With the war over and the Sonic Works at West Drayton disbanded, George was no longer compelled to apply his Sonic principles to engines of destruction. He had for many years shown an interest in motor cars and other vehicles for road and rail transportation. The first practical indications of this interest appeared in his patent for a monorail system in 1910, followed by others relating to components of internal combustion engines and transmission elements. Among these there was a paraffin vaporizer patented in 1911 which was tested in a car fuelled by crude paraffin oil on a journey from London to Brighton and back - at a cost of one shilling for the fuel for the round trip! During the war years George had already made an in-depth mathematical analysis of infinitely variable transmissions and had arrived at a concept for a mechanical Torque Converter. He claimed that such a device could be universally applied in industry, motor vehicles, railway locomotives, ships, military tanks and agricultural tractors. The exigencies of the war had interfered with further development, but now there was an opportunity to return to the theme.

Of the various possibilities which came to mind George concluded that the most fruitful and marketable line of approach to start with should be improvement in the efficiency and ease of operation of the transmission train of motor cars. He was concerned at the wasteful use of fuel inherent in existing engines and their transmission systems, the inconvenience of manual gear changing, and clutch operation, with consequent shocks to the engines and transmissions. Also there was the high price of motor cars, which only a minority of people could afford to buy. His idea was to produce a low cost one hundred guinea "peoples' car'' which would travel 100 miles on one gallon of petrol at the most commonly used road speeds of 30 to 40 miles per hour. George arrived at this figure after conducting a comprehensive survey of average car road speeds and designed his car to benefit the most people, rather than a car of higher speed which would only benefit a minority. He considered that this performance and low cost could be achieved by using a cheap 500 cc single cylinder two stroke air cooled engine together with his Torque Converter transmission which would eliminate the conventional gear box and clutch. Experience in this field could then be applied to the transmission of much higher powers in heavy vehicles such as railway locomotives.

George admitted that the mechanics of his automatic variable transmission, or Torque Converter were not easy to explain because the theory of conventional mechanisms did not apply to this invention. It will be recalled from previous chapters that George used a hydrosonic system to operate the aircraft machine guns and the injector system for Diesel engines. His Torque Converter was a sonomechanical application of his theory on the transmission of power by vibrations where the impulses are transmitted through solids instead of liquids. Power is transmitted from the engine to the output shaft through a system of oscillating levers and inertias¹ arranged in such a way as to split the alternating motion derived from a primary crank rotating with uniform angular speed and torque into two components. The first component oscillates a mass or any form of inertia. The second component oscillates a set of mechanical valves. The valves are arranged in pairs out of phase at 180°, so that positive and negative impulses are rectified into a unidirectional torque on the secondary. The angular speed and torque on the secondary varies automatically within wide limits according to the resistance to be overcome and the revolutions of the engine.

The Torque Converter was typical of the way in which George Constantinesco worked, the first model emerged from a purely mathematical analysis of the problem to be solved, followed by a mathematical solution translated into a working drawing. The model made from the working drawing performed as predicted without "trial and error'' or modification.

A great deal of interest in the invention was aroused in the popular and technical press, but many of the statements made were inaccurate or misleading. In an attempt to clarify the situation, the Editor of the Automobile Engineer invited George to discuss the whole concept of variable transmissions for motorcars in a series of five articles in that publication from November 1923 to July 1924. These articles, together with the correspondence which arose, were reprinted at George's request, in a volume entitled "Variable Transmissions for Automobiles'' [9], and issued in quantity from his offices at 7 Grosvenor Gardens in London. This was an indication of George's confidence in the accuracy of his analysis of the problem and his solution provided by his invention, in spite of the criticism and sometimes heated arguments generated in the correspondence. Further criticism and disbelief was silenced when George published a letter in the July 1924 issue of the Automobile Engineer offering a prize of £100 ''to the first individual who will prove that my mathematics are wrong or the interpretation of my formulae are not in strict accordance with logic".

To illustrate the basic principle of the Constantinesco Torque Converter, consider figs 16a and 16b. The impulses are produced by a crank connected to a point distant from the apex of a pendulum, or lever with a weight on the end. This apex is connected by a short link to a fixed point; the apex of the pendulum is connected with links to unidirectional ''mechanical valves" on the secondary shaft, which operate like ratchets, but much more smoothly. When the primary crank rotates slowly, figure 16a, the pendulum swings to and fro about the apex, or fulcrum of the lever, as in a clock and no energy is imparted to the secondary shaft. This corresponds to the ''neutral" position in a conventional gear box with the prime mover ticking over.

