* ++

When I started this series of articles last February, I had already spent eight months thinking about a new homebuilt for my own use. I had also found talented partners, Larry

by JOHN G. RONCZ, EAA 112811 15450 Hunting Ridge Tr. Granger, IN 46530-9093

Lombard and Mike Dilley, who were willing

done a superb job of reducing the drag of the engine installation. Then they had come up with a tricycle-gear version of the airplane, which included a very clever nosewheel installation. Mr. LoPresti had also convinced Lycoming to do a special 200 horse engine for him, which has the fuel injector head on the front of the engine, instead of on the back. This permitted a clever ram-air installation. On the standard version of the IO-360 engine, the fuel injector head faces the firewall. This means you have to take the induction air, and turn it 180 degrees to feed the fuel injector. On my Rockwell 112's, this resulted

to work on the design with me, and later build whatever we designed. They came to Indiana, and we laid out a forward-swept wing 2-place design, using a Lycoming 180 horsepower engine. I had already obtained the engine performance data and installation blueprints from Lycoming. Larry, who is also an artist, drew the first print (Figure 1). The design objectives set way back then were: two place 180 HP four cylinder engine

roomy cabin stall speed of 55 knots standard (large) tires speed matching the Duchess (160 knots +) exceptional visibility good high altitude performance Back last May I showed you revision 14 of the airplane, and told you that it was the final revision. Boy, did I lie! Larry and Mike are currently building the tooling - for revision 20! Let's see what happened to the airplane since then.

in a 2 inch loss in manifold pressure - this means the engine always thought that it was 2000 feet higher than it was, with a corresponding loss of performance. On the LoPresti version of the engine, the injector can be force-fed by the propeller, making the engine think it is at a lower altitude - a gain in performance. By stealing the LoPresti engine, nosewheel, engine mount, and all related hardware, we feel that we will save an enormous amount of time and eliminate a lot of development work. Curt LoPresti supplied drawings and all the weight and balance information for their installation. Since it weighed more

than our original estimates, it required some additional work on our airplane to make it balance. Our homebuilt, which we recently named

TWO PLACE

We now have a four-place airplane, although the weight and balance won't allow two big guys in the back. The first cut at this

was drawn by Mike and Larry, who felt that only minor changes would be needed to accommodate the extra people (Figure 2). The rear seat passengers faced aft, because they had a lot of leg room that way, and we thought that the center of gravity of the airplane could be kept in a smaller range.

FIGURE 1

Once I had the design drawn on the computer, I found that turning the back seat passengers around, to face them forward, did almost nothing to the aircraft's center of gravity. So now we have a four-place airplane, with everybody facing forward. 180 HORSEPOWER ENGINE

Last summer I approached Roy LoPresti, of Grumman, Mooney and LoPresti-Piper fame, about stealing everything firewall forward off the Swiftfury. His son, Curt, had

FIGURE 2 SPORT AVIATION 29

however, to survive an engine seizure at night in such a beast - an event which I myself did survive in a more sensible airplane. If you really want to do yourself in, couple high stall speed and high sink rate with little bitty tires, so that they'll dig into the unpaved

ground and flip you at high speed if you manage to put her down somewhere. If you choose such an airplane, please make me the beneficiary of your life insurance policy! Larry, Mike and I all value our continued existence. For those reasons we decided to compromise top speed for lower stall speed, and insisted on 6.00 x 6 airplane tires. My

challenge was to do these things without a severe performance penalty. The wing airfoils I designed for our bird sacrifice a small amount of drag at cruise in order to gain a lot of lift for landing. Coupled with semiFowler slotted flaps. Forte should meet its

stall speed goal.

EXCEPTIONAL VISIBILITY

FIGURE 3

Forte, now is powered by a Lycoming 200 horsepower engine, 20 more than our design goal had stated. ROOMY CABIN

The fuselage is 49 inches wide at the front seat. It is 54 inches high at the same place (outside dimensions). A large cabin does not have to slow down the airplane, if it is shaped correctly. Cessna's 310 and 340 share a common wing. The 340 has a much bigger cabin-class fuselage. Yet they both have essentially the same drag! So larger doesn't necessarily mean draggier. Even if it did slow the airplane down by 5%, I personally would prefer to be comfortable for 5 hours than uncomfortable for 4 hours and 45 minutes. Think about it. Last May I thought I had the Great Tire Crisis resolved, with the tire parked merrily inside the wing. This scheme didn't work. My friend Rob Schirtzinger arrived in June, armed to the teeth with finite element structural analysis software. We put together spreadsheets to calculate the spar thickness and shear web layups, then made a finiteelement computer model of it. The computer thought about all of this for awhile, and decided that yes, indeed, forward swept wings are structurally unstable! In a nine g pullup, the wingtips deflected up as you'd expect, but they also increased their angle of attack in the process. That's a no-no. The second

spar arrangement also failed this test. The third worked like the proverbial charm - however, Rob had parked his spars right through the middle of my tire! This is why I have a love/hate relationship with structural engineers.

