Die Entwicklungsalternativen fr Grand Prix Motorradmotoren

Design Alternatives for MotoGP Engines

Gordon P. Blair







1. Abstract
From 2002 the FIM (Federation Internationale Motocycliste) sanctioned the use of
990 cm
3
, naturally-aspirated, spark-ignition, multi-cylinder engines operating on petrol
for racing in the prestige and fastest Grand Prix category, i.e., MotoGP. This 990
cm
3
classification replaced what had been a 500 cm
3
class since the dawn of
motorcycle racing and where in recent decades the two-stroke engine of some 3- or
4-cylinders and producing some 200 hp had eliminated all competition from four-
stroke engines which would struggle to produce perhaps 150 hp from a 500 cm
3

engine. Consequently, the designers of racing engines had to learn new design
techniques and much, if not all, of the theory of tuning learned the hard way on 2-
strokes had to be rapidly unlearned. Many companies, such as Honda, Yamaha, or
Ducati, to name but three, already competed in World Superbikes where all engines
were historically 1000 cm
3
four-stroke twin-cylinder, or 750 cm
3
four-cylinder, or 900
cm
3
three-cylinder units. Within such organisations much experience already existed
on designing and developing four-stroke multi-cylinder racing engines. However, in
MotoGP, the choice of engine layout is much wider and the selection of the final
design layout for MotoGP racing often devolved to personal choice, copying of that
which is deemed to be currently successful, or to some long-held but as yet unproven
theory. This paper attempts to set out some logic for any mechanical design to be
selected for MotoGP racing and, in particular, shows the advantages and
disadvantages of designs with two to six cylinders all nominally producing the same
power output of some 230 hp at 15000 rpm.
2. The Geometric Layout Design Parameters
In a statement of the obvious, the engine has to fit into a motorcycle and be mated to
a (probably six-speed mechanically foot-operated) gearbox. This means that width,
centre-of-gravity locations, and weight of the engine are all significant design
parameters. Almost without discussion, the engine will be liquid-cooled and also
without excessive disagreement will probably have four-valves per cylinder in a
pentroof-type combustion chamber. The valve operation will be double-overhead
camshafts and the valve follower actuation may be desmodromic, finger followers,
but more commonly will be flat direct-acting bucket tappets. The valve control may be
desmodromic, but for bucket tappets and finger followers at this juncture in racing
engine design this will be done with steel or titanium coil springs and the valves will
also be of steel or titanium. The use of pneumatic or gas springs, as used in F1 car
racing, may come along eventually to permit engine speeds in excess of 17000 rpm,
but that stage would not appear yet to have been reached as 230 hp can be

achieved at 15000 rpm in a 990 cm
3
engine and 230 hp is felt to be enough to win a
MotoGP race at this time provided that the shape of the back-up torque and power
curve gives competitive acceleration out of slow corners.
Consequently, if it is assumed that 15000 rpm is a comfortable maximum engine
speed for coil-spring controlled valve actuation and that an engine can run at a mean
piston speed of 25 m/s without piston lubrication failure, and that a four-valve
pentroof design can breathe air and fuel adequately, and burn it successfully, to
produce some 14 bar bmep on gasoline at a 14:1 compression ratio, the following
simple deduction can be made as to the basic engine dimensions and performance
[1]. The cylinder stroke is 50 mm and the engine will produce 232 hp at 15000 rpm.

Figure. 1 Cylinder bore sizes for two to six cylinder MotoGP engines.
The bore sizes shown above correspond to bore:stroke ratios of 2.245, 1.836, 1.59,
1.42 and 1.3 for two- to six-cylinder units, respectively. A bore:stroke ratio of 2.245 is
typical of a F1 car engine, although there bore sizes are a fraction less than 100 mm.
The translation into engine bulk is best illustrated in Figure. 2 where the above
dimensions are transposed into a plan scaled view showing in-line and vee formats.

Figure. 2 The bulk of MotoGP engines with two to six cylinders.