When the revolutions of the prime mover are increased considerably, the frequency of the crank oscillations increases and thus tries to increase the frequency of the pendulum oscillations. At this point a new set of circumstances arises. Due to its inertia the pendulum weight tends to remain stationary fig. 16b. Under these conditions and when the load on the secondary is moderate, the fulcrum of the pendulum, which was at the apex, has been transferred to the position of the pendulum weight. The result is that the apex of the pendulum oscillates instead, to the maximum to and fro motion, permitted by the design. This causes the links to oscillate the valves, which in turn rotate the output shaft to the maximum angular speed permitted by the design. Under these conditions the system is operating in ''top gear" with a 1 to 1 ratio. (The remarkable similarity to the operation of the hydrosonic system for the aircraft firing gear is apparent. As soon as the aircraft engine reached a predetermined number of revolutions, the inertia of the liquid column diverted the high frequency pulses down the pipes connected to the trigger motors.)

At intermediate angular speeds of the input crank the effective fulcrum will take up intermediate positions on the pendulum rod. Consequently there will be more or less swing (or amplitude) of the pendulum weight and more or less travel (or amplitude) of the valve links according to the speed of the input crank and the torque on the secondary shaft, fig. 16c.

In other words, the operation of the mechanism is dependent on the frequency and amplitude of the input impulses to the mechanical valves which automatically increase or decrease according to the load to be overcome on the secondary shaft. The mechanical valves rectify the alternating nature of the impulses into unidirectional impulses which act on the output shaft in a cumulative way. One valve rotates the output shaft on the forward stroke and the other valve rotates the shaft in the same direction on the reverse stroke.

The first model ever made to illustrate the principle of the Torque Converter is shown in fig.17. It will be observed that the mechanism is upside down in relation to the previous description with the ''weight" of the ''pendulum" at the top instead of at the bottom. The construction of this model follows the arrangement shown in fig. 18, where the primary shaft is connected by a rod to the centre of a floating lever. This introduces more elasticity in the system, but in all other respects the model demonstrates the basic principle of the Torque Converter as described.

Other models to demonstrate the principle of the Torque Converter were made in Meccano by schoolboys and other enthusiasts during the 1920s, an example of which is shown in figs 19 and 20 applied to a model car chassis. This example follows the same arrangement as in fig. 18, but ratchets had to be used as it was not possible to construct mechanical valves with Meccano parts. This mechanism occupied the space of the former conventional gearbox in the standard Meccano model of a motor car chassis. The Torque Converter mechanism was driven by a Meccano electric motor connected to the input eccentric by a chain.

George first successfully tested a Torque Converter in a car in May l 923, using an experimental model that had been built only for bench tests. He obtained an old Sheffield Simplex chassis and replaced the big 45 hp engine with a 10 hp "light car'' Singer engine, and built a platform on the chassis. This car was driven around the outskirts of London loaded with 10 people, including the inventor at the wheel, fig.  21, and later towed a lorry up a steep hill. As a further demonstration of the capabilities of the car, six inch wooden blocks were placed in front of the wheels. When the accelerator was depressed the car climbed over the blocks smoothly and without hesitation, to the astonishment of the bystanders. This test was merely to demonstrate the possibilities of the Torque Converter, because even a 10 hp engine was much larger than necessary for the production car envisaged.

The next development was to apply the idea of a low cost 500 cc single cylinder two stroke aircooled engine coupled to a Torque Converter installed in a light car chassis. This was exhibited in the Palace of Engineering during the British Empire Exhibition at Wembley in 1924, fig. 22.

The chassis performed as predicted under test but the disadvantages of the aircooled engine and balancing problems associated with the one cylinder were recognized and subsequently overcome in a completely new design concept for a production car. This comprised an integral balanced power unit of about 500 cc capacity (bore 67 mm, stroke 70 mm) with the Torque Converter mounted between two water cooled cylinders and an improved carburettor, also one of George's patents. The RAC rating was 5.58 hp and the tax was £6. A cross section of the power unit through the converter mechanism is shown in fig 23, where the arrangement of oscillating inertias, links, and valves can be seen. The diagram in fig. 24 shows the position of the inertias in the extreme working condition equivalent to fig. 16b.

The general arrangement of a chassis fitted with this power unit is shown in the drawing in fig. 25. A chassis of this design, together with two prototype cars were exhibited at the Paris Motor Show in 1926, fig. 26.

