THE TECHNICAL JOURNAL OF THE IHPVA

Vol. 7 No. 4 $5.00 KUE Summer, 1989

The physiological limits of long-duration human powerproduction-lessons learned from the Daedalus projectby Steven R. Bussolari and Ethan R. NadelINTRODUCTION

In 1985, a team of engineers at theMassachusetts Institute of Technologybegan a three-year research program withthe goal of flying an advanced human-powered aircraft a distance of 119kilometers from the Greek island of Crete.The project was named Daedalus, incelebration of the mythical Greekinventor who constructed wings of waxand feathers with which he flew fromimprisonment some 3500 years ago. Theproposed distance of the Daedalus flightwas rather ambitious, considering theexisting record for human-powered flightwas 35 km, set in 1979 by Bryan Allen,flying the Gossamer Albatross across theEnglish Channel. Preliminary estimates ofthe Daedalus aircraft performanceindicated that the flight would last four tosix hours. Because an aircraft engine mustcarry its own weight, its performance isoften expressed as the ratio of mechanicalpower to weight and the estimates ofpower required to fly Daedalus rangedfrom 3.0 to 3.5 watts/kilogram. Before theDaedalus could be designed in detail, weneeded an answer to the fundamentalquestion: what are the limits of thehuman's capacity for long-durationpower generation and how do thoselimits affect flight duration and range?The first part of this question is certainlynot new, nor is it unique to human-powered flight, for it is applicable to anysystem that derives mechanical powerfrom the human body.

The production of mechanical powerby humans has long been the subject ofstudy by a considerable number of

investigators. A good summary of theavailable data can be found in Whitt andWilson (1982). We were particularlyinterested in the relationship betweensteady-state power produced and themaximum length of time that powercould be sustained, but were disap-pointed that the methods employed in thevarious studies we reviewed variedwidely, as did the results. In particular,the almost complete lack of data fordurations in excess of one to two hoursmade it impossible to generate a quantita-tive model for the expected duration ofthe Daedalus flight. It is important tonote, however, that measurements of thistype are extremely difficult to perform,simply because each measurement mustbe carried out until the test subject isexhausted. The physiological preparationand psychological motivation of the testsubject becomes an important experimen-tal variable that is difficult to control in arepeatable fashion. A further limitation ofthe reviewed literature was the fact thatimportant parameters, including test-subject body weight, level of training, anddetails of the measurement techniques arenot uniformly reported. The result wasthat the existing data, while useful forestablishing rough bounds on theproblem, were of little help in establish-ing the engineering feasibility of theDaedalus flight.

THE PREDICTION OF LONG-DURATION POWER OUTPUT

The metabolic cost of flight. It is rela-tively easy to determine the metaboliccost to the pilot of maintaining a constant

mechanical power output. Potentialenergy is converted into mechanical workby the oxidation of stored fuels within themuscle itself (Nadel, 1985) with anefficiency that is generally accepted to be24% (Astrand and Rodahl, 1986). To putthis in quantitative terms: for every 24watts of mechanical power delivered atthe pedals, 76 watts are generated as heatfor a total metabolic cost of 100 watts. Analternative means of expressing themetabolic cost is by oxygen uptake inmililiters per minute per kilogram ofbody weight. Exercise physiologists referto oxygen uptake as "aerobic power", inview of the assumption that the composi-tion of the oxidized fuels remainsconstant for a given individual andtherefore oxygen uptake is proportionalto the metabolic power generated. Inorder to predict the maximal poweroutput that we would expect from a pilotduring a long-duration effort, we neededto determine experimentally whetherhumans reached their limits in theoxygen, fuel, or heat-transport systems ata sufficiently high power output on anergometer. The maximum oxygen uptake,or maximum aerobic power of anyindividual is an objective index of thatperson's functional capacity to generatepower. Middle-aged, healthy adults havea VO2 mx of around 35 to 40 ml/min/kg

and are able to increase this maximum byup to 20% within three months ofbeginning a moderately serious programof physical conditioning (Nadel, 1985). Inelitfp Pndurance athleteS the maximilm

oxygen uptake (VO2max) may be as high as

(continued on page 8)

The typesetting and layout of this issue was done by Kim K. Griesemer, 8204 South Bettina Lane, Sandy, Utah 84093, 801-942-3333.

Human PowerThe Technical Journal of the

International Human-PoweredVehicle Association

David Gordon Wilson, Editor21 Winthrop Street

Winchester, MA 01890-2851, USA(617) 729-2203 (home)(617) 253-5121 (MIT)

Associate EditorsTheodor Schmidt, Europe

Rebackerweg 19CH-4402 Frenkendorf

SWITZERLANDPhilip Thiel, Watercraft4720 7th Avenue, NE

Seattle, WA 98105 USAIHPVA

P.O. Box 51255Indianapolis, IN 46251 USA

(317) 876-9478

Marti DailyAdam Englund

Bruce RosenstielPaul MacCready

Doug MillikenGlen Cole

PresidentSecretaryTreasurerInternational

PresidentVP WaterVP LandVP Air

David Gordon Wilson Executive VPAllan Abbott Board Members

Mike BurrowsMarti DailyPeter ErnstChet Kyle

Gardner MartinMatteo Martignoni

Dennis TavesDavid Gordon WilsonHuman Power is published quarterly by the

International Human-Powered VehicleAssociation, Inc., a non-profit organizationdevoted to the study and application ofhuman muscular potential to propel craftthrough the air, in the water and on land.Membership information is available bysending a self-addressed, stamped business-sized envelope to:

IHPVAP.O. Box 51255Indianapolis, IN 46251-0255 USAMembers may purchase additional copies of

Human Power for $3.50 each. Nonmembersmay purchase issues for $5.00 per copy.

Material in Human Power is copyrighted bythe IHPVA. Unless copyrighted by theauthor(s), complete articles or representativeexcerpts may be published elsewhere if fullcredit to the author(s) and the IHPVA isprominently given. We welcome contributionsof articles, notes, reviews and letters sent toany of the editors. Contributors' guidelines areavailable from David Wilson.

Special thanks to the authors, Marti Dailyand Kim Griesemer, without whom this issuewould not have been possible.

2 Human Power 7/4

EditorialsEngineering standards

Three recent events emphasized thedesirability of adhering to engineeringstandards in most cases and of departingfrom them in others.

The first was a series of accidents topeople in truck-mounted "cherry-pickers"-person-carrying baskets onarticulated arms-in which severaltelephone line repairers have beensuddenly dropped to the ground, givingthem serious injuries in many cases. I wasasked to serve as an "expert witness" inone of these cases, so that I had access tosome of the design data. The bucket isheld up by several strong steel strandedcables going over pulleys or sheaves. Theengineering standard for the minimumdiameter of sheaves for wire rope is 72times the rope diameter. The designer ofthe cherry-picker had used about a tenthof this safe diameter. Consequently it wassimply a matter of time before every ropefailed in fatigue, no matter how "strong"it was. The stress induced in the rope bybending it around a small pulley was farhigher than the stress needed to hold upthe basket. It might have been preferable,in fact, to have used much-smaller cablesoperating at higher mean stresses withmuch-lower bending stresses around thepulleys. It would also have been prefer-able to have used either much-largersheaves, or to have used chains instead ofwire rope.

