Aerodynamics of open cockpit, FF single track vehicles V.2

Definition FF

A single track vehicle with a seat base less than 20 inches above ground level at ride height, with a seat back capable of fully supporting the rider. The front suspension should not be steered.

Application

Bodied, open cockpit FFs.

Sources and definitions

Observations of the aerodynamic performance of the Voyager designs are the source of this article. Tthere are photos of all the vehicles described on the bikeweb site with links to specific photos. There was considerable evolution of general vehicle aerodynamics during the Voyager project and it's important to understand these recent developments.

Streamlining

is often taken to be a case of getting as close as possible to the smooth cigar shape of an aircraft - minus the wings. But that's a shape designed to fly in clean air. A vehicle is a shape designed not to fly in disturbed, turbulent, air. Turbulence is what you get at ground level if there's even a breeze. It's a given on a any road with traffic.

It turns out that clean air, lamminar flow, shapes are less efficient and stable in turbulent air than shapes that separate the airflow from the surface.

So although some parts of the shape needs to be smooth and curved, to direct the airflow, quite a lot of the bodywork is about filling space, not actually touching the airflow.

Separation

is the key to achieving this. It's the business of ensuring that airflow separates where it's supposed to. Once airflow separates from a shape, a layer of turbulent air forms between the surface and the airflow. This insulates the surface from aerodynamic effects (pressure changes) caused by airflow direction changes or turbulence.

Here are some general rules about separation;-

Airflow will separate randomly from any surface with an angle of attack below plus 2 degrees, with propagation of turbulent separation from any surface protrusion above .05mm

It will not re-attach untill the angle of attack gets to at least plus 5 degrees.

Separation can be enforced on a smooth surface by any section change greater than 5 degrees, or any ridge or ledge greater than 2mm. String or wire, under tape is a common experimental method.

Re-circulating bubbles

are what happens when there's a a sheltered space with airflow on one side of it. The air in the space rotates, powered by the airflow drag. These bubbles can be quite violent, typically providing a cold air supply to the necks of sportscar drivers. To prevent this effect as much open space as possible should be filled with bodywork - like the head fairings used on Voyagers. Where there must be an open space, like the cockpit, it should be 'filled' with ram air.

It can be seen from the above that streamlined 'cigars' are unneccesary and may be less efficient than shapes with carefully controlled separation and filling.

PTWs and the Sidewind issue

It is widely believed that bodied PTW's are susceptible to sidewind disturbance. This is sometimes cited as the reason for their ban in motorcycle racing. Sidewinds may disturb an FF if it has the 'wrong' characteristics but it is actually quite easy to achieve excellent cross-wind stability.

Sidewinds act on the whole vehicle, tending to roll it, on it's contact patches, out of the wind. This effect is modified by the shape of the vehicle and normal PTW steering geometry.

The shape of the vehicle has two effects. If the shape has more side area behind the CG than in front of it, it will "weathercock", turn into wind, and this is a desirable effect. The other effect is that if the side area is high, especially at the front, it will increase the tendancy of the vehicle to roll out of the wind. This is an undesirable effect.

This is the basis of the 'notched nose' shaped used in the Voyager and Cmax designs.
(http://www.bikeweb.com/node/2023).
In the course of development it also became clear that flow separation devices damp and reduce these 'shape' related effects.

PTW steering geometry modifes the effect because the steering rotates away from a sidewind, the trail element in the geometry providing the lever. This is the same effect as normal counter-steer. The vehicle will then roll into the wind with gyro and tyre-coning effects resisting the wind-induced counter steer to provide self-balance at some resultant angle to the wind.

This effect can be modified by bodyshape - countered by a high frontal side area, or assisted by a suitable tail area.

It can also be modified by PTW geometry. Dynamic trail provides directional stability, so large trail figures will generate a higher resistance to deflection of the steering by a given sidewind - and reduce the steering angle of a given deflection. High trail figures therefor reduce the automatic counter-steer effect.

