Please note: Building your own wind tunnel is very dangerous. In these modern times when litigation takes the place of natural justice, it is therefore necessary that those wishing to read the monograph must realise that they do not, nor will ever hold the author responsible for any damage or injury. The reader is only reading this for interest, and must never intend to build such a device. This is because the author simply does not want any one to get hurt and certainly does not want a parasitic lawyer coming on heavy because some damn fool thinks he or she is some brilliant engineer and aerodynamicist, or stubbed their toe and wants a fortune because they happen to be stupid.
I don't mind putting my knowledge on the web, after all, it's there to help make a better world, but not so that some damn fool can sue me. J.P. Please use your vote to keep assholes out of politics.

Introduction.
The author wishes everyone to have a chance at design as the big manufacturers and race teams. For vehicle design, the biggest toy in town is often the wind tunnel. With a little thought and effort, it is not so out of reach as one might imagine. Whether you have a bicycle, motorcycle or larger machine, aerodynamics is not difficult to do. The only hard part is interpreting the more subtle aspects. The poor areas of airflow are easy to recognise and solve for reasonably good aerodynamics, and certainly better than most cars. You may even find that fitting some 'go-faster' parts may slow your machine down. It is only when you decide to enter the world of fluid flow will things get hard to solve. Nevertheless, the bottom line of having a low drag factor, then modifying your machine until is shows the lowest drag figure, is not exactly difficult to understand.

Do not be put off by 'experts'. For example, the 2005 Reith lecture by some gov't 'Sir', stated that in Britain today, innovation can only be done by large groups and the individual no longer able to advance technology or science. I consider this utter crap.

Although getting funding is important, (but impossible in Britain) a single person can still do a vast amount of excellent work without money. This 'Sir', shows the potential downfall of Britain as an engineering and science based country, gradually becoming a land of accountants and bankers, who do nothing to make a better world. (Just a fatter one).
Thatcher may have destroyed industry, thinking we are all shopkeepers, while Major wanted us all accountants, and Blair wants us all to be badly educated tax cows, but we are not all soap and soccer fodder, yet.
So never be put off by 'experts' who tell you that you are useless and must only work for some big company. These arses are like all the experts, part of the money system and may not be after a better world.
Despite the corporate and political scum always floating to the top, always work for a better world.
Have faith in yourself.

Please compare the 2005 Reith lecture with the 2005 Dimbelby lecture, both care of the BBC Radio 4, as they are poles apart. One from a 'Sir', who puts down individual, and the other which makes a far better argument for innovation and a better world.
Please read them to see why Britain is going down the pan.

Have faith in yourself.
This monograph uses commonly available technology applicable on a daily basis, for those who wish to learn with a hands-on approach. This empirical approach has the intention of making maximum effort with minimum equipment. You don't have to spend money. Hopefully I have included a reasonable balance of encouragement towards innovation, without detracting from the safety and technical problems. Technical descriptions are comparatively easy to write, simply re-hashing work of others.
But encouragement and guidance are also important requirements of learning the art of design, all too often left out of reference books. Read on.
The aims and objectives of this monograph and its companions, is not just to get the job done, but also to encourage the aspirations, develop innovation (and to minimise the all too common 'sheep' approach to design). Design and innovation comes from the mind, not from books or drawings. If this monograph does not offer what you seek, then glean what you can and seek further.
Texts such as this are merely guides to a higher process. This higher process is an art and no amount of copying will inspire the reader to better things. There seems however, a gaping hole between theory and practice.
It is hoped this monograph will help towards bridging the many gaps, hopefully to inspire and help others to get their hands dirty and build their own aerodynamics testing equipment on a very small budget, or if necessary, purely by scavenging for free. You can build a complete, full size wind tunnel with computers for nothing. It is not rocket science.

(If I was a technology teacher in my local school, - a so-called 'technology college' - with just token gestures towards science, I'd have a full wind tunnel working there within six months. Building it would also be part of curricular activities. I've a B.Ed and a B.Sc, but they couldn't even offer a part time cleaning job. This is VERY common across much of Britain, as many honest friends have subsequently learnt. Britain is still a case of who you know, not what you know. If I took the advice of a Freemason friend, he said he would find 'no problems' in finding me a teaching job. - Yes, the scum float to the top.)

It's brains that's important, not money.
No excuse is given for the somewhat blinkered view of building as an art, rather than technology or science. This approach is for two reasons: First, not all hi-tech is a matter of sitting down with a calculator and following specific rules. Calculations do not take into account the variables of hand building, nor does it allow the study of the technology with test pieces, which often offer much more accurate feedback. Second, what is today's hi-tech materials is tomorrow's everyday kit. Materials are just that: materials, nothing impossible, nothing hyped. Most of us have lost the art of flint knapping, yet quite happily use steel, plastics and modern adhesives. We all move on.

There is nothing here that probably has not been done before. There is nothing difficult in building a wind tunnel; it is based on theory developed from the works by Burnoulli and Boyle over a hundred years ago. It's the way that they are brought together which can make a better whole. For some, the sheer audacity of thinking of building your own wind tunnel and to do it on a small budget is to some people, just too far. Never be put off by the high tech equipment of the ultra expensive wind tunnels of aerospace establishments with their multi million budgets. There is no black art in basic aerodynamics and wind tunnels are just a big hole with a controlled draught.

To be true, when understanding the very precise nature of airflow over a specific surface, there can be a lot of black art in refining aerodynamics and the various methods to be employed. When understanding the finer subtleties, this is fraught with many hours studying airflows trying to decide just what is happening and how best to resolve the problems.

When building bikes and motorcycles, the reader may be surprised at what can be done at home. While designing and building, don't follow blindly, but always be open to inventing and developing new, radical or simply lateral styles and methods to see where they lead. No one has the right answer for the perfect machine. Therefore someone has to make the necessary giant steps into the future, it can be anyone. It is hoped that innovation is the driving force for the reader.

This monograph does not assume the reader has an engineering or scientific background. Many good machines have been created by ordinary people who would shy away from maths and theory associated with advanced materials and engineering. Some absolutely awful machines have also existed, so optimism must be tempered with pragmatism and common sense. What this monograph cannot supply is the time and effort to build up the skills, nor the time and patience to study both the theory and the art of design. When aiming high, occasional failure will naturally ensue. Always realise that careful study of failure is an important part of the learning process and an excellent way of focusing on development. The reader will, however, have started on the path toward their ultimate machine. The longest walk starts with the first step.
Many will be pleasantly surprised even with their first step.

Quote:

This monograph takes a two stage approach to aerodynamics.

The first stage is straightforward aerodynamic assessments without a wind tunnel. This is simple common sense, as there is no point building a wind tunnel when such a great amount of testing can be done without one. Therefore the basics are given for those who want most of the applications of a wind tunnel without the need to build one.
The second stage is when the machine under study is judged to be in need of further refinement, whereupon the wind tunnel can then be applied, and of course, how to build your own wind tunnel.

As mentioned in all my monographs, don't be put off by all the hype. The reader should realise that hi-tech is not always high-difficulty or high expense. Most engineering can be done by those unemployed like the author. There is nothing difficult, the only hurdle is getting past the thought that wind tunnels must be multi million pound installations needing overpaid specialists.

Aerodynamics.
There is no point making a better machine for it to loose some inherent attributes to poor airflow.

There are many machines which are an aerodynamic nightmare and the cycle and motorcycle are but one form. Many machines have shown that getting the maximum from a rider demands cleaning up the design for good effects, but not always.
Never allow the weight penalty outweigh the aerodynamic advantage, so if riding a HPV predominantly in hilly country, then carefully consider the effectiveness of low weight and also the weight penalties of aerodynamics. In windy flatlands or if racing, then spending some time on aerodynamics usually becomes worthwhile.
For motorcycles and record breakers, poor aerodynamics are always a hurdle to those few extra miles per hour.

For efficient machines, reducing frontal drag is just a starting point.

I prefer the low recumbent which is aerodynamically better than many cycles, as it has a small frontal profile, so less volume of air has to be disturbed. Having the pedals at the front is not going to help matters, by disturbing the flow greatly at the front of the process. The churning of air around the pedals is best kept to itself. For many HPV's, this sometimes means a clear plastic screen, but can be almost anything including papier-mache on a bamboo frame. On touring machines, where headwinds may be a problem, then the design can be favourably improved with suitably shaped front and rear luggage.

For many motorcycles, the classic riding pose is a barrier to speed. Therefore it is better to start with a fundamentally more aerodynamic machine before applying aerodynamic refinements. (See companion monographs). The classic way is to clean the flow, by applying some external form of smooth profile for the air to be coerced aside smoothly with minimal friction. The fairing still seems to be the best solution for most situations, although other solutions are also possible.
Formula one is blatant in it's lack of smooth form or fairing, yet manages the airflow with consummate control.
On HPV's, tail winds can also be put to good use with various styles of tail. Wet weather can be ameliorated with a lightweight cover designed with aerodynamics in mind, as wet weather often arrives in the company of high winds. Luggage should be seen not as a problem, but as an opportunity to improve the aerodynamics of the design. For motorcycles, full wet weather and safety protection can follow the path trod by the Ecomobile, JP7 and others.

Most standard cycles and motorcycles are framed designs, with aerodynamic panels fitted onto such structures. External panelling makes aerodynamics much easier. Adaptation of various panels is an excellent path to tread for initial studies. It can also lead to highly refined and idealised aerodynamic forms, but separation of the chassis structure and the aerodynamic profile will require a heavier machine.

Much of the following can also apply to cars, speed skaters, and many other systems. Let's face it, if you can clean up a pedal cycle and a motorcycle, then you can clean up just about anything, especially as they are very sensitive to external forces.

The subject of aerodynamics can be read in a vast number of books on the subject, but never confuse aerodynamics as simply the use of a wind tunnel. Aerodynamics is the general approach to airflow over a body. Even simple improvements can be considered aerodynamics, although some simple 'improvements' from catalogues may not always turn out to have such positive attributes as expected.

The wind tunnel and it's associated testing systems are mostly applicable and many of these practices are available without real expense. The first economical approach is to use the full scale working design itself.
Testing at full scale in real conditions may seem less than ideal, especially in this age dominated by high tech visions of wind tunnels, but an 'ordinary' cycle does not live in an ideal world. Wind tunnels are good for some things, but simply cannot afford cycles with all the more serious and varying situations they encounter in the real world.
Initial small scale aerodynamic tests can give a general view of how the airflow will behave over and through a design. The main problem with scale are the much higher airflows or pressures required to get reasonable data for mathematical reasons. It is rarely worth modelling a machine to scale, if a full size machine is already available !

For many cycles and motorcycles, the main aerodynamic areas of concern are overall speed effects from the traditional analysis of frontal airflow. Many other problems arise from strong, buffeting side winds at medium and low speeds, where aerodynamics can be of great help. Side winds are usually tested in the real world. A wind tunnel is capable of simulating bow waves such as when passing lorries, buffeting or intermittent and side winds near high interrupted buildings. This could be done in theory, (see later monographs) but bikes will always remain a balancing act and all endeavours at aerodynamics should adapt accordingly. Road testing in high wind regions with 'constant variability' still remains the true arbiter of such situations.

Bikes are not aircraft generating positive lift, nor a formula one car generating negative lift, but are usually considered a narrow block moving forward, generating the minimal resistive forces. Because the unfortunate conventional bike and rider are quite tall and therefore directly susceptible to side winds, minimising the problems from enclosing such designs is not always worthwhile. Enclosed machines suffer in side winds, the nearest equivalent being an aircraft tail fin. Tail fins have very strong effects on an aircraft. The recumbent has a better starting point, with the low form often being considered as good as one may get. Whatever the forces that must inevitably be generated, they should act positively or be as neutral as possible.

