Veteran to Classic - Roesch and the 105

The Brilliance of Georges Roesch

When I asked to see the engine of one of the elegant Talbot 105s pictured above, its owner said to me, “Which side? One is tame, the other very tame.” The mechanical minimalism of the six cylinder 3-litre engine was quite a revelation; on one side a flat wall of steel, and a row of six spark plugs, on the other a parsimonious-looking single carburettor, and an equally frugal exhaust. And yet for all that it had a graceful, understated elegance that was the hallmark of the whole machine.

Its designer, Georges Roesch was not a boastful, or articulate, man. Born and brought up in Geneva he was hardly at home with his own language, preferring like his father, to let his work be his self-expression; the situation was only magnified when he came to England to work for Daimler in 1914, after training under Barbaroux of Delaunay-Belleville, and also under Louis Renault, in Paris. His cars could only take after their creator in this respect, and their modest appearance when stationary belied a smoothness, fluidity, and mechanical perfection when on the move.

The Talbot 105 stands at the summit of a line of development masterminded by Roesch during his years working for Clement Talbot Ltd, and taken as a whole the series of cars he designed there represent a quest for his own automotive ideal.

It must have been a double shock for the young Roesch launching himself into a foreign country, the language of which he hardly knew, when shortly after he arrived in 1914 war broke out with Germany. Not only were his hopes of a swift rise through the ranks of a thriving industry temporarily dashed as manufacturers turned their attention to the production of armaments, but he was followed by an ominous spy novel type character for a whole year, before the authorities finally gave this recent immigrant with a German surname and accent the benefit of the doubt.

It was an advert in a 1916 newspaper, by a forward looking Clement Talbot Ltd who were looking for a designer of a new car in anticipation of the end of the war, that rekindled Roesch’s hopes of realising his ambitions. He applied for, and was given the job, and set to work on the design of an impressive 1750cc touring car. At the same time the Darracq syndicate bought the Clement Talbot Company, and later the Sunbeam Co Ltd, and became known as the Sunbeam, Talbot and Darracq combine. The car assigned for Talbot to produce was a small two-seater called the 8/18, designed in France. It was not a success, and Roesch was given the task of designing a four-seater to replace it. The 10/23 was a car which momentarily restored to the Talbot factory some of the prosperity that had ended it with the 8/18.

This was not to last, however, as Roesch was called over to Paris to work with Darracq and the London factory lapsed into a state of inefficient production, building a machine that became increasingly over-priced.

Roesch was called back to Talbot in 1925, and in many ways the dilapidated state that he found the factory in was the opportunity that he had been waiting for. The bank would only guarantee credit for a further 12 months, and so a new model had to be designed for the 1926 Motor Show, and it had to be an instant success or the firm would collapse.

Moreover, it needed to be perfect straight from the drawing board; there would be no time for experiment or testing. The task clearly called for a designer of exceptional ability, with a painstaking attention to detail, and Roesch approached the project with immense enthusiasm.

One lesson taught by the 10/23 was that the factory was too small to cater for the high volume popular market. The new car was aimed above that, where it could price itself more competitively, and as a large, elegant, four-seater tourer, with a straight six engine of unprecedented power and smoothness, the astonishing 14/45 (the first figure corresponding to RAC hp rating, the second to bhp output) was to become the progenitor of all subsequent Roesch Talbots. It was an immediate success, and the star of the 1926 Motor Show, with orders soon exceeding production capacity, which was consequently raised to 100 cars per week.

Roesch was not one to rest on his laurels, however, and despite the fact that the new car was technically excellent, by midsummer 1927 customers had managed to clock up sufficient mileage to expose one or two weaknesses in the design. Roesch decided to see for himself how the car would perform on the most rigorous test he could imagine; a high-speed 3000 mile journey across France, and through the Alps which was to constitute his summer holiday. It was to become a regular event, each new model subjected to the same gruelling ordeal, each breaking the previous record by a considerable margin.