A demonstration model of the power unit and a similar chassis is housed in the Science Museum in London. A two seater model of the car, with the inventor at the wheel, is illustrated in fig. 27. The performance of the car was exactly as George had predicted, 100 miles per gallon of petrol at 38 miles per hour. Only the originally estimated sale price of 100 guineas was found to be too optimistic for a car fitted with the improved two cylinder power unit and had to be revised to £215, and £315 for the saloon version.

In addition to the economic virtues of the car, the other outstanding feature was the ease of control, which George often demonstrated in a convincing manner. Here was a car which even a child could drive after a little practice at steering, as proved by his small son Ian, who was taught to drive and demonstrate it at the tender age of eight. Then there was M. Antoine Bourdelle, the famous sculptor, who drove the car around the streets of Paris after only a few minutes tuition, although he had never handled a car before. Another of George's demonstrations was to have somebody leading the car with a thin string attached to the throttle lever on the carburettor. In fig. 28 his wife Sandra is performing the demonstration and George is following on behind.

In order to appreciate how easy it was to drive this car a description of the method of operation is appropriate. With the car at rest and the handbrake on, the engine would be started and allowed to tick over. To move off, the handbrake would be released in the usual way and the right foot would depress the accelerator pedal. At about 1200 rpm. the car would start to move away smoothly and progressively gather speed, but at the same time the engine revolutions would gradually decrease until maximum speed was attained in the equivalent of "top gear''. Speed control thereafter would be entirely through the accelerator pedal. On reaching a hill it would be noticed that the car would tend to slow down, but the engine revolutions would not decrease as in the case of a conventional gearbox. By depressing the accelerator pedal fully the car would regain speed and continue to climb the hill. On a very steep slope, say 1 in 3, the car would slow down but continue to climb, without loss of engine revolutions. In other words the Torque Converter would automatically select the correct "gear ratio'' under all road conditions encountered. Starting on a steep hill was also very easy, because when the brake was released the car was automatically spragged by the mechanism and could not run backwards. No matter what the conditions were, the engine would never stall due to overload. For example, it was possible to place the front of the car against a wall and fully depress the accelerator, but the engine could continue running at maximum revolutions and exert maximum torque on the driving wheel.

In order to slow down and stop the car, the accelerator pedal would be released and the left foot would depress the brake pedal. The car would soon stop if required, as it had brakes on all four wheels, and then the handbrake would be applied. The only other control in the car was a lever, which actuated a reverse gear in the back axle.

Another novel feature was that the back axle had no differential. It was unnecessary because there was only one driving wheel and the other was free. Since the propeller shaft rotated anticlockwise looking from the rear, the torque on the propeller shaft gave a considerable increase of loading on the driven wheel. Furthermore, as the propeller shaft torque was about five times that commonly encountered in the orthodox chassis, the single fixed wheel had a road adhesion comparable with that provided by a conventional differential axle driving two wheels. A further important feature was that as the direction of rotation of the propeller shaft was the same for both forward and reverse, the road adhesion was the same for both directions.

In addition to motorcars, the railway locomotive offered fertile ground for the application of the Torque Converter transmission. The steam locomotive was still the best practical solution available for long distance heavy haulage, but was admittedly uneconomical. The internal combustion engine could provide greater economy but it had not been possible to employ it at reasonable cost. Although electrification seemed to provide the ultimate long term solution, the conversion from steam to electricity would entail enormous capital outlay and it was thought that many years would pass before the full benefits could be made available. Even then, it would only be applicable to a small percentage of the world's railways.

The use of George Constantinesco's Torque Converter transmission in a locomotive appeared to offer an immediate and economical solution to the problem because of its ability to cope with heavy starting and acceleration torques without overloading or stalling the engine, while the wide and automatically variable gear ratios would enable light weight high speed internal combustion engines to be used. Although George's main effort in the early 1920s was directed to development of the Torque Converter car the prospects for a Torque Converter locomotive appeared to be equally compelling. Consequently, he fitted a locomotive chassis with a six cylinder 250 hp petrol engine and Torque Converter and demonstrated it on his stand at the Wembley Exhibition in 1924, fig. 29. The chassis of this locomotive was a former Great Western Railway ''Armstrong Goods'' 0-6-0 No. 395, which George converted to a 2-4-0, using the leading jack shaft for transmission of power to the other four coupled wheels from the Torque Converter.