Whatever aerodynamic tools are available, the first tool is common sense, so the rider should be positioned to advantage, not only for comfort and handling ability, but to maximise the interaction with the machines airflow. There are many variables such as the classic bike racing position with the head tucked into the fairing.
As for powered and unpowered recumbents, consider the most comfortable chair in your house. Sit in it and put your feet up on a stool, nice and comfy. Now consider if this is more aerodynamic than sitting upright. See website for associated monographs for a full study of this form. Various methods of refining the basic rider and machine are mentioned in the appropriate companion HPV cycle and motorcycle monographs.
On racing motorcycles, when the head and upper torso is lifted as an airbrake it can transfer a little pressure to the rear wheel under braking. For touring, the rider is usually sat in an upright position and occasionally a huge fairing used to barge it's way through the air. Although fairings are getting better, they have oh so far to go.

Without a wind tunnel, many airflow studies need not be compromised. -

You are probably building your machine for the real world, not drag racing or formula one. The real world of aerodynamics is mainly for the upper reaches of the bikes performance envelope, and this is where your testing should be directed.
If the real world of biking cannot work in the wind tunnel, then turn the problem upon its head.
Like JPseries crash testing, if one can't get help from the experts, then study their work, and do it anyway. See later monographs and please be careful when crash testing with real riders. Crash testing first hand, will always give much greater insights to the design needs and the possibilities which can lead to radical new ideas in crash design. Always be painstakingly careful when crash testing with real riders and always prepare for the worst.
(It is ardently hoped that computer crash analysis will be made available to all vehicle builders as soon as possible. In the same way Volvo gave the their safety belt patents openly to the world. Thank you Volvo. Will those with safety software please make the basics available free to all, as the safety of many test riders, including the author are directly affected.)

The many forms of wind tunnel testing equipment can be taken to the real world.

Assessments can be done in real time on real machines on real roads. Pressure gauges, airflow sensors, smoke probes, various pattern assessors and a few of the authors ideas awaiting patents, can be backed up with various data collection and video. Many people now carry a digital camera which can take small video sequences. These allow most appropriate wind tunnel practices to be applied and more importantly, in real time and at full size.

Most cycles are framed mechanical devices with airflow control devices added later. Therefore the adaptation according to feedback can be done easily without upsetting the underlying structure. For example, if low profile hub centre steering is employed, this can be used to advantage to interact with the aerodynamics to improve the handling such as in side winds. See also Tony Foales website. Some cycles can be designed from the outset to be aerodynamic as possible, from a minimal frontal area, to the overall chassis and rider layout. If aiming for total aerodynamic perfection, consider designing the next machine to integrate the aerodynamics from the outset. If building a lightweight touring machine, consider making the luggage areas as part of the chassis and aerodynamics as discussed in companion monographs.

Wind
Unless you live in the Doldrums, wind is free. Choice of landscapes can be chosen to maximise the best and the worst conditions available by the airflow over the countryside and in cities. Using local test sessions and scenarios gives regular, fast and interactive changes to refine the aerodynamics and there is no need to book wind tunnel time. Check local weather records. Some winds are so reliable that they have their own names.
The testing schedule can be based on the highest, bleakest moorlands during winter and in the windiest towns. This can also give a fine selection of both fast, twisty and poor roads with the highest winds normally encountered, with a choice of both (reasonably) clean air and maximum turbulence plus a wide variety of buffeting scenarios.
Once the machine is fettled and road worthy, bleak sections of main roads in winter with high sided lorries can be used to assess intermittent buffeting, bow waves and whatever else could be thrown at the design.
Wind tunnels may give 'tight' data, but the excellent scenarios of riding poor, twisty roads in high, buffeting winds will probably give a form of feedback than no wind tunnel could ever hope to achieve. Road testing in bad conditions allows the rider to feel the true effects needed for improving the actual riding parameters.
When testing in winter, there is no embarrassment in using a car to get there, especially if there is a warm flask of coffee to take the edge off the effects of shivering in the snow.

As many of my motorcycles are radical, there is no way I can get them road legal from the outset and many months of testing are needed before I am happy to present them for their first MOT certification.
(For information on enhanced SVA contact: The Vehicle Inspectorate 91/92 The Strand Swansea SA1 2DH Tel 01792 458888)

If such scenarios are used, take advantage to ride the machine on a level, ice covered car park or similar, as this vastly exaggerates the control possibilities. Tip: As part of testing, the author has ridden over frozen ice fields in the Lake district with the front brake on. It takes a lot of practice, but helps develop riding and control skills. Likewise mud and sand. After riding ice sheets, the riders study of control is far more refined and 'tuned in', so you get to notice the many small aspects of stability and general driving needs. (Never ride on ice covered ponds or rivers.)

Once the machine handles adequately, but before making major changes, it may be preferred to simply asses the airflow over the machine.
When parking the machine on a windy ridge, many of their quieter access roads curve around the hill, so the section with the optimum air flow can usually be achieved. Always place secondary transport downwind. Position the test machine clear of scrub and rocks, preferably on tarmac or ice for a clean flow just before the machine. Bleak carparks, usually in high scenic areas are ideal, especially in windy winter conditions, when the tourists have all long gone. I live near Dartmoor and have a nice selection of primary test sites.
Walk the area in front with the stick with wool, to assess the surface airflow. A stick with a piece of wool can be pushed into the ground as a basic wind guide, so any changes in direction can be recognised during a test session. Once the machine is positioned about right, you can begin the study. Although fairly reliable, do not assume the same place will give the same airflow each and every time. Sticking a cycle with a large lump of blue tack onto a trig point (survey point) to be free of ground buffeting, is not recommended, as a simple steel bar offers less buffeting. But if cycling past such a scenario and the conditions are right, and a long length of wool is teased from barbed wire fences and stuck on a thin stick, or lots of dry grass can be chopped up, then any machine is in the right place for a five minute assessment. This is not always possible, but gives the reader a simple, open approach to basic aerodynamic assessment.

There are many options to discovering airflow.
With a rider pedalling the machine on rollers, or while stationary with the chain derailed, then chopped streaks give only a small view, whereas smoke probes are much better for understanding the longer disturbed and intermittent airflows around the feet.
When emulsion is used, the surface flow is visible on screens, coverings and shells. Cotton tufts stuck on the surface allow external and internal flows to be studied just above the surface layer. These can be used while riding normally.

Conducting the tests in various windy areas at various speeds with video or digital camera sequences, allows the various changes of airflow with speed to be studied at leisure. This full scale testing not only gives objective data, but also the very important subjective assessments only assessable by road testers.
The areas of subjective testing is also much more important on two wheels than any other form of road transport.
A variety of testers will also improve riding assessments.

Particles.
The ability to apply many techniques to real life scenarios does not limit the study which is normally associated with wind tunnels. Videoing of airborne particles flowing over the machine gives direction speed and orientation of the whole airflow, which can be measured in single and sequential stop frames. Pet stores can supply a small bag of straw, so buy a good pair of scissors. Do not use polystyrene chips as these damage the countryside. In lighter winds, cheap confetti may suffice. Tissue paper or paper towel rolls cut into various sized chips is easier and damages the countryside in a minimal manner by degrading faster, but can clog up when put through paper shredders. Light weight, bright, contrasting paper such as kitchen paper towel roll is best, as it flows with the air much better and a choice of various sizes can give more accurate readings.
Always transfer the photos or video for storage on a computer, as there will be much time spent analysing single or sequential frames, which would otherwise damage a tape. This also allows you to make short animated sequences which can be run continuously for closer study.
The length of each strand of straw or paper will be streaked, allowing the distance to be measured and hence the speed, as calculated by the exposure time of each frame. If this is difficult, use a good camera and note the exposure time, or just simply measure the relative lengths of streaks. Usually relative streak lengths will suffice, as the direction and relative speed of the streak is the more important part.
Each streak in a still photo frame will be offering the viewer the directon and the speed. If you mark these by length and direction, then you have a good deal of information eaily available for assessment.

Clothing.
To recognise the paths over the shell and rider, various simple methods can be employed. A slightly loose fitting, light weight rain suit over the rider will highlight smooth and flapping airflow as well as highlighting the positive and negative pressure zones. Disposable decorators overalls will also suffice. Loose light clothing easily shows turbulent areas and negative and positive regimes. Unfortunately it's very ability to deform in accordance with the airflow disturbs it with the similar problems to Shroedingers cat. To know the result is to disrupt the experiment. Therefore keep the study to small sections, tailor or tape the clothing for minimal flapping.

Cotton tufts.
On conventional cycles and motorcycles for road use, cotton tufts can assist and which have been used before wind tunnels existed.
These should be small strands of wool or cotton wool, taped or lightly glued or blue tacked at one end to various parts of the surface to be studied, to show direction and any turbulence in their movement over the surfaces. (Drafting tape is like masking tape, but less sticky for those with pristine paint jobs.)
Cotton tufts can be assessed statically or while riding. Blue tack or tape until they start coming off motorcycles at high speed, then glue must be used. If the fairing is nicely painted or on screens, then use tape to anchor the wool or cotton. It is difficult to glue to tape, so I tease the wool through a small hole in the tape. Carry a few extras, as they are cheap and easily pull off at high speeds. If following with a video in another vehicle, this should be done at a reasonable distance to eliminate disturbing airflows, with zooms taken to any areas of particular concern. The advantage of cotton tufts is their simplicity to show the airflow close to the surface and can be easily removed.

A variation part way between smoke probe and cotton tufts is the long thin strand of fluffy wool on a thin wire, which can be played over the surface to see the airflow. Unfortunately, the longer length will straighten and thus distort the forward part of the strand due to drag on its length, so only trust the end of a long strand.

Cotton tufts will require photographs to record the state while riding, as they do not retain the information after use, unlike slime. If using long wool strands over clear bubble fairings, be aware that constant use will cause eventual scratching, especially if testing on dusty roads.

Slime.
On partially or fully enclosed machines, use of surface layer 'slime' over panels should be water based, so the machine can be hosed clean afterwards.
Streaks can be painted across the expected airflow to highlight airflow direction and strength. Almost anything will do, a light smear of dyed olive oil or water based lubricants and many adaptable foams, gels and such like at the discount section of the beauty counter of most chemists. Airflow streaks are more easily seen when simple poster paints are used to create straight lines in the slime prior to a run. If covering the whole surface and painting grid lines into this, always allow the slime to settle a little before making the lines, otherwise false readings may occur. Many slimes will be affected by the cooling effects of high airflow on cooler days, so choose the materials according to the weather. Carry an old towel for cleaning after testing. Not all the surface needs covering, just a few lines painted across the airflow will usually suffice.
If expecting rain, oil based is better and disposable cheap weather proof clothing may be applicable for the outer layer, available from industrial outfitters. Rain should not be seen as a problem, as it will help the process with its greater mass of the drops to deform the heavier oil based slimes.
Always take photos immediately after a test run, with the rider in the test position.

Smoke probe.
Use of the classic smoke boom and probe can be done statically in most winds. This is the method often connected with wind tunnels, but can be applied anywhere.
The smoke probe is not the answer to all solutions, but it has many advantages. It can be used anywhere, in any condition, can last for a few minutes, the smoke does not upset the surface flow unlike cotton tufts or flapping suits, is easily seen, and when videoed, gives a very close reading to the actual situation and is therefore easier to assess.
When testing on the road, smoke generators should be fitted on the machine itself and vented in various positions for much easier testing and analysis. Most plumbing suppliers offer smoke generator sticks for testing gas flues. These are ideal for stationary, still-air overheating tests, and in long hallways to check belly pans.
If a handful of these do not supply enough smoke, consider using a soft firework, or a small hand held pyrotechnic flare as used for emergencies at sea. Use pyrotechnics in a safe, lightly pressurised metal container with low pressure vent, and duct via a selection of pipes. On motorcycles, it may be preferable to pressurise the container from the bow wave, so the smoke flow is sufficient for the speed involved. If an enclosed container is used to contrain the smoke generator, always use a downwind vent pipe to prevent excess pressure generation and never use on public roads. Never use parachute flares, just the basic hand held items, or use gentle fireworks as a starting point.
If attempting to disassemble a flare past its use date, or to remove the parachute and reduce the smoke material to more manageable units, then always do this in the garden and close to a big water bucket. Preferably on the end of a wooden table, so it can be swept into the water and a lid fitted. Do not to use spark making tools or metal bench surface. Always use a wood surface, leather gloves, face mask or crash helmet with the visor down, and a plastic picnic knife is ideal. If too much smoke is supplied, carefully deconstruct pyrotechnics outdoors using safety clothing and a bucket of water nearby. It is not usual to make pyrotechnics for the occasion, but various mixes are available to maximise the smoke and minimise the flame, so consider making special 'recipes' if using this means of testing for many years. Most pyrotechnics manufacturers keep their recipes secret, but may be able to make materials to suit and in various colours. As emergency flares are available in various colours, choose a colour to contrast with the machine to highlight the airflow.