The weaknesses of the 14/45 boiling on mountain passes, a stiff brake pedal, and front wings that shook loose were included in a number of modifications which resulted in the next Talbot, the model 75. Roesch had decided to change the model number system because it had been rendered meaningless by a host of other car manufacturers giving numbers to their cars that bore little relationship to RAC ratings or bhp. He reckoned the new car would comfortably reach 75mph and so the model number was aptly chosen.

The new 75 saw its engine capacity increase to 2267cc within the same length block, by arranging the cylinders equidistant from each other, widening the bore by 8.5mm to 69.5mm and the stroke by 5mm to 100mm. The crankshaft was given ample support by the provision of seven bearings, all of generous proportions. This work was included in a host of modifications, including a pressurised cooling system, valves of austenetic steel, heavier springs, and modified push-rods, that when assembled into the new engine and subjected to a full power test, soon showed that Roesch had turned a fine piece of engineering into a masterpiece. The engine was uncannily smooth, spinning effortlessly away at 4500 rpm, and producing 60 bhp; an increase of 50% for a capacity increase of 40%. This was an astonishing figure from a road engine of this capacity, not easily matched by its contemporaries, and certainly not if they were to possess the same grace and refinement as the Talbot unit.

In fact the performance of the new car was so good as to attract the attention of an enterprising Talbot dealer and aspiring race team manager, Arthur Fox of Fox and Nicholl Ltd. He believed they would stand a good chance at the Double Twelve hour race at Brooklands in 1930, and he soon persuaded George Roesch that here was an opportunity to develop and demonstrate the cars as never before.

Roesch agreed, but felt that the 75 was not good enough to race, and should be able to withstand a considerable power increase without sacrificing its reliability or refinement. The 75 unit had produced 60 bhp on a modest compression ratio, and gradually to increase bhp was clearly the path to an increase in output. Raising it to 7.1 yielded a reasonable increase of 5 bhp to 65 bhp at 4250 rpm, but at the same time exposed weaknesses in the breathing department. The inlet valve was too small, and yet was already close to the end of the available space in the ‘bath-tub’ that ran across the combustion chamber situating the valves in-line. This problem was soon overcome by the simple expediency of undercutting the head between adjacent cylinders. The valves were enlarged by 1mm each, and with stronger valve springs the engine gave 70 bhp at 4500 rpm.

At 8am one morning the engine was started up, accelerated up to full revs, and held there to see how long it would last. Everything ticked away happily for the first few hours before the power gradually fell off, and the exhaust lost its smoothness. Examination showed that the exhaust valves had overheated, the valve heads distorting under the sustained load. They were redesigned accordingly with more metal in their centres, and very soon Roesch had achieved his target of an utterly reliable 70 bhp at 4500 rpm.

Fox would have been happy indeed had Roesch left it there, but no-one was quite ready for the astonishing announcement one morning that Roesch intended to raise the compression ratio to 10.1. This was an unheard of figure for a car in 1930, other than for a few do-or-die specials that whizzed around Brooklands with the speed and duration of a firework. But the original engine had been designed to stand diesel compressions for future developments, and Roesch felt confident that it would withstand the increase. Withstand it it did, and with remarkable results. It was just as smooth, just as tractable, but accelerated up the revolution scale in one eager surge of power. With its light parts and superb balance its appetite for high revs was insatiable; and it produced 85 bhp.

Thus was born the new model which Roesch called the 90. This development of the 75 was put in a 9ft 3in chassis. On the day of the Brooklands Double Twelve the Talbots were sensational, astonishing more or less everyone with a speed that belied their seemingly effortless and silent running. The event was to end in disaster, however, when Rabagliatti, one of the Talbot drivers, made a mistake at the end of the pit straight trying to pass the leading Alfa. He braked too hard and too late, the car snapping suddenly to the right into the path of the following Talbot driven by Hebeler. The violent accident instantly killed Rabagliatti’s co-driver Ted Allery, and one spectator, and many others were injured when the Talbot ploughed through the iron railings of the spectators’ enclosure. The event continued, the atmosphere soured, but the third Talbot was withdrawn.

