In that very enjoyable book Cosworth, by Graham Robson, reviewed last June, the author makes the important point that the great racing-engine designer and tuner Keith Duckworth, whom his book is mostly about, was the first engineer to use a narrow-angle, four-valve cylinder head with the valves operated by twin overhead camshafts.
From time to time MOTOR SPORT has looked at the different ways of driving overhead camshafts, various valve arrangements, and how ingenious designers have contrived hemi-heads without the complication of upstairs camshafts. Now Robson has drawn attention (for the details, read his interesting book) to how Duckworth went for a narrow included, or opposed, valve-angle in his Cosworth-Ford FVA and DFV racing engines of 1966/67, which were financed by Ford of Britain and were extremely successful, DFV-powered Formula One cars winning 154 races between 1967 and 1983.
Duckworth's four-cylinder sixteen-valve FVA engine had its valves at an included angle of 40-deg, the 32-valve DFV V8 an included valve-angle of 32-deg. Robson as we all do, traces the origin of the race-winning twin-cam multi-valve engine back to the Henry-designed (or drawn) 1912 GP Peugeot, with which Georges Boillot won that year's 956 mile French classic at Dieppe at 68.45 mph. That was enough to cause many other racing car designers to use twin overhead camshafts to actuate four valves per cylinder and Peugeot to refine its concept for the purpose of winning the 1913 French Grand Prix and the 1913 Coupe de L'Auto races. In the 1920s the true hemi-head, using two valves per cylinder gradually took over, but today the multi-valve twin-cam engine has returned, for both the racing car and the production car. It was Duckworth, however, who, as Robson says, "set a new benchmark" with his narrow-angle valves.
It is debateable how the valves in the 1912 Peugeot GP engine were angled. Laurence Pomeroy, as Robson remembers, quotes 45-deg, Griffith Borgeson (in his book The Classic Twin-Cam Engine) says possibly 60-deg, but drawings studied later suggest 90-deg. As no GP Peugeot of this kind has survived, we may never know. (Incidentally, Borgeson's profound knowledge of old engines does not seem to extend to new ones, as a photograph caption in his aforesaid book fails to identify Keith Duckworth in a group of people standing round one of his own DFV power units!) Incidentally, Robson notes that a modern Cosworth racing engine betters the efficiency of the 1912 Peugeot engine by a factor of eight or ten; however, the power output of that Peugeot engine has been variously quoted as 130, 148 and 175 bhp.
When WO Bentley was working on his 3-litre Bentley engine in 1919 it was given a valve included angle of 30-deg, which as Robson points out, "was almost the benchmark set by Cosworth five decades later!" On the other hand, perhaps the lower compression ratios, lower power outputs and lower fuel octane ratings make any comparisons invalid.
It is well known that WO was assisted by Fred Burgess, who had designed the 1914 TT Humbers, when planning the 3-litre Bentley and that he had one of the Humbers (which one of the team remains an intriguing mystery) to dissect and study. The TT Humber engine followed closely the design of the 1913 3-litre Coupe de L'Auto Peugeot, so might be expected to have had a wider valve-angle than the Bentley, that of the Peugeot being 90-deg. In fact, the Humber's valve-angle is 90-deg.
I think therefore, that the Bentley engine was given this very narrow valve angle simply because, with a single overhead camshaft, rockers were necessary for prodding the valves, and they could be kept decently short if the valves were not too widely inclined. Indeed, the valve-gear of the production 3-litre Bentley, using a forked-rocker to open the two inlet valves and separate rockers for the two exhaust valves, with triple cams per cylinder, was notably compact. There were other engines, both push-rod and overhead camshaft, which had a small valve angle, and even in some case near-vertical or vertical valves, no doubt in the interests of short rockers, a narrow cam cover, and perhaps reduced costs in head casting, etc. Even with twin overhead camshafts, the less the valve inclination, the closer together could be the camshafts, resulting in a reduced chain-run, or fewer gears, in the camshafts' drive.
Robson remarks, I think quite rightly, that if the designers (of the 3-litre Bentley) had only realised it, they were onto an efficient combustion chamber shape, for which reason Duckworth opted for narrow-angle heads 50 years later. Or did WO know about this design factor which was later employed in Cosworth engines? Are we, in fact, in thinking otherwise, doing Walter Bentley a grave injustice?
In the case of the Bentley and other fixed-head engines, another reason for a narrow valve angle could be that, as Graham Robson himself reminds me, machining the valve seats etc up such a casting, a matter of over 150 mm in the case of the Bentley, might have been difficult if these had been acutely angled. The 1914 GP Mercedes engine, which WO had looked at before planning his 3-litre Bentley, had valves inclined at 60-deg as had the Bentley and would have had the seats machined up a bore length of more than 165mm. WB