It will take a little time to get a good smoke probe, as it must be designed to minimise any disturbance upstream to ensure a clean flow downstream. Old brake pipe or similar tubing may often suffice for the boom and nozzle, although for greater airflow, metal car fuel pipe is larger and available from the same suppliers. Make sure the tip is a smooth aerodynamic transition to the airstream. Swage the tip around a drill bit for a gentle nozzle stream. Practice making the probe in front of a domestic cooling fan using a cigar. Smoke booms can be built from car piping for greater flow, then bent to give the perfect downstream flow with minimum turbulence. This usually means making the tip of the boom curved to prevent upstream disturbance at the tip.

If riding the bike and wanting to run the smoke probe when the ride is perfect, it is necessary to ignite by a switch. To ignite a soft firework at the appropriate time, simply wire a thin piece of steel in the touch paper and short it via two thicker wires and a switch via a battery. Simply choose a piece of wire so it glows when shorted across a large enough battery, most torches (flashlights) or the motorcycle battery will suffice for ignition supply. Also consider using wire wool and the ignitors from model rocket motors and also a few deconstructed match heads. Done properly, a smoke generator can be triggered from a simple handlebar switch. Allow a little time for generation, so trigger well before the test run area. If triggering is problematic, the fuse section of bangers or firecrackers and the heads off waterproof matches will help to liven up the ignition process.

Smoke generators using a heating element and special oil are also possible as used in discos and adaptable for the hot exhaust system as a generator. Not unlike a particularly smoky two stroke. This will require an exhaust box clamped over a hot area, plus a fast oil drip feed and such like, plus an upwind pressurising system to ensure reasonable smoke flow. This is preferable for stationary testing. It is now possible to buy party smoke in an aerosol.

Although predominantly for dynamic airflows, smoke can be used in a quiet garage while testing for overheating in a simulated summer traffic jam. This is particularly useful for air cooled machines. Place smoke candles under the motorcycle while it is running hot, to simulate traffic jams on hot days in still air. Remember where the most overheated vehicles are seen, usually part way up a long hill on a bank holiday on the way to the seaside. Always check your machine for all forms of aerodynamic problems. This test enabled the JP7 to have a better cooling system for slow speed, inner city work, than would be possible otherwise with a more traditional motorcycle radiator system.

Pressure gauges.
Although problematic, on faster machines it may be preferable to reduce the build up of pressure zones, especially an excessive bow wave, or the creation of high pressure and low pressure zones. These high and low pressure zones all require energy to create and maintain.
The simplest pressure gauges are waterbombs (small balloons) from kiddies toy shops, attached to plastic tubing to the required area for study and the balloons stuck with tape under the fairing for the rider to assess. This will give relative assessment of positive or negative pressures. If they stand up, then positive pressure. If they are limp, then neutral. If sucked empty, then a negative pressure regime. (Just how difficult or expensive can aerodynamics get!)

Where a few pressure probes are needed, cheap and simple water manometers can be made from long lengths of clear plastic tubing as used for windscreen washer pipes, but do not expect easily read differences in pressure. These can be terminated by bending into U tubes with coloured water, then mounted on contrasting side panels with blue tack, allowing the video on the attendant machine to read them in zoom mode while riding alongside. Keeping the water level in the probe and vent pipes close together reduces false readings and makes comparison easier. This eliminates any need for data loggers and makes life very easy by using simple components and ensures the data is kept with the images. Although it is easier to measure variations in high winds while static, the differences will be small. Static testing allows the manometers to be positioned at an angle for easier reading. Angled manometers cannot be easily read while riding, as they need neutral positioning or zero acceleration. Carburettor vacuum gauges are rarely sensitive enough. Mechanical or digital barometers are another option.
Low pressure zones are often deliberately required for radiator airflow, while some negative and positive pressure zones can improve some handling traits.

Cleaning up the airflow is usually aimed for lower drag or greater comfort, usually both. Top speeds on HPV cycles is not going to set the world alight, but every ounce of effort saved is effort available for going faster, further or easier.

When aiming for top speed, the overall aerodynamic drag can be assessed with a spring gauge when aligned into a headwind whose speed is known. A cycle ridden in still air at 30 mph can be easily simulated in a 30 mph wind across moorland or a local hill. (if not too turbulent.) Modifications then applied until the lowest drag is possible. Choosing an appropriate season and location can give surprisingly constant conditions. Local weather forecasts and local knowledge of wind paths for many years are usually available for those without wind tunnels looking for appropriate locations. It is rare to get higher winds for testing motorcycles, but even the use of aerodynamic assessments at lower speeds will usually highlight the main offenders, to help reduce drag at higher speeds.

In a headwind, it is possible to get a general assessment of drag.
This is not efficient without a wind tunnel, as the wind will vary and possibly have gusts from the topology and scenery. What this can offer is comparative assessments for various aerodynamic shells and rider poses.
For static aerodynamic comparisons, the bike with rider must be free to roll. Remove all rolling friction such as brake drag and placing on level plates with spirit levels and adjusters to check. Therefore it is recommended that the brakes are slack and not dragging, any disc pads are pushed back a little by flexing the disc sideways a little. The tyres need not be pumped up harder, but should be on flat, level surfaces, as the inherent flexibility of the tyre allows for free movement. A spring gauge pulling against the wind will give a rough assessment of drag. It is much simpler to create an extremely minimal backwards slope so the machine and rider are just balanced to overcome whatever wheel and bearing friction remains - while checking in still air. Do not expect to use a very strong spring. Always choose a spring which will give plenty of easy measurement.
Simple bungee or string side restraints are required which must allow the spring and machine to move freely. - A long cord, a couple of tent pegs and gaffer tape will suffice. (Gaffer tape is the wide, cloth backed, super-strong adhesive tape used by roadies for sticking wiring to the stage at gigs, and for motorcyclists to get home, when the clutch cover is partiality trashed.)
Although this set-up cannot measure drag at all speeds, it nevertheless offers a good comparative assessment between the differences applied with the intention of improving a fairing. Measure the rest length of the spring, then the initial set-up, and then with the wind acting upon the spring with bike and rider. Adjust the fairing until the lowest drag on the spring is attained. Photograph from all four sides and use this to adjust or modify the intermediate form of the fairing. This is further modified by scissors, cardboard and more gaffer tape.
This set-up is good for initial cardboard and duct (gaffer) tape fairing and panel comparisons. Always rest any panels on the machine, using panels on poles to move about the machine to reduce drag. Holding the panels off the machine will give false readings if the forces on the panels are absorbed by the person holding it, rather than the machine absorbing these drag forces.
Measuring the windspeed and load on the spring can give only give an approximate drag factor. The relative adjustments allow assessment of the improvements if the wind is fairly constant, as the improvements will be mainly comparative. A long, light spring with simple pointer and scale is quite adequate, as the intention is to reduce frontal force when trying various fairing profiles and positions. If possible, plot curves for different wind speeds and compare and project as needed, (within reason). It is possible to measure the wind and plot the various drag forces. On other days, the windspeed will be different, so plotting these on a the same graph, of drag force against windspeed will give a curve. If the curve is plotted across 20, 30, 40 mph, or something approximating this, and with the common zero, then a generalised projection is possible for higher speeds.
For HPV's it is worthwhile to compare with and without a reasonable pedalling cadence across all speeds.
It is most probable that buffeting or other fluctuations occur, giving difficult readings, whereupon a simple damper unit, based on a modified and shortened cycle tyre pump can suffice. This should be free of friction by modifying the piston. It can use simple displaced air across the piston area to dampen movement, but does not work well. Water is better as it does not compress, thus preventing bounce. Water is difficult to seal without friction. Use rubber bellows around the shaft, with a small water reservoir such as a balloon to eliminate the friction of a seal. The piston needs a reasonably slack fit in the cylinder, or a couple of small holes for damping purposes. A digital sensor can give a more easily discernible difference and accurate reading, but fluctuations will make reading difficult unless it can be used in an averaged mode over readable time slices. Some lab scales can take average readings over set times, such as weighing scampering pets. Never use live animals for testing.

As can be seen, the basic approach to aerodynamics is easy, but the devil is in the detail.

Cheap wind speed measurement devices are commercially available, usually made of clear plastic using a venturi effect plus a lightweight ball in tapering measurement cone. Electronic hobbyist wind speed kits are often available from electronic outlets. Electronic anemometers are now becoming popular. Although the two shown here at the same instant gave different readings these two eventually gave identical readings, but they must be used correctly. A third was wildly different and was considered suspect as the shaft was a tad noisy.

Alternatively, a simple anemometer can be built from three small cups rotating on a small electric motor driving a volt or milliammeter with potentiometer to adjust. If using a computer fan, potentiometer and meter, remove the mounting lugs to clean up the airflow and mount on a handle with a piece of wool to correctly align into wind. Calibrate with known wind speeds at low, medium and high speeds.
A way to calibrate high speeds is to mount on a motorcycle or car and adjust at 70mph max deflection, or what is appropriate and legal, doing so in still air conditions, then calibrate the lower and intermediate speeds. Best mounted just in front of the bonnet. (Hood.) If not in still air, average the readings from runs in both directions. Always keep the same wire and connections after calibration. The more accurate methods is a magnet and reed switch, then into a frequency to voltage converter chip (LM2917), to a meter. Try to keep to 0 to 5 volt output, so the output will interface with a PC. See later. For those who enjoy using electronics, consider redesigning the airflow sensors of disused fuel injection systems, which often use a temperature compensated heated probe.

Here is the JP7c belly pan, it took six months to develop and refine. It is designed to deflect and control the radiator cooling air flow, but also tap some airflow to cool the engine compartment and exhaust system and also emplyed a low drag rear vent area for high efficiency. The final version worked well at high speed, in towns and also while starionary in traffic jams in cities in high summer.

If the ground airflow is to be studied, such as around the front wheel, belly pan or underslung fuel tank or panniers, simply duct a set of smoke risers or place cut-down smoke candles in still air and ride the machine through the risers while videoing the effects. Best done on balmy days in car park, or in a hallway. Also consider on-board smoke probes for quiet roads or disused aerodromes.
Simple stationary probes are a simple piece of plastic tubing with evenly spaced holes. Always fit a tapered strip either side of the plastic tube, so that the action of the wheels will not squirt the smoke out of the tube when it passes over the tube. This may initially seem to be used to advantage, if a sealed bag of smoke is used, so the action of the bike will squirt the smoke out at the appropriate time. But the assessment is flawed: it must never take into account the direction of the smoke moving vertically very fast, rather than being a representative part of the passive surrounding airflows.