Conversly, low trail figures generate a relatively large steering rotation for a given sidewind, and the dynamic trail effect is less. Therefor short trail will produce steering that is more reactive to sidewinds, and this allows adjustment of the speed and energy of the self-balancing effect.

HCS systems readily tolerate rapid self-balancing, which minimises the detectable disturbance and is desirable.

To summarise,

Tail area helps automatic turn into wind. Good.
High nose side area helps roll out of wind. Bad.
High trail reduces and slows, while Low trail increases and speeds, sidewind-induced counter-steer.
Fast sidewind-induced counter-steer reduces disturbance.

Short trail, small nose side area, and a good tail side area (at least as much as the nose) will produce a strong response, rolling firmly into a side wind. When flow separation features are added the result can be good 'self-correction' which does not intrude on the driver except in truely heroic winds.

For reference, the early Banana bodywork, without a tail, could be blown out of winds, but with later bodywork, including a tail and nose separation features, it did not, self-correcting adequately.
(http://www.bikeweb.com/node/120)
(http://www.bikeweb.com/node/391)

Aerodynamic Efficiency

low drag, is usually regarded as the main reason for bodywork. In practice it has a minor role in the efficiency gain provided by the FF layout. At road vehicle speeds, simple frontal area is the main determinant of drag.

Almost all FFs that run as well as the vehicle they share an engine with, return fuel efficiency gains of around 20%. Many, especially high powered ones, achieve similar increases in top speed. Even 001, where drag was hardly considered, comfortably out-performed the Ducati motorcycle. It's always worthwhile keeping frontal area in mind when designing bodywork.

There are detailed ways to improve efficiency. None of these are mysterious. Smooth, clean surfaces and separations, slow surface direction changes, filled voids especially behind the rider, can add up to quite surprising fuel efficiencies - when the vehicle is being run fast enough for these details to matter.

The actual nose shape is least important for efficiency, although it must be carefully shaped to 'set-up' the separated airlfow over the rest of the vehicle.

Fat Jogger was optimised for efficiency more than any of the other shapes. In addition to it's clean lines it also has a sharply cut off 'Kamm Tail' and the resultant low pressure bubble is filled by the hot radiator outlet flow
(http://www.bikeweb.com/node/577). The side and centre stands are part of the shape, with the centre stand, when retracted, forming the 'chin' of the radiator inlets.

It impossible to say whether this detailing is 'worth' the effort. Although it has reached 90 mpg it has also returned worse figures than Ian's Production Voyager travelling together Vagaries of old engines and states of tune probably have a greater impact. However the three Voyager shapes are all fairly clean and a computer generated prediction of the Cd for 002 was .3. This level of efficiency is clearly worth having.

Indifference

Extreme efficiency can lead to extreme problems. In general the cleaner a shape and the closer the airflow over it, the more sensitive it will be to disturbances in that airflow.

For FFs these problems chiefly amount to sensitivity to turbulence, where the vehicle is noticeably buffeted by turbulence, typically on motorways, at speed, in traffic. Ideally an FF should be “Indifferent” to these disturbances.

Poor indifference is caused when attached (lamminar) airflow over a surface changes direction or pressure. Sideways lift generated across a smooth nose may be balanced by sufficient tail area but the sudden cessation, or reversal of the airflow will instantly cancel the side load and this will be felt in the steering or attitude of the vehicle.

This is avoided, on cars as well as FFs, by enforcing separation of the airflow from the surface at chosen points, rather than allowing the separation line to wander about the surface generating unpredictable lift effects.