The 1930 Le Mans vindicated Roesch’s and Fox’s faith in the Talbot as a competition car, when they finished third and fourth behind two 6½-litre Bentleys, the third place car of Lewis and Eaton coming first in the Biennial Cup for Efficiency. The Talbot 90 completed the season with first Places in the Irish International Grand Prix, the Ulster Tourist Trophy and the Brooklands 500-Mile race.

Fox and Roesch then turned to face the 1931 season, and to exorcise the ghost of the Brooklands Double Twelve, and thus was born the 105, the most perfect expression of the genius of the Roesch Talbots. If any lesson was obvious from the 1930 racing season it was that a 2¼-litre engine was severely handicapped by having to compete in the three-litre class. Roesch’s first priority was to improve the power-to-weight ratio by stretching the engine to three litres. But his target was, rather typically, an increase in power disproportionately greater than the increase in capacity. Roesch’s eye was always on development, and he proposed to give the 105 engine a more efficient breathing capacity, and higher operating speeds. As well as the capacity increase it was therefore also necessary to modify the crankshaft, the valve sizes, and the valve gear. The new engine was to be no longer than the 90, the increase in capacity being accommodated by lengthening the stroke from 100mm to 115mm and widening the bore from 69.5mm to 75mm. The crankshaft was already adequately stiff, but Roesch intended to sacrifice none of the car’s refinement in his quest for higher revs and power; therefore the bearings were enlarged, pushing the main bearing from 60 to 64mm and the big end from 43 to 48mm. The con-rods were made of exotic nickel chrome steel, of very thin but stiff H-section, that were stamped and finished, like the valve spring retaining washers, in a coining die. This had a double advantage of making them cheap to manufacture, and also of giving them a hard surface skin that was resistant to fatigue. The 105 engine was initially designed with an aluminium block, which was used in the racing cars throughout 1931. Whether they continued to be used into the next season is not known.

The valve gear underwent a similarly extravagant modification procedure to insure it against failure at higher revs. The camshaft was reprofiled to lift the inlet valve by 11/32 inch, but the timing was retained with no overlap whatsoever. The camshaft was then case hardened, being plated with a layer of hard chrome, one of the earliest examples of this wear resisting technique. The knitting needle like pushrods were thickened slightly, retaining the same barrel section, but the rockers were completely redesigned. The broad wedge of the old fulcrum pivot was replaced by a ball and socket-type joint. The alignment was preserved by the forked rocker end that fitted over the valve stem, but now adjustment was simply a case of screwing the rocker pivot up or down. Lubrication came from a pressure fed gallery in which the pivots were set, and reached the ball and socket through a tiny feed hole in the ball end.

But the main problem facing Roesch in the construction of his new engine was how to enlarge the valves. The in-line valves of the 90 engine had already used up all the available space across the width of the cylinder head, and had even gone so far as to overlap the width of the cylinder bore, by undercutting the head. The solution had been discovered by Wolseley in the late twenties, and Roesch adapted the idea to his cylinder-head; he angled the valves across the head in a zig-zag formation. He was then able to install the biggest valves that had ever been used in such a short engine; the inlets were 44mm, the exhaust 40mm. This staggered formation caused other problems that Roesch soon managed to turn to his advantage. The rockers had to be of different lengths, and by making the exhaust rocker the longest, he was able to reduce the side thrust on the valve stems, a complaint to which the exhaust valves are more susceptible. Moreover because carburation system demanded a hot spot and therefore the exhaust and inlet manifolds to be on the same side, Roesch was left to decide which port to make the shorter. He opted for the exhaust port, to reduce the heating effect of the gases on the cylinder head, and the cooling system was arranged so that cool water would come up from the block and circulate the valve seats. This arrangement would also allow for a smoothly curved inlet port, feeding the mixture down into the cylinder.

