Last month we disclosed Griffith Borgeson’s skillful defence of Leon Michelat’s 1914 GP Delage design as the first to employ effective desmodromic valve-gear (positive valve closing by mechanical means) on the high-speed internal combustion engine.
Greatly admiring Borgeson’s research into this subject in the American Automobile Quarterly, we quoted him as establishing that this design breakthrough rightly belonged to the advanced team at Delage racing cars which took part in the dramatic French Grand Prix at Lyon, on the eve of the First World War.
At the time, racing (and indeed, other high-speed) engines turning at up to 3,000rpm, were often plagued by valve spring breakage or loss of tension. Thus designers were tempted to duplicate the valves, in hope of reducing stresses with lighter components, or at least enabling a car to carry on with one or more valves or springs hors de combat. So it is easy to see why total elimination of the valve return springs was not to be despised.
After the war, engine speeds continued to increase and hemispherical combustion chambers were deemed the most efficient. So large valves, two to a cylinder, became the norm and valve springs were even more highly stressed. Furthermore, if they didn’t wilt under engine heat, they were unlikely to be able to control valve-float, particularly with large valve masses and racing cam-shafts imposing rapid valve-opening and closing.
Even if the valve springs were up to an engine’s demands, heavy loadings on other parts of the valve mechanism were imposed by springs strong enough to prevent valve flutter or bounce, which if present altered the timing plot of the cam. Desmodromics were attractive, but difficult to attain.
The way Leon Michelat got around the problem on the 1914 GP Delage (an advanced car which had followed Peugeot’s celebrated 1912 “first” of twin overhead camshafts operating sixteen valves in pent-roof heads, and had a five-speed gearbox and lever applied four-wheel brakes) was to open each valve with one cam and close it with two more cams flanking this opening cam, all working within a stirrup-shaped tappet. The cams were keyed separately to the camshaft and very rapid valve-opening and closing was achieved. There was provision for adjustment and small springs on a bridge-piece ensured valve adhesion to the seats, should expansion or distortion require this.
Close limits were involved, but the prize was ridding the engine of the danger of valve spring fatigue, maintaining the designed timing, and obviating the power losses which might arise from compressing strong valve springs.
Whether the end justified the means is another matter. In that fateful 1914 468-mile Grand Prix, won by Mercedes, two of these Delages retired (with respectively 94 and 70 miles left to run) and Arthur Duray’s could do no better than eighth place. Kent Karslake and TASO Mathieson suggest that the novel Delage valve gear had been re-adjusted before the start, adversely affecting the cars of Bablot and Guyot, but Borgeson says last-minute changing of the magnetos caused the trouble.
Although the Delage team did not perform well in the 1914 GP, Borgeson shows that they ran quite well afterwards in American races so assuming they continued to use their original valve gear (one engine blew-up and was replaced by a Miller power-unit) Delage desmodromics were fully proved. It was apparently from looking at the 1914 Delage engine that Harry Miller evolved his own interpretation of positive valve opening and closing, using one cam per valve and a forked rocker, for his 1916 4.7-litre four-cylinder sixteen-valve racing engine.
In that same 1914 GP, Theophile Schneider also entered a team of cars, designed by Michaux, with desmodromic valve-gear. So, while these achieved rather less success than the Delages, two retiring and the third finishing ninth, credit for the adoption of positive valve-closing must surely be shared between Michelat and Michaux?
Delage, incidentally, had used valves inclined at 90°, but those of the Schneider were inclined by only 10° and were operated by a single ohc, the desmodromic action being obtained by a second cam acting on an extension of each rocker.
It is always dangerous to try give credit for “firsts” in engineering. It is said that Aries had desmodromic valve-gear in the 1908 Grand Prix des Voiturettes, but none of the three cars lasted more than two laps.
After the war, smaller racing engines, built to the 3-litre Formula and using small multiple valves in small multiple cylinders, temporarily eased the valve-spring anxiety; but by 1922, when 2-litre GP engines were approaching a peak crankshaft speed of 5,000rpm, it appeared again. Yet few seemed convinced that desmodromics were a means of defeating it.
Although enthusiasm for a positive valve-control was at a low ebb, and after trying it for a racing engine, Roland-Pilain quickly abandoned it, another engineer was ready to give it a whirl. M Causan of Automobiles Bignan was known for the careful development of his ohc engines, first of two-valve, then of four-valve configuration.
There had been a side-valve 31/2-litre Bignan in 1920 in which a detachable water-jacket covered the valve caps, as on the first 12/20 Calthorpe in 1923 which was attributed to Hugh Rose and not proceeded with. Whether Causan was responsible I do not know, but he gained a considerable reputation in 1921 with his 3-litre four cylinder sixteen-valve ohc racing cars built for the Corsican Grand Prix. This turned out to be a very tough 2741/2 mile race, which Duray won in the Bignan Sport at 45mph after his team-mate Delaunay had been killed. Then, for his new 2-litre, Causan went desmodromic.