A no-tunnel wind tunnel.
Let's take a look at just what a wind tunnel is to measure.
For experts it means speed and lots of data, for cyclists it means less effort.
The two main variables are wind speed and drag force.
Supposing we could keep the drag factor constant and measured the amount of airspeed needed to balance this drag. This would highlight an efficient aerodynamic form over a less efficient form. The higher the speed with a constant force, the better the aerodynamics.
For cyclists and motorcyclists who ride in constant condition, the force applied to push the machine along the road can be measured while riding. This force is proportional to the drag. A pressure sensor can be applied to the mid point of the top run of the chain. This can be applied in a set gear to measure the force to deflect the chain at mid point. The aerodynamics then adjusted to reduce the force applied by the rider to the drive chain at a set speed. To test the aerodynamics on the same day in the same road, in the same headwind and at the same speed, then the force will decrease if the aerodynamics are better.
Unfortunately for HPV's, there are many problems as the gearing is variable and the force applied is distinctly pulsed, so this may not work very well, if at all to any significant degree. To measure force, perhaps fit a simple pressure roller on the upper motorcycle chain run, with a sensor to tell when its applying a constant force. This can be a simple switch or a potentiometer with a readout on the dash. An alternative on a motorcycle is to fit a throttle stop which will give a constant power output in a set gear. Another method with a (slight movement between a) rubber swing arm mount is a pressure sensor or strain gauge between swing arm and frame. Therefore a constant force can then be applied to the rear wheel and this is the equivalent of a constant drag force in a wind tunnel.
This is working backwards, with a variable wind speed and a constant drag factor. The speed will therefore increase if the aerodynamics are improved and the road and wind speed is constant. At the simplest, a simple throttle stop and a clear road with different aerodynamics and plenty of cardboard and gaffer tape should be able to add five miles an hour or more at fifty miles an hour. At higher speeds, you will need to use a lot of gaffer tape, but the final higher speed may be well worth the effort for a racer.

A special device which can measure and display force applied vs roadspeed, will effectively give a comparative drag factor measuring device. Unfortunately, measuring the force applied is the real problem, as this will require sophisticated sensors in the drive train. Probably sensors similar to remote tyre pressure gauges, but inserted into the cush drive, or positioned between frame and swing arm on the drive side and then calibrated for each machine. This area is still under study by the author and may be offered as a drag factor display device for motorcycles and cycles in the near future if funding permits. Will probably include a smoke probe kit and video camera fairing mount and such like, plus documentation under the trade name of 'Pocket Wind Tunnel'. (C) J.Partridge. 2003. Many other ideas also under study.

A vast amount of aerodynamics can be studied without recourse to wind tunnels or electronics. Airflow is a fundamentally basic physical phenomena, which can be measured by equipment developed a couple of centuries ago.

Not even a formula one machine leaves the computer labs and wind tunnel ready to run perfectly. It will leave in a 'ball park' condition, but much of the refining work is done by road testing. This applies to all machines.

For higher speeds, such tests can be used on motorways or disused airfields. When testing on fast public roads, always use a video camera securely fitted to another vehicle or have a passenger operate the camera. Always be aware of possible accidents, especially when assessing lorry bow waves. Always ride in the same, safe manner for ordinary riding, allowing the study to be done only by the camera operator.

If testing side flow when cornering hard, then choose a long, smooth U bend and position the camera at the axis of the curve, with the zoom set perfectly, so the camera can move smoothly with the bike without having to adjust the focus or zoom. Preferable to pan the viewfinder to keep the edge of the frame on the front wheel, so videos transferred to computer can be sequenced in a neat manner, without any jumping between loops. If the bend is on a hill, then angle the tripod pivot to be level with the angle of the hill.
If the camera is very light, and hand-held shakes occur during motorway rides with cotton tufts, then blue tack and strap the camera onto a large block of lead or similar to dampen the movement.
As mentioned earlier, when testing on the motorway, some slimes will be affected by the cooling effects of high airflow on cooler days, so choose the materials according to the weather. If expecting rain, oil bases are better and rain is not a problem. Always take photos immediately after a slimy test run.

With this open minded approach, aerodynamic study can be easy. For example, if on holiday and suspecting a problem, or simply wish to know the airflow pattern when about to cross a mountain pass or suspension bridge, simply pull into the nearest lay-by, then mix some olive oil and colouring, gravy granules or jam (jello), then 'paint' stripes along the side of the bike, or whatever is to hand. After driving through the 'test area', video or photograph for future reference. If not at all happy about driving across the Alps on a bike striped in jam, then use a clear medium such as olive oil, or engine oil, or whatever is in the luggage, then afterwards dust with flour to highlight the airflow. Simple, ready to use and effective, but will require parking next to a stream for a good clean-up.

How the reader approaches science depends upon view point. As an example, for some people, the following list of 'slimes', according to flowrate may seem ridiculous, for others, it can work perfectly well. Brown sauce, treacle, ketchup, hairgel, jam, crunchy peanut butter.
Science can be fun, - but only if you let it !

It is hoped that the above has illustrated that aerodynamics is not high tech.
It is the interpretation of the information available which will highlight the situation. The above can reduce drag and improve overall efficiency and engine or cyclist cooling. Unfortunately the above cannot easily measure drag, which for some is the holy grail of efficiency and speed freaks.

I hope you will see that much of this stuff can be done by 11 to 16 yr olds in ordinary schools during technology classes. If there are any tech colleges or schools interested in developing their own wind tunnels please email for assistance. No school in the area of the author (B.Ed, B.Sc), of Plymouth England has shown any interest. This despite three colleges / schools purporting to be technical establishments.
Britain rightly deserves to become a second rate engineering country. Britain may eventually loose Rolls Royce and Bentley to countries which probably respect engineers. Britain can no longer build our own aircraft carriers nor trains. A sad and sick country indeed.

Having got your hands sticky and covered in fluff, your hair blown about and thus served the above apprenticeship, yet still wishing to push to the limits, then the second part of the aerodynamics programme can be applied.

Part two.

HPV cycles and motorcycles are not very large and rarely take up much room in a normal garage. For those wanting everything, then it is quite acceptable to consider building a wind tunnel in your garage or garden shed. (If room and neighbours permit).
Location is important, as the inlet air intake must be steady. The largest wind tunnel in the world has a massive intake, allowing the airflow to be smooth and controlled. This is not always possible in a domestic garage or garden shed, but a little thought can clean the intake up a little. Removal of plant pots and ladders is a must, as is a deflector if noise is a problem. Dual side opening doors are better and adaptable to give a controlled intake zone. If dual, full width folding doors are used, or beach shelter type of shielding, then this is useful.
You must only consider using the tunnel when the airflow is advantageous. Running on a still air day is ideal, but will surely upset the neighbours relaxing in peace and quiet. Therefore consider test runs on windier days, choosing the direction of the prevailing winds to work with the wind tunnel, not against it. This can be done with a fully wide open garage on a typically windy day, with a ping pong ball on a fine thread on a stick for checking airflow. Such considerations are not always possible, but anything which helps should always be considered. If wishing to be ecologically sound, then a simple wooden garden shed, covered with earth and turf to help reduce noise is quite acceptable. Please note that removing both front and rear end walls will cause a weak structure to collapse sideways, so external and minor internal bracing may be needed. Always reinforce before removing any component. - The Romans invented the arch, so use it to advantage. Covering in turf works quite well for many buildings, including the Australian Parliament building. If you find yourself lucky enough that a twenty mile an hour wind can regularly pass through your garage, then make the most of it.

Temporarily block off part of the side of the garage for the control room if you need one. This will also require removal of part of the rear wall if a side door is not available on a single entry garage. See later.
Use parts from disused double glazed patio components or similar to make the rest of the garage into the test cell. The walls inside this test cell must be fairly smooth and flush, covered with lining paper or tape where needed to make a smooth flow. Use perspex over the viewing area or disused double glazing patio parts are useful as they allow viewing from many angles and cut down noise. They are available cheaply from various sources. Cheap or discarded plywood is often available from the protective covering of large bundles of wood supplies to DIY suppliers. Supermarkets are a good supply for strong, large cardboard boxes.

Make sure there is plenty of room around the bike to allow a clean airflow, to reduce turbulence off the walls. The walls leading into and inside the test cell must be flush, covered with card or tape if needed to make a smooth flow. It may be useful to fit angled fillets over the wall to ceiling and wall to floor joins to reduce the cross sectional area in the chamber without upsetting the core airflow around the test object. The larger corner fillets which make the cross section octagonal, are also a good place to insert fluorescent tube lighting.

As intakes are usually much larger than the test cell, the cell is best designed for 'adequate' clearance of the cycle or motorcycle, which can use a smaller cross section than most wind tunnels which are for cars and larger items. To encourage clean airflow, the corners of the cells are filleted so the cross section looks like a squashed, tall octagon with wide sides and angled corners. If a few fluorescent strip lights are available, then these can be placed in the corner fillets behind perspex sheets for excellent illumination which is easy to fit and remove. The art of flat-pack has evolved greatly in the last few years and a little study pays dividends.

If electrical power distribution permits, consider an old industrial electric motor as it is quieter than an engine and more acceptable, but check the working revs and many are not up to the job of supplying 60 HP or more. For those who do not wish to annoy the neighbours, electric motors are ideal, especially if three phase supply is available. This is unlikely, unless the garage is part of an industrial area. Usually only domestic wiring is available and the power available is limited. Using domestic electric motors is only useful for low speed airflows, which is ideal for HPV cyclists, as we also enjoy the peace and quiet.
Only build up enough banks of electric fans if the domestic supply allows it. Check the main fuse rating, and don't go any nearer than 70 percent.

If only a few neighbours, or a garage without neighbours, rig up a car engine on a basic metal or brickwork stand so that its crankshaft is about half way up the rear wall. Retain and restrict the rubber engine mounts and always fit the silencer. Use a basic engine as the revs should be kept low, with plenty of torque. Bolt on an aircraft propeller from a small domestic two seater aircraft or similar. Mount direct onto the flywheel as the outer ends of the prop do the main work. Ensure clearance for the starter pinion.
If a small engine, or if using a high revving engine, such as a motorcycle or a small car engine, it is preferable to fit the propeller onto the end of the gearbox output shaft, as this allows larger, lower revving, more efficient propellers to keep tip speeds below the speed of sound. Retaining the gearbox allows the revs to be controlled for the ideal propeller speed. Simple tapped holes into the flywheel will not always suffice. If on the gearbox output shaft, a wider mounting flange will be needed. Simply build up with steel, run the mounting in top gear and apply an angle grinder to true the mounting face. If mounting on the flywheel, place bolts from the rear of the flywheel to mount the propeller and use large plates onto the wood and always use locking nuts, spaced to clear the starter motor. Use the engine as it's own lathe for centring the mounting devices. It's best just to use a basic car engine and gearbox.
Second hand, worn wooden propellers will often do. Check direction of rotation. Propellers can also be hand made, but be careful and never let anyone near when running if not certified. Always balance them. Aircraft propellers rotate much slower than car engines. It is better to have a multi blade propeller running at reasonable revs through the cars gearbox than to have a large dual blade propeller, running at lower revs, which takes up more room. This is one reason why many wind tunnels have muti blade fans. Running the propeller in the required duct also improves efficiency, but good fundamental design is best. A two or four bladed propeller mounted on the gearbox, running at mundane car crankshaft speeds is pretty close to normal propeller usage.

If the garage is small, then the engine can be situated outside, with a hole cut in the rear wall to take the propeller and a tarpaulin to protect the engine. As the machine for study is a surface device, mount the engine so the bottom of the propeller is near ground level. Then brick or block in the engine area and sound proof as much as possible. Where the wall is structural, the original Roman empire developed an excellent semi circular design of arch which suits this purpose very well. The semi circular arch also makes a good safety device to prevent the blades flying away should they break. This Roman arch can be used as a doorway as well, allowing the engine and prop to be slid or pivoted out and stored close by. If the airflow may cause excess noise, split and deflect the air upwards with a few baffles from whatever noise absorbing materials are available. Try old carpets. This deflector can also be turned upside down and used as an engine cover after use. If daft enough to use polystyrene chips to assess airflow, ensure the exit is netted to prevent the polystyrene chips from covering the neighbourhood and for re-use.

The airflow through the wind tunnel will depend upon use. If it's for HPV cycles, then forty miles an hour is usually adequate, for record breakers, a hundred miles per hour and for motorcycles, whatever is deemed suitable and safe to use. Even if the top speed of the machine in a head wind cannot be obtained, the aerodynamics at lower speeds will still improve the overall package, including drag, comfort, handling and fuel economy. The data can then be projected to the higher speeds with some degree of reliability.