At the front this is usually done almost immediately, certainly within 500mm of the leading edge. 001 used immediate separation devices on all the leading edges and some of these sharp direction changes, can be seen in the picture.
(http://www.bikeweb.com/node/494). 002 had less separation on flat surfaces, as nose lift was not seen as a problem but has clearly visible channels running up each side of the nose to enforce separation of any airflow across the nose surface
(http://www.bikeweb.com/node/501). The ledge in front of the lights is intended to enforce separation of the flow over the lights and features on all the Voyagers in various forms.
(http://www.bikeweb.com/node/1078)

002 however had a minor indifference problem. Strong turbulent side winds could be detected as tail buffeting. The probable reason for this became apparent when wool tufting tests were studied. 002 has no separators on the upper tail for cross flowing air. It's tail ridge is a gentle curve. .
(http://www.bikeweb.com/node/501)

The production Voyagers are a simple step on from 002. There is more tail area, chiefly due to the use of slides on the head fairing and a complete head fairing. The separation feature on the nose is somewhat more 'styled' and the nose is shorter due to packaging improvements in the basic design. The result was very encouraging with excellent stability and indifference, while efficiency seems as good as any other of these shapes.
(http://www.bikeweb.com/node/549)

Enforced separation of flow across the tail is just as important as across the nose. 001 and FJ both use separation enforcing edges on the upper edges of the tail. Subsequent study revealed that other vehicles have suffered from cross-flow generated tail lift and buffet and this device is a fairly common solution. The sharp upper edge of FJ's tail may be extreme but some device like this is important.
(http://www.bikeweb.com/node/392)
FJ's nose shape paid only lip service to enforced separation. The shut line gap of the nose opening panel was supposed to bleed air into the airflow over the nose to provide separation. It's pressurised by the air entering the headlight opening. There is no evidence that this works and indifference was initially so bad that strip separators were retro-fitted, very crudely, to the nose.
(http://www.bikeweb.com/node/578)

Indifference was still only moderate in some conditions and a new nose was made with explicit separation features and other changes.
(http://www.bikeweb.com/image/tid/86). This shape improved indifference and flow into the radiators.

Finally the plan view of the vehicle should be considered. If the shape tapers too sharply after the point of maximum width then there may be too much separation towards the tail leading to poor indifference in some conditions. FJ has a slight problem in this respect, whereas the fatter-tailed production shape is markedly more indifferent. Remember the airflow will separate below 2 degrees positive angle of attack, so the tail should be almost parrallel untill the cut-off (Kamm-style) or whatever final separation point is chosen.

Style

The aesthetic qualities of all these vehicles have been criticised at length. There was only one
(disastrous) session with a 'real' stylist and everything else was done for functional reasons.. The challenge for stylists is to achieve similar aerodynamic qualities in a shape that meets their aesthetic requirements.

Although each shape produces a unique reaction to airflow there is no need to copy the notched-nose 'Voyager' shape which evolved in the course of learning about FF aerodynamics.

Similar stability will be achieved by any shape that also has sufficient tail area, low enough frontal side area, good void filling and reasonably fast steering. Neccesary separations can be achieved anywhere desired, using features that can be part of the aethetic styling - or hidden by paintwork.

Once stability, efficiency and indifference have been combined in a beautiful shape there are other important things to consider.

Cockpit environment

In some respects an open-cockpit FF is similar to an open-topped car. There is a need for a tail, or at least a head restraint panel to eliminate the recirculating bubble of air that both vehicle types tend to generate. This is massively inefficient and gives the rider a cold neck. It is possible to stop this effect with a tail that doesn't reach fully to the height of the rider's head. 001 was quite successful in using a cut off tail, otherwise rather similar to FJ.

Recirculating bubbles also tend to to run on either side of the cockpit opening. The cut outs needed to give leg clearance allow airflow running alongside the keel to roll up into the cockpit low pressure area. It has proved quite difficult to avoid some airflow through this gap.

At lower speed, particularly urban speeds, a smaller recirculating bubble may form in the upper cockpit in front of the rider's face. This is unpleasant and if it continues at any speed head buffeting will cause control problems.

A number of detailed devices have been used to try and eliminate these problems. The most effective increase the pressure in the cockpit. Strong ram air directly into the cockpit works very well. The trick is to select an intake that won't carry water, or litter, is directed to the right parts at the right speeds and, most important, is heated. All the Voyagers have heaters, with fans, and this is a crucial advantage in anything but hot weather.