I couldn't bring myself to using toy tires on

30 FEBRUARY 1991

Forte, which would have solved this dilemma. So we reverted to the original scheme of putting the tire inside the fuselage, under the front seats. The fuselage had to be deepened to accomplish this. You can see the difference between versions 14 and 18 of the airplane in Figure 3, which the tire caused. On the version 14 drawing, you can see two sizes of pilots in the front seats. Forte is going to have seat tracks just like the big boys, and we needed to find the range of adjustment using a small and large person (2.5 and 97.5 percentile). We are also considering a gear-up landing's effects on the pilot's spine. I have a German friend and fellow Ferrari worshiper

who has painfully recovered from the paralysis caused by a forced landing. He is now an expert in seat design, and will be helping us with this. One of the main reasons that version 18 has a much deeper fuselage cross section is to house the tire inside the body, and provide clearance for the padding that will protect the pilot and passenger in the event of a hard gear-up landing. STALL SPEED OF 55 KNOTS AND LARGE TIRES

As I've told you, I have experienced two

engine failures in my 1700+ hours of pilot-

ing. I therefore know that every aircraft engine will fail eventually, in spite of whatever tender loving care you give it. For that reason, I will no longer fly a single engine airplane which has both a high stall speed and high power-off sink rate. This rules out many currently popular homebuilts. There is certainly no magic in taking a small airplane with tiny wings, stuffing in a monster engine, and making it go fast. It vv/7/take a magician,

Once I was flying my Duchess to California, and passed over a meteor crater in Arizona. I decided to ask Air Traffic Control for permission to circle the crater, so I could see it better (I was on an IFR flight plan). After they said OK, I circled around. All I could see was the top of the wing or the side of the engine nacelles! Other similar experiences made me long for an airplane I could see out of both up and down! This requirement led to the forward sweep of Forte's wing. It created many problems, but I don't regret the amount of work. If you look at the side view of the airplane, and draw a line from the pilot's eye to the leading edge of the wing, you'll see that we have excellent visibility for a low wing airplane. You'll also notice that the passengers' eyes are completely behind the wing, so they'll be able to see up and down also. I also have a personal hatred for windshield comer posts. So my own copy of Forte will have a post down the center of the windshield curving up between pilot and copilot (a roll-over bar), with a one-piece canopy on either side. I find the center post easier to see around than a windshield corner post. CRUISE SPEED > 160 KNOTS

As an airplane accelerates, the drag increases and the thrust decreases, until the two are in balance. Making your airplane faster therefore depends on making more thrust and/or having less drag. The conventional solution is to stuff the biggest motor you can

afford into the smallest airframe you can tolerate. This indeed gives you a fast airplane which works great as long as the big motor continues to operate. Another more elegant solution is to reduce

the drag as much as possible. This is the approach Roy LoPresti used on the Grum-

man Tiger, the Mooney 201 and the Swiftfury. This was the approach I used on

Forte as well. Before you can reduce drag, it helps to understand what drag is. Imagine that you are sitting in your living room listening to Dmitri Shostakovich (probably the greatest musical mind who ever lived). Suddenly an airplane flies through the room. If nothing was disturbed by its passage, the airplane would have zero drag. In reality, the drapes would flutter, the newspapers would go sailing, and the dog would hide under the couch. The airplane's

