PORSCHE 917-10 (in 1972)


MARK DONOHUE'S PORSCHE 917-1Os aren't the only ones racing in the U.S., but at the present are the only ones actually running with the turbochargers described in Pete Lyons' story in this issue. Peter Gregg of Florida has an unblown 917-10, and in our own neighborhood Porsche-Audi dealer Vasek Polak has another, which driver Milt Minter piloted to fourth place at Mosport and third in the Road Atlanta Can-Am.

The car is essentially the same with or without turbos but there are some important differences. In general, it looks the same as the original 917-10 shown in the Penske Porsche story rather than the much changed Penske car. Its chassis structure is the same configuration as that of the turbo car but it is probably a little heavier because of the use of less exotic materials here and there. Its brake calipers are not finned, though it does use the cross-drilled rotors like the faster car.- And its 12-cyl engine is bored out to the full SA-liter displacement currently possible with the 917 engine design. This means a 90.0-mm bore, and it's likely that this unblown engine's 90.0 x 70A-mm dimensions are the same as next year's 6-cyl production Porsche 911 2.7-liter engine is going to be. The unblown 12-cyl engine produces 665 bhp DIN @ 8300 rpm and a peak of 470 Ib-ft torque @ 6800 on its 10.5: 1 compression ratio.

Vasek Polak invited us out to Willow Springs Raceway, where Chet Kastner runs the west-coast Jim Russell Driving School, to watch his 2nd-string driver Jim Cook break in a new engine in his 917-10, inspect the car, and finally take a few laps in it. Polak is running an unblown car currently because he wants to hold off a while and see how the blown one works before investing the extra $34,000 it'll take to go the turbocharging route-can't blame him for that, for though Polak has raced for many years and has a most impressive collection of Porsche racing cars, $34,000 isn't an amount to be spent lightly. If he does go for turbocharging later, he can modify his present engines by changing the cylinders and pistons to bring the displacement down to 5 liters. Porsche is not yet confident about engine durability with 5.4 liters and turbocharging.

At least when we tested it, Polak's car didn't even have the airfoil extension that had been added to the original 917-10, so its wing and supports extend very little rearward of the main body form and altogether it's about 1112 ft shorter than the Penske car.

Polak's chief mechanic is Alwin Springer, and for six months Polak is enjoying the special support of Gustav Nietsche from the Porsche racing department in Stuttgart, a l5-year employee there who has been with the 917 project since its beginning. Polak will continue to campaign the unblown car for the entire Can-Am season and hopes (with good reason, I think) to finish the season in fairly high standing. He was sorting out the car and running in the engine. for Road Atlanta when we went out to take our look.

Driving the 917-10

VASEK WAS understandably reluctant to actually let anyone except his own driver go out in the 917 but after a little coaxing he was persuaded to let me take it out for a few slow laps-without the bodywork in place. He no doubt reasoned that there would be less temptation to go fast and that, in case something did happen, at least the fragile fiberglass bodywork would be spared.

So I climbed in "Climbed" is the appropriate word, too. First you have to put your foot in precisely the right spot (marked by tape) on the sheet-metal tub that becomes a mock doorsill, then climb on down from there, carefully threading feet and legs into. the space around, between and through structural members, battery, steering linkage, pedals, wiring and fluid reservoirs. Once in place, you are sprawled out with your legs nearly horizontal and it is amazingly comfortable. The seat has a surprising soft cloth covering and as it was dimensioned to accommodate Milt, who is heavier than my 145 lb, there was plenty of room to be comfortable while still not rattling around. I was strapped in with the double upper-body straps and crotch strap, given instructions about the minimal instrumentation and switches (especially the master switch) and told to limit the revs to 6000. The redline is at 8400 and there's an ignition cutout to make sure it's observed.

Although the tractability of this l2-cyl engine has been commented on before, it was nevertheless surprising to me to push the starter button, have the engine catch immediately and then idle at only 1000 rpm. That's real tractability. Into 1st gear with the right hand lever. It's short and ball jointed, has a wooden knob and works almost as easily and precisely as that in a good road car and there's about 1.5 in. throw between speeds. To get underway, you ease the clutch out, feed in a little throttle and simply drive off.

The laps I had around Willow Springs were, of course, at speeds far below the 917's potential. But even driving it as if it were a road car, it was possible to sense its capabilities simply by the ease with which it trundled around. The pedals are perfectly placed for double-clutching, though this is not something a racing driver ordinarily does. The steering wheel feels very small in the hands, there are only 1.5 turns lock to lock, and steering effort is moderate, but not light. It is super-responsive to steering input, of course, even when compared to the most obedient road sports car. The braking was . terrific and though we did not have the car instrumented, the stopping capability was obviously over a "g."