A slightly longer chassis of 9ft. 6in. did involve some modification, to the suspension and running gear. Stronger clutch springs were fitted; the steering lock was reduced, and the brakes were improved. The latter by widening the drum, and fitting a finned aluminium ring to the outside to assist the cooling. A further modification was the adoption of the new Luvax hydraulic dampers, instead of the old friction dampers, the new articles possessing softer springing at low speeds, but as firm at high speeds. Roesch’s absolute faith in them as being perfectly adequate, even for racing purposes, was to have serious consequences in the Brooklands Double Twelve, for as on the previous 90 model only two chassis modifications, (a reduced steering lock, and a close ratio gearbox) differentiated the road cars from the racing machines.

The Talbot 105s brought together for MOTOR SPORT by Ian Poison are road-going models and represent the three styles of bodywork supplied for that machine by Vanden Plas coachbuilders. The elegant lines of the three vehicles differ mainly in detail work on the doors and the wings. The 105 Talbot Sports Tourer model came with a smooth flat curve to the top of the door and side of the front and rear wings. The 105 Sports Tourer deluxe had a rounded door, with a stepped top edge that matched the line of the rear wing. The front and rear wings had deeper sides. The blue car, owned by Mr Poison, is a 105 Speed Model 4-seater ‘Coupe des Alpes’ body with even more deep sided front and rear wings and a flat bottomed door with a steeply angled leading edge.

While these cars have a more prosaic elegance than the rakish splendour of an Alvis Speed 20 for example, on the road they are probably even more civilised and every bit as quick. There is an enjoyable parity between the lively yet smooth noise of the exhaust at one end of the car, and the almost uncanny silence at the other. And while the engine is perfectly tractable and will pull from very low revs, it really did seem to enjoy the top end of the rev scale, spinning away at over 4000 rpm with effortless enthusiasm.

The pre-selector gearbox worked beautifully and had obvious competition advantages, but also provided for a very smooth journey on the open road. Perhaps the most surprising aspect of the performance of the 105 was the accuracy with which it could be placed on the road, with direct, responsive steering, and very neutral handling characteristics. Even more impressive given that the chassis was essentially from the 14/45 of 1926.

The engine, again with a modest compression ratio of 6.5:1 for the road car gave out a remarkable 100 bhp at 4500 rpm, at which speed it was able to spin ceaselessly without complaint. Once fitted with a 10.2:1 compression ratio the power rose even more markedly to 140 bhp at 4900 rpm. It has a trace more vibration than the 90, and the performance was more peaky with the power rising steeply only after 2000 rpm. But it was clear that for the 1931 racing season prospects were rosy indeed, especially considering much of the main competition had folded with the economic depression.

Fate was to deal the Talbot team an unjust hand, however, in the form of a 2.3-litre Alfa Romeo, and although this on its own would not have been insurmountable opposition, coupled with a handicapping system that hadn’t caught up with the performance of the little MGs and favoured racing cars against tourers, Talbot proved to have a harder season with that superb 105 than they had done with the 90.

The Brooklands Double Twelve was the start of the 105’s racing career and although they were 1st and 2nd and 3rd on the track, and in class, running at a speed and for a distance unmatched by any opposition the leading Maserati retiring after a fast and furious lap by Ramponi they were so hopelessly handicapped against the swarm of diminutive MGs that outright victory was simply impossible. However success was not to elude them forever, a 3rd place at Le Mans, a 2nd in the Brooklands 500 Mile race, and in 1932 another car 3rd at Le Mans and 2nd in the Brooklands 1000 Mile race, being eventually crowned by the famous victory in the 1932 Coupe des Alpes in all classes above 2000cc, with the team prize going to a British team for the first time since Rolls-Royce had been awarded it nearly twenty years beforehand.

But the real victory for the Talbot 105 was in its all-round ability; its speed stamina and refinement in a simple and reliable construction. In the words of the late Anthony Blight whose book on Roesch and the Talbot is one of the most entertaining motor histories I have ever read; “If the Talbot 105 was not the fastest car of its time, it was undoubtedly one of the most versatile that the world has ever seen. CSR-W.