X-ray spec -- Ferrari 640

Three wins may not sound earth-shaking. But, as its designer explains to Keith Howard, Maranello’s ’89 car moved the goalposts

Designed by Briton John Barnard following his departure from McLaren in late 1986, Ferrari’s Tipo 640 Formula One car of ’89 earns a prominent place in the annals of the sport not so much because of its race results as for the technology it pioneered. Although the car — to the surprise of everyone involved, not least its driver Nigel Mansell — won the season’s opener in Brazil and triumphed again in Hungary and Portugal (the latter with Gerhard Berger), the dominant car that year was the McLaren MP4/5. Ferrari finished third in the constructors’ championship, and Mansell and Berger fourth and seventh in the driver rankings. Only after Barnard had left at the end of the season did his work for Maranello bear full fruit when the mildly reworked 641 very nearly took Prost to the title in 1990.

The technology which made the 640 a trendsetter was its semi-automatic gearbox, operated by the now ubiquitous paddles behind the steering wheel. It is natural to imagine that Barnard dreamt up the idea as the result of some simple arithmetic: time saved per shift multiplied by the number of shifts per GP equals the time gained throughout the race. But as Barnard — who comments here on various aspects of the car — explains, that was not the original motivation at all.

“The project started because I wanted to make the monocoque really narrow. In those days we had to fit a gearshift run through the monocoque, alongside the engine and back to the gearbox. It was a real pain to find a route for this and make room in the cockpit for the selector and the driver’s hand. I thought, ‘Surely, instead of a gear lever, I can have a switch.’ So it was a packaging imperative. The gear linkage affected lots of things, and of course the driver was taking his hand off the wheel during changes. The time-saving advantage came afterwards.

“Vittorio Ghidella, running Ferrari post-Enzo, was terrified of the ‘box failing and had a manual version built; Mansell tested it and said ‘forget it’. But the effort was a danger to the project.”

Originally intended to appear in ’88 on the 639, the electrohydraulic ‘box and normally aspirated V12 were held back until the following season once it was clear that the equivalency rules for ’88 made it vital to run a turbo to be competitive. When it did at last race in ’89 the new transmission was initially unreliable, but it eventually became clear to everyone in the pitlane that here was a technology they would all have to copy.

“Berger’s Imola accident was a big moment for me,” says Barnard. “If he had been badly injured or even worse I would have stopped. I remember standing there and thinking, If he’s gone I can’t do this anymore.’ In many ways it was the rules that caused it. In those days we could run the lower edge of the endplates level with the bottom of the car, and of course you wanted them closer to the track to make the wing more efficient. The FIA ruled out flexible endplate undersides, so we were forced to make the endplates stiff. All our strength calculations had been made assuming download on the wings but it transpired that, with these enforced stiff endplates and Berger riding the kerbs, it was an upward load that caused the main support tube to break. With download the wing then went straight on to the ground and under the front suspension. With flexible bottoms to the endplates, I don’t think the accident would have happened.”

“The 639 had conventional spring/damper units on top of the chassis, but because the 640 monocoque was so narrow I drew up a torsion-bar arrangement instead which started the short-torsion-bar fashion that continues today. It kept the installation as compact as possible, and also I didn’t like coil-over dampers. The springs were never well enough made to avoid side loads on the damper rods and consequently added friction. I designed a lower-friction package with the torsion bar, which ran on ball bearings. It was a really good solution.”

“The engine internals I left to Maranello, but I did influence its external features and how it fitted to the car — how the four main bolts were picked up in the engine; up to then that was not how they hung their engines from the tub. I was also very keen to get ancillaries beside the engine as small and tucked in as possible and on keeping the crank low — I drove them hard on those features and on keeping the engine length down, with it being a V12. That resulted in a cast-iron block with cast-in liners, which they cursed because these gave problems with honing the bore — the surface finish is very important to how the liner works with the piston rings — but they mastered it in the end. They thought the four-bearing crank would benefit power but years later they realised that its whipping actually lost more power than was gained by reduced bearing friction.”

“I originally thought the gearshift might be actuated by buttons on the steering wheel, but somebody in the drawing office, I can’t remember who, suggested paddles on the basis that karts used something similar. So we didn’t try buttons in the end, we went with the paddles. I tended to do that: make a decision and go for it. Unlike one of today’s F1 teams, with 600 or more people you couldn’t go through all the options. Today they can make three decisions and chase them all, but you couldn’t do that then.”

“At McLaren we were still using top exits on the radiators. But I knew this had an effect on the rear wing, so I thought: what better than to take that flow all the way inside the car and out the back, away from the wing. If you look at the 640 and 641 you’ll see they both use a completely ducted flow. When it was very hot we had to fit another small duct in the side of the car to extract a little of the flow (illustrated on the cutaway) but it wasn’t run all the time — it would be open or closed depending on conditions. Packaging the fuel centrally was also a concern. We carried a lot in those days — 220 or 230 litres — so the monocoque has pronounced panniers on the side around the fuel cell. These were bonded on to the monocoque, which was cut away inside to produce one large volume.”

“The semi-auto gearbox was slagged off early on for being unreliable, but that was unwarranted. A lot of the retirements in early 1989 listed as gearbox failures weren’t at all — they were due to loss of power to the ‘box. The alternator was driven by a belt from the crank and this kept falling off. It took a long time to find out why, using high-speed photography on the dyno. At this time the V12 only had a four-bearing crank which started to whip at certain revs, causing the front pulley to shed the belt. The alternator would stop and so would the gearbox electronics. We didn’t have any fundamental problems with the ‘box itself. It was pretty reliable. It was mostly standard inside and the hydraulic system was simple and robust.”