The blown 917-10 is probably not this tractable, but it's probably not far off. It is incredible to me that a Can-Am car could be so easy to drive. Is it equally easy to drive fast? Cook says yes.

Porsche 908

POLAK HAD also brought along one of his two 908s and with this he wasn't nearly so cautious as with the 917. Although he specified a rev limit to be observed, he put no restrictions on the number of laps and the bodywork stayed on-which I later had reason to regret.

The 3-liter 8-cyl 908 is of course smaller and lighter than the 917. This particular example had very short gearing and it had a 5-speed box rather than the 9l7's 4-speed. This gearbox, whose shift pattern is like that of the early 911s, was less precise than the 4-speed but still nice. Its handling and braking seemed basically similar to the 9l7's and after building up speed on successive laps, it seemed almost as though it was impossible to make a mistake. Then. . . .

By this time Cook was running the 917 around again, easily going faster than the 908 even though I was pressing on. Bill Motta was riding with me, taking some of the pictures you see here. Cook came up from behind on the back straight between two fast bends, I moved left to let him by and suddenly found the left front wheel off the edge of the pavement. Poor Bill was really concerned as we bumped along-he wasn't belted in since there's no belt for the "passenger's" seat-and though I worked the car back onto the pavement, the damage was done. The low-riding front end, which-Vasek had warned me might hit the pavement on turns, was ripped and new fiberglass would have to be formed to repair it. But Vasek, who must be one of the nicest guys in racing, was hardly bothered. My thanks and apologies to him.

It was a superb treat to be allowed to tryout these two Porsche racing machines. Even at Road & Track we don't get a chance to drive this sort of equipment every day.

Our best wishes to Vasek Polak and his team for a successful

Can-Am season.

A MERICANS ARE ADDICTED to The Noble Experiment. For years the SCCA tried to run the Can-Am with as few car rules as possible, a noble experiment dedicated to the proposition that a good spectacle could be achieved with essentially Formula Libre machinery. Unhappily, in six years the series has never really reached its potential. It is the premier form of American road racing but it hasn't attained the stature of various oval-track or drags trip shows. Thus the Can-Am has been open to criticism, and over the past few years several different forces have been chipping away at the original Formula Libre concept. Aerodynamics, for one large instance, are under rigid control, and more and more the movements of an innovative designer's pencil have been laid out for him by one regulation or another.

Still free, though, are two areas: overall weight and engine size and type (except that, to pick nits, you can't use a gas turbine, nor an engine under 2500 cc). To those of us who still wistfully subscribe to free formulae, then, it is gratifying that in finally making their long-anticipated North American assault the Porsche engineers have concentrated on the two free areas. The new Can-Am Porsche 917 is made of elaborate and expensive lightweight materials, and by exhaust supercharging their 12-cylinder engine they have made their first solid foray into a new realm of road racing power. The best part, though, is that this complicated, potentially troublesome new concept works. In the car's debut performance at Mosport Park Mark Donohue was fastest. . .

Naturally, the critical forces bayed like hyenas. Unfair! Too Expensive! Ban It! Write New Rules! Outlaw Turbocharged Aircooled Multicam Engines! Hold off, pack. Walk out to one of the corners with a stopwatch. Check cornering speeds. That's where Donohue was making it at Mosport. His turbopower-800, 900, whatever-was hustling him up the straights pretty well but he was a lot faster than anybody else in the corners. By a factor of 5%-which is a bunch.

No, the thing about the L&M Penske Porsche is that it is a total automobile. It's not a trick with one good feature, like a dragster. It does the whole job. Banning one aspect of it would be silly, a thoughtless gut reaction.

Porsche's Group 7 project takes the form it does for practical reasons. Even those engineers who are intrigued by the difficulties of developing a turbocharged road circuit engine must recognize that it's not the best way to generate lots of horses.

The simplest way, which in this context as in so many others is the best way, is to build big engines. The 500-cu-in. Chevies are the best Can-Am engines because they are simple, light, compact and relatively cheap. The most straightforward way to challenge these attributes would be to design a better Chevy. However, to do that Porsche would have had to start from scratch, and that would have been far from cheap (especially the way they do things). It was much more practical to draw on their roomful of existing 5-liter 917 engines, virtually useless after the end of the 1969-1971 Group 5 regulations. These can be taken out to about 5.4 liters and it might be anticipated that careful development work and higher rev limits could squeeze out something approaching 700 bhp. However, in terms of low-speed performance (torque) such an engine would be left behind by a good Chevy. But if you supercharge an engine, you in effect increase its displacement. In round figures, stuff 50% more air into a 5~liter and you get 7.5 liters. Of course, there will be problems-dissipating 50% more heat, holding 50% more BMEP inside-but solving them should be cheaper than developing a whole new engine.

