10 years of naturally aspirated BMW enginesThis article is an extract of "Ten of the best", an article written by Ian Bamsey and published in issue 057 of Race Engine Technology. If you wish to read more, you can buy the issue at www.highpowermedia.com and put 'f1technical' as voucher code to benefit a 10% reduction on your purchase price.
I first met Dr Mario Theissen at the launch of the BMW-Williams Formula One partnership in January 2000; I interviewed him at least once a year through the ten years that BMW raced in Formula One. Dr Theissen came from an engine background, so it was always fascinating to discuss development of the BMW 3.0 litre V10s, and later the 2.4 litre V8s, with him. Clearly there was much that he had to keep confidential at the time but a paper (see Reference) published in 2010 provides further insight into a great technical adventure. This article draws together my findings in the field, information from that paper and additional questions that BMW Motorsport has kindly answered.
The baseline: E41/4
The BMW-Williams Formula One alliance began with testing in 1999 using a contemporary Williams chassis converted from its usual Supertec engine. The new BMW E41 was of similar configuration, size and weight to that 72º V10, which was a derivative of the Renault engine designed in 1997. In fact, design of the E41 had also begun in 1997, when the BMW board gave the green light to a return to Formula One, and it was the third iteration that started racing in 2000. By 2000, the top rival engines were smaller and lighter with a lower centre of gravity, to the benefit of the overall car package. “We won’t have the most powerful or the lightest engine this year,” remarked Dr Theissen at the January 2000 launch of the Williams-BMW. “This season we will not have engine power at the level of Ferrari and McLaren-Mercedes, or the same engine weight. They can take risks in establishing a new benchmark. Our aim for this year is to have an engine which is stable, in order to establish ourselves in Formula One.”
In mid-season Dr Theissen added, “If you are in a situation that we were in, as in to build an all-new facility, get the team together and design the engine, then the teambuilding process is the most difficult one and the one that takes the most time. It will take us all of this first race season to really get everybody to know what they have to do and when, and what not to do. This is a process you cannot speed up.
“I have learned that Formula One is not so much about driving quicker, or speed – which is what I had expected it to be – but that it is predominantly about precision and perfection. The whole field is very competitive, very close, and precision and perfection is what decides if you finish the race and whether you are in the points or not. This is something closely linked to the teambuilding process. It really takes time to get this understanding, to get the right attitude and have the team really act as one, with a clear target.”
Earlier, at the launch, Williams’ chief aerodynamicist Geoff Willis had observed, “You can’t design the engine in isolation from the chassis. Straightforward problems like vibration, weight distribution, stiffness, fuel consumption, heat rejection – the trade-off between what temperature you run the engine versus the size of the radiators – these are all things that engine manufacturers need to be aware of. I think BMW is going through a big learning process, of absorbing all this information.”
Thus the E41/4 was understandably a cautious engine design, although mid-season Dr Theissen said, “You cannot be sixth on the grid if you do not take risks! But we have not been too adventurous, more on the conservative side. Starting with an engine that is too much on the edge wouldn’t have helped us finish a race. This is especially important when you have a team which is still growing.”
In 2000, Ilmor-Mercedes was exploiting the potential of aluminium beryllium, which the FIA had effectively banned as of the end of the season. Dr Theissen noted, “We didn’t go for aluminium beryllium for this year’s engine as it doesn’t make sense to develop something that you have to remove after only one season.” The author asked him about the use of MMC. “In my view it will be used increasingly in the future.”
BMW can now confirm that it didn’t ever use MMC pistons through these naturally aspirated Formula One engines, nor did it ever run DLC-coated piston skirts. Back in 2000 Dr Theissen didn’t give away much about the E41/4 internals, other than to confirm the use of conventional Mahle pistons, Goetze rings and Pankl con rods. We are now informed that the rods were titanium throughout these engines, all of which had “liners integrated into the casting, the cylinder bores using BMW’s own nickel-dispersion coating”.
The BMW paper tells us that the E41/4 had a bore of 94.0 mm, and with four titanium valves per cylinder it had 40.5 mm intake and 31.2 mm exhaust valves. BMW used compound valve angles throughout but the valve angles remain undisclosed. The timing drive was always at the front of the engine. Valve operation used finger cam followers, and intake valve lift was in the region of 16 mm. Throughout, the compression ratio was in the region of 14.5-15.0:1. The E41/4 was based on an aluminium block with 107 mm bore spacing and 20.5 mm bank offset. The engine was 620 mm long, 524 mm wide and 395 mm high. A notable feature was the lower crankcase, which formed a bedplate that stiffened the structure and housed various ancillaries [Fig. 1]. The E41/4 was equipped with variable length intake trumpets, and this technology was used until the end of the V10 era, after which it was banned. The engine ran to a maximum speed of 17,500 rpm, and following development through the season eventually made 810 bhp with maximum torque of 350 Nm. As probably the most powerful engine of 2000, the Ilmor- Mercedes made a claimed 820 bhp at 17,500 rpm with a maximum speed of 18,000 rpm.
Although engine reliability was not yet a strong suit, in terms of speed the Williams-BMW was the best of the rest behind Ferrari and McLaren-Mercedes in 2000. Reliability improved through the season, as BMW Motorsport improved both “engineering and processes”, as the paper (see Reference) put it. Nevertheless, the plan was to have a brand-new, state-of-the-art engine for 2001.
From E41/4 to P80
As Dr Theissen put it to the author at this time, “With last year’s E41/4 engine we went from zero to 90% of what is possible; with this year’s P80 we have gone from 90% to 98%. You cannot go from zero to 98% in one step – you don’t know how close you are to the edge. The worst thing you can do is cross that line, as you then end up with nothing. To us it was the perfect first-year experience, to learn how close we were to the edge, where we needed to improve, where we had to be careful. Only on this basis were we able to go one step further and make the big step to this year’s engine.”