Making propellers.
If making a propeller, then a fairly close airspeed is needed at the proposed revs. As most small aircraft produce around 100 hp at 120mph with max power at take off, then this is more than enough for most purposes. Having wind speeds more than this through a garage is asking for trouble, or asking for a brick or metal structure.
If making a propeller, there are many texts on the subject. They are basically simple to do with power tools but require more than just a little theory and a good piece of wood. If propellers are hand made, be careful and never let anyone near where they may disintegrate in use. The area around the propeller should be strongly ducted.
Solid wood laminate about 1/4inch, up to an inch thick planks for big propellers. The layers must be well clamped together along the whole length, and not such that inherent twist or other forces are applied by the clamps as they are bonded. The wood must be evenly clamped with minimum distortion. The wood must be well seasoned and free of any imperfections. The wood is then drilled with the centre point hole. This then allows it to be trimmed to size and checked for initial balance.
The profile and trailing and leading edges of the aerofoil are marked in pencil. Rough shaping can be done with many tools such as chain saw, angle grinder with chain saw disc, power plane, spokeshave, then belt sander or other power sanders. By the sanding stage, the balance will be important, and a snug fitting bar placed though the centre and the prop balanced on parallel edges, usually a pair of spirit levels on planks will suffice. If the centre bar is not a perfect fit, then a pair of cones are used, so the prop will always align perfectly against the mounting face, and concentrically with the centre axis. If the engine mounting boss is made, then the prop should be made to fit, and this then becomes the best centring device. The output flange of a gearbox can be welded with a larger flange to take the propeller then the prop mounted on this for balancing.
Always check the prop as it is turned during manufacture, it should just rub a marker evenly on each face of each of both blades. This is similar to checking the run-out of a bicycle wheel and will require the front and rear mounting faces to be dressed first, while on the centre bar.
The ends of the props can be epoxied for extra strength. The rest of the propeller varnished against the ravages of the weather.
Type and quality of materials is important. As larger props are heavier, or other props rotate faster, then the forces increase, requiring suitable strength materials. For slow props, where the diameter in inches multiplied by the revs is less than 170,000 then spruce can be used. Up to 210,000, then walnut, mahogany or white oak. Over this, then the wood must be birch or hickory. Up to 240,000, then the limits are reached and thicker propellers are needed. It should not be necessary to get this far.
The design of the blade depends upon the power available, the revs (usually below 3,000 rpm for good efficiency) and the airspeed. Without resorting to maths, an average propeller can have an efficiency between 50 to 95 percent. This depends upon many variables, so a closer study is often needed if the first couple of attempts are not adequate. Some fine information including the facts that a Harley 74, 20hp, will swing a 51x40 prop at 2,000rpm, resides at Propeller Articles Archive and Links at http://www.wood-carver.com/articles.html.

Just to place things into context, there is a small, homebuild single-seat, all-metal aircraft in the USA which is powered by a 20hp Briggs and Stratton V-twin lawnmover engine and flies at 120mph.

Propellers are designed with constant pitch across their length. This means that if the tip of the blade were to move through one revolution, then the angle of the tip would tend to move forwards through a set distance of air. This is caused by the angle of the blade to the airflow. Likewise, the increased pitch at mid point of the propeller blade would also need to be designed to move through the same distance. But as the inner sections of the propeller rotates through a smaller diameter, so the angle of the pitch must be larger.
As the largest path swept by the propeller is at the tip, it has the least acute angle, gradually working down through the different circumferences of the path of the blade, to maintain the pitch. Near the root of the blade, the angle of the blade is steepest. Naturally, the part of the blade near the root is running at low airspeeds, due to the small circumference, and is of a fatter, low speed aerofoil section. Whereas at the tip, the airflow will be extreme and have a high speed airfoil section. It is best to study propeller design before making them. The actual physical making of propellers is not difficult, but the devil is in the detail. Many propeller makers use the eye to get perfectly acceptable blades, and a poor blade is better than none, but only if it is safe.
Mounting the blade must be done between two strong flanges. Simply drilling two big holes into the root of the blades is not good, they must be well spaced to allow the forces to be resolved without splitting the wood. The wood must be perfect grain, not too heavy, and the blades must be beautifully balanced in weight, position, pitch and efficiency.
Most professional wind tunnels use multi bladed fans with twenty or more blades. If a set of three or more blades, then they must be chosen and mounted as balanced sets.
As this is to be a general purpose propeller, for a wide range of speeds, it should still be most efficient at max airspeed, as this is where the power will be needed. It may be necessary to replace with a larger engine, but this is straightforward scrap yard engineering. Overall efficiency is not so important like an aircraft, which is designed for maximum efficiency at best cruise speed. Wind tunnels are best designed for top speeds and can afford to be less efficient at lower speeds.
Purely for example. 100mph = 8800 ft/min. 8800fpm/3000rpm = 2.9 ft pitch (33 inch pitch). So for one turn the propeller angle will sweep through 33 inches. In half a rev, 16, in quarter turn 8 inches. This gives the builder the basic working angle for the pitch. The angle the pitch required will depend upon the particular part of the propeller. At the tip, it will be shallow, at the root, much steeper, but running at a lower speed.
If the prop is six foot diameter, (72 inches) then the tip will move through 3.14x6x12=226 inches per rev. Ten degrees will move the tip through 226 x10/360 inches=6 inches. This gives a pitch angle of 1 inch forward, for ten degree of movement at each part of the prop blade. Unfortunately, this assumes perfect efficiency. Any reasonable efficiency assumes perfect airfoil for the speeds, excellent manufacture and a host of minor attributes.
Speed of sound at 740 mph = 65248 ft/min. A six foot prop circumference = 6ft x 3.142 = 18.852 ft/rev x 3500rpm = 65982 ft/min. So around 3,300 rpm should keep things just subsonic at the tip. Speeds should be kept much lower than this for propeller efficiency.
There are many aspects to propeller design, but a simple single piece, two bladed propeller is not overly difficult. Carefully choose the best wood for as many blades as required and drill the mounting holes as one, using a jig. Always mark out the root and tip pitch, then carve with a chain saw if skilled, or power plane or saw cuts and a chisel if sensible, followed up with a power band sander or an angle grinder with sanding disc. If making separate blades, make them identical, as mentioned earlier, then weigh each, then weigh each at the tip, by resting on light weight scales, then weigh each at the root, again resting just the very end on lightweight scales such as the kitchen scales. This is then followed by final static balance checks prior to fitting to the engine. They must be balanced tip, root and whole. If a four bladed design, fit two on the mounting and balance, then add the other and balance these.

Propellers are dangerous.
Propellers in confined spaces are very dangerous. They can break off poorly welded mountings, the engine mounts must be protected from moving too far, as they are not designed to take fore/aft forces, so a compression engine mount must be fitted to resolve this force. The engine mounting frame must be securely bolted to good, solid foundations, the test rider must be protected, and the propeller mounting shaft must be secure.
Never run the propellers outside of a cage or a strong wind tunnel cowling. If the cowling is weak, then build up some metal banding or build it in brick. Then a few protection bars made from strip metal on the upstream side of the propeller.
Always make sure the 12volt feed wire to the ignition coil passes in front of the propeller, with a tensioned bullet connector, so it will easily break in an emergency should anything untoward happen. If required, a fail-safe kill switch arrangement can be connected to the upstream edge of the propeller mounting, so that initial movement from excessive play or bearing failure will also kill the engine. Running the 12volt supply wire to the kill switch, then to the engine coil, by placing it around the edge of the propeller duct so that it is easily cut in emergency will also help. The propeller should break the wire if the main bearing should fail or collapse.
Your safety is in your own hands, so a little thought and safety practice can save a load of tears. As the propeller will be making a lot of longitudinal strain on the end of the crankshaft, expect much faster rate of wear on the crankshaft and end-float shells. If the gearbox is used to get max power at lower revs, first check the shaft on which the propeller is to be mounted will be able to handle the forces generated by the propeller. If in doubt, always have the engine upstream, so any failure of thrust bearings, or failure of the gearbox end bearing will be fail safe, as the propeller will be pulling itself into the engine, not out of it. The propeller or its shaft must never be allowed to leave the engine. It should preferably be positioned to pull in towards the gearbox rather than pull out of the gearbox. Mounting a safety barrier or grille over the propeller intake is not sufficient.

An electric-start car engine is easiest, about 1000cc to 1400cc to two litre average domestic car engine from a scrap yard is cheap engineering for very little cost. A basic engine tends to rev lower for more efficient propeller use. A fancy, high revving engine is definitely not wanted, but preferably a nice, mild, quiet water-cooled motor with simple carb and a single wire ignition system. An in line engine offers less drag.
Air-cooled engines are noisy. A simple header tank in place of a radiator is applicable for most water cooled engines, which should boil off any excess heat. Water cooled engines are also quieter and will suffice for short runs.
Always fit the exhaust and preferably an even quieter exhaust system with an extra tail silencer or two, tucked away behind the rose bushes or shrubbery. A fine wire mesh prevents furry animals crawling inside and blocking the exhaust.
Rig up the engine to run off a battery. Extend the starter wire, kill switch, throttle and choke cables outside the cell. Rig the engine starter and a basic friction throttle to be simply controlled by the operator, plus one for the rider. A small lawnmower fuel tank is more than enough for most uses if gravity can deliver the fuel flow required for the carburettor. If the engine uses a mechanically operated pump, simply stick the fuel pipe in the small fuel can. (If you chose fuel injected - why?)
A simple wooden lever and piece of string will suffice for throttle and choke. A piece of string can pull a rubber banded starter cable onto the battery.

Position the engine to extract air from the garage. Extracting rather than pushing the air in, will ensure the air flow is not disturbed by the rotating propeller. To keep the airflow clean, place several long thin planks of plywood between the test cell and propeller, horizontally and vertically to prevent any swirling effect upstream in the cell. Old planks with smoothed edges will do. For low speed wind tunnels, heavy cardboard or boxes will suffice if secured with gaffer tape.
For the intake, also use thin plywood or cardboard slats to straighten the airflow into the test cell. If the local airflow is horizontal, then vertical slats are all that's needed, but if it is turbulent, then a few vertical horizontal slats will also help. A few horizontal slats prevent vertical slats from moving and flexing in higher wind speeds, while allowing the whole front slat unit to be removed for positioning the bike. Old plywood as used for protecting decent wood supplies is available from most wood suppliers and is often thrown away.
Alternatively consider a variation on stacking reconstructed cardboard boxes with tape to smooth the edges. A selection of strong cardboard boxes strongly taped together, so they will fold flat for storage. - Industrial origami.
Ideally the inlet cross section should be much larger than the test cell cross section, but is unlikely in most home wind tunnels. Ensure both inlet and exhaust airflow are clear of obstructions, with no reflective walls or other surfaces.
As the round propeller opening is not a good blend with the test cell, then where the propeller opening is, the round surface can be attached to four pieces of strong cloth, and these stretched forward as triangles to a bungee hook to make a smoother airflow at the rear of the test cell, but without loosing garage space when not in use.

Make sure the throttle lever has plenty of movement, as aircraft propellers do not rev very high, so most of the control will be at low revs, although max throttle may be needed for a well matched propeller. Rig the engine throttle with a simple friction lever controlled by the operator, using a calibrated air flow meter to adjust to the airspeed under study. As a long throttle cable is difficult to find, and the engine is static, then simply use pieces of string for throttle, choke and to touch the starter cable onto the battery for starting. Always fit ignition kill switches for both the operator and any rider or boom operator in the test cell.

If the engine is upstream of the propeller, keeping the airflow clean can be further improved by shrouding the engine with a little cardboard and gaffer tape. If the propeller is ahead of the engine, (not recommended) a spinner may be useful to cover the larger flywheel. Just a simple conical piece of card and gaffer tape, held in place by a small metal frame.

Once the wind tunnel is built, it should be tuned.
Use a ping pong ball on a light cord and pole to test and get a clean airflow in the wind tunnel.
Generally tidy up the airflow and get the engine revs neat and stable, also check the airflow into and out of the test cell. Airflow must be as clean as possible, and a smoke probe should give superb indication of any turbulence. Best to stand to one side, while holding the probe upstream on the other side of the test cell. The windspeed can now be calibrated and the best position for the airspeed indicator mounted.