In This Issue-

The physiological limits of 1long-duration human powerproduction-lessons learnedfrom the Daedalus projectby Steven R. Bussolariand Ethan R. NadelEditorials by David Gordon Wilson 2Letters to the editor 3Some ideas used on Hydro-ped-- 5a hydrofoil pedal boatby Sid ShuttThe baking of the Bean 11by Miles KingsburyReviews 15

This is an engineering standard thatis also ignored by most manufacturers ofbicycle hand brakes and derailleurshifters. Although I try to buy the mostsupple cables available, to keep them wellgreased and to avoid kinks, I know thatmy cables will fail every so often becauseof the appallingly sharp corners and smallradii around which bicycle cables arerequired to operate. The failure of a gearcable is irritating but not life-threatening.A brake-cable failure is potentially veryserious. Here is an engineering standardthat we should urge bicycle-componentdesigners to adhere to as soon as possible.(Incidentally, the hydraulic brakes of mycar-perhaps I should say my ex-car-have just suffered a virtually total failureas I was braking for a red light. There isan engineering standard to ensure thatone could never lose braking on morethan two wheels out of four, but thedesigner apparently did not conformclosely enough to this. Bicycle designersare not alone in their cavalier avoidanceof good engineering practice.)

A second example, this time of less-virtuous standards, was revealed in theintroduction to a delightful book aboutthe audacious transatlantic steamersdesigned in Britain in the last century byIsambard Kingdom Brunel: "The IronShip". The Royal Navy had been sosuccessful in earlier times that engineer-ing standards had been issued to fixvarious ratios, such as of beam to hulllength, at those of the best ships of thepast. But technology advances, and at theend of the eighteenth century Americannaval and commercial ships, unhamperedby such archaic regulations, couldoverhaul British ships in almost allconditions. (The iron steamships thatBrunel designed, without recourse tocrippling rules, helped to restore somedegree of international supremacy toBritish powered vessels. Sailing vesselsremained in the second rank behind theUS.)

My third example is more related toHPVs, and is of what is merely a de-factostandard: that of using half-inch (13-mm)chain for transmissions. Most HPVs haverecumbent riders and very long, andtherefore heavy, chains. There was anunsuccessful move in the early years ofthis century to introduce smaller chainsfor bicyle drives, chains which wouldhave been fully adequate for the duty.

Designers of HP aircraft have generallyused "chain" of conventional pitch butcould not afford the weight of regularsteel roller chain, and have frequentlyemployed a Berg lightweight driveconsisting of two fine steel cables bridgedby nylon molded-on cylinders, spacedand sized to resemble the rollers of aregular chain. Berg tried to produceversions that would work on derailleurs,but gave up. MIT's Daedalus groupfound the Berg drive over the long spansrequired for HPAs unreliable enough todevelop a gear drive for its record-breaking long-distance aircraft. AlecBrooks and Allan Abbott reintroducedsmall-pitch (1 believe of 3/8", 10-mm)chain for the Flying Fish, and their lead isbeing followed by other builders of HPBsand HPVs in general. May the designersof derailleurs develop gear changesystems to use this smaller, lighter andprobably more-efficient size of chain.

Aluminum forks?In the March issue of Bicycle Guide,

Doug Roosa reviews the new generationof aluminum bicycle forks coming on tothe market. Charles Brown sent me acopy, stated his concern for the integrityof these components, and suggested that Iwrite a letter to warn potential usersabout them. I confess that I do not quitehave the conviction to do so. I'm takingthis easy way out.

The problem is this. Almost all thealuminum components of my bicycles,however robust their appearance, have atone time or another failed without anywarning. Three handlebars, a handlebarstem, a crank, wheelhubs and rims areamong this group. They did not getflexible and creak, as steel componentsusually do before failing. They broke as ifexplosive had been inserted somewhere. Ihave survived these failures morethrough luck than skill. If one wereriding a diamond-frame "regular" bike,no amount of luck could save one from ahigh probability of injury if a front forkfailed in this manner. One of the majoradvantages of recumbents is that compo-nent failure, even of a front fork, is far lesslikely to result in serious injury than ifone were poised head first over the frontwheel.

Roosa includes a warning: "Alumi-num forks should be treated as high-performance items. That includes regularinspections for any signs of fatigue at thefork crown and dropouts". But myexperience of aluminum components is

that they do not show any evidencebeforehand of incipient failure.

The ndvortisomonts for aluminum

forks and frames are, however, reassur-ing. They have been rigorously tested forthe equivalent of several bicycle lifetimeswithout failure. I am an engineer, and Ishould not stand in the way of newtechnology. It wasn't long ago that steamboilers with pressures under one atmos-phere were exploding with sad regularity,killing many. Nowadays we have steamgenerators running at hundreds ofatmospheres and high temperatures thathave run safely for decades.

So let us hope that the fork designersand developers have done their workwell. As far as I am concerned, I would be

happy to have one on my recumbent, butI do not have the pioneering passion towant one on any "regular" bike anyone inmy family rides.

(continued on page 10)

Letters to the editor

In defence of the Moulton-theauthor's response

I cannot imagine the reason for DerekRoberts' indignation [over our article in7/1/88/11. I have shown only the resultsof measurements following an interna-tionally standardized method. On mostsurfaces the AM7 was the best commer-cially available HPV. What more does hewant?

The difference between the roadster

and the AM7 is small, but the AM7 hasthe great advantage of a much lowerrolling resistance. The vehicle weightswere noted in the vehicle descriptions.The AM7 is my own, used daily forcommuting, equipped with a rear carrier,generator and lights, a Citadel lock andsome modified components, so that thebare mass of 10.9 kg was increased to ahigher figure to include these neededextras.

I agree with Roberts' personalopinion on the AM7: that he prefers it toall his other bicycles. I felt the same wayabout mine, too. Just lately I have builtmy own bicycle with my own design ofsuspension.

The measurements show that on allsurfaces the vibrational stress at thesaddle of the AM7 is much higher thanfor the hand-arm system. I do not knowthe exact reason but I guess that the

spring rate of the rear suspension is toohigh and/or that the frequency responseof the rubber leads to a stiffer spring ratefor higher frequencies. I may be wrong in

assuming an effect of the internaldamping of the rubber as the reason forthe bad results (relative to the results forthe front suspension).

Swinging of the rear suspension (andthe front suspension, too) on the AM canbe ascertained when riding on smoothsurfaces. It would be better-and it can bealmost realized-if pedalling had noeffect on the suspension.

The aim of my article was to showhow very high the vibrational stress onbicyclists can be. I wanted to encouragediscussion of the quality of cycle tracks inGermany, and I wanted to encourage thereaders of Human Power to take suspen-sion into consideration when designingan HPV.

I did not want to assert that the AM7had a deficient suspension, but the resultsof our measurements show that a morecomfortable bicycle suspension is neededfor some conditions. First experimentsindicate that the goal can be reached.

Rainer PivitFachbereich 8Physics, Universitat OldenburgPB 2503D2900 Oldenburg, WEST GERMANY

Hydrogen-air fuel-cell-poweredvehicles

I am writing to ask for information,developments, and suppliers of equip-ment for the development of a hydrogen-air fuel-cell-powered vehicle based onexisting HPV technology. Most hydrogen-powered vehicles to date have usedconventional internal-combustion enginesin their design, which simply changes onefuel for the other within a system plaguedby high Carnot losses and inefficientoperation... Utilizing high-pressurehydrogen to several-hundred atmos-pheres in a Kevlar- or carbon-fibre tankfor light weight, a small fuel cell fromUnited Technologies, and a high-qualityelectric motor as used in the Sunraycerwould illustrate the practical andimmediate potential of hydrogen as a fuelfor transport. The hydrogen would bederived from photovoltaic cells, or from awind generator, electrolyzing water athome and compressing hydrogen into aspare tank in the home.... The onlybyproduct of the fuel cell is pure water.Surplus hydrogen could be stored at

7/4 Human Power 3

-

- -

home to be used for cooking, heating, etc.I await your reply. I have further

information to share with IHPVAmembers should (they) request it.