CP

FIGURE 4

drag may not be responsible for the dog, but it did cause all the other commotion. Looking at it objectively, your living room contains so many cubic feet of air. Each cubic foot of air has a certain number of air molecules. Each molecule weighs something - it has mass. So before the airplane rudely interrupted Shostakovich, a mass of air was at rest. After the plane whizzed by, this mass was put into motion. It took work to put a resting mass into motion. That work falls into the category of drag. Now imagine that all the air in your living room was colored blue (you changed to Gershwin's Rhapsody in Blue). Video tapes of the event taken by your nosy neighbor clearly show that the airplane was surrounded by a blue halo as it left your living room. Why? As the airplane passed through the room, its skin friction "kidnapped" some of the blue air and took it along for the ride. The mass of this kidnapped air, which was at rest, was accelerated to nearly the speed of the airplane. It took work to do that, also. Finally, your wall thermometer, which of course is accurate to millionths of a degree, shows that it is a bit warmer now than before the airplane flew by. Whenever you stir molecules, the temperature rises. All the energy to accomplish the kidnapping, temperature rise, and putting the air into motion came from the aircraft's propulsion system. If you added power, the airplane could then accelerate the kidnapped air to a higher speed; accelerating the stolen air to a higher speed would use up the extra energy available. So the greater the energy you have available to spend, the more havoc the airplane can wreak on the mass of air in your

Your job as an airplane designer is to

shape the airplane and its components so that you leave the air as undisturbed as possible. Laminar airfoils kidnap less air than turbulent airfoils. Fuselages which are curvy everywhere disturb the air less than fuselages with flat sides. Exposed antennas kidnap air, while buried antennas don't. At high speeds, the kidnapping of air done by skin friction is the greatest crime the airplane commits. At low speeds, the downward acceleration of the air, which blows the

newspapers around is the biggest crime the airplane commits. We'll look at how to reduce the crime rate and make the airplane faster. However, the quickest and easiest way to

make the airplane go faster is to move the pitot tube and/or static ports around. It's very

easy to gain 20 knots or more this way. A

couple of factory-builts and several homebuilt kits use this method. The fact that this

is merely instrument error doesn't appear in

the sales brochures, I guarantee you! If you've got one of these, you'll be amazed at how often you seem to be flying into a headwind, after you compare airspeed to groundspeed.

CRIME REDUCTION

AT HIGH SPEED

There are two approaches to reducing the crime rate at high speeds. The logic of the first goes like this: each square foot of surface area is kidnapping so many air

UPPER SURFACE PRESSURES

t

tower pressures LOWER SURFACE PRESSURES

VERSION 18 UNMODIFIED AIRFOIt

hic^er pressures

VERSION 20 MODIFIED AIRFOIt MODIHIO IUSUAGE

MODIFIED AIRFOIL BY II SELF (NO FUSELAGE)

living room. This is the physical picture of

what drag really does.

FIGURE 5 SPORT AVIATION 31

DLR20B flLPHfl-0 FIGURE 6

molecules. The answer, then, is to reduce the number of square feet doing the kidnapping. This approach leads to airplanes with small wetted areas. Our homebuilt finished the race with 430.55 square feet of wetted area. A lot of work went into keeping this as small as I could, while keeping the big cabin and a wing big enough to have a low poweroff sink rate and a 55 knot stall speed. By comparison, this is 16.9% less than the wetted area of a Defiant, 11.6% less than a Wheeler Express, 23.9% less than a Cirrus, or 6.6% less than the original 2-place Swift. It is, however, a whopping 53.8% more than

the 2-place Questair Venture (!), and 8% more than a 5-place Catbird. Considering the

cabin volume, tire size and stall speed, I'm happy with our final number. I kept meticulous records of the wetted area and internal volume of the fuselage as I went along. Since you have pictures of versions 14,18, and 20, you might be interested in how these numbers changed during the design refinement process: VERSION WETTED AREA

VOLUME

14 18

195.57 FT'2 197.19 FT'2

154.74 FT'3 158.04 FT"3

20

202.84 FT"2

168.51 FT"3

These numbers are for the fuselage +

spinner + canopy only, and don't include wings or tails. It's interesting to note what 32 FEBRUARY 1991

the extra wetted area did to our cruise speed. The final version 20 should cruise at 184 knots (212 miles per hour), which is 24 knots more than our original goal. If we could have built version 14, which has less wetted area, it would have cruised at 185.1 knots (213.3 miles per hour), assuming that the drag per square foot of wetted area stayed the same. This is a false assumption, because the wing/body intersection drag would have been higher in version 14 than it is in version 20. Nonetheless, 1.1 knots isn't much, considering that we gained 13.8 cubic feet of space inside the fuselage, which is quite a lot.

The logic of the other approach to reducing drag in cruise goes like this: I will reduce the number of air molecules being kidnapped by each square foot of area. This approach depends upon being able to predict, at least in a general way, how changes to the airplane's shape will affect the number of

molecules kidnapped by each square foot of wetted area.