The first experimental trials of this locomotive hauling a load of goods wagons took place on the Southern Railway in its Longhedge Yard, Battersea, on 30th June 1925, coinciding with the celebrations of the Centenary of Railways. At the same time, it was inspected by members of the International Railway Congress. This considerable publicity and expense did not result in this Torque Converter system being adopted on British railways, but it was adopted by the Romanian State Railways for railcars on their branch lines a few years later.

Apart from cars and locomotives, the possibility of the universal application of the Constantinesco Torque Converter in all cases where automatic adjustment of speed and load would be useful was given due weight in George's advertising material and demonstration models. Some of the more important applications considered included ship propulsion, auxiliary machinery on board ships, winches, cranes, haulage gear, machine tools and heavy duty starters for powerful engines.

Fig. 30 Bench model of torque Converter application fo ship propulsion

Bench models of some of these applications were demonstrated on George's stands at Wembley in 1924, Paris in 1926 and at a special conference on George Constantinesco's work at the French Society of Civil Engineers in Paris on 16th December 1926. Fig. 30 shows a model of a marine application to replace clutches, reduction gears and reverse gears. In this model a very high speed petrol engine is driving a slow running propeller through the gearless transmission. The only control is the lever at the top of the converter, which enables forward, neutral and reverse to be obtained without changing the speed of the engine. Fig. 31 shows an example of the use of a constant speed cheap A.C. electric motor to drive machines of any kind, with the same single lever on the converter to obtain the desired variation of speed and load control.

At the Paris Motor Show in 1926, George had another convincing demonstration of the capabilities of the Torque Converter in the shape of a starter for heavy engines operated by a small electric motor. The output of the starter was connected to a lever with a tractor seat on the end. Members of the public, the heavier the better, were invited to sit on the seat and to their astonishment were lifted effortlessly aloft by the small electric motor, without gears!

The Constantinesco Torque Converter aroused intense interest in the popular and technical press in many parts of the world. The stand at the 1924 Wembley Exhibition alone generated over 300 enquiries from firms and individuals and over 200 articles in magazines and newspapers. Frequent requests came in to 7 Grosvenor Gardens from individuals and firms wanting to be appointed as Agents for the hundred guinea car from such diverse locations as North and South America, Europe, India and Australia, as well as from the British Isles. Unfortunately all these enquiries were premature in that it had not been possible to develop the car to the production stage due to lack of resources. Thus, the enquirers had to be informed that the car was still in the experimental stage, but they were placed on a priority list for eventual delivery on a first come, first served, basis.

As in the past, with wave transmission and the aircraft firing gear, George had the greatest difficulty in obtaining adequate financial backing for the development of the Torque Converter, and the reaction of the motor manufacturers varied from lukewarm interest to active opposition. A case in point was the refusal of the Wembley Exhibition authorities to allow Constantinesco to exhibit the chassis among the cars, due to objections from The Society of Motor Manufacturers and Traders on the grounds that he was not a member of the Society and manufacturing the chassis in quantity. George sued the authorities and lost the case with costs, but was eventually allowed to exhibit in the Palace of Engineering. In the absence of support from the motor industry and following the demise of the Romanian based Company, Industria Sonica, a series of small British Syndicates and Companies were formed to finance research and development of the Torque Converter, the costs of the exhibits at Wembley and Paris and ancillaries such as carburettors, speed indicators and liquid level indicators.

One of the more long standing of these Companies, formed in 1922, was Constantinesco Torque Converters Ltd, with offices in 40 Grosvenor Gardens and works in 130 Wilton Road, London. This Company, with a share capital of £75,000 acquired world rights in the invention. Under the agreement George Constantinesco was appointed the Consulting Engineer at a fixed salary, out of which he was to bear the cost of obtaining British Patents, continue research on behalf of the Company, provide drawings and designs, act in a consultative capacity and supervise development and experimental work. By today's standards and costs for research and development, the funds available were modest and fell far short of requirements. Another Company, Engine Power Ltd was formed to boost finances by disposing of a large stock of 250 hp Ricardo engines stored in the Slough premises, but sales were disappointing.

A breakthrough appeared to be in sight when in February 1925 the General Motors Corporation of Detroit took an option on a licence to manufacture the Torque Converter for use in cars under an agreement with Constantinesco Torque Converters Ltd and George Constantinesco. About $100,000 were advanced against future royalties to cover further research and development costs to meet specifications and $3,000 per month were paid in consideration of the option. The option was to be exercised within three years upon payment of $3 million, or $4 million within four years and a royalty of $2 was to be paid for each converter sold in the U.S.A.