Gregory SpryP.O. Box 83876Fairbanks, Alaska 99708 USA

News from JapanI organized a small lecture about

solar power and human power on March4, 1989 in Yokohama. The lecturer was aMr. Peter Ernst, of Future Bike, Switzer-land. He talked to us about HPVs, theHPA Daedalus, and the Tour de Sol.Over sixty people came to the lecture.

Toshio Kataoka914-6 Mamedo, KohokuYokohama, Kanagawa, JAPAN 222

(Toshio Kataoka gave us several other pieces ofnews that I sent on to Jean Seay for HPVNews-we are grateful for his contributions).

Arthur Baxter comments on hisarticle

It has improved my ego to have aletter from the USA arising from myarticle. As I have not done any morepractical work on recumbents since 1939,I am sorry that I have no basis for anupdate. The paper you sent [on thehistory of recumbents, from an IHPVAScientific Symposium] is of great interestto me, as I had not known of the manyactivities described in it. Perhaps somecomments on these, and on non-circularpedal motion would be of interest?

I agree that for general use the seat ofa bicycle or tricycle should not be lowerthat those of a normal automobile;otherwise the three "D"s-danger, dirtand derision-are unavoidable. You maynow question why my '39 trike had theseat a mere six inches or so (150 mm)above ground level. This was worked outas follows. Minimum frontal area andshort wheelbase were considered to beessential. These could not be combined ina two-wheeler. A tricycle had to be as lowas possible to prevent capsize on bends(the rider cannot lean in to counteract thison a recumbent, as he can on a normaltricycle). In the '39 trike, first drive, (HP7/3/23) the pedals are level with the seatonly at the lowest point of their travel,and 13" (330mm) higher at the top point.This is why the rotary drive was so tiring,the feet being above the hips most of the

time. The pendulum motion (HP 7/3/24)was much better for both comfort andreduction of fatigue. An attractive featurewas the ability to freewheel with the feetopposite each other on each side of themachine.

In the Avatar [and many similarrecumbents] the relative positions ofpedals and feet are very different to theabove, and I can well appreciate that in1his case, rotary pedalling is quitesatisfactory. I am surprised that you findthe [LWR] Avatar 2000 no harder to pushthan the shorter [SWR] Avatar 1000.About the only thing that keen cyclistshere are agreed about is that longer bikesare harder to ride up hills. Also the[LWR] would seem to be less safe inaccidents, when the experience might belike sliding down a staircase handrailwith a knob at the bottom (long top tubeand steering-head top). The escape routefrom a [SWR] seems to be much lesshazardous! The long bike will take upmore storage space, but then you havemore room in the States than we havehere!

Even if pendulum crank motion isnot needed on the [LWR] Avatar, it maybe interesting to consider why I did nothave such difficulty with this type ofmotion as you and your friends have had

:'as described in the paper mentioned].I he configuration of the pendulum on mytrike was such that, when one of thependulum cranks, via its connecting rod,pulled its rotating crank to top deadcentre, the other rotating crank had, by afew degrees, passed bottom dead centre(see sketch). Even so, I had to ease off thefoot pressure towards the end of travel sothat the momentum of the cranks-chainwheel assembly helped past thepoint of possible "lock up". One does thisin normal rotary pedalling. [This type ofsystem is called a "quick-return mo-tion"-ed].

[The following comment is on tireconstruction]. When the first pair of(heavy Dunlop) tires wore out,.. Iordered a low-price pair of hand-madeConstrictor tires. I was horrified when Isaw them. It looked as if there was notenough rubber on them to last 100 miles,and as if they would puncture very easily.But what a revelation! Much fasterunning, and all the smaller bumps

completely absorbed, their flexibilitysaving them from rapid wear andpunctures. Since then I have believed thathard narrow tires are a hindrance toprogress except on perfectly smoothsurfaces, and where are they? The power

Pendulum configuration

from one's legs has to be shared betweengoing forward and going up and downdue to bumps, so the less there is of thelatter the more is left for forward prog-ress. The rolling drag of fatter flexibletires is a small loss for a large gain.

A. D. Baxter74 SouthgateScarborough, Y012 4NBENGLAND

Rear suspension pivot pointIn HP 7/3/89 Robert L. Price wrote

an article on HPV steering and suspen-sion design. What he wrote about Fig. 19can be interpreted as that the bestposition for the pivot of a rear wing-armis in, or a little above, the line of the chain.I have constructed six recumbents withrear suspension, and by the old-fashionedmethod of trial and error found that thebest position for the pivot is 20-30mmbelow the chainline. Meanwhile myfriend Wilfried Schmidt has worked theo-retically on this point and has found thesame result. I'm sending the article andsomething I wrote about suspension. I'mafraid that the two are available only inGerman.

Werner StiffelHubschstr. 23 D 7500Karlsruhe, W. Germany

(I'll copy these and send them within the USfor $2.00. I'm embarrassed to have had anexcellent how-to-build article by WernerStiffel on one of his suspension recumbents,translated by Theo Schmidt, but have lost thediskette. I will ask someone to transcribe itagain so that we can have it in the nextissue--ed.).

(continued on page 16)

4 Human Power 7/4

1' I' ' r .1 J

10

8O

Drag(N) 6C

40

20

1 2 3 4Velocity (m/s)

Figure 3. Displacement performance

w

at

5

be made with this power using a kayak-type displacement hull. The additionaldrag of the hydrofoils reduced the speedto 3.0 m/sec at the same input power, andthis speed was selected for takeoff.2. Light weight.

Light weight is important in reduc-ing the hull displacement for low hulldrag, in producing a smaller load thehydrofoils must lift, and less mass toaccelerate. The Hydro-ped I includinghydrofoils weighs 20 kg.3. Low induced drag.

Below takeoff speeds the hydrofoilsneed not produce lift, so their lift coeffi-cients are made low resulting in lowinduced drag. In displacement operationthe main hydrofoil is set to have a lift

coefficient CL of 0.35, an aspect ratio, A,of 33, and a hydrofoil shape and depthcoefficient, k, of 2.6 giving an induceddrag coefficient, Ci = CL2 /(A*k), of 0.0015,an acceptably small value.

Before takeoff the bow hydrofoil loadis its own weight, a force of 0.4N. Justbefore takeoff its induced drag coefficientis less than 0.0002. The additional drag of

80

Drg(N

40

20

the hydrofoils without significantinduced drag is then small enough topermit good speed before takeoff.4. Large hydrofoil area for takeoff.

The load that the main hydrofoilmust lift, L, is 780N. A lift coefficient, CL,of 0.90 at takeoff is realistic. The mainlifting hydrofoil area at takeoff was madeto be 0.21 sq. m. These parameters

:w ·NN500 W

200 WattsInput Power

10 t/sec

1 2 3

Figure 5. Hydrofoil performance

---- L ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~ h~~~

4 5 6 7 8Velocity (m s)

produce ample lift when takeoff isinitiated at a velocity, v, of 3 m/s. Therelationship of these parameters isL=530*CL*A*v2.At takeoff

At takeoff several ideas are used togive deliberate, quick, and easy takeoffsto lift the hull out of the water using theparameters established before takeoff.These are:

1. bow raised to increase mainhydrofoil lift;

2. momentum to overcome largetakeoff drag; and

3. reduced hull drag.1. Bow raised to increase main hydrofoil

lift.At a takeoff speed of 3.0 m/s the bow

of the boat is raised by pulling the bowhydrofoil down relative to the hull bow.The bow hydrofoil then produces a lift of112N which lifts the hull bow about 0.4mabove the water surface. This pitches theboat up by 6 degrees increasing the mainhydrofoil lift coefficient from 0.35 to 0.91.The load the main hydrofoil is able to liftat thes ronrtioins is t-ha 1 "11NI-,. -.. ',VAIL~VIO so IIt J zz..l-