Although these simple pictures do not tell the whole truth about bulk and weight,
some interesting conclusions can be drawn. The twin-cylinder engine, be it in-line or
vee, is compact and simple. The in-line three-cylinder is wider, but not by much, and
is certainly narrower than the in-line four (a very popular Japanese format). The vee-
four is narrower than the in-line three-cylinder engine and also narrower than the vee
five- or vee six-cylinder units. Interestingly, the vee-six is potentially narrower than
the vee-five, which latter is currently the engine of fashion in MotoGP. From an
exhaust tuning standpoint, collections of exhaust runners from engines firing evenly
give the best tuning when there are three or more cylinder exhausts being so
collected [1]. This leaves the in-line four cylinder as the number one option, followed
by the in-line three and vee-six engines. The vee-four is also in good shape for
exhaust tuning if the crankpins are stepped and the cylinders fire at 180 degree
intervals. The vee-five is the most awkward to exhaust tune as it is imperative that
the front three cylinders fire at 240 degree intervals and the rear two cylinders fire at
360 degree intervals. This imposes design constraints on the crankshaft and the
ensuing torsional oscillations must be carefully designed out by the almost-inevitable
addition of a torsional damper at the non-drive crankshaft end.
3. The Gas Flow and Combustion Design Parameters
The next design step is to determine the physical dimensions of the valves and their
location within the cylinder. A four-valve pentroof arrangement will be employed for
all cylinders of each engine. The software package used to do this is called 4stHEAD
[2]. The outcome is a determination of the valve sizes which can logically fit within a
chamber; which chamber volume provides the required compression ratio; the
surface area of this chamber is determined for future engine simulation requirements
[3]; the valves can be moved on-screen dynamically with piston motion and the valve
cut-out depths on the piston crown determined. For all of the engines types analysed
here a valve seat plan width of 1 mm, and a piston to head minimum clearance of
0.75 mm, is assumed. The maximum lift of the valves is set at 37% of the inner seat
diameter. The upshot of this design process is shown for the two-cylinder engine
below. It is drawn to scale and the actual dimensions and lift of each valve are shown
on the sketch, Figure. 3 below, for the two-cylinder engine.


Figure. 3: The two-cylinder engine with the piston at bdc.
The maximum depth of the cut-out in the piston for the outer seat edges of the
exhaust and the intake valves are shown on the sketch but the valve face extension
beyond those seats must be added to the dimensions shown above so as to

determine the full depth of the relevant cut-outs. The angle of inclination of each
valve on the pentroof, and the included valve angle, is also illustrated on the sketch.
The following Figures, Figures. 4-7, show the equivalent cylinder layouts for engines
with up to six cylinders. When examining Figures. 3-7 remember to take into account
the optical illusion of the printed Figure; it is the stroke which is common at 50 mm
and not the cylinder bore!


Figure. 4: The three-cylinder engine. Figure. 5: The four-cylinder engine.


Figure. 6: The five-cylinder engine. Figure. 7: The six-cylinder engine.

The surface area of the head, the squish band and the piston crown are computed by
a fine mesh procedure and calculated as a surface area ratio, K
sr
, and a surface to
volume ratio, K
sv
[2]. The graphical output for each engine cylinder head is shown
here and set very approximately and relatively to size on the printed page.


Figure. 8 The cylinder head of the two-cylinder engine.

The surface to volume ratios, K
sv
, for each of the engines of two to six cylinders is
0.545, 0.554, 0.558, 0.565, and 0.567, respectively, giving an overall shift of 4%

reflecting the increasing included angle of the pentroof with engines having larger
cylinder numbers. On the other hand, it is quite clear from these pictures, Figures. 8-
12, that combustion in the six-cylinder engine will be much more rapid and efficient
by comparison with that in the two-cylinder engine and will most probably be reflected
in a shift of ignition advance by as much as 15 deg, and a reduction of burn period by
some 10 deg, over that same cylinder range [1]. The in-cylinder tumble
characteristics will also be enhanced in engines with decreasing bore:stroke ratios.


Figure. 9: The three-cylinder head profile. Figure. 10: The four-cylinder head
profile.


Figure. 11: The five-cylinder head profile. Figure. 12: The six-cylinder head profile.