To use a wind tunnel, the machine under study must be held lightly.
To keep the machine lightly constrained vertically, it can be supported in the upright position with a cross cord. A fixed spring on one side of the test cell, with the other end of the cord connected to an external spring balance or gauge will read left and right side loads on the cord. A light fisherman's spring scale will suffice, to graphically highlight any offset pressures, as common with cycles. If cyclic loads such as pedalling will confuse the side load readings, then a heavy pendulum can highlight the overall side loads.

The rider and machine must be set-up to balance perfectly vertical in still air, then fixed to the cross cord. This is easily done with gaffer tape over a large knot or loop in the cord, taped across the cycle frame or motorcycle fuel tank. For a cycle frame, the cord can have one turn around the frame tube, then adjusted for balance and taped. To ensure the steering does not upset the balance which can cause the rider to compensate while assessing asymmetrical airflow, then rider and machine will require setting up prior to a run. Ensure the rider and machine are perfectly balanced upright in still air and the steering remains neutral during the test. Use an elastic cord or bungee stretched between rear and front wheel to maintain the front wheel in the straight ahead position, otherwise changes in steering due to the rake angle can upset the side balance. Rest the riders hands lightly on the handlebars. Get the rider perfectly balanced, then tape the cross cord in position.

Alternatively, when testing handling in winds, a HPV cycle can be kept upright by riding it on rollers, but this will not allow easy assessment of drag, unless the ground roller assembly is faired flush and mounted on low friction device.
Mounting a rider on rollers can be difficult, but a set of small pivots, or the whole suspended on cords from the side walls, or simply on three or four marbles on flat plates. Whatever is used, it must have zero fore-aft friction to allow a good reading of the drag. If there is drag from the roller, then calibrate the assembly without the rider or machine first, then eliminate the drag of the roller assembly from the readings at various windspeeds. This is also possible for checking ground airflow of motorcycles at high speed, but best used only for studies of belly pans and very low motorcycles such as recumbents like the JP4, 5, 7 and 8 series.

When checking ground airflow, such as belly pans or underslung fuel tanks or panniers, it is preferable to use rollers on the wheels and a fabric dummy rolling road surface driven by the rollers, with the front wheel directly above a free roller under the fabric. For HPV's an old running machine may suffice, but may not last long at the higher speeds.

The main purpose of riding while balancing on rollers is to assess any upsetting influences from the fairing or other items affecting the airflow which cannot be assessed with a restrained machine. This may include belly pan aerodynamics, but usually means checking the cooling airflow around any low radiators from front wheel and road turbulence or the turbulence from pedalling with high speed recumbents.

The machine must be free to roll in the direction of the airflow. Therefore it is recommended that if the test stage is not on friction free rollers, then the brakes are slack and not dragging, any disc pads are pushed back a little by flexing the disc sideways a little. The tyres need not be pumped up harder, but should be on flat surfaces. The flexibility of the tyre allows for free movement.
Adding a loose safety chain bolted to the floor and around the wheel so the bike does not migrate too close to the propeller is also recommended.

Once the machine is delicately set-up and balanced neutrally in the test cell, the drag can be measured.
To measure drag, the machine must be allowed to roll freely in the airflow on a perfectly level floor, constrained only by a light cable to a remote spring balance to measure the drag in pounds or kilograms. Where spring gauges are not sensitive enough, simply use longer or softer springs or levers to multiply the forces, but do not allow too much movement on the machine to be measured. If the propeller looks menacing, use a slightly slack, secondary restraining metal anchor wire or a metal chain. Gradual modification of the machine should then be able to reduce the drag factor across a wide range of speeds. There may be a small step in the graph before readings are measured, caused by the friction in the wheels and bearings. See also above.

As the readings are made, have a table of intended data to be captured. Writing it in pencil on a pre-printed sheet saves time and ensures all data is captured. To gain a vast amount of data to smooth the graphs, then the spring and windspeed sensor should be integrated using a sensor to a computer, to allow real time data to be collected. This will give a set of readings for each test run for windspeed, drag, axle loadings and whatever is needed to be recorded as a set for analysis later. A large collection of well spread data will tend to give a better overall view of the parameters under study. For digital data capture, see below.

For more interesting studies, or to understand the overall balance of the machine during high speeds, a pair of bathroom scales can be mounted flush with the floor, or built up to reduce turbulence with card and tape. As most wind tunnels reduce in cross section over the test cell, then simply use this to advantage. Scales study how the weight balance of the axle loading changes with speed. Most garage floors are fairly level, needing just a little work along the centre line to give a good floor area for the wheels or two identical scales. If the floor does not offer a smooth airflow, simply cover with some lining paper or yet more cardboard boxes. Be careful when using scales during drag measurements as they can create an inaccurate floor level with subsequently incorrect drag readings. Strain gauge scales with minimal movement can eliminate this problem. Digital scales allow the readouts to be extended to outside the cell and make study much easier. For heavy loads such as bike weight, use bathroom scales.
Where a low load digital read out is needed, such as the front wheel of a recumbent, consider modifying kitchen scales. If wanting to measure fluctuating forces, then more fast acting sensors will be needed.
Always calibrate prior to the run.

As can be seen, there is no need to buy expensive equipment.

Understand what is being measured, and the degree of calibration required. If building a fighter aircraft or formula one with unlimited budget, then shop 'til you drop, otherwise, use sensible choices of equipment suitable for the purpose.

Computers.
(A free computer will suffice - 286 to 486 without hard drive etc.)
Where data capture is to be automatic then mechanical graph plotters are possible, but need a high degree of mechanical skills and tinkering to get reasonable accuracy. Computers are gradually removing the skills of the watchmaker from engineering.
The simplest and cheapest data capture device is now the humble discarded personal computer. An old IBM compatible will suffice, with standard game port which is usually on the sound card. Yes, anything from a 386 with 1 mb ram, game port and just a floppy will suffice. You don't even have to have a hard drive or the rest of the sound card working to use the game port.

If the computer is a bit more modern, thus able to take a cheap video capture card or accept digital camera video sequences. A reasonable specification is needed for video capture. Otherwise just a basic PC with floppy drive, minimal monitor, printer, game port and a couple of joysticks is all that's needed.
Minimal DOS, Qbasic and a programme or two can fit onto, and run from a single 720 floppy !
This should cost nothing, as this stuff is regularly discarded. I still stand by my ethos of making wind tunnels for next to nothing.

Getting the data into the computer is fundamentally simple, but the devil is in the detail. The art of making sensors is a case of modifying what is available. There is a lot of test equipment available at very high prices, but a lot more at incredibly low prices, or for free - if you keep your mind and eyes open.
Start by deconstucting an old games joystick and checking the resistance of the joystick pots, then use a simple linear potentiometer fitted across a spring to measure the drag or whatever is to be measured. At the most simple, use the joystick directly as a spring loaded lever to a tension cord. Joysticks do not use the whole of the potentiometer track, and also have abysmal pivots. Deconstruction is highly recommended. It is simple to use home made interface boards for connecting to the joystick controller port, but the standard lead is usually quite acceptable. Old analogue joysticks are two a penny.

Specialist wind tunnel kit is unnecessary, wastes money, takes time to set-up and learn, and needs backup when it goes wrong.
Home made interfaces are more easily adaptable and allow channels to be modified and adapted with minimal cost. For those who cannot understand the philosophy of alternative engineering, then specialist equipment is also available for silly money. Alternatively, the nice folks at Pico make a nice simple and cost effective unit for schools called DrDAQ and includes software. They also have an eleven channel analogue to digital unit with software for data-logging and such like to make a really profession set-up for about a hundred quid. (Nice one Pico.)

On most personal computers, and all IBM 8086 / 286 / 386 / 486 and compatible computers since the original back in the early eighties, two joysticks are possible. Therefore, four analogue channels are already available and waiting for measuring drag and airspeed, and two other analogue sensors, plus four buttons are available. The usual potentiometers are 100k ohm linear.
Linear means the resistance of the resistive track had a constant variation across the movement. Logarithmic potentiometers should not be used as they have their resistance relative to position in a logarithmic variance. Linear potentiometers are available in many forms and can be straight or rotary, or even multi turn if required. Also possible are light sensitive devices, which will need extra circuitry to be used.
Joystick ports can supply enough current to fry small wires and permanently damage the joystick port. So go carefully at first, placing a minimum of a 100 ohm, 1/4W resistor in the line of anything that is fitted. If in doubt, use a cheap disposable sound card with a joystick port.

Simple BASIC computer programmes will suffice for the reading and storing of data. Generate the data as sets, to measure each input and to store them in a usable format, followed by a short wait so the data flow rate is sensible, then repeat.

To prove that computing is extremely simple, search out a copy of DOS and Qbasic. Qbasic came free with Dos 5. For the simplest interfacing of joysticks, use the Qbasic 'stick' and 'strig' functions. Qbasic with DOS can be run from a 720 kb floppy, with plenty of room to spare.
Make the floppy as a boot disk, then add Qbasic. When booted, type 'qbasic' to run. Primitive on-screen help is available, type 'help'.

Stick and strig returns simple values of up to 256 divisions, so is not ideal, but often suffices. Writing the software is kids stuff. As an example, the following programme is simple Qbasic. This is done using the examples shown in the Qbasic help files. Other programmes offer up to 65,000 divisions and are given at the end of the monograph. More complex progs for wind tunnels will be posted on the website updates should they be requested.

Simple BASIC programme for reading joysticks.
Simply load the programme into Qbasic, then run.
The first few runs will be to calibrate the sensors and to check they offer sensible values. They can be calibrated at home while building the sensors.

An old computer with a single floppy, a cut down veriosn of Dos5 and Qbasic is all you need. It is data you are after, not posing science or overinflated science budgets.

____________________________

REM prints out joystick values while joystick button is pressed.
ON STRIG(0) GOSUB toscreen
STRIG(0) ON
PRINT "Press Esc to exit."
DO UNTIL INKEY$ = CHR$(27): LOOP
END
toscreen:
temp% = STICK(0)
PRINT "S1 "; STICK(0), "S2 "; STICK(1), "S3 "; STICK(2), "S4 "; STICK(3)
REM use LPRINT for hard copy
FOR j = 1 TO 200: NEXT j: REM delay loop adjust as needed for sensible print speed
RETURN

________________________

This prints out the sensor readings during a run and 'esc' stops the print.
The above simple programme displays the data only when the button is pressed, allowing for a series of data only when required. Such a scenario is one press for calibration before starting the wind tunnel, then press for a few readings at ten mph, then another set at twenty mph, then thirty mph etc. In this manner, the steps in the readouts will be easier to read. There are many options, such as using one of the joystick buttons to enable the test rider to start the data run when everything seems settled, or perhaps during a problem scenario.

Many may just want to measure the drag, but with four channels, you can include the airspeed. The other channels can be used for anything, perhaps you may wish to measure the weight change on the front wheel, or measure the bow wave pressure, or the airflow through the radiator.

Adding a suitable storage format code will allow inputting to some spreadsheets or databases to allow graphs and charts to be automatically generated.
Raw data can be calibrated into force, airflow and such like, using the sensor and a test load for calibrating the input data. If needed, raw data can be converted directly into airspeed or kilograms or whatever is being assessed during the test run. This allows easier analysis if the sensors are reliable and rarely go out of adjustment.
Preferably take separate calibration data before and after a session, checking the zero, mid and max readings, so the sensor readings are safe and accurate. If controlled by the rider, the various joystick buttons allows a choice of programmes to be chosen, perhaps a simple calibration run, with the calibration loads in place, a standard test, a fast test and a slow test, using different timing variables.

