To the best of my knowledge, the feckless Life F1 team came closest of anyone to date of putting a 'W' configuration engine on the F1 start grid. They began the 1990 season running a 3.5L W-12 designed by former Ferrari engineer Franco Rocchi. The idea behind the 'W' was that it would be nearer the overall length of a V-8 engine but would have V-12 power.
The Rocchi-Life F35 'W' engine essentially was a V-8 with a third bank of four cylinders with a 60° bank angle. It used a conventional V-8's flat plane crank with three connecting rods per crank journal but its configuration otherwise was typical for the era; DOHC & 5-valves/cylinder.
Bad enough Life were running a dated chassis (bought from the stillborn
First team), but what the F35 lacked in horsepower, it made up for by being overweight and unreliable. When it managed to complete a lap in anger, it typically was 20 seconds or more off pole pace. At Hockenheimring's long straightaways, it was slowest through the traps by 64kph/40mph.
After failing to qualify at any of the first 12 races of 1990 (back when the term "qualifying" was taken literally), Life switched to a V-8 Judd mill, but for naught. The Life F190 never made its way to the starting grid.
A couple of years earlier, aerodynamicist Guy Nègre (of
Moteurs Guy Nègre) built a 3.5L W-12 he hoped would find a home in F1. IIRC this was his first attempt designing a petrol engine, and he did it from scratch. That MGN today is better known for "zero emissions" compressed air powered cars is a clue how successful Nègre's first and only F1 engine was.
Despite Renault having introduced the pneumatic valve to F1 in 1977, which was the genesis of today's bullet-proof 23,000 rpm valve trains, for some reason, in 1988, Nègre thought there was still another, better, approach. His W-12 used rotary valves.
The rotary valve (also sometimes called a 'sleeve' valve) is one of the simplest possible, but it is so functionally different from the normal automobile engine poppet valve, it can be difficult to conceive how it might work. This is one interpretation of the rotary valve, the most comprehensible cutaway view I could find (reproduced from
Auto Technology magazine):
Instead of the cylinder's dome having several round openings for multiple valves, there's only the one square-ish window, or 'port,' quite similar to the port in piston port 2-stroke engines. Except in the case of the 2-stroke, the ports are in the cylinder wall, with separate ports for intake and exhaust. In the 4-stroke, the lone port in the dome allows for both ingress of fuel-air mix and egress of spent exhaust.
The 'valve' is cylindrical, and hollow, with windows or ports of its own, which interface with the port in the cylinder's dome. The valve cylinder rotates at one-half the speed of the crankshaft, switching from intake to exhaust and back, once every two revolutions of the crank. In the position shown, the intake port in the rotary valve is aligned with the port in the dome, allowing the cylinder to receive the incoming fuel/air charge (blue side) through the hollow body of the rotary valve itself. When the valve rotates 180°, it opens the path to the exhaust manifold (red side), and the exhaust gasses are expelled through the rotary valve by the rising piston and their own their energies.
The MGN W-12:
Conceptually, it is brilliant. It has no reciprocating components. Rotating in general is preferable to reciprocating because of the added stress, wear and energy involved in cyclically reversing the direction of the reciprocating bits. Apart from indexing the timing chain, it requires no adjustments. It eliminates cam shafts, rocker arms, tappets, the poppet valve, valve seats, and valve springs. There is zero chance of valve float, valve bounce or dropping a valve, all of which are the particular province of the poppet valve.
But the fact that you never have seen a (4-stroke) road automobile that uses rotary valves is a hint that the brill design is not so easily translated into reality. In the case of Nègre's interpretation, there were problems with the effectiveness of the seal of the rotary valve to the port. It wore aggressively, which exacerbated the sealing problem. And the turbulence caused by the opening to the cavity within the rotary valve proved disruptive to flame front propagation.
During the summer of 1989, Nègre personally financed the testing of his engine in an
AGS JH22 chassis, which was the closest it ever came to F1. ASA Armagnac Bigorre put the MGN W-12 in their
Norma M6 1990 LeMans entrant, but they were unable to get the engine running long enough to get it out of the garage. Literally.
The MGN W-12 powered AGS:
Thanks to the rotary valve, WHEN it ran, the MGN W-12 could be revved to well more than 12,000 rpm. It produced horsepower on par with the (admittedly lighter) Judd V-8, but Lamborghini's 1990 engines turned just as many revs using a conventional valve train. And the Lambos, on occasion, would run long enough to at least get clear of the garage. And sometimes around the entire circuit, as well.
.
). That helped to make his mind up so he bought one and stripped it down, or as he puts it "disassembled and the heads and cylinders carefully measured".