The patent showed that the overhead-shaft was retained, but now used to operate the two vertical valves per cylinder via plates or cam-faces driven by bevel gears. Rollers on the valve stems engaged the contoured edges of these swash-plates or cams. At first, springs had been fitted to assist the closing of the valves (Delage had used small springs and Schneider normal ones, for the same purpose) but Bignan, like Mercedes-Benz after WW2, discovered they were unnecessary.
The desmo-Bignans ran in the French Touring Car GP, retiring, but not because of their advanced valve-gear. Gros was third at Spa, the engine running at 4,000rpm, with a peak speed of 5,000rpm.
Jacques Bignan was so confident in the new desmo-engine that it was put into production. A 2-litre sports-tourer with this valve operation was shown at the 1922 Paris Salon and then came to England where it was exhibited on Stand 173 at White City, the Olympia overflow. The more sporting models were guaranteed to do 80mph.
To publicise the advanced valve-gear, Gros drove one at the 1922 Armistice Brooklands meeting, and the rasping car had no difficulty in winning the 90mph handicap at 79.9mph. He also set up three Class B Records. However, perhaps it was all too costly, for, whereas the standard chassis cost £675, you had to pay an additional £200 for the springless valve-gear. A year later Bignan abandoned the project.
Although the valve-spring problem did not go away, it was to some extent alleviated in the vintage years by improved springs, developed by Terry’s and others, the use of double valve springs to reduce valve-bounce, and better understanding of cylinder-head cooling. Nevertheless, Parry Thomas preferred leaf springs, the motorcycle world had its hairpin springs, away from the heat, and there were attempts to eliminate springs by having rotary-valves.
About the last use of desmodromic valve-gear in vintage times was on the little six-cylinder 745cc Vagova sports/racing car from France. One might have thought its valves would be so tiny as not to cause much trouble, but speeds of 5,500-6,000rpm were envisaged. It was deemed prudent to employ Godet & Vareille desmodromics, which operated the oh-valves positively with two camshafts and fulcrum rocker-arms, the lower camshaft opening and the upper camshaft closing the valves.
The most convincing employment of desmodromic valve-gear came when Daimler-Benz, the world’s greatest automotive engineers, returned to motor racing in 1954. Experiments late in 1952 had convinced Hans Gassmann that, unless conventional valve springs could be dispensed with, the desired high cr of the new Mercedes-Benz racing engines could not be achieved, owing to valve bounce. Valve springs proved unreliable in any case.
So he thought up positive operation, apparently without reference to any previous designs of this kind. The concept was one of two cams (one opening a valve, the other closing it through the medium of an L-shaped rocker) with a light spring to ensure final seating. They called it Zwangsteurung, or Z-drive.
This system appeared on the straight-eight 32-valve M196 engine in the successful W196 Mercedes-Benz Grand Prix cars which gave Fangio the 1954 and 1955 Drivers World Championships. At Silverstone, Fangio was going to 9,800rpm, and apart from being reliable at this speed, the high valve lift, quick valve acceleration and maximum opening with a minimum of overlap contributed to the great power of these cars.
The Z-drive was gradually improved; the return springs were deleted after being apt to break, a single-piece rocker replaced the two-piece rocker for the deceleration cam (the Y-arm of which engaged grooves on the valve stems to provide the mechanical closure), and eventually double in place of single eccentrics provided for D-cam clearance.
The equally-successful 300SLR sports/racing Mercedes-Benz cars which won the 1955 Mille Miglia, TT and Targa Florio races also had this Z-drive valve-gear, enabling the 3-litre engines to run normally up to 7,800rpm and go to 8,400rpm without harm. On the last of these engines further improvements were made to the desmodromic valve operation — the cams being moved closer together, resulting in stronger but lighter followers, oil-feed being provided for the cams, and closing rocker-clearance being made constant over the whole cam surface without recourse to relieving high-spots by hand. Moreover, moving the A-cams (acceleration cams) and D-cams closer together made possible an individual locking-clamp and hold-down screw for each set of serrated eccentric collars.
This was the ultimate proof that positive valve-closing paid off, enabling Mercedes-Benz to dominate Grand Prix and sportscar racing until it retired from such expensive activities.
There was spasmodic adoption of desmodromic valve-gear by Osca, Lance Reventlow’s F1 Scarab, Stanguellini, and by Ducati and to a lesser extent Norton in the motorcycle world. But its most effective employment and development was undoubtedly by Mercedes Benz. WB