Good combustion is not aided by excessive valve cut-outs in the piston crown and
from Figures. 3-7 it can be seen that the two-cylinder engine has cut-outs of some 5
and 7 mm deep. The depths decrease with the number of engine cylinders with the
result that in the six-cylinder engine they are minimal at only 1.8 and 2.5 mm.
The general and approximate conclusion from the above comments on combustion is
that under conditions of equal charging efficiency there could be a 10% increase of
indicated torque over the range of cylinders from the two- to the six-cylinder engine.
4. The Breathing Design Parameters
The breathing of air into the cylinder, or of exhaust gas out of it, is very much a
function of the number of valves, the valve and seat dimensions, the valve lifts and
durations, the cylinder dimensions, and the compression ratio. While the valve sizes
and lifts are also numerically shown in Figures. 3-7, they are graphed separately
below in Figures. 13-14. It can be seen that the valve size, graphed as outer seat
diameter EDOS, decreases with number of cylinders and the maximum valve lift,
calculated at 37% of the inner seat diameter, EDIS, has the same trend. On Figure.

14, there is also a graph showing exhaust valve lift as a function of 40% of the inner
seat diameter and its inclusion will be made clear later in this section of the paper.
All of the valves in all of the cylinders of all of the engines are, as an initial design
gambit, assumed to open and close according to the following schedule at a set cam
tappet clearance of 0.3 mm; the exhaust valves open, EVO, at 80 deg bbdc and
close, EVC, at 50 deg atdc, for a lift duration of 310 deg; the intake valves open, IVO,
at 60 deg btdc and close, IVC, at 90 deg atdc for a lift duration of 330 deg.


Figure. 13: Valve diameters. Figure.14: Valve lifts.

The relative intake to exhaust valve size ratio, K
ie
, is set at between 1.17 and 1.2 for
all engines [1, 2]. The valve lift profiles are set to reasonable equality for each engine
and the actual profiles for each of the exhaust valves are shown as graphed by the
preparation software [2] in Figure. 15 below; that for the intake valves is almost
identical but is 10 deg longer in camshaft angle terms. The valve ramp lifts are each
0.3 mm. It will be observed, by those experienced in the art and science of valve lift
design, that the profile of the acceleration diagram selected is quite mild in terms of
the ensuing profile of the jerk diagram it creates, particularly around the nose of the
cam which will provide it to the valve.


Figure. 15: Valve lift characteristics for all exhaust valves.


To estimate the breathing capabilities of these engines, before inevitably turning to
an accurate engine simulation [3] for a definitive answer to the question, the time-
areas per unit swept volume have been shown to be a valid and realistic empirical
means of conducting such a study [1].
For the exhaust breathing process, the basic elements of time-area for the exhaust
blowdown period, the exhaust pumping period, and the exhaust overlap period are
shown in Figure. 16 below.


Figure. 16: The exhaust time-areas for blowdown, pumping, and overlap
valve periods.

For the intake breathing process, the basic elements of time-area for the intake
ramming period, the intake pumping period, and the intake overlap period are
shown in Figure. 17 below.


Figure. 17: The intake time-areas for ramming, pumping, and overlap valve periods.

While the sketches, Figures. 16-17, show the relevant total valve areas in mm
2
.deg
units, they are converted to time-area values per unit swept volume, TASV, in s/m
units by dividing by SV (the swept volume of any one cylinder) and the relevant ratio
of crankshaft angle to absolute time with respect to the crankshaft rotation rate (rpm).
It is shown [1] that TASV values are directly related empirically to the delivery ratio
proportion of gas which will flow through the valves, to or from the cylinder, during the
periods indicated. In short, it is a much more sophisticated means of empirical
comparison of engine breathing characteristics than the traditional mean gas
velocity or mean Mach index or Lovell factor values seen elsewhere in the
literature because each time-area element for each segment of the valve duration for
each valve must have a numerical value within a specific numerical range. If this is
not the case, the required airflow/gasflow breathing will be stunted within that
segment and hence affect the total breathing behaviour through that valve of the
engine. Hence, all segmental numerical criteria must be satisfied or the desired
torque (as bmep) will be unattainable, or so runs the empirical argument.