If you choose what has emerged as the most efficient supercharging method, in which the compressor(s) is (are) driven by exhaust pressure, you introduce the additional problem of a lag between the time the driver's foot asks for more power and the time- the engine grants it to him. At Indianapolis the engines are small and the boost pressures are high, so the throttle lag is sizable. Drivers cope with it there because of the relatively simple (which word emphatically does not mean easy) nature of the cornering. Occasionally such engines are tried on road courses but there hasn't been any discern able rush to adopt them. However, if you were to start with a relatively big displacement and didn't ask for too great a boost, maybe the lag could be kept within manageable bounds.

Porsche engineers were developing a turbocharged unit all through last season. Occasionally a driver would come back from a test session with his eyes big and his mouth quavering out stories about giant wheelstands and 1000-horsepower spyders. Yes, but how about the throttle lag? "Well, yes, there is some throttle lag. . ."

Other aspects of the coming Can-Am assault were not being neglected. Jo Siffert brought a new spyder to the series to try out some chassis ideas. This 917-10, lighter than the earlier coupes and the spyders derived from them, had its own space frame and its own aerodynamic peculiarities. It was hearteningly reliable, Seppi finishing all five races he entered and in fact scoring two second places.

Meanwhile, the redoubtable Roger Penske Racing organization was itching to get back into something they'd done well enough to tantalize themselves (Donohue won two races, in 1966 and 1968) but which they knew required the resources of full factory backing. After negotiations with three other European factories, they entered into a 3-year liason with Porsche. Mark began to spend week after week in, Germany.

"I've got a unique relationship in this project," says Donohue, "a kind of relationship I've never had before Helmut Flegel is, oh, sort of 'the boss of the 917' and he's a man who is capable in every respect. I can talk to him, I can say things to him and he to me and we understand each other perfectly. We speak the same language, and I've never really had a relationship like that before. I've always been used to doing it all myself. This is fabulous for me. I've really learned a lot."

Good lord, a double Donohue. McLaren took this new threat so seriously they pirated Jackie Stewart away from Lola and laid down their first all-new car since 1967. Their Detroit engine facility began playing with turbocharged Chevies. Can-Am fans, used to spending the winter with crossed fingers, began to have real hope. The 1972 confrontation might be the real thing at last.
Early in the year the chassis sent to America as a development hack allegedly went around Atlanta at record speed but later it wasn't so impressive at Riverside. At this stage it wasn't using the turbo engine; that was still obviously not ready. By April there was no hard evidence that the Porsche was going to be a contender. Visitors to McLaren's Coin brook plant were given the impression they no longer worried about it very much; their own M20 prototype was going really fast at Goodwood. Apparently their own experiments with blown engines had shown them there were big problems to be solved. They probably wouldn't go that route.

This was all more or less confirmed by Donohue at Mosport when he remarked that only a short time before he'd expected to miss the first event. But then the development team made some big breakthroughs. In the last couple of weeks the car changed a lot, both in the engine compartment and the overall body shape. Pictures that had leaked out into public gaze suddenly looked outdated.

In the week before official Mosport practice Donohue got in some solid testing at the track-at this point the McLarens weren't even in the hemisphere-and cut about three seconds off the existing Group 7 lap record. (That had been Dan Gurney's, from 1970, at 1: 16.4.) In fact, he was about 1.5 seconds better than the best Fl time. We'd have to wait for the McLarens to arrive to know, but it looked like they had some real catching up to do at long last.

There's a lot of familiar 917 in the Can-Am car. The basic dimensions are about the same, the suspension layout and pieces are very similar, it retains the space-frame construction concept, and the flat-12 engine and 4-speed transaxle look much the same. A lot of the details are special, though: the frame tubes are of some secret material which is probably a magnesium alloy; the body is not glued to the frame but the metal sheets forming the fuel tank sides are welded into the structure for stiffness (the rubber cells merely rest in these without being enclosed-the inquisitive finger can depress the exposed top of the cells); the suspension links are blue-anodized aluminum alloys; there is lots of magnesium and titanium and honeycomb fiberglass everywhere. Going on the announced weight of the normal-issue 917-10 (driven at Mosport by Minter and Gregg), which is about 1565 lb, the Penske car probably scales about 1520. That makes it very close to what the McLarens weigh.