Since 780N is needed to support thehull out of the water, a force of 132N isFigure 4. Hydro-ped-hydrofoil configuration

6 Human Power 7/4

�

I '-

--

_��d

r

14;��

tz: 5

available to accelerate the boat massvertically. The initial upward accelerationat liftoff is greater than 1.0 m/s2 . In lessthan a second the hull is out of the water.2. Momentum to overcome large takeoff

drag.The lift coefficient at takeoff increases

from 0.35 to 0.91 giving an induced dragcoefficient that increases from 0.0015 to0.01 and a profile lift coefficient thatincreases from 0.010 to 0.018. The bow-hydrofoil drag coefficient similarlyincreases from 0.012 to 0.04. The in-creased profile, induced, and wave dragof the bow hydrofoil and the mainhydrofoil added to the hull drag atinitiation of liftoff is considerably largerthan can be maintained at 3m/s at aninput of 200 watts. But, because the pitch-up occurs rapidly, the momentum of thevehicle adds the needed additionalforward force to overcome the increaseddrag at liftoff to get the hull out of thewater before it slows significantly. Oncethe hull leaves the water the total drag issmaller than it was before takeoff wasinitiated.3. Reduction of hull drag.

The hydrofoils lift the hull out of thewater and the hull drag is reduced to airdrag. And, of course, this is the main ideaof the hydrofoil vehicle.After takeoff

After takeoff some ideas are used toadjust the parameters to minimize inputpower over a broad speed range, go fast,be stable in all axes, and allow a light-weight structure that has sufficientstrength using available materials. Theseideas are listed:

1. reduced drag of hydrofoils;2. height and main-hydrofoil-area

control;3. force balance;4. bow-hydrofoil operation;5. roll stability; and6. structural considerations.

1. Reduced drag of hydrofoils.As the hull increases in height above

the water surface the hydrofoil angle ofattack decreases, lift coefficient decreases,and the main hydrofoil area decreases.Profile and induced-drag coefficients aredecreased as the boat accelerates tosteady-state velocity. At top speed themain hydrofoil area is near 0.06 sq.m, areduction of 3.5 from the liftoff area.2. Height and main-hydrofoil-area control.

The bow height above the watersurface is set to match conditions ofoperation. If a fast sprint is desired a highsetting is made giving small main-hydrofoil area for high speed. If duration

is desired a low setting is used giving alarger area. The height can be adjusted togive the main hydrofoil area that willminimize the input power at a givenvelocity.3. Automatic force balancing.

Operating at above takeoff speed thehydrofoil lift forces equal the weight ofthe craft. If the boat decreases in speedthe hydrofoil lift is less causing the hullcenter of gravity to move down. Since thebow foil holds the bow at nearly constantheight, the main hydrofoil angle of attackand its area increase until sufficient lift isgenerated to balance the boat weight. Ifthe boat increases in speed the CG lifts updecreasing the area and angle of attackuntil a new balance is reached. This actionis constantly occurring maintaining thehull at an average preset height.4. Bow-hydrofoil operation.

The bow hydrofoil has a similaraction as the main foil in maintaining aforce balance. A horizontal hydrofoil of0.03 square meter area is attached to thebottom of a vertical fin that is also therudder. The top of the vertical fin isattached to a horizontal forward arm. Atthe forward end of the horizontal arm is aplaning surface that rides on the averagewater surface. As the load on the bow foilis increased the hydrofoil is pusheddown, causing its angle of attack toincrease resulting in more lift until thegenerated lift balances the load. Thisconfiguration is very stable with a goodresponse to input disturbances. Detailedcontrol analysis of this design will showits stable operation with lead compensa-tion giving a well damped transientresponse to input-water-surface irregu-larities. I have found this design workswell on my hydrofoil sailboats wherestable, fast response to rough conditionsis absolutely essential.

This design has been patented by meand used elsewhere. The first time myHydro-ped ands Allan Abbott's and AlecBrook's Flying Fish were together, in thespring of 1988, the Flying Fish had itstrailing-surface-sensor, elevator-linkedbow-hydrofoil assembly. At the 1988IHSPC at Visalia they had changed to myconfiguration for their front hydrofoil.5. Roll stability.

The surface-piercing main hydrofoilsnot only give variable hydrofoil area, butalso prevent the high center of mass fromtipping in roll. The dihedral angle of thehydrofoil at the water surface alsoproduces a stable bank in turns.6. Main-hydrofoil structure.

Eighty-five percent of the load when

on the hydrofoils is carried by the mainhydrofoil. Since ends of the mainhydrofoil extend above the water surfacethe tips are connected to an aluminumtube that carries much of the load in bothbending and torsion without addingmuch drag or weight. The hydrofoilcenter is supported by the drive strut.This structure allows hydrofoil shapes tobe used that are determined by hydrody-namic rather than by structural considera-tions for the under-water portion.

CONCLUSIONMost of the ideas used in my

hydrofoil boat configuration have beenaround for a while, and some are new,but I do believe the total combination ofall the ideas applied to the Hydro-pedcreates a new and unique water vehiclethat allows me, with my limited energy,to enjoy operating it on the hydrofoils asit is intended. It is hoped that some ofthese ideas will help stimulate others todesign and build water vehicles tocompete in the International Human-Powered Speed Championships in futureyears.

REFERENCESRichard Von Mises, Theory of Flight,

1945 Dover Publications,Inc., New York,NY.

Ira H. Abbott and Albert E. VonDoenhoff, Theory of Wing Sections, 1959,Dover Publications, Inc., New York, NY.

Sighard F. Hoerner, Fluid-DynamicDrag, 1965, published by the author, P.O.Box 342, Brick Town, NJ 08723.

C.A. Machaj, Sailing Theory andPractice, 1964, Dodd, Mead, & Co., NewYork.

Lindsay Lord, Naval Architecture ofPlaning Hulls, 1963, Cornell MaritimePress, Inc., Cambridge, MD.

John K. Vennard, Elementary FluidMechanics, 2nd Ed., 1947, John Wiley andSons, Inc.

S. Timoshenko and D.H. Young,Elements of Strength of Materials, 5th Ed.,1968, D. Van Nostrand Co., Inc.., NewJersey.

(Sid Shutt is an electrical engineer and aregistered professional engineer in mechanicalengineering and control-systems engineeringin California. He operates an engineeringcompany specializing in new-productdevelopment.--ed.)

Sidney G. Shutt612 Briarwood Dr.Brea, CA 92621 USA O

7/4 Human Power 7

Lessons from Daedalus(continued from page 1)

70 to 80 ml/min/kg (Astrand andRodahl, 1986) and they are able tomaintain energy production without theproduction of excess lactic acid duringexercise that demands up to 80% ofVO,.... We hypothesized that theseathletes, who have induced adaptationsto physical activity in both the oxygen-delivery and oxygen-acceptance systems,would be capable of sustaining, for fourto six hours, a power level that producedoxygen uptakes of 70% of their VO 2m,x.We chose the figure of 70% arbitrarily toprovide a margin of safety from theexcess-lactic-acid threshold.

Measurement of oxygen uptake. Webegan a series of laboratory measure-ments of oxygen uptake and mechanicalpower production as part of a screeningprocess in order to identify a pool ofsuitable pilot-athletes. Our plan was toselect those athletes who could meet thecriterion of a mechanical power produc-tion of 3.5 watt/kg at 70% of theirmeasured VO2max and then verify ourhypothesis that they would be capable ofproducing that level of power for a periodof time comparable to that anticipated forthe Daedalus flight. Announcements ofthe search for pilots for the DaedalusProject produced 300 applications, anumber of these from Olympic-caliberathletes. We invited 26 (24 men and 2women) to be tested, all of them eithercyclists or triathletes. We had determinedthat athletes who have been training onthe bicycle would be expected to be themost efficient at power production for theDaedalus aircraft because they had beenusing the specific muscle groups forcycling in their training regimes.

