When the fashion for putting the engine behind the driver really got under way among the special builders, the problem arose of what to do for a gearbox and final drive unit. In the days of the front-engined special builders there was no problem, for engines like Bristols came with their own gearbox attached, and the ENV pre-selector gearbox could easily be coupled to most engines. Jaguar gearboxes were also available and Salisbury or ENV could supply final drive units. With the 1,500 c.c. Coventry Climax engine in the rear, Cooper utilised a modified Citroen front-wheel-drive unit, turned round through 180° and this led to a development that ended up with Jack Knight making a complete Cooper gearbox with the final drive combined within. I am not sure who it was who had the bright idea of taking a Volkswagen Beetle gearbox/final-drive unit and turning it through 180 ° and upside down, but it may have been Mike Hewland. Certainly it was Hewland who set the ball rolling with improved internals for the VW casing and later improved casings, oiling systems and so on, all of which developed into the Hewland gearbox/final-drive unit which is used by something like 95% of the racing cars built these days. The Americans refer to a combined gearbox and final drive as a “transaxle” which is a fairly neat description and though a lot of American racing cars use Hewland units, they also have some domestic ones.
In Italy Valerio Colotti, who used to be with Maserati, produced a “transaxle” for Formula One and in Germany Zahnrad-fabrik (Z.F) also made a suitable transmission, which Lotus used extensively. Through all this the serious “constructors” as distinct from “assemblers” were making their own transmissions for the rear engined phase, notably Ferrari and BRM, but subsequently neither of them produced as neat a package as Mike Hewland. Today in Formula One Ferrari stands alone in having a transmission that owes nothing to Mike Hewland, all the other teams either using a standard Hewland “transaxle”, a modified one, or one of their own design that has developed from the Hewland VW concept.
The major teams, like Williams, Tyrrell, Lotus and Brabham have their own designs of gearbox casing, made for them by such firms as Kent Alloys, into which they fit certain Hewland components with their own ideas on crownwheel and pinion and limited-slip differential. These teams develop their gearbox casings to fit in with the basic design of the rear of their car, incorporating their own mounting lugs, suspension pick-up points, their own oiling system and so on. A year or two ago McLaren squeezed a sixth set of gears into a Hewland casing, the normal one having five, but it was not a line of development that has progressed very far. Renault and Ligier rely on parts from Hewland and some of the less afluent or less well-equipped teams, rely on a complete Hewland transmission, as supplied. When Brabham were using the Alla Romeo flat-12 engine they started off with a Hewland package, but gradually it developed into an Alfa Romeo unit, albeit still to the Hewland layout. Lotus have experimented with a gearbox of their own manufacture, with assistance from the German Getrag firm, but it still owes its layout to Hewland, which in turn owes its layout to the Volkswagen which all comes down to old Dr. Ferdinand Porsche.
This basic layout consists of the crownwheel and pinion, and differential being situated just behind the clutch, with the gear clusters behind the axle line; the drive from the clutch runs under the differential, into the gearbox and forwards to the pinion. This is a layout that Dr. Porsche used on his original P-Wagen of 1933 that became the Auto Union and while it is a perfectly reasonable and reliable layout, it does mean that you have a great lump of weight mounted behind the centre-line of the rear axle. For any designer who believes in low polar moments of inertia this gearbox layout is not a good thing. It has to be suffered and allowed for by other means, but it is by no means ideal. As in so many things to do with racing design Enzo Ferrari and his designers have other ideas, and in 1975 became up with the “Trasversale” or transverse transmission. In this the whole gearbox is ahead of the axle centre line and to one side of the centre-line of the car. The input shaft feeds in through bevel gears so that the gearbox mainshaft and layshaft lie across the car instead of parallel to the car’s axis. The drive to the differential is through spur gears, doing away with crownwheels and pinions. Maserati used a similar layout on their 250F in 1954, with front engine. The Ferrari transverse gearbox and final-drive has proved enormously successful, extremely reliable and remarkably compact, but no-one has seen fit to copy it.
Similar in principle, but totally different in layout is the “transaxle” designed and built by Pete Weismann in California, for Ecclestone’s Brabham team. Whereas Ferrari was interested in not having a great lump of weight out the back of the car, Weismann was more interested in compactness, narrowness (for under-car air reasons) and having space behind the transmission in which to put the suspension units. The Weismann transmission unit is tall and narrow, taking up space that is not needed for anything else. Small bevel gears turn the transmission line at right-angles on entering the box and then gearbox shafts and final drive pinions are all in the same transverse plane.
While Ferrari shows no sign of deviating from the successful “trasversale” unit, and Hewland has no reason to change from the Volkswagen concept, there are other thoughts in designers’ minds. Colin Chapman still has the idea to develop a fully automatic transmission and Renault are thinking along the same lines. With the narrow r.p.m. band of the modern racing engine a close-ratio five-speed gearbox has such close ratios that it is not far from an infinitely-variable transmission, and with a turbo-charged engine, such as the Renault, that does not enjoy changes in r.p.m. it would seem logical to develop some form of automatic and infinitely-variable drive in what would be, in effect, a torque-converter. Some form of hydraulic drive would be the obvious answer, but space, weight and complications are against most systems as far as the racing car designer is concerned. In spite of what some people may think the power and torque of the racing engine is still the heart of the problem in racing car design and as more and more horsepower is squeezed out of the 3-litre atmospheric engine or the 1 1/2-litre forced-induction engine, r.p.m. continue to rise and power-bands become smaller and smaller. A usable r.p.m. band of 2,000 is about the average in Formula One, such as 9,000 to 11,000 r.p.m. as in the Cosworth DFV or 10,500 to 12,500 r.p.m. as with the Ferrari, so that gear ratios either have to be very close, to prevent “falling off the cam” or there have to be a lot of ratios, which is tiring for the driver, or some new form of transmission is called for. It is quite possible that we shall see something new in the coming years. When you look at the 1933 Auto Union and the Hewland gearbox of today we do not seem to have made much progress over the years. When Antique Automobiles resurrected a 1938 Auto Union that was devoid of a gearbox unit, they grafted in a Hewland without much difficulty. From a design standpoint that cannot be right, for what was good in 1933/38 should have long been superseded by 1980. — D.S.J.