The headlight opening is an excellent intake for cockpit air, being well out of the spray and exhausts nearer the ground. This realisation came after the creation of the Voyagers which bleed air out of the wheel bay with reasonable success. Both options line up well with a good heater location spilling warm air, ideally, into the riders lap. It will flow over the hands, knees and face without further assistance. The cockpit pressure change caused by turning on the heater fan in FJ results in a detectable reduction in the speed at which rain is kept out of the cockpit.

A duct ramming air up the inside of the windscreen is very effective - and needs no heating. It increases the height of the aerodynamic effect of the screen by several inches at 'main road' speeds. FJ and some of the production Voyagers have these ducts, running from the headlight opening, up to the rear of the screen. These work best at speed and are quite common on motorcycles fairings
(http://www.bikeweb.com/node/1036)
(http://www.bikeweb.com/node/1997)

Further ram air can be taken from the front of the keel, on either side of the front wheel. In the Voyager layout the engine heats this air, some of which is exited into the cockpit along the leg clearance cut-outs to counter the airflow mentioned previously tending to take cold air through this gap. These outlet ducts can be seen on the panels running from the top of the radiator inlets to the rear of the footboxes
(http://www.bikeweb.com/node/1439). The radiator outlets on the Production Voyagers were intended to do the same job.

Care has also been taken to seal the rear of the cockpit, to prevent through flow, all the air entering the engine bay is directed into the radiator intake except for a bleed through the gap between the seat base and seat back.

Each of these devices produced a noticeable improvement in FJ's cockpit environment. Some development of features like these should be expected before an optimum environment is achieved in any prototype open cockpit FF.

It is impossible to overstate the benefits of a good cockpit environment. Staying warm, dry and comfortable allows an FF user to fully exploit the dynamic and efficiency advantages of the layout. Even in winter.

Lift

If a sharp edged tail like FJ or 001 is used it is extremely unlikely that the tail will generate meaningful lift. In any case re-attachment of the airflow onto the tail will not be laminar so even a flat surface is unlikely to generate much effect.. The shelf on the rear of the Production Voyagers, caused by the use of seat slides, was given a slight positive angle of attack to eliminate any possibility of lift. However the actual angle was set so that a pint of beer would not slip on it when parked.
(http://www.bikeweb.com/node/1800)

Lift at the nose would probably have been achieved by the clay shape offered to the project by a professional stylist, which was why it was rejected
(http://www.bikeweb.com/node/574). The 'Notched nose' shape used throughout the Voyager project is principally intended to reduce side area at the front but it is easy to provide for positive angles of attack over the entire upper surface.

In practice most prototype shapes generated too much negative lift, or download. Malcolm's SEV Z1300
(http://www.bikeweb.com/node/402) and the original Banana, both accelerated fast enough for the arrival or this download at around 80 mph to be visible, with the body settling onto it's rubber mounts. This can produce instability at high speed, where the vehicle apparently hunts about behind the huge pressure wave. It can also increase sensitivity to side winds, possibly due to variations in download according the the airflow angle over the nose.

It has proved better to seek to eliminate positive lift, by suitable separators, positive angles of attack, etc., than to attempt to achieve download.

Cooling ducts

All vehicle body shapes need entries and exits for cooling air. The most efficient form of intake is the 'pitot' type. Basically a hole in the front. Such intakes work well on FFs. The Banana demonstrated that with only 10 degrees of lock each side, a hole in the ('Dustbin') nose just big enough for wheel clearance, would provide ample airflow for cooling the engine. Until the space around the radiator was blanked off it acted as a vacuum cleaner, sucking litter off the road surface and throwing it into the cockpit.

If an air-cooled engine is used the front wheel bay is necessarily the intake and this works very well. 001, with close cowled fins, seriously over-cooled. In winter use water droplets appeared in the oil and no oil temperature was indicated between October and March.

Water cooled engines will have thermostats and 002 placed a radiator in this intake, immediately behind the front wheel. In terms of cooling this is perfectly satisfactory. There are problems with this approach however.