E

EE

c

'U

60 - 1

50 i

40 o upright* semi-recumbent

30 .i2 3 4 5

Mechanical power output (W/kg)6

Figure 1. Typical results of a test formaximum oxygen uptake on a semi-recumbent cycle ergometer.

8 Human Power 7/4

The test for maximum oxygen uptakewas given on a semi-recumbent cycle er-gometer with which the athlete's me-chanical power output could be preciselymeasured. During the test, the athletebreathed through a low-resistance, two-way valve, into an apparatus with whichwe could measure the mass flow rate ofrespiration and the amount of oxygen andcarbon dioxide in the expired air. Weused an incremental load paradigm forthe measurement of maximum oxygenuptake (Astrand and Rodahl, 1986).Figure 1 shows the results for a typicalergometer test. The relation betweenoxygen uptake and mechanical poweroutput is linear except for the final value,which diverges from this linear relation-ship due to the fact that, at this highmechanical power output, a significantfraction of the power is producedanaerobically. This anaerobic powerfraction cannot be accounted for byoxygen uptake; therefore, as we increasethe demand for mechanical power, we donot see a corresponding rise in oxygenuptake. When a straight line is fitted tothe data (without including the highestvalue which includes the anaerobiccomponent), we obtain the equation:oxygen uptake = a + b x (mechanicalpower output), with the value of arepresenting the oxygen uptake contribu-tion to metabolic processes independentof mechanical power output, and with thevalue of b representing the ratio ofoxygen uptake to power output. Wedetermined the maximum aerobicmechanical power by substituting themaximum oxygen uptake measured forthe athlete into this equation. Calculationof the mechanical power output at 70% ofVO2 max is accomplished by substitutingoxygen uptake at 70% of the measuredmaximum into the equation and solvingfor mechanical power.

THE EFFICIENCY OF MECHANI-CAL POWER PRODUCTION

Definition of efficiency. The mechanicalefficiency of a given athlete may beestimated from the slope of the linearrelationship between oxygen uptake andmechanical power production if onemakes suitable and consistent assump-tions about the composition of thecompounds oxidized in the metabolicprocesses. In particular, the ratio ofglucose (sugar) to lipid (fat) "burned" isquite important and differs amongindividuals. Remarkably, the mechanicalefficiencies among the 25 athletes testedranged between 18.0 and 33.7%, indicat-

ing a maximum oxygen uptake alone isnot a sufficient measure of performancewhen predicting mechanical poweroutput. As Figure 2 illustrates, twoathletes with nearly identical maximumoxygen uptakes can have widely differentefficiencies, 20.1 and 26.6% in the exampleshown. At 70% of maximum oxygenuptake, the values for mechanical powerproduction are 3.31 and 4.21 W/kg,which means that the more efficient of thetwo is able to generate 27% more me-chanical power at this level of uptake.Another way of looking at the differencebetween the two, which is applicable tocases where the task demands a constantpower output (human-powered flight is agood example), is that the oxygen uptakenecessary to generate a power output of3.5 W/kg is 74% of maximum in the less-efficient athlete and only 59% of maxi-mum in the more-efficient athlete. Themore-efficient athlete will therefore beoperating further from anaerobic limitsand would thus be expected to havegreater endurance.

.c

-E!)

a

c

1o0ID

0

2 3 4 5 6 7Mechanical power (W/kg)

Figure 2. A comparison of the relation-ship between oxygen uptake and me-chanical power output for two athletes.Though they have similar maximumoxygen uptake, their relative efficienciesin mechanical power production differsignificantly.

The reasons for the large range inefficiency are not entirely clear. Certainly,a small part of the variability can beattributed to differences in the recruit-ment of auxiliary muscle groups notdirectly involved in mechanical powerproduction. These muscle groups willdemand oxygen, but will not contributeto the mechanical power output. Themajor part of the variability is probably aconsequence of differences amongindividuals in the metabolic machinery,perhaps the content of certain enzymesresponsible for muscle fiber contraction.In view of the wide range of efficienciesthat we have observed, the distinction

that we have made between thephysiologist's definition of "aerobicpower" and mechanical power output isquite important, particularly since VO,m,,is frequently given as a measure of anathlete's performance. Although it wasbeyond the scope of our study, it wouldbe particularly interesting to explore thepotential for changes in efficiency in anindividual, particularly as an indicator oflevel of training and perfection oftechnique.

Upright vs recumbent. A questionabout mechanical efficiency that con-cerned us and was not resolved by theliterature was whether the semi-recum-bent position entailed a lower mechanicalefficiency than the standard, uprightcycling position. The semi-recumbentposition has significant advantages forhuman-powered aircraft because of thelow frontal area of the pilot and the factthat the hands and arms have morefreedom to manipulate flight controls. Wetherefore tested several of our athletes inboth positions and a typical result isshown in Figure 3. We were able to findno significant difference in either maxi-mum oxygen uptake or mechanicalefficiency between the two positions. Wewere somewhat surprised that there wasno difference in maximum oxygenuptake, because of the potential forupper- body-muscle recruitment in theupright position, particularly during thehigh-power portion of the test. Perhapsthis can be explained by the fact that oursubjects were experienced cyclists andthey have optimized their pedalingmotion for a wide range of power output.

HEAT, HYDRATION, AND FUELHeat dissipation, sodium, and water loss.

A person's ability to dissipate the largeamount of heat produced during heavyexercise is compromised by increasingdehydration. We calculated that a 68-kgpilot with a mechanical efficiency of 24%would produce about 900W of metabolicpower during steady flight. Since onlyabout 225W will be in the form ofmechanical work, some 675W need to bedissipated through evaporative, radiative,and convective heat loss. The rate of heatloss from the body due to radiation andconvection was estimated as 60W,assuming an average skin temperature of33°C, a cabin temperature of 28°C, a totalheat transfer area of 1.2 m2, and a heat-transfer coefficient of 10 W/m20C)(Gagge, 1972). This leaves somewhat over600W to be dissipated by evaporation ofsweat. Taking the heat of vaporization

E

Ea)

Co0,

0

IV

60

50 i

40 o upright· semi-recumbent

302 3 4 5 6

Mechanical power output (W/kg)

Figure 3. A typical comparison of oxygenuptake and mechanical power productionfor a single athlete in the upright and thesemi-recumbent cycling position.

into account (0.7W per gram of waterevaporated per hour) we estimated thatthe pilot will lose about 900 ml of waterper hour. Since the ill effects of bodydehydration begin to occur when the lossof body water exceeds 3% of body weight,the ability of the pilot to sustain flightdepends upon the restoration of bodywater at a rate comparable to its loss. Weassumed that the sodium concentration ofthe evaporated sweat was 20 meq/liter, inwhich case our pilots would also lose theequivalent of 0.4 grams of sodium perhour.