4.1 The Intake Valve Breathing Design Parameters
The time-area characteristics at the nominal peak power engine speed of 15000 rpm
for all engines for the intake valves are shown in Figure. 18; the plot units are s/m x
10
4
. It is clear from discussion elsewhere [1] that these intake time-areas have a
strong correlation with airflow delivery ratio and hence bmep at the engine speed
selected for peak power.

Figure. 18: The intake valve time-areas for ramming , pumping and overlap.
The design criterion for all of the 990 cm
3
engines is for a bmep of 14 bar and
examination of the numerical criteria published [1] shows that the relevant values of
intake time-areas required for this torque output at peak power are 19 (+/- 1.0), 128
(+/- 5.0), and 39 (+/- 2.0) s/m for the intake ramming, pumping, and overlap valve
periods, respectively (again quoting in the more convenient units x 10
4
). Examination
of the graphs in Figure. 18 shows that the two-, three-, and four-cylinder engines
meet or exceed these time-area criteria but that the five- and six-cylinder engines are
more marginally acceptable. The engines with less than five cylinders could be easily
re-optimised, with slightly lesser valve lift, or period duration to precisely meet the
criteria, and vice-versa for the engines with five and six cylinders. However, it should
be remembered that this is empiricism and such decisions of fine detail are best left
to optimisation within an accurate engine simulation program [3]. Nevertheless, and
in a reversal of the conclusions of the combustion discussion, the basic trend for
superior intake valve breathing is for the engines with fewer cylinders, or, perhaps to
re-state it more technically, for engines with the greater bore:stroke ratios.

4.1 The Exhaust Valve Breathing Design Parameters

Figure. 19 (a) & (b): The exhaust valve time-areas for blowdown and pumping.

The time-area characteristics at the nominal peak power engine speed of 15000 rpm
for all engines for the exhaust valves are shown in Figures. 19 (a-c); the plot units are
s/m x 10
4
. The design criterion for all of the 990 cm
3
engines is for a bmep of 14 bar
and examination of the numerical criteria published [1] shows that the relevant values
of exhaust time-areas required for such a torque output at peak power are 15.5
(+/- 1.5), 100 (+/- 5.0), and 31 (+/- 3.0) s/m for the exhaust blowdown, pumping,
and overlap valve periods, respectively (again quoting in the more convenient
units x 10
4
).

Figure. 19 (c): The exhaust valve time-areas for the overlap valve period.

Examination of the graphs in Figure. 19 (a-c) shows that the two-, three-, and four-
cylinder engines exhibit the same general trend as with the intake valve. When the
valve lift is at the standard value of 37% of the inner seat diameter, EDIS, and the
valve lift profile is that shown in Figure. 15, then only the engines with the greater
bore:stroke ratios approach the required empirical flow criteria. In particular, the six-
cylinder engine is deficient in valve flow area. To redress this, new time-area criteria
are calculated when the exhaust valve is lifted higher, to the same lift profile and
duration as in Figure. 15, but now to a maximum value of 40% of the valve inner seat
diameter. The outcome is also graphed in Figure. 19 (a-c) and it can be seen that
this improves the situation considerably towards the acceptable for all engines, with
the exception of the overlap time-area in Figure. 19 (c) for the six-cylinder engine.


Figure. 20: More aggressive exhaust valve lift profile used for the six-cylinder engine.

To show that all engines can be eventually raised to an acceptable exhaust time-area
level prior to commencing engine simulation studies, a more aggressive valve lift
profile is employed for the exhaust valves in the six-cylinder engine. This while
retaining the same 37% of EDIS as a maximum lift value and also the same opening
and closing valve timings of 80 EVO/50 EVC already specified above. The outcome
of the analysis for the six-cylinder engine only is shown plotted in Figure. 19 (a-c)
above and marked as special. The more aggressive valve lift profile to achieve this
effect is provided through the 4stHEAD software [2] and is shown in Figure. 20. The
valve lift profile in Figure. 20 raises the valve lift-duration envelope ratio, K
id
, from
0.557 to 0.595, i.e., a gain of 7% which translates into the roughly equivalent time-
area changes for the point marked as special in Figure. 19 (a-c) [2].
The intake and exhaust valve lift profiles can be directly exported to engine
simulation software for the immediate commencement of simulation studies of any of
the engines.