First is that there also has to be an outlet. If this is the engine bay, as in a car, the bay will become extremely dirty, to the point where this may compromise function. It also places the hot exit air in the riders lap and this is unacceptable in hot weather. These problems can be solved, as Yamaha has with the Tmax, by ducting the air into the keel. Yamaha use this technique on their race bikes and it clearly a good low pressure area. However, It can be difficult to manage this ducting when a collection of unmatched components have to be packaged. It was incompatible with a forward mounted Reliant engine.

The 2010 Cmax conversion uses a pitot intake in a full (dustbin) nose fairing with the stock keel air exit and this works very well.
(http://www.bikeweb.com/node/1997)

For 002 the biggest problem with the front radiator was the rearward movement of the CG, caused by the radiator and fan fitted between the engine and the front wheel. The weight distribution was too far to the rear and moving the engine forwards as was done in the production design, by 3”, was a major dynamic improvement. This was achieved by fitting two radiators either side of the engine and the front wheel, fed from the wheel bay and exiting into the leg clearance cut-outs on each side. .

This was not very successful, with insufficient airflow through the radiators One Voyager subsequently fitted a third radiator across the top of the wheel, exiting into a hole in the shelf in front of the headlights
(http://www.bikeweb.com/node/997). This provides sufficient cooling but is costly in components and plumbing. Much later development has indicated that the poor sealing between the wheel bay, the radiator intake, and the outlet areas may be a large part of this problem.

FJ took twin, larger, radiators to the rear. The intakes were placed below the riders seat in attempt to capture the strong airflow turning up into the cockpit mentioned earlier
(http://www.bikeweb.com/node/1995). The exits, as also mentioned earlier were fed into the low pressure area of the Kamm tail. There were initially problems with water flow, due to over-elaborate plumbing but this system works well, with the engine temp falling below 'Normal' (Thermostat fully open) at high speeds. Fans are needed in these long ducts at low speed, but heat balance is reached well within urban speeds.

An attempt was made recently on another Voyager to replicate this system, using a single radiator on one side. Although designed from the outset for a clean duct it has proved very difficult to make it work. It was immediately clear that air was re-circulating through the radiator - out at the top, back in at the bottom, a perfect recirculating bubble.
(http://www.bikeweb.com/node/1117)

It seems likely that the problem is very poor airflow into the intakes. This probably relates to the different nose and keel shapes. The production nose is intended to cause a low pressure area behind it, like a dam and the resulting low speed air blanks the intakes. FJ's nose is cut away and shaped to minimise the interference with air flowing either side of the front wheel and keel. Near full airspeed is maintained some distance into the intakes.

Basically the standard pitot intake works well, but needs work on the exit if the engine is to stay clean and the rider cool. Rear radiators solve most of these problems and are a good packaging solution but need careful airflow optimisation from the outset.

Beyond open cockpits

Many FF proponents are interested in greater degrees of enclosure, Roofs and full enclosure are popular features. Every shape produces a unique reaction from the air and it would be extremely rash to extrapolate anything learned with open cockpit FFs to roofed or fully enclosed vehicles.

However the general rules apply. The more bodywork in contact with the airflow the more important it is to get the stability balance right and to ensure that separation is predictable rather than dependant on airflow direction. Efficiency and cockpit environment should be better than an open cockpit unless something is seriously wrong!

There are several other vehicles on this site with roofs or enclosure and the NSU Hammocks are worth careful study. Not least because it was discovered they take off at around 200 mph... The Quasar shape, colossally heavier, also lightens up surprisingly when driven by other engines to speeds above 120mph. It also appears to generate too much download over the nose initially and small spoiler strips across the nose and at the top of the tail were tried, proving to lighten and quicken the steering on 'fast main roads'.

In general I hope I have made clear that open-cockpit FF aerodynamics are quite complex, but that it is quite easy to arrive in an acceptable window. Providing the basic rules for stability are followed and attention is paid to indifference it is probable that most designers will find the details of the cockpit environment most challenging..