The author asked him: next year 99%? “Next year the 100% line will have moved on. We have to do something just to stay at 98%!” What is the 2% that is holding you back? “It is small details in each and every component. But that is the art of making such an engine. I would say that nobody will ever be at 100%. As soon as he is, suddenly the limit is pushed beyond.” Thus it was that BMW moved to a nocompromise solution for 2001. In the process it evaluated V8 and V12 options (as it had done before 1997) and, according to the paper, “business plans for high-cost parallel developments were prepared”. However, the FIA then mandated V10 engines from 2001. The next decision was the choice of bank angle. While 72º provides even firing intervals and keeps the unit slim to aerodynamic advantage, 90º provides less disadvantageous vibration characteristics and lowers the centre-ofgravity height. Ferrari pioneered the 90º V10 configuration in 2000; all others came to follow suit, except for Renault, which went to an even wider angle, but without success.
Seeking higher revolutions, the bore of the new 90º V10 was up from 94 mm to 95 mm, with corresponding reduction of stroke to maintain displacement at 2998 cc. Thus stroke went down from 43.2 mm to 42.3 mm, the corresponding stroke:bore ratios being 0.46:1 and 0.45:1. The Ilmor-Mercedes V10 that set the pace in 2000 is known also to have had a bore of less than 96 mm, while the rival Ferrari – which likewise ran to 18,000 rpm and produced a claimed 817 bhp at 17,500 rpm – had a bore of 96 mm. With that it had a stroke of 41.4 mm (for 2997 cc) and as such it experienced mean piston speed (MPS) of 24.74 m/s at 18,000 rpm (with maximum piston acceleration, MPA, of 8890-g). The E41/4 had originally been targeted at maximum speed of 17,500 rpm, its 43.2 mm stroke then implying an MPS of 25.1 m/s . Had it run to 18,000 rpm it w ould have experienced an MPS of 25.8 m/s. By contrast, the new P80 could run to its target speed of 18,000 rpm with an MPS of 25.35 m/s (the rod length has not been disclosed so we cannot calculate the MP A). The paper adds, “Cylinder head integrated barrel throttles, revised ports, modified rocker arms and valve lift curves allowed maximum speed to be increased to 18,000 rpm.”
The E41/4 had used butterfly throttles, whereas from the P80 onwards a barrel-type throttle was always used. The P80 did retain the proven pneumatic valve return system of the E41/4. With the increase of bore, the sizes of the titanium valves were correspondingly increased, from 40.5 mm intake, 31.2 mm exhaust, to 41 mm intake, 32.65 mm exhaust. Thus intake valve area as a proportion of bore area rose marginally, from 37.1% to 37.25%. By comparison, having 40.4 mm intake valves, the 2000 Ferrari had an intake valve area of 35.4% of bore area.
Although the bore had been increased, the bore spacing was reduced, from 107 mm to 103.5 mm, hence there was now 8.5 mm rather than 13 mm between adjacent cylinder walls. The bank offset was also reduced, from 20.5 mm to 19 mm. The engine was now 32 mm wider but 22 mm shorter and 54.5 mm lower. The reduction in height was not solely due to the increase of bank angle but also to a reduction of installed crankshaft height, from 76 to 65 mm. The paper says this was “with the aid of now separately bolted heavy metal balancing weights and a greatly reduced clearance between the con rod path and the crankcase”.
Interestingly, the ‘bedplate’ lower crankcase/ ancillary housing had now been abandoned. The paper reports, “It was now clear that the extreme structural stiffness provided by the E41/4 was not actually necessary” and remarks that there was now a “deep skirt design with separate but geometrically and structurally integrated bearing caps. The auxiliaries were externally mounted, giving a considerably increased flexibility with a decreased error rate by engine build.”
The P80 also had a new design of cylinder head, and instead of the head covers carrying the chassis mounts, those were now connected to the main body of each head, which then had a lightweight cover. The upshot was that the centre-of-gravity height had fallen from 167 mm to 145 mm, while the weight was down from 117 kg to 105 kg. Dr Theissen remarked that to have gone even smaller and lighter would have been “high risk for small reward”.
For its 2001 car, according to Willis, Williams “increased the cooling capability a bit, on the expectation of having more power”, and it wasn’t disappointed. There was no specific qualifying engine, although sometimes new developments were evaluated first in qualifying. Nevertheless, the P80 quickly gained a reputation as the most powerful engine of 2001, and by mid-season there had been two victories and only one engine failure (a rod bearing failure). By this stage the P80 had been developed to produce 850 bhp, and it reached 880 bhp by the end of the season, still not exceeding 18,000 rpm. Enjoying a power advantage thought to have been in the region of 30 bhp, Williams-BMW went on to win two more Grands Prix in 2001.
“It is very much more compact,” remarked Williams’ chief designer Gavin Fisher of the 2002 BMW P82. “Substantial progress has been made in terms of volume and weight.”
BMW hadn’t been satisfied by attaining the most powerful engine in Formula One, it also strove to have the most compact and the lightest, to the benefit of the overall car package.
Thus there was another all-new engine for 2002, and titanium was even considered for its structure but was rejected in favour of an improved sand-casting process at BMW’s Formula One foundry.
The full article is published in RET issue 57 and can be purchased with a 10% reduction if you pass 'f1technical' as a voucher code on your purchase at highpowermedia.com.
This article was partially based on a paper released by BMW internals.