Fuel. The body's fuel stores contain asufficient supply of potential energy tosustain activity for days, yet most of thatfuel is in the form of triglycerides (fat), aform that is not metabolized at a greatenough rate to provide all the energyneeded for the Daedalus flight. The otherfuel source is glycogen, the storage formof glucose (sugar), which can be trans-ported and metabolized at the requiredrate. However, the amount of glycogenthat can be stored in the liver and musclesis limited and would be depleted withinapproximately three hours of the start ofexercise at 70% of maximum aerobicpower (Coyle et al., 1986). The rate ofdepletion of glycogen stored in the bodymay be reduced by ingesting glucoseduring exercise and this phenomenon isexploited by endurance athletes whoconsume drinks or food containingglucose during exercise. To estimate therate of glucose uptake by the body duringflight, we assumed, from the data ofCoyle et al., that the rate of glucoseoxidation would be on the order of 50% ofthe total fuel-oxidation rate. Since thetotal oxidation rate yields an estimated900 W, we used the energy conversion of4.6 W per gram of glucose oxidized perhour to estimate that our pilots would

require approximately 100 grams of-lrr1cin rnr hrnir fn fly -ho Dnf-l-,-

aircraft.The Daedalus drink. Our first approach

to developing a drink to address the fueland hydration problem was to conduct aliterature search, including a comprehen-sive review by Murray (1987) of thecharacteristics and physiological effects ofa number of carbohydrate-electrolytedrinks currently on the market. However,none of these drinks had sufficientcarbohydrates and electrolytes to com-pensate for the losses we had estimatedfor the Daedalus flight. We proposed torehydrate our pilots at the rate of one literof fluid per hour and designed our drinkwith 10% glucose and 18 meq/litersodium, concentrations estimated toreplace glucose and sodium at ratesequaling their losses. We tested the drinkwith four of our pilot-athletes during six-hour ergometer flight simulations at amechanical power of 3.1 W/kg and foundthat blood glucose and hydration levelsremained stable throughout the tests("N,,] l nnd R,zrlqri 98R)

CONCLUSIONAs the technology of human-

powered vehicles becomes more ad-vanced, the challenge of long distancesand durations push the limits of ourknowledge of the physiological limits ofexercise. The study that we have made ofthese limits was subject to the constraintsimposed by the need to field anoperational vehicle in a reasonable timeperiod and therefore we have perhapsraised many more questions than wehave answered. We have determined anddemonstrated, however, that elite athletesare capable of maintaining a continuousmechanical power output at an oxygenuptake of 70% of their maximum for aperiod of greater than four hours, given

proper supplementation with a glucose-sodium drink. Our flight simulationshave demonstrated that an effort of 3.1W/kg at the pedals may be sustained forsix hours under the same conditions.

On April 23, 1988, Kanellos Kanel-lopoulos left the shore of Crete in theDaedalus aircraft. Nearly four hours later,after having flown 119 kilometers, he wasforced into the water by gusty surfacewinds a few yards from the beach onSantorini. During the flight Kanellosdrank almost four liters of the glucose-electrolyte drink. His heart rate neverexceeded 142 beats per minute and at notime during the flight were there anysigns of impending fatigue.

7/4 Human Power 9

sophisticated section. He had two inmind, one being the NACA 66 wingsection and the other a section designedfor axi-symmetric bodies. The latterturned out to be rather too slender at thenose and tail, so an NACA 66.028 waschosen. (The last two digits are thepercentage thickness, so that NACA66.028 with a chord length of 100 mmwould be 28 mm wide).

At this stage Jeff suggested that a1/3rd-scale model would produce moreaccurate results, so I set to work. Thistime the vehicle could be designedknowing more or less that the screenangle would be 15 degrees to the horizon-tal giving an overall height of 1 m.

The second model was made in asimilar fashion to the first, the onlychange being that this time seventy-foursections, each 6 mm thick, had to be cutout. This sounds a lot more time-consum-ing than it really was: to speed things upwe built an adjustable pantograph. Thismeant the aerofoil had to be plotted onlyonce and all other sizes could be scaledup or down from this one shape. A jigsaw was used to do the cutting, with avacuum cleaner attached to suck awaythe sawdust.

The plug was completed and twomodels taken from it. Jeff was given oneof them and left to experiment.

Time was running out; MIRA wasgetting close (i.e., races at the MotorIndustry Research Assoc.,-ed.) and therewas little point in going for the DuPontPrize in the pedal car.

At this stage I had no idea of whetherthe machine was going to be rideable ornot, as it had a very low seat height ofabout 250 mm. I drove up to the localdump and purchased a 14"-wheelkiddie's bike and proceeded to modify itinto a tubular-steel-framed, single-speed,front-wheel-drive recumbent with an

0,0Coos

Z.. _FLN

Z _ I::,* eEAZ WW.EL

.t FAJ/

10ELI bATA

W-Ee, , , , --

'r4EL, 2 T0 A,-'EL 2 A,.

0.5 1*0

Z,-, FULL rCL

1.5

Figure 3. Wind tunnel data for both models

12 Human Power 7/4

adjustable rear wheel giving seat heightsof between 250 mm and 400 mm.

It was 'built dirty', but finished inone weekend. The seat height was set tothe minimum of 250 mm and after a fewfalse starts, I pedalled off with remarkableease. A couple of days later it wasdecided to test the device at 'speed' at alocal industrial estate (park) where thereis a half-mile section of road with a slightdownhill gradient. I reached about 25mph (11 m/s) which was not bad on a 50"(1.27 m) gear, and went home satisfied.

When the results for the secondmodel came in we were delighted: theeffective frontal area (full size) was downto about 0.025 m2at an overall height of970 mm. This gives a drag coefficient ofabout 0.07 for a frontal area of 0.35 m2.Wind tunnel data for both models isgiven in Figure 3.

The time had come to build the realthing-MIRA was only six weeks away.

This time there was very little to re-design. The height had to be reducedfrom 1000 to 970 mm and windows,central spine, seat, etc., added; the basicshape remaining the same.

The plug was made in the same wayas the previous two, but this time therewas the added problem of weight. Therewere seventy-four sections, each 19-mmthick. To reduce weight and cost, themiddles were cut out of the largersections and used for the smaller ones.

A very large pantograph wasrequired this time along with threetemplates of various sizes. All the sectionswere glued together and the fin wasfitted. The plug must have weighed about300 kg. The next stage was to rub itdown-this part alone took about threeweeks of evenings and weekends. Thesurface was then sealed with polystyreneresin and polished to a shine. The plugwas sawn in half and each half laid onaluminium sheet. The whole lot waswaxed and applied with release agent.The mould was constructed with about 4mm of glass fibre and a tubular-steelsupport frame to stand it on.

When the two moulded halves wereremoved from the plug, a number of airholes were found near the surface. Thesehad to be filled and rubbed down, whichwas a week lost.

MATERIALSIt had been decided at an early stage

in the project to make a sandwich-typeconstruction using a nre-nrpo l]asS-fihr

..- C - r v r--o o .....mat with an aluminium honeycomb toform the shell of the Bean. The shell was

Figure 4. Mould layup stack

the main structure, the only other load-bearing member being the central spine.This was made from 10 mm-thickFibrelam which is a glass-fibre-paperhoneycomb sandwich used to makeaircraft flooring. Figure 4 shows themould layup stack.

The pre-preg (epoxy pre-impreg-nated fiberglass) sheet comes on a roll inan uncured state and has a similar feel toit as PVC sheet. To get its full mechanicalstrength, it has to be heated up to about140°C for two hours.

To cure the pre-preg reant buildingan oven and this was done in the cornerof a room using the walls and floor asthree of the sides of the oven. Theremaining walls and top were made fromchipboard. The front section had threesmall windows in it and was removable.The oven measured 10 ft x 4 ft x 4 ft (3 mx 1.2 m x 1.2 m) and by experiment it wasfound that about 6 kW was needed to getthe temperature up to 1400C. The mainheat source was five cooker rings at-tached to one end.

Additional heat was provided by anumber of spot lamps on the roof of theoven. Air was circulated using a 1/2-hpmotor outside the oven, with an extendedshaft and a 12" (300 mm) fan on theinside.