Anyone with no formal or empirical experience of aerodynamics would do well to read widely on the subject. A recently published book “Road Vehicle Aerodynamic Design” ISBN 0954073401, from www.mechaero.co.uk is a good starting point with many references and a non-mathematical approach. Single track vehicle are not mentioned.

Anyone at all who wished to gain a feel for good shapes will do well to look very closely at modern cars. There is virtually no limit to the real complexity of FF aerodynamics. One might well build an FF purely to study this subject.

Royce Creasey
Jan 2011
Copy free for credit.

Subsequent to the above piece I wrote the following (originally as an email) on the practical application of these principles that I used to produce the generic 'Voyager' shape (also Cmax, et al).

"Lets see if I can break it down a bit.

You start with the basic teardrop. Th proportion of this can be as short and fat as the basic dimension of the vehicles demand. This doesn't matter as the speeds we're talking about are pretty low. The tail will still be much too long so we can cut that off to make a 'Kamm tail', which we know works as well as a full tail at road speeds. Ideally this should be not later than one third of the length of the teardrop past maximum chord (the thickest bit of the teardrop) but you can cheat - see next para.

Then to prevent excessive separation over the rear of the vehicle we can reduce the taper from max. chord, ideally where you sit, to the cut off. I think this should be, at most, 3 degrees negative angle of attack, but parrallel works fine too. Once the air has separated it won't re-attach below five degreeds positive angle of attack, so there's no danger of buffet turbulence back there. As FJ seems to have demonstrated the problem is likely to be too much taper, leading to eddies big enough to disturb the tail.

So now you got something looking a bit like a bullet.

But we need to reduce frontal side area, increase rear side area and make sure the centre of pressure is moved back - unlike a bullet we don't have spin stability, so we need to use 'weathercock stability' - turn it into something more like an arrow.

One way of doing this is to make the basic teardrop smaller, so it only encompasses the rider up to the shoulders, or hand controls. Then we can add the top half of another teardrop, starting well back, to just cover the riders top half and taper that back into a fin-like blade at the rear. We know from American streamliner cars ("The Leading Edge") that this should have a sharp top edge to prevent attached transverse flows, this was also shown by 002 which had slight problems with it's rounder tail. This produces the 'conning tower' shape I've used.

This moves the centre of pressure back reduces frontal side area and increases rear side area. It's also starting to look a bit familiar, but blame the air for this!

Finally we can move to the front, cut off the top of the frontal teardrop to absolutely prevent any pressure wave coming off it (FJ's original problem). with some sort of separation along the top of the leading edge, which itself should be high, close to the horizontal centre line. This also reduces frontal side area a bit more. We can then cut the sides of the basic bullet off to flat(ish) surfaces, to reduce frontal area and tighten the shape up to just clear the rider (leave some crash protection depth here) This sort of automatically produces separation lines at the sides of the nose, as the flat sides intersect the rounded corners of the nose.

Last thing is to cut off the ground clearance shamfers. AIrflow isn't a problem here, the front wheel separates (mashes) this airflow into chaos, it isn't going to attach to anything.

The rest is detailing. A big 'behind the screen' duct to make the air think there's a roof and flow over the riders head to the tail. Air intakes (don't forget the heater!) Lights, indicators, mirrors - all these can be used as separation features. Don't forget to separate the sides of the conning tower. The production Voyager shape used big car indicators, in cut-ins, to do this, FJ has blunted edges to the light cut-outs.

There are rules for the actual nose shape. The top deck cut-off produces an upside-down aerofoil shape. This is good. It can be very blunt. This radius should be reproduced over the corners at the sides of the nose. If this results in a 'chisel' nose, where there's actually a portion of straight leading edge across the centre of the nose, thats good. Any sort of point is bad. Should be a very gentle curve at most. Top deck can have a positive angle of attack, up to five degrees. Transverse separation features along the sides of the top deck are important - and can be seen on virtually every modern every car er..'Hood' (wot we calls a 'bonnet') styled in a myriad of different ways.

That should do it. Wish I'd known this thirty years ago etc. "

Royce Creasey August 2012