Sensors.
From the outset, decide what is being measured. Is it objective or subjective, comparative or relative, as wind tunnels can be used in many ways.
All sensors will need to be calibrated, but may drift out of adjustment from wear, poor components or many other reasons. Regular calibration is recommended.
For those who get really accurate, then temperature, humidity and other variables are also possible to refine the readings. Again, simply buy a digital clock, barometer and humidity probe. They are surprisingly cheap and some cheap domestic items are supplied with more than one probe, so the test bench can tell temp, humidity, and the time as well. As interfacing them to the computer is problematic, simply do what everyone else does, write the date, time, humidity and air temp at the top of each run in pencil !

The wind speed sensor, as mentioned earlier, can be a simple venturi device. For direct connection to a computer, the joystick will detect a voltage between 0 and 5 volts which is difficult to get from such a device, nor easy to refine. The voltage output can be adjusted, but usually is not enough for full deflection. Therefore a frequency to voltage chip is recommended. Where an airspeed rotary potentiometer is to be used direct from the joystick, then a very basic square tube with gravity controlled flap acting on a rotary potentiometer may suffice but is not recommended. If using this, then consider a rotary logarithmic potentiometer.

The classic drag sensor is a spring.
This gives an increasing force with linear movement relative to the load. As drag is measured in pressure, this directly translates to lb pounds force or Kg.
The main problem with some sensors is that a large force is to be measured across a small part of the load, such as 10 lbs +/-1 lbs. Ways around this are to have a sensor with a preload of, for example, about 70 percent of the full measurement, and the sensor starts reading from about 70 percent to 130 percent of the mid value. This eliminates much of the excess potentiometer movement, to give much more accurate reading across the range of the potentiometer movement. Placing this on the test machine, then pulling with a fisherman's scale to simulate drag, the sensor can be calibrated on the computer prior to running up the propeller, and any notes added.

As the various forces often increase with speed, then a number of sensors are normally used for the range expected. A pair of wires to the drag sensor can be easily connected to a selection of gauges by using simple two pin connectors. Poor connectors can cause resistance, so use reasonable items. Having the drag measuring cable positioned next to the observer, allows sensors to be easily swapped without stopping the test run. This is particularly useful for motorcycles, where runs at many different speeds may need two or three preloaded sensors.
For early experiments, it is simpler and better to make an adjustable sensor, which allows the start of the drag to be adjusted by preloading the spring with a threaded adjuster. Use of elastic tends to give inaccurate readings unless accurately set-up. Metal springs tend to be constant if kept within their elastic limits.

The range of forces on a cycle at ten mph is small, at around 2 lbs of force, and at thirty mph, up to about twenty five pound force. For motorcycles, this increases dramatically, as at a hundred miles an hour, the drag on a level road in still air is around twenty to forty horse power, depending upon efficiency and cross sectional areas of bike and rider, so the drag will vary vastly. This will require much higher spring gauges and require initial test gauges across a wide selection of spring rates. For initial drag checks at the proposed speeds, simply use a cable to the bike. The cable leads outside the test cell and is gradually loaded until it balances. This can then be weighed and sensors made to suit. Initial assessment done by a large mechanical spring scale as used by fishermen to weigh their catches. If you don't think 20 horsepower is enough for a hundred miles an hour, then just to place things into context, there is a small, homebuild single-seat, all-metal aircraft in the USA which is powered by a 20hp Briggs and Stratton V-twin lawnmover engine and flies at 120mph. Admittedly it does not have the drag of wheels, but neither do road vehicles have the excess drag of a set of wings and control surfaces.

The drag sensor is totally arbitrary, as shown in the picture: A scale, but no values. In this case, it was set up to read percentages of a 20 Newton reference load, but unless noted, can be used to read anything, - from bags of peanuts or feathers to elephants. It can be mechanical or as increasingly popular, electronic in actuation, but either will do. Such items are simply a measuring device to give a reading of the sensor applied. If needed, simply apply the overall load, such as 20 Newtons, then mark the edge of the gauge with a pencil, so that the reading and a few strokes with a calculator will give the actual drag loads.
To load a drag sensor in situ, simply make up a small frame with a roller. Then using a piece of string, hang your reference weight over it, to pull on the sensor, under the influence of gravity.

In most cases, simply reducing the overall drag is the primary aim of the test and only needs relative indication or measurements to show when drag is being reduced.

To make a sensor, buy a selection of potentiometers which match the joystick. Almost anything will do.
As a flap windspeed sensor tends to give a geometric output, decreasing the movement with speed, then a log potentiometer may be preferred.
For most other sensors, linear potentiometers are often used, to match the action of springs against which most forces are calibrated. Buy a selection of springs, either tension or compression, then use a linear scale reading potentiometer. Springs can be reshaped and still retain a fair degree of acceptable, constant and reliable movement. Springs with constant pitch between turns act in a linear manner. If the coils are gradually changing in pitch or diameter, the spring is not linear, and the sensor must be calibrated very carefully.
Rotary potentiometers are possible, which can be fitted to a drum or pulley, upon which the tension cord is fitted. This allows a tension spring or a torsion coil spring to be mounted, which can be pre-tensioned to a wide range.
A ceramic potentiometer is more accurate, as the cheaper carbon track type are built down to a price. Cheap carbon track potentiometers can be deconstructed and the tracks reprofiled with a knife to remove some of the outer track, or built up with carbon pencil to modify the characteristics of the output, but this is rare.
Any damping of the reading is best done by the mass of the machine under study, possibly with a silicone greased bar inside a tube as a slow damper. There must be no effective friction preventing a slow movement, but damped enough for movement to be controlled. See also water filled bicycle pump, mentioned earlier. Silicone grease around the shaft of a rotary potentiometer often works well.
The simplest sensor is a spring with linear pot beside it, with one end fitted to the restraining bracket, the other end of the spring to a cord, and it's deflection measured by the potentiometer wiper arm. Always ensure the spring can't move further than the wiper arm. This is easily constrained by a piece of string or wire, which must not upset the action.

Once the sensor is made, it can be simply calibrated by applying a load, reading the first deflection of the track and weighing the force, possibly using a weight on a bathroom scale to check intermediate points if suspect. For drag, the sensor is deflected by a set weight over a simple pulley and the potentiometer reading then calibrated.
Then run the joystick programme, a few clicks of the button and mark the zero load applied on the print-out for reference. Then again at the other end of the full load on the potentiometer and few more clicks and write the load next to the reading.

Drag is a relative force, so a single sensor will do the job of informing the tester whether the drag is lowering or increasing. It is when the graph is compared across various speeds and sensors that accuracy becomes relatively important.

Computers are not mandatory.
Where a very sensitive, purely visual analogue device is needed, then simply fit a pivot on the moving cord and fit a cheap key-ring laser on the pivot to play its beam on the floor or a far wall. The wall can be marked with chalk of the measured reference values, to give a very large scale. This is ideal when adjusting riding position on a conventional machine, to instantly see the effects of tucking in the head, knees and elbows. This allows the rider to adjust themselves by looking at the 'display' and give instant readings and constantly adjust rider position to get the lowest possible drag in a very short test run.
Lotus modified Mr Boardman on the Olympic wining composite cycle design, then used a wind tunnel to check. Having the rider in the tunnel, with the rider being able to directly see when the minimum drag is happening, will allow even faster assessments and adjustments without the need to do too many test runs. It would be even better if Mr Boardman were to be able to adjust himself to give the lowest readings by simply looking at the sensor. (Perhaps it is a case of wind tunnels needing to make money per hour of use. But I would far prefer a wind tunnel to run for just a few minutes at a time, especially in a built up, or quiet area.)

Drag is an objective and easily measured force, but can also be used in a relative way to best effect. The intention is to reduce the force during the test session, so as a relative force, then a simple sensor will do the job of informing the tester whether the drag is lowering or increasing.
The lower the drag, the faster the machine can go.
If your wind tunnel can only manage 60 mph, but you want 100 mph, then plot 20, 40, 50, 60, then predict the rest. It will not be guaranteed, but gives a ball park of the possibilities, especially if the airflow regime over the machine is well controlled and gives a smooth curve.

Wind tunnels are often run for only a few minutes, then the results studied. This keeps the neighbours happy.

Not all assessments are from sensors.
Cotton tufts, or slime can be used. Small openings upstream in a narrow test cell allow smoke probes to be used and videoed. The video can then be transferred to computer for assessment. Always try to video from many different points, including side, front and top. It is often quite acceptable to have an assistant leaning against the opposite wall with the smoke probe, if the test cell is not too cramped, and does not upset the area under study. A long smoke probe can be played across the whole of a machine from a small hole in the test cell wall.
For those wishing to, a small remote camera, such as the guts of a web cam can be fixed to the smoke probe for a closer look downstream and linked directly to a computer to grab stills or video sequences. Never allow a webcam to upset the airflow around the smoke probe exit.
The advantage of using a camera is that the neighbours are not upset when studying the information at leisure. Professionals with wind tunnels away from the public can run wind tunnels whenever they wish, but in a suburban English garage with rose bushes, this is not possible.

Much time is taken simply studying the information, so borrowing a video camera or suitable digital stills camera which can take animated sequences is highly recommended.
A cheap option is to use a video capture card for a computer. The costs of buying or borrowing a cheap second hand computer and a budget video capture card are much less than a new video camera. Video capture cards start from 25 quid, but need a camera, so an old USB video camera can be used, as they are often thrown away, or left collecting dust.
The video can then be copied to computer for lengthy analysis. Analysis direct from a video will soon damage a camera or player. Any 150mhz computer is adequate for collecting many still pictures and an optimised 400mhz machine will suffice for video, but have a reasonably large hard drive. If poor like me, consider using video compression or using grey scale if colour is not important.

Real engineering is doing for pennies, what an 'expert' takes thousands of pounds to accomplish.

Video can be transferred to computer for closer study. When videoing, always hold the video in each position for a few seconds to get a decent sequence. If using a hand held video camera in the cell, then using zoom will allow less disturbance of the airflow around the machine. If using a web cam, strip the outer covering and reprofile it to reduce turbulence if mounted on a probe.
Always work out what is to be videoed before use, and then add further sequences in light of feedback and experience.

When setting up important components such as a fairing, such items can be flexibly positioned on the machine and adjusted externally with fine cords until an optimum position is created. This allows a real time assessment of the fairing to be refined without stop start sessions.
Allowing the rider to make the adjustments will eliminate all other spurious inputs such as external forces tugging on the panels. The rider can also make body and panel adjustments while watching the drag display. Always try to make aerodynamic components adaptable in real time to save time and to get direct and thus more subtle assessments and feedback. See also Yamaha Morpho and my monograph on adpative aerodynamics developed for the JP series.

It will surely be quite laughable to look back on this computer specification in the next few years. The BBC B computer is still greatly missed for such jobs.
But always, as mentioned earlier, remember what information is being gathered, then choose the appropriate equipment for the job. It does not cost money, just common sense. (See 2005 Reith lecture, then compare with the far truer 2005 Dimbleby lecture both c/o BBC Radio 4.)

Pennies vs experts: Contrary to popular opinion, a wind tunnel can be cheap and is often recommended for those who live in the country. Recycled patio door and window components, discarded plywood and hardboard and an old car engine and simple welded engine mounting are fairly cheap engineering. The test cell and ducting can be stored flat against the wall or in the roof when not in use. The engine and prop can be slid to one side when not in use and covered with the rear deflector box. You may want to include a few plant pot shelves to disguise it when not in use. ( What, me ?, a wind tunnel in my garage, here in Surbiton ? )

The controls are all easy to extend and easily mounted on a plank or desk outside the test cell, then easily adjusted and read during the test. If videoed, most tests should only last a short while to keep the neighbours happy or confused. A simple table-top control system should be able to stop and start the engine, adjust wind speed via a friction throttle, read the windspeed, drag, side loads, front and rear axle loadings. Digital videoing with a cheap TV capture card can be recorded in grey scale for those who use an older computer. A second engine kill switch can be positioned on the test machine for the rider, also a head phone with boom microphone, cheaply available for most computers. This can be extended to headphones for the boom operator and controller if noise is a problem during a run. Rider feedback via a speaker and microphone will help highlight unusual experiences which may need closer study, such as a problems at a particular speed. Also record the chat for later reference, if it does not upset the data collection of a old computer, otherwise use an old tape recorder.