MOULDINGThe moulding was done one half at a

time. The pre-preg and honeycomb werelaid into the mould as shown above, thenthe release layer, the air-bleed felt andfinally the rubber. A vacuum was thenpulled between the rubber and themould. This was done using the inlet of astandard 1/2-hp compressor. A pressuregauge was also fitted to measure thevacuum.

The effect of the vacuum produced apressure of about 10-12 psi (70 kPa) onthe sandwich, ensuring a good bondbetween the three layers when the pre-preg cured.

In practice this part of the processturned into a new game-hunt the hiss.

Z .o

The moulds were rather porous, so it isjust as well I don't build canoes. Themoulds had to be stood on their sides andthe backs painted with resin until thehisses had stopped.

Once a pressure of 300 mbar hadbeen reached and maintained, the wholelot was put into the oven and the heatersturned on. This was a very worrying andexciting two hours, with one eye on thepressure gauge and one on the ther-mometer. The Bean began to cook.

The vacuum began to fail severaltimes during the two hours, but fortu-nately it was traced to a leaky bit ofdouble-sided sticky tape each time. It wasa very unpleasant feeling flinging openthe door and probing about in air at140°C. Have you ever smelled chipboardat that sort of temperature?

Finally the two hours were up, theheat was turned off and the mouldremoved. Extracting the finished compo-nent from the mould proved to be rathermore difficult than expected because thesandwich seemed to be acting as acapacitor. Huge sparks were jumpingeverywhere as the two parts wereseparated.

The next thing I did, as would anynormal HPV builder, was to grab thespring balance-it weighed 7-1/2 pounds(3.4 kg) which seemed very good for halfthe structure.

Once the other half was complete, thetwo were taped together and a woodenmock-up of the seat and centre spine weremade, to which a pair of conventionalcranks and pedals were fitted. At thisstage the canopy was not removable, so Ihad to climb in through the windscreenhole. Once inside I noticed that there wasno room to spare anywhere, so theeventual positioning of all the compo-nents was extremely critical. After a fewadjustments of seat angle and height, Iwas happy that everything would go inas planned. Now came the point of noreturn-to glue the two halves together.

The glue used was Redux 410, a two-part epoxy which was absolutely superb.It will stick just about anything and isparticularly attached to the palette knife,which it glues to the floor, the table or the

ALUMINIUI PLATES WIEEL BUSHES

Fiur 7

Figure 7. Section through rear-wheel fin

mixing bowl while you are admiring yourhandiwork.

The two halves of the shell were firsttrimmed and lightly sanded with glass-paper along with the edges. They werethen taped back together and the gluewas poured into the small gap from theinside. (The masking tape stopped theglue from oozing out and gave a neatjoint.)

ING HOlSING

Fl RELPM

Figure 6. Cross-section of spine

When the two halves were stuck, thecanopy was marked out, and sawn alongthe pencil line with a fine-toothedhacksaw blade. (The dust produced bythe glass fibre was extremely irritating tothe skin and whenever I worked with it Iwore gloves, goggles and a mask.)

THE INNARDSNow that the canopy was off, it was

much easier to get at the inside of theshell and all the internal componentscould be fitted. The centre spine was thefirst to go in. This was constructed byfolding 10-mm- thick Fibrelam to producea deep 'U' section. As shown in Figure 5,

xz= rrt360 °

L Ci'Zt JM fI1llllll llllll1 1 ITMI lllllVMy

Figure 5. Folding Fibrelam

foldability is a very useful asset ofFibrelam. By cutting a length from one ofthe skins it is possible to produce a bendof a given angle with a nice radius on theoutside.

The centre spine has to support thecranks, the crossover gearing and thefront forks. The forks were fitted into analuminium tube with a conventionalheadset; the other two parts weremounted to aluminium bearing housings.The housings and tube were glued intoposition when the spine was assembled.Figure 6 illustrates a cross-section of thespine.

The spine was then complete andready to be fitted into the shell. It wasdesigned to extend under the seat andlocate accurately into the inside of the fin.Before it was fitted, two 2-mm-thickaluminium plates were positioned oneither side of the fin. These were designedto spread the load of the rear wheel andare shown in Figure 7. The spine was thenglued into position.

Next to go in was the seat which wasmade from Fibrelam. First a pattern wasmade from cardboard and the outsideshape cut out. It was then folded andglued and the centre section removed forcomfort. Kevlar cloth with a zip was thenbonded to the front face. The zip was foraccess to the rear wheel. The seat isillustrated in Figure 8.

zip

R CLOTH

C UTOUT-----

Figure 8. The seat

Finally, the front forks and drivechain were fitted. The front forks weremade from Renolds 531 with a standardcrown and steerer tube. Campagnolo reardropouts were used as a gear hanger wasneeded, and the forks spread to take a six-speed block, as shown in Figure 9.

7/4 Human Power 13

OWNX

Figure 9. The drive chain

The wheels were based on standard14"-diameter rims with our own design ofhub. There were no spokes, but instead24-swg aluminium discs were spun overthe rim and glued to the hub.

Front-wheel drive was chosen tokeep the rear of the machine clean, andalso to reduce the amount of chainrequired. With 14" wheels, gearing was aproblem. In order to get a top gear ofabout 200", a two-chain system wasneeded. The final drive relied on thechain being more or less parallel to thecentre line of the forks. If it is not, thenthere will be a tendency to pull thesteering to one side. The chain used was8-mm-pitch industrial roller which is nolighter than standard 1/2"-pitch, but doesmake the chainrings and sprockets morecompact. With the arrangement shown inthe diagram above, the top gear is 200"with the 120T chaining fitted and 150"with the 90T.

Before the front wheel could befitted, two cutouts had to be made. Thesmaller was to allow normal wheelmovement, and the larger to enable thewheel to be fitted and removed. Thislarger cutout could also be used asventilation during longer events. The twofoot flaps were also cut out at the sametime.

This just left the steering, whichcaused quite a few problems. It wasdecided not to try to design the levers andlinkages until the vehicle was nearlycomplete. The original idea was to have alever on each side of the driver whichwould be connected to the top of thesteerer tube by a push-pull rod. This wastried but because the linkages could notbe straight, the whole thing was toospringy.

The second attempt at the steeringwas a central joystick which was con-nected to the steerer tube by a universaljoint. It was made telescopic for safety.This arrangement was better but theredidn't seem to be enough leverage.

14 Human Power 7/4

By this time I was beginning to get abit worried as I had still driven the Beanonly about 50 yards without falling over.

The third and final version wasbased on the Avatar system with a pair ofhandlebars under the seat, connected tothe steerer tube by a pair of Bowdencables. I thought the problem was solveduntil I tried it and still couldn't balanceproperly. The front-wheel drive waspulling the steering to the left and it wasfound that this was being caused by thechain being out of line. The size of thesecond drive sprocket was increased andthe problem was solved. I could now stayupright for long periods at a time.

SrnNAN RD RIM

Figure 10. The wheels

TESTINGGetting going in the Bean has always

been the greatest problem when it comesto testing. It is quite easy to start off byyourself without the canopy on as youcan sit upright and move the torso aboutmaking balancing easier, but with it inposition it is much more difficult. If youemploy the help of someone to push youoff, a lot of confusion can arise. Forexample, if you are falling in one direc-tion, then you tend to steer into the fall-the pusher cannot understand why youare going off at 90 degrees. There is also atendency for the pusher to give a littlehelping shove just as he or she lets go.The result in both cases is a horizontalBean as demonstrated three times atThamesmead.