4.2 The Intake and Exhaust Duct Design Parameters
Effective intake and exhaust tuning is only possible if the ducts have the correct
lengths and diameters [1]. The first essential criterion is to set the area of the relevant
duct some few diameters away from the valves, usually at the manifold face at the
cylinder head, to such a value as will provide the correct amplitude of pressure wave
to be propagated away from the cylinder through that duct towards the atmosphere.
This pressure wave, of expansion in an intake duct or compression in an exhaust
duct, will subsequently be reflected to provide the optimum strength of intake
ramming or exhaust suction as to constitute effective engine tuning [1].


Figure. 21: Intake and exhaust manifold areas (as equivalent diameters).

For all of the engines studied here, the intake manifold-valves area ratio, C
im
, and
exhaust manifold-valves area ratio, C
em
, are set to 0.95 and 1.3, respectively, using
the advice on such numerical values published elsewhere [1]. It is calculated by the
4stHEAD software as the time-area analysis described above proceeds [2]. The
outcome for all engines is shown above in Figure. 21 where the area required is
expressed in terms of an equivalent diameter. The general trend, as could be
expected, is for reducing engine duct sizes with increasing number of cylinders.
Advice on exhaust and intake duct lengths to tune the engine effectively at 15000
rpm, with collector exhaust systems where appropriate, for all of the engines with two
to six cylinders is also available but will not form part of this discussion [1].


5. The Mechanical Design Parameters for the Valves
The design debate above makes the major assumption that the valves designed for
all cylinders can be operated by overhead camshafts actuating tappets and valves for
the lifts and durations already specified and using coil springs to control the dynamic
movement of the valves. The Figure. 22 below shows a snapshot of the cam design
for manufacture from the 4stHEAD software for the exhaust valve of the six-cylinder
engine at its higher 8.0 mm lift specification. It is shown actuating a flat direct-acting
bucket tappet. For this particular design, using the mass of the valve, its tappet, its
spring retainers, and the coil springs stiffnesses that will control it, the software
predicts that at 15000 rpm, and using a SAE 10W-40 oil at 90 deg C, the minimum oil
film thickness will be 0.93 micron and the maximum Hertz stress at the cam and
follower interface will be some 780 MPa. In short, it will operate successfully and
such a cam could be ground directly from the relevant 4stHEAD software output files.
However, such success may not be possible for bucket tappet actuation and coil
spring control of all of the valves in all of the engines discussed.


Figure. 22: The six-cylinder exhaust cam and its bucket tappet.

From the software, for all engines and for all valves, it is possible to calculate the
valve mass and the mass of the bucket tappet and the spring retainer. Assuming that
the bucket tappet and the spring retainers are in steel, but that the valve could be
either steel or titanium, the information so gathered is plotted in Figure. 23 below.
Needless to add, the data plotted in Figure. 23 is not definitive in that the numerical
band will vary from designer to designer and their choice of valve stem thickness,
tappet wall thickness, etc., etc., but it is doubtful if that variation is more than +/-10%
of that quoted in Figure. 23. With that caveat in mind, it is clear from Figure. 23 that
the variation in valve and bucket tappet and spring retainer mass over the range of
engines from two to six cylinder is very considerable, as is the range of their valve
lifts, to the point where it is highly improbable that the valves of the two-cylinder or
three-cylinder engine could be operated in this fashion. It should not be forgotten that
it is the combined mass of the valve, tappet and spring retainer which reciprocates
and must be controlled by the coil spring. From the two- to the six-cylinder engine,
the higher is the reciprocating mass and the valve lift, then the thicker in wire section
and the larger in outside diameter becomes the wire of the springs to control that
movement. The upshot is that some of the coil spring designs will have excessive
dynamic movement at the 15000 rpm operating point and high surface stresses
leading to failure.

The experimental evidence from various racing engines, and corroborated by
theoretical advice from the 4stHEAD software, is that a total reciprocating mass of
some 50-60 g can be successfully controlled at 15000 to 16000 rpm by coil springs.