Although not perfected yet, for those who may wish, consider using a cheap laser and a more generalised smoke area for study. Then scan the machines surface with a laser which can give a planar effect, possibly with an oscillating mirror or a beam splitter, or simply oscillating the laser fast enough to give a sense of being a flat plane. This may require removal of the body, to oscillate just the laser chip mounting. A cheap laser keyring can cost just a few quid, or free if the owner has dead batteries. Use an oscillation frequency which will scan the area fully for each frame of the camera. Mounting six small mirrors to a nut, then spinning it in the laser path will work very well if the mirrors are well aligned. The laser should create a visibly vertical 'slice' through the smoke, if not too dense. Adjust the density of the smoke by moving the probe relative to the laser scan. The highlighted scan can highlight the various pressure zones, where the generalised smoke area is denser or less compressed. Using a video or web cam to record the slices as it moves along the machine can give sequenced info which can be run on a computer as animated sequence. Again, this is not expensive, not even a wind tunnel is needed, just a windy day, a video camera, a cheap laser and suitable smoke. It doesn't even need a smoke probe, just an area of evenly dispersed amount smoke or mist. If in a wind tunnel, where the area of smoke should not be disturbed by a person in the test cell, then preferably fit the laser and camera on simple rollers, (with the oscillations isolated from the camera), then slide it along a wire strung clear of the machine, drawn by string. If the wire is gently angled upwards, into the wind, then gravity and airflow will allow the string to move the camera and laser run back and forth across the areas of interest.
If re-directing the engine exhaust with two stroke oil injection into the exhaust to create smoke, ensure the rider has fresh air breathing equipment, such as a snorkel mouth piece and some plumbing.

The careful study of aerodynamics is a big subject, so the builder must find a useful balance between understanding all the theory or simply refining a working design.

Getting the drag low, or the cooling airflow through the radiator is basic stuff, but fraught with variables. The first step is simple guess work, based on the wide range of the better designs of others. See companion monographs for making rider and machine aerodynamically efficient before making aerodynamic shells.

If making a fairing from scratch, simple cardboard and gaffer tape is the first step. This is then followed by simple wind tunnel tests to adapt the fundamental shape to the lowest drag or whatever is needed. Simple modifications using adjustable cardboard shapes while in a test wind will give constant hands on feedback. If doing so in high windspeeds, then this must be done in stages, as cardboard easily flies out of the hands in high winds. Running across Dartmoor in high winds keeps you fit.

Once the basic shape is made, this is then refined by smoothing the profile and modifying the inside ducts and other air passages. This is important for cooling and rider access and in various riding positions. A racing motorcycle may enjoy high speeds in the tuck position, but would prefer to slow down really fast when the head and torso are raised above the fairing, just before a sharp bend. Applying this airbrake force to the rear wheel to prevent it skipping can also help. (Unless wanting to drift or 'wind up' the rear wheel for setting up into a racing corner.)

Applying liberal amounts of modelling clay, a piece of wool on a stick and a good scraper and later a smoke probe can work in concert with the wind tunnel to make for fast and efficient aerodynamic adjustments of the finer aspects of the airflow. This is important for reducing turbulent areas, increasing and wanted down forces, refining the airscoops for engine cooling and other ducting. This can allow the negative pressure areas to draw airflow through the radiators or any of the many thousands of other variables which make for an efficient design. Using a smoke probe and a bucket of modelling clay and a scraper keeps run times down, to minimise annoyance to the neighbours. If there are no neighbours, then an afternoon will need warm clothes and gloves, but lead to a very refined machine. If an afternoons work is likely to use lots of smoke, consider making a large smoke chamber from many strong polythene bags, similar to a small hot air balloon. Once the pyrotechnics are extinguished, a light load such as a blanket can be placed over the massive poly bag to pressurise the container, and a simple clamp used to apply smoke only when needed. Stand by with the sticky tape to block any (easily seen) holes in the bin bags.

Doing other things with wind tunnels.
It is often assumed that a wind tunnel is used for optimising the aerodynamics at high speed. For most people, this is as far as they think.
As a cycle has to pass busses and traffic in general, then it is possible to simulate such occurrences where the airflow is less than ideal. If building a machine for commuting, then take the opportunity to simulate these situations. If is not difficult to simulate the bow wave of a lorry or bus in a wind tunnel, or any other upsetting influences.
Even if the machine is not to be modified, the designer can test for any unusual effects, then modify if unwarranted problems are apparent. When placed on simple rollers, it also allows the test riders to assess the general handling properties in such situations prior to road testing. This can be particularly important if developing a fully faired design, or refining steering geometry.
ALWAYS take the time at the end of a run to use imagination, try to upset the machines airflow or handling, to see where potential problems may lie.

A wind tunnel is also of interest to model aircraft and kite builders, speed skaters, skiers and many others. This may or may not be a good thing, and will depend upon one's neighbours. Other applications are as mentioned above, regarding the ability to test elsewhere.

Scale models.
Not all aerodynamics in on a high, windy ridge, during normal use or in a wind tunnel.
Much of the initial study can be done at home in the bath. Small scales can be useful for initial development of unusual or high speed machines, where the initial testing of potential designs cannot be done using real machines. This is important where the whole machine and fundamental chassis design are awaiting a reasonably close approximation of the final aerodynamic form.

For example, a final selection of models such as for the fully enclosed JP7, is shown opposite in two blue foam (skinny and fat) and a plasticine model. Plasticiene is commonly available, kiddies waterproof mouldable modelling clay, ideal for initial moulding and modifications as the early tests underwater are gradually modified to reduce problems and improve general airflow.
The low mass, blue foam models were then used to assess pressure effects on the shell, to see if the front end dipped too far under bow pressure, or if the tail causes snaking. The lighter models are mounted on pivots to allow them to move slightly, (I used various piano wire mountings which flexed slightly) to highlight the effects of the airflow, used to asses the centres of pressure, both forward and from side winds.
In water, use flakes for general airflow around the machine and striped slime on the model to show the airflow over the surface of the machine. If you have ever taken a class of 12 year olds in making paper, then you know how flakes are made in water.
Underwater testing is done head on, perhaps with a sideways component to mimic the various airflows at speeds. again, use a camera to assess at leisure. When the airflow is about right, then this is tested for balance effects, such as buffeting, and pressure zones, using basic forms which can be modelled in rigid foam and mounted on pins so they can show the effects of the pressures on the model, by deflecting it.
If you end up doing a lot of underwater testing, then borrow a couple lengths of rails tracks from your kiddies train set, and blue tack these to the side of the bath, then make a wooden frame mounted on a couple of rail trucks for an instant mini hydro test tank, as used for testing ships hulls and such like.

From this data, the designer can build up a working picture of the overall effect likely to be found in real use, but without the hassle for major redesign later in the project. This data will not be accurate and can only be considered as a general representation of the effects. But initial tests using models can save a massive amount of time, effort and cost in the long run, by allowing easy testing at home.

Scale models can offer some information in a stream of air, but this is not perfectly applicable for many physical reasons. Models of eighth size or less are much better tested in a denser fluid than air, and models can be easily carved in plasticine, then lighter versions in rigid foam such as blue foam, (see composites) and can then be initially tested through a test tank or bath. An initial alternative is plasticiene or waterproof modelling clay, which then allows fast modification of the initial shape to greatly reduce drag while testing in the bath.
Water tanks give a basic understanding of the way the machine will behave and help decide the most effective choices or allow modifications to be done before starting a full size build.
If many shapes are chosen then such assessments will highlight the low drag versions far easier and with minimal hassle. Simple assessments can be followed by easy modifications with plasticine and a bath of water.
Then run the shape though the denser fluid using rails and a simple force sensor such as a bendy bar to assess the drag, until the best design is found, and then drag is then further improved by refining of the preferred design(s).

The behaviour is not only the flow around the machine, but also the pressures and their directions acting upon the model. It is often important to include internal airflows such as the cooling flow of radiators. This may require holes to be carved so the flow is reasonably accurate. For greater accuracy, the model should be run along the bottom of the bath, to represent the road surface. This will only require twice the height of water of the model's height. Having the model just below the surface will give false readings.

The equipment is not complex, just a simple plank along the bath, with a smooth roller or slider. A firm metal rod is attached to support the scale model. For basic drag, the model should run along a straight path. For a neat, temporary rail system, simply use two lengths of scale train rails and a couple of trucks or carriages with a plank across the bath between them. To allow close camera work through the flecks, and minimal bending of the rail support, fill the bath fully.
Turning the model at an angle to mimic side wind buffeting, or by using a block to represent a lorry in front of the model. This is not always a good representation of the true situation on the road, but can give useful insights to the design. Mounting the model so it can rotate mid point between the wheelbase will give an indication of side wind pressure effects and how the front to rear will react. It really is not rocket science.
To assess the centre of pressure, the model can be pivoted to allow the relative pivot positions to be adjusted at set speeds so the effects of drag can highlight the lift or compression of the front and rear. This is particularly useful where a handful of similar models are available, allowing back to back testing. For good assessment of a series of similar designs, it may also be useful to pin pairs of models on the rig for back to back assessments, and all the models at the same effective point, so they all act equally to the drag effects.
The easiest alternative for single and multiple designs is to mount the support arm on a pivot far behind the model, on thin piano wire, allowing the arm to flex on a soft rubber bush. Position the model so it is aligned correctly to the direction of travel while in the water. A small steel rod is easily bent to the appropriate angle, but a foam model will tend to float if not neutrally balanced in the water. A little lead wrapped around the support arm can be slid up or down the support arm may be needed to balance various models. This will highlight the effects in both vertical and horizontal planes without unduly modifying the orientation of the model. Relative drag can be assessed by employing kitchen scales into the linkage. Always compare drag of models at accurately set speeds. This can be done with a metronome and set spacings along the track if you want to get really accurate.
The rig and model is then pushed steadily through the water while the flow is studied. The pressure (relative drag) on the rig at set speeds can also be measured.

To see the flow, place lots of small neutral density balls, such as balatini balls or preferably find a piece of plastic which just about neutral in water then shave flakes off it using a cheese grater, or use flakes of shredded paper plus a good stir to allow the flow to be observed. Using tissue paper shredded in a food mixer may cause too many particles which prevents visibility. Always use discreet pieces of paper which will not turn the water into a murky soup. Allow the water to become still prior to a run. The particles will highlight the flow over a small scale design. The relative speeds and densities are problematic, requiring some maths, but the overall effects are often useful for studying initial designs. As water if far denser then air, the speeds used can be a great deal slower. Because the model should be at least its own height below the surface of the water, visibility will be difficult, so it is important to get the density of the particles not too dense so that it can be easily seen. As mentioned later, this can be improved with a video or a web cam attached to the rig, viewing from just above the water, or fit a waterproof camera level with the model, with the flow separated from the model by a flat screen, or use a clear sided long tank made from old planks and a few long sheets of clear plastic.
Before draining, use a kiddies shrimp net to remove the particles, or place a wire or cloth screen over the plug hole so the plumbing does not get blocked. For easiest drainage and to keep the particles, I find that a long cloth screen or towel can be slid slowly over the plug hole so the water can flow more easily.

As a lot of particles are needed in a tank and visibility can be a problem, this may not be appreciated by some people. The model can be painted with an emulsion which will spread across the surface of the model. This will highlight the surface flow paths, but cannot show the overall airflow, which requires particles. If the test tank is disposable, then ink can be used. Never use inks in a white enamelled bath. By placing a fine nozzle ahead of the model, ink made from various sources such as unwanted ink or thinned poster paint can be allowed to leave a line in the water ahead of the model, similar to the smoke probe. Placing the ink tank at water level or adding a piece of foam in the line will reduce the need for a control tap or ink bung and allow the nozzle to introduce the ink as required.

Placing a simple manometer at various positions on the model will indicate relative pressure zones and give an indication of how these can be reduced or modified to requirements. It is very rare to design a high pressure zone, except for stabilise high speed stability of a