Finding somewhere to practise hasalways been a problem with the high-speed machines as they are not reallypractical enough to use on open roads.The Bean's first test was at Slough CycleTrack which is concrete, 500 yards longand very bumpy. It probably had notbeen used for racing in about 20 years.After studying the local Ordnance Survey

map looking for nice straight and flatroads, a disused aerodrome at Chalgrovenear Oxford was decided upon. It was a2-1/2-mile perimeter road with one sidevery nearly a mile long. After somenegotiations, permission was given to tryit out.

We drove out to the aerodrome earlyone Saturday morning. I managed twolaps of the perimeter track with thecanopy off to have a look and get the feelof the surface. The track was generallygood, but there were about six rathernasty bumps. I then tried the Bean outwith the top on and got a shock-it didn'thalf go! I flicked through the gears andsoon found myself cruising in a 150" gearwith very little effort. I was doing 45 mph(20 m/s) down the straight and free-wheeling around the corners at well over30 mph. The bumps made things veryexciting, with both wheels leaving theground by three or four inches. On thefifth lap I hit a large stone at about 35mph and the front tire exploded, but Iwas able to park the Bean safely whichwas very reassuring. (The same thinghappened to the rear wheel at a later date,and again I stopped safely.)

Figure 11. Front-wheel cutouts

RACINGThe first outing for the Bean was to

be Thamesmead which meant only twoweeks left for finishing and further tests.At this stage the body shell was still in the'natural' state-a dull yellow. It wassprayed and the go-faster stripes werestuck on the day before Thamesmead.

The weather was near perfect on thefirst morning of the sprints with hardlyany wind-in marked contrast to theprevious day. The first run was disap-pointing. I went through the gears too fastleaving me in top gear and pedallingmuch too slowly. I clocked 10.40 secs. forthe 200 m giving a speed of 43.01 mph(19.3 m/s), putting me in fourth place. Onthe second run I didn't use top and itwent much better. I took the lead with a

__ - __ - - - __ - - 0

L·

Ties-Technology, Innovation andEntrepreneurship for Students

TIES is the title of a nicely producedDrexel University magazine for students.Marti Daily contributed a exciting survey,with excellent colored photos, of recentIHPVA activities in the January/February1989 issue, emphasizing the entries fromschools and colleges. Chet Kyle followedthis with "Designing efficient HPVs", andthere was a third HPV article, to tap thehumanitarian impulses of constructors,about the work of Ken Hughes and theInstitute for Transportation Developmentthat, among other things, suppliesbicycles to Africa, Latin America andother Third-World areas.

Designing and building the three-wheeled human-powered vehicle

by Tom McGriff and Jim WolpertThe designers of Hudyn Vehicles

have put their thoughts, opinions andsome of their secrets down on paper in awell-illustrated and racily written book-report. They take the would-be designerthrough the stages from initial concept tofinal construction. They force the readerto ask questions to ensure that s/he hasthought through the real purposes of theeffort. They discuss the principal alterna-tives at every stage.

I heartily recommend this book to allenthusiasts starting on three-wheeledHPVs. I saw an early edition of the bookand found some places where I wouldhave recommended differently (forinstance, I believe that a general-arrange-ment drawing should be made beforedetail sketches) but I know that revisionsand corrections have been made in laterversions. I don't know the price andavailability: write to Tom McGriff at P.O.Box 22444, Indianapolis, IN 46222.

Alternate Energy TransportationThis newletter reports that GM has

announced "GM Sunrace USA" for solar-powered vehicles from Florida, July 8,1990, to the GM Technical Center inDetroit, for 32 N. American schools andcolleges. The winners will be awarded anexpense-paid trip to Australia to competein the 1990 World Solar Challenge (qv,below). Proposals should go to GMSunrace USA, Proposal Review Commit-tee, 825 Myrtle Avenue, Monrovia, CA91016 (AeroVironment).

The World Solar Challenge will start inDarwin, Australia on November 25, 1990.The AET editor had some influence inscheduling the International Electric

Vehicle Symposium, EVS-10, to be held inHong Kong, so that people could attendboth events. However, the currentschedule, December 3-5, would not allowlate finishers time to get to Hong Kong,and the editor is asking for EVS-10 to beheld just before the race.

Letters to the Editor(continued from page 4)

Stability or control?In HP 7/3, Doug Milliken describes

some fascinating experiments, and thensuggests that control is more importantthan stability-but there really is nocontradiction. Statements such as ". . isthe condition required for stability" or".. vehicle configuration is unstable" areincomplete. For completeness, thequestion of stability or instability mustalways be associated with a particulardynamic mode of motion, with input andoutput variables defined.

Regarding page 9, center column,fourth paragraph, the "aerodynamicallystable" configuration with center ofpressure aft of center of gravity comesfrom an analysis in which steer angle isheld constant, very similar to the analysiswhich requires the neutral steer point tobe aft of the CG to avoid oversteer. Thus,it is more complete to say that the CG of a(multitrack) vehicle must be forward ofthe CP for a vehicle with fixed steerangles to remain on its path in cross-winds. If steer angle can vary, the abovestability analysis is less relevant, and acomplete stability analysis would includethe rider's control response to the sideforce. Without pretending to understandtheir results completely, I note that theexternal side force can be cleanlycancelled if it is applied near the steeredwheel, ie at the head tube, so Doug's andMax's observations make sense and donot particularly contradict the constant-steer-angle stability analysis.

In the same issue, Robert Price'sarticle makes an erroneous statement onpage 20, first column, last paragraph. Acar's pivot point during cornering lies onan extended line which originatesbetween the CG and rear axle only belowabout 10 m/s. At faster speeds, the pivotpoint moves forward, even well beyondthe extended front axle line. Thus the rearwheels always have a greater turning-circle radius during high-speed cornering,

so figure 11 is correct for front steeringonly at low speed.

John C. WhiteheadJCW Engineering3322 BiscayneDavis, CA 95616 USA

(Doug Milliken responds: '"Thanks toJohn Whitehead for clearing up an areathat I'd left slightly vague in my article"From Rob Price: "Mr. Whitehead iscorrect in that after about 10 m/s (25 mi/hr) the steering pivot point movesforward from between the c.g. and rear-axle centerline as shown in the car case inFigure 10. Just where the pivot point willbe at any instant is dependent primarilyon sideward tyre loading, which is afunction of vehicle speed and the radiusof the turn. Generally when the pivotpoint moves forward of the front axle thedriver has established what is known inautomobile racing as a 'four-wheel drift'or just 'drift.' It requires careful powerregulation and skillful minute steeringinputs to maintain this condition througha turn. It takes considerably more powerthan can be generated by a human tomaintain a drift, unless on ice or loosedirt, and is a condition which would beanomalous if encountered in an HPV, sodoes not need to be designed for.

Figure 11 does illustrate the slow-speed case for the bicycle. The rear wheelwill track with or outside the front wheelduring hard cornerning, as Mr.Whitehead states, but if the steering pivotpoint moves forward of the front-wheelcenterline, establishing the machine in adrift as described above, chances are greatthe bicycle is about to crash!")

Recumbents and suspensionRegarding Rob Price's article in HP

7/3 on steering and suspension design,could we get someone to write on how tobuild a simple recumbent-bicycle frame,and a simple suspension system? I haveseen everything from fancy systems likethe Moulton to springs-or even bungeecords-on a modified fork. I have alsoseen rear suspended frame trianglesusing both springs and bungees, as on theBowerbike. I figure, as a nontechnicalperson, may be able to build one, butaesigning is another matter altogether. (Iget so much out of HP articles and I reallyappreciate the knowledgeable authors fortaking the time to write. Thank you!)

Robert J. Bryant16621 123rd Ave, SERenton WA 98058 USA

(Mea culpa! See my note re Werner Stiffel'sarticle--ed). [

16 Human Power 7/4

-

-I_~

I