The body shape breaks ground in that at long last (although this comment also applied to the Chaparral 21) some aerodynamicist has had the courage to produce an ugly shape when his tests showed it to be best. The sharp-bottomed, concave, louvered nose is very. short on esthetics-but if you visualize the airflow and pressure distributions over it, you'll see it must work very well. The dished contour produces downforce; then when air reaches the wheel domes and must bend back to horizontal flow, a condition that would otherwise produce lift at that point, the copious venting fills in the low pressure with air from underneath. It goes deeper: since there is in fact no belly pan to this nose section, any low pressure generated below the visible surface is carried right down to the road. Presto, a "sucker" effect.

It is noticeable that, compared to earlier photos, the new nose has allowed the rear airfoil to be enlarged and mounted farther aft. It may be this "giant-hand" aerodynamic system that accounts for the very high cornering power (at 80 mph on Porsche's Weissach skid pad the car reportedly has generated
.1.54g). At Mosport the Porsche was using the identical Goodyears as fitted to the McLarens and the Shadow, so it wasn't just a matter of tires. The high engine power is able to pull the necessarily high-drag shape up the straights as fast as or faster than anything else, too.

Which brings us to the most glamorous, and most secret, part: the blown engine. Each bank of six cylinders has its own blower system, although there are interconnections between the two sides at certain points. The six exhaust pipes per side merge down into one, which feeds a smallish German-made (Ebersbacher) turbine wheel. The two blower units are mounted in available space to either side of the transmission. Just before the turbine the two exhausts are connected by a cross pipe over the top of the transmission. In the middle of the pipe is the single blow-off valve. As on any Indy installation, this is pre-set to relieve pressure at a certain value and is probably the primary "horsepower screw."

The pair of compressors draw air from twin NACA inlets on the body top. and force it forward to a pair of log manifolds atop the two cylinder banks. So far so normal, but from here it gets tricky; careful study of earlier photos (see some of the L&M ads in the magazines) indicates the following details are fairly recent additions. Atop each log are four devices which look like spring-loaded relief valves; these "steam whistles" may serve to eliminate the chance of manifold explosions, or they may have to do with equalizing pressure along the length of the log. On the other hand, there is a theory that they are inlets which open when the manifold pressure drops and thereby convert the engine temporarily to normal aspiration. This could reduce throttle lag. There aren't many who know just what these eight things do and those who know aren't talking.

Farther aft, just after the air comes out of the compressors, each pipe has what is pretty certainly a straightforward pressure blowoff valve. This is also something Indy engines don't have. The most intriguing gadgets are between these two blowoffs and the logs. In a T-shaped housing just before each log lives what has to be a little butterfly valve, spring-loaded to the closed position and located so that when it opens it opens the manifold to outside air. A sliding-link connection to the main throttle linkage allows this butterfly to close when the throttles are opened more than about half way, but when the driver lifts his foot right off this butterfly is caused to snap open. Now, what do you think is the purpose of all this? Could it suck air in at low throttle openings, once again to make it normally aspirated? Or does it serve to dump inlet manifold pressure when the driver reaches a braking zone, canceling any tendency of the engine to drive on for an instant? Could it also eliminate most of the back pressure in the intake manifold which might develop on a suddenly closed throttle, "stall" the compressor, slow it down, and therefore create throttle lag when next the driver wants power?

All we know is that in the Mosport race one of these butterflies jammed in the closed position on the 17th lap. The car, which was leading, became "undriveable" in Donohue's words he lost several seconds on the next lap and came in to the pits to have the mechanics discover what was wrong and free it up. It absolutely cost him the race.

Mark and the other people behind the Porsche are not lighthearted types. They take the whole thing very seriously, they have worked exhaustively to solve their problems, and they are determined to keep their secrets, their "unfair advantage," as long as humanly possible. They discourage prying eyes with tarpaulins, flatly refuse to answer certain questions, and are positively rude to photographers. Fair enough; we survived that sort of attitude in the gruesomely tense Trans-Am 1970, when six Detroit firms were locked in bitter bloody battle. The CanAm used to be relaxed and friendly, but if we need glint-eyed enmity to create a good competitive series, fine.

"Besides," admits Mark with a sudden grin, "it's more fun this way!" He really enjoys turning aside sudden rapier questions about his latest toys. He has come right out and told us the name of the secret frame material, though. He's granted us that little:
It's "Unobtainium."

back to index

first 2 rounds 1972

can am index

Author: ArchitectPage

a development the 930