Encouraged by this arrangement and in anticipation of early marketing of the Converter George committed all available funds to research and development, not only for the car but also for new applications. The maintenance of his life style and generosity to family and friends was commensurate with the expected rewards for a hard working and successful businessman, but in the event the potential returns never materialized and the period was marked by a series of domestic and financial difficulties which undermined his health. His first marriage had broken down during this difficult period and ended in divorce. At the same time he was being hounded by the inland Revenue for tax assessed on royalties received during the war before his inventions were taken over by the Government, including the Government award for the synchronizing gear, which he had expected to be free of tax. To add to his frustrations it appeared that General Motors were reluctant to exercise their option within the specified dates but willing to service the fees. This did not suit George, because it meant in effect that the use of his Converter in cars was blocked. He was impatient to see the invention used and personal financial rewards were a secondary consideration, or so it seemed, as the agreement with General Motors lapsed and he tried to make alternative arrangements for the manufacture and marketing of the car.

Finally, this approach failed and towards the end of the decade George again found himself to be in a precarious financial position and without funds to proceed further with the motorcar project.

In the meantime George had met with Eva Litton who was to become his second wife and constant companion for the rest of his life. Eva, with two sons, Richard and Michael by a previous marriage, was of independent means through a legacy from her father, a Lancashire textile mill owner. As well as being highly intellectual and a proficient pianist like George, she had inherited her family's business sense and thrift and soon advised George how to extricate himself from his difficulties in London and Weybridge. He wound up his affairs in London, including the sale of his patent rights in the torque converter car, sold his expensive house 'Carmen Sylva' in Weybridge and with his young son Ian joined forces with Eva at Oxen House, on the shores of Lake Coniston in the English Lake District.

The Romanians had shown interest in the Torque Converter for use in railcars and here surely there was new hope for the further development of the invention. There was a need for inexpensive and cheap to run railcars on branch lines and the use of relatively small internal combustion engines coupled to the Constantinesco Torque Converter appeared to meet requirements. The development work would need George's presence in Romania for some time, so he soon converted the outbuildings at Oxen House into offices and a laboratory and completed plans for the locomotive Torque Converter. Eva accompanied George on his trips to Romania on the development of the railway project.

Testing, development and manufacture was carried out at the former Malaxa Ironworks of Bucharest during the 1930s. Initially, successful tests were carried out with a 10 hp engine and converter mounted in a 10 tonne railcar. These results were very encouraging considering that similar cars in other countries were using 100 to 140 hp engines. The next stage was the establishment of production lines and special machine tools for the building of 30 tonne passenger railcars with 60 seats. These were propelled by two engines of 20 hp each mounted under the chassis. These railcars ran without trouble at about 40 mph, a performance which appeared to be quite suitable for conditions in Romania at the time. Looking to the future more power and faster speeds would be required on main lines as well as branch lines to meet the needs of rapidly growing industry and infrastructure in all countries. Bearing this in mind, George made a survey of requirements and solutions proposed in several other countries, including Germany, Britain and the United States.

In spite of the rapid progress in electrification and the development of Diesel electric power units, George maintained his confidence in his Torque Converter system as a simple and economic power unit but conceded that further development work was needed to cope with higher powers and speeds. The problem was not so much with design parameters as with lack of suitable materials and methods of manufacture. This work was in progress at the Malaxa Works in the testing laboratory together with the production work, when a change in railway policy led to withdrawal of funds. Thus George had to abandon the project in Romania and return to England, where lack of funds and advent of the second world war prevented any further development of an invention meant for peacetime conditions.

George Constantinesco was not the first inventor to have failed to bring an outstanding invention to commercial fruition. The Torque Converter was quite unique and introduced a new concept in mechanics, but perhaps it was too far ahead of its time. The inertia and scepticism of manufacturers committed to conventional transmissions was understandable, particularly during periods of depression between the wars, while economy in fuel consumption was not regarded as of consequence as fuels were plentiful and cheap. Looking to the future, there is an ever increasing need for economy in the consumption of fossil fuels, and for moderate speeds, safety and driver comfort on congested roads. This inevitable situation is resulting in research workers, inventors and manufacturers taking renewed interest in variable gearless transmissions coupled to high efficiency engines for motor cars. Better materials are now available, as well as improved design and manufacturing facilities with computer assistance.

1

 The inertias are weights used by the Torque Converter and its action depends on their resistance to a change in movement - their inertia. A pendulum bob acts as an inertia for a clock.