Figure. 23: The mass of valves, bucket tappets and spring retainers for all engines.

By this criterion, and taking data from Figure. 23 above, it is probable that the five-
and six-cylinder engines would dynamically survive operation at 15000 rpm using
steel valves and bucket tappets. The four-cylinder engine valvetrain might survive the
same engine speed by using a steel exhaust valve but would need to use titanium for
the intake valve. Close to the 50g mass borderline, the three-cylinder engine would
probably need to use titanium for the intake and exhaust valves; however, successful
bucket tappet actuation now becomes less probable due to its relatively high mass
but the designer could opt for (lighter) finger followers for the valve actuation. For the
two-cylinder engine, it would appear from Figure. 23 that coil spring control of
titanium exhaust and intake valves, which already alone weigh some 40 and 50 g,
respectively, by whatever valve follower method, is really not a design option.
Consequently, the use of desmodromic valve actuation or valve control by a
pneumatic valve spring is the only way forward for the successful operation of the
valvetrain of that particular power unit at 15000 rpm. However, desmodromic valve
actuation does come burdened with high friction losses at very high engine speeds
and high valve lifts, so it may not be so attractive an option as it might appear to be at
first sight.
6. Summary and Conclusions
From the discussion above it is clear that the best combustion characteristics are to
be found as the number of cylinders increases and, as this analysis evolves it, to the
engines with decreasing bore:stroke ratio. Breathing, which is the ability to induce
fuel and air to high levels of delivery ratio from which torque and power are
maximised, is more easily accomplished in engines with higher bore:stroke ratios,
i.e., as this analysis evolves it, to the engines with fewer cylinders. As usual, the
design engineer is faced with a compromise which is equally heavily influenced by
engine weight and bulk and also power spread over a wide speed range. However,
even here a further compromise appears as the engines with fewer cylinders will, in
general, have the least bulk and weight. On the other hand, the engines with better
combustion and a lower bore:stroke ratio will normally breathe, and certainly burn,
their charge better over a wide speed range. However, and this is a most important
point, it is shown above that one can change the valve lift and valve lift profile,
without even changing valve lift duration, to raise the time-area or breathing
characteristics of any engine to another, higher, level. There is no equivalent

procedure to rectify the inherently inferior combustion characteristics of engines with
high bore:stroke ratios. Finally, as if the above was not enough to worry about, there
is always a future racing season for the designer to bear in mind as peak power
demands inevitably rise and the engine speed to attain them does likewise.
It is not the purpose of this paper to declare any particular engine layout as the
ultimate design to win MotoGP races. Indeed, the design brief is to examine the
potential for two- to six-cylinder engines to produce the same power output of 230 hp
at the same engine speed of 15000 rpm and it is hoped that the reader has observed
that the design path is not one suited to personal choice, copying of that which is
deemed to be currently successful, or to some long-held but as yet unproven theory,
expressed above as methodologies seen to be much in vogue. If nothing else, this
paper illustrates that todays designers have available software tools to aid the design
debate in a fashion that was not possible when four-stroke engines last raced in
MotoGP some thirty or more years ago. As the great man once said, ..until you can
put numbers on your problem, you are not yet at the beginnings of a science...
However, history is also a great teacher and the history of motorcycle racing provides
the circular argument that ultimately shapes racing engines; more power always
wins; more power is produced at higher engine speeds; higher engine speeds
come from engines with more but smaller cylinders; lightweight valve control
mechanisms permit the highest engine speeds. Basically, with numbers attached,
that is what this paper says.
References
[1] G. P. Blair, Design and Simulation of Four-Stroke Engines, Society of
Automotive Engineers, Reference R-186, 813 pages, 1998.
[2] 4stHEAD software suite, Cylinder Head Component Design for Four-Stroke
Engines, Prof. Blair and Associates, www.profblairandassociates.com.
[3] VIRTUAL 4-STROKE Engine Simulation and Optimisation Software, Optimum
Power Technology, www.optimum-power.com.


Autor/ Author:
Prof. Gordon P. Blair
Senior Associate
Prof. Blair and Associates
Northern Ireland
&
Professor Emeritus
The Queens University of Belfast
Northern Ireland