a low order mode.

　　Lasers in general are inefficient. In a slow gas flow CO laser, about 90% of the power from the discharge is not converted into laser power and goes into heating the gas, and as its temperature rises, the laser process becomes less efficient. The gas is cooled by conduction through the gas into the glass walls of the water-cooled discharge tube. Increasing the tube diameter in an attempt to increase the laser power doesn't work because it causes the temperature in the centre of the tube to increase. Hence the expression of power output in watts per unit length of discharge, rather than in watts per unit volume. The literature at the time suggested that the discharge could be effectively cooled by forced convection, by flowing the gas through the discharge at much higher speeds. It appeared that in this regime, power would be proportional to the mass flow of gas through discharges and would not be limited by the discharge length.

　　High gas flow could be achieved by pumping gas across a shallow channel with mirrors at each side. In principle electrodes could also be placed at the sides of the channel, but in practice it was more effective to have them at the top and bottom of the channel so that the gas flow, laser beam axis and discharge were all mutually orthogonal. Gas could be circulated using a high volume flow fan. Serious discharge problems with this approach were anticipated, however, and it would have been very difficult to produce a large volume homogeneous discharge which effectively filled the mode volume of the laser. This type of laser could well be very inefficient and could have a very asymmetrical transverse mode, making the beam less effective for welding. Given the other difficulties of producing as much as 2kW of laser power, it seemed prudent to ensure that whatever power was produced was in a low order, symmetrical mode, which could be focused to a small, high intensity spot. The cross-flow approach was, therefore, not pursued at TWI.

　　The requirements for a high quality output led the researchers at TWI to attempt to develop a fast flow CO laser with discharges running axially in cylindrical tubes i.e. a fast axial flow laser.

　　First attempts to produce a fast axial flow CO laser used the simplest possible optical arrangement with a mirror at one end and a gallium arsenide output window at the other. Two discharge tubes were arranged in line with their cathodes earthed. The high voltage anodes were insulated from earth by two lengths of glass tube through which the laser gas flowed before entering the discharge tubes. Gas was pumped through the discharge tubes by a large Rootes blower, backed by a rotary vacuum pump. The Rootes pump, similar to an automobile supercharger, was considerably more expensive and much bigger and heavier than a fan, but unlike a fan it was capable of developing the relatively high pressure differences which were necessary to force the gas through the relatively narrow fast axial flow laser tubes. The gas flow geometry was straight through, for simplicity, without gas recycling. As a result, a large cylinder of helium would be consumed in about 5 minutes, which led to the speedy development of rapid manual tuning techniques. First attempts to produce laser power were very disappointing. The problem seemed to be in the difficulty of producing a homogeneous glow discharge. Instead, it tended to cling to the sides of the tube in line axially with the electrodes. As the current was increased above a few milliamperes, the discharge became increasingly constricted into a number of unstable, long, thin and bright filaments, which moved about rapidly in the turbulent gas flow and consequently came to be called streamers. Getting rid of the streamers quickly became the researchers' one aim in life. Many arrangements of electrode, combined with different electrode materials, produced very little improvement and each time the current was increased the streamers appeared.

　　Since the discharge always clung to the wall of the discharge tube, the gas was finally given a radial component of velocity (i.e. towards the axis of the discharge tube) by inserting tubes of the same diameter, inside the larger diameter insulating tubes upstream of each discharge section. The gap between the inner tube and the discharge tube was set at a few mm to force the gas to enter the discharge with a radial component of velocity. From the first moment the laser was turned on with this new arrangement in place, the discharge behaved in a completely different manner. The plasma was seen as a soft homogeneous glow, with no sign of the streamers which formerly had always accompanied the raising of the discharge current. The laser power, which had always been negligible, quickly rose to several hundred watts, as the mirrors were rapidly turned before the helium bottle was emptied. It was found only later, that the introduction of the restriction into the gas path at the entry into the discharge tube, was causing the gas to reach supersonic velocities and it was the shock wave produced by this that was forcing the discharge to run homogeneously. In addition, the isentropic expansion also cooled the gas locally and allowed more power to be fed into the discharges.

　　A considerable amount of development using the test bed was still required, especially in the areas of the tube and electrode systems. In addition, a gas re-circulation system with the aim of providing a run of 8 to 10 hours on a single large bottle of helium was needed, if a usable laser was to be realised. The number of plasma tubes was doubled to four to reach the required 2kW power output. When optimised, each tube was 38mm diameter and 600mm long. The overall length of the laser was kept within about 4 meters by introducing a single optical fold containing two 90 degree mirrors. The maximum power was approximately 2.5kW, and 2kW could be maintained for extended periods. The first prototype 2kW laser is shown in Fig.7.