For the Forte, this analysis took three forms, each of which required rather sophisticated software. The first finds the distribution of pressure over the skin of the entire airplane. The pressures can be color coded, as in Figure 4, to make the plot easier to interpret. You can think of pressure as push-

ing or pulling on something. The dark blue color represents the highest pressure pushing on the airplane. The pink color represents the lowest pressure pulling on the airplane. If you put your car in neutral, you can pull on the front or push on the back to move the car forward. An airplane in flight has pressures pushing or pulling on all the surfaces of the plane. I try to lower the pressures on forward facing surfaces (pulling on the front) and increase the pressures on aft facing surfaces (pushing on the back). I try to avoid pushing on front facing surfaces or pulling

on rear facing surfaces, because these forces would try to slow down the airplane. If you study airplanes which have benefited from a drag cleanup, like the Mooneys or the Swiftfury, you'll find that formerly blunt

windshields have been replaced with more steeply angled windshields. The blunt

shapes had high pressure air pushing on their forward faces. Slanting the windshields more reduces this effect. You'll see the same thing on modern auto designs, for the same

reason. The sleeker windshield shapes have

more wetted area, so this is an example of how making something bigger not only doesn't slow down the plane, it actually speeds it up. The reason is that the reduction in pressure drag more than offsets the in-

CP

FIGURE 7

crease in skin friction drag. The pressure signature of an airplane's wing also carries over onto the fuselage. On many airplanes, the fuselage begins to contract near the trailing edge of the wing. This contraction means that the fuselage skin begins to face rearward. You'd like to push on these rear-facing surfaces, in order to help push the airplane forward. It becomes obvious, then, that you can help out the fuselage

drag situation by using a wing section shape which slows the air down (raises its pressure) before the air hits those aft-facing sur-

faces. The wing root airfoils used on the Starship, Catbird, and Triumph all used this approach. It's quite easy to see the difference in wing airfoil shapes near the fuselage on the Catbird, so take a close look next summer at Oshkosh. This is also one of the reasons that some people using recent high-pitching-moment

laminar airfoils find that the plane goes faster if they deflect a trailing edge flap up in cruise.

Based on reports of this published in Sport Aviation, it seems that this gain in speed occurs in spite of the fact that the amount of deflection they're using is killing the laminar

flow on one side of the airfoil, and therefore

increasing the airfoil's drag!. This shows that marrying the wing pressure distribution to the

fuselage pressure distribution is one of the most useful things you can do with a pressure analysis. Forte was the first forward-swept airplane I've done. I was quite surprised at how intense the hatred was between wing and fuselage on this design. It took three redesigns of the fuselage, and eight redesigns of the wing root airfoil to make them love each

of the plot, higher towards the bottom. There are two curves plotted in each color. One shows the pressures on the upper surface, the other for the lower surface of the wing. The leading edge of the wing is to the left. The dark blue color shows the pressure

distribution for the original wing, for a slice taken one inch outside the fuselage. In general, if the curve is sloping up as you move to the right, the airfoil will have laminar flow.

The lower surface plotted in blue shows a small area of low pressure just behind the leading edge. This "spike" is large enough to change the laminar flow to turbulent flow on the lower surface. To fix this, and to make the pressure distribution conform more to the

original intent, both the airfoil and the fuselage were redesigned. The end result is shown in black. Notice that the lower surface no longer has the spike near the leading

edge. In order to show you how much influence the fuselage has, I took the final airfoil section, whose pressure is plotted in black, and ran it without the fuselage. This pressure result is shown in red. By comparing the red and black plots, you can see that the fuselage makes a dramatic difference to the pressures, and hence to the performance of the wing near it. Another tool I use is to paste electronic tufts of yarn onto the airplane, and watch which way the wind blows. This shows the path that the air will take over the fuselage. Such a diagram is shown in Figure 6, which shows a closeup of the wing/fuselage intersection, at the trailing edge of the wing root. If the wing and fuselage hate each other, which they usually do on the first attempt,

other. This is the most work on this design

juncture problem. I didn't show you 18A thru

G, and 19A thru E, which were intermediate steps along this road. I use several tools for examining the relationship between wing and fuselage. One is another kind of pressure analysis, showing the pressures along a spanwise slice of the wing. Figure 5 shows this kind of pressure

diagram, called a pressure distribution, for the wing root airfoil on 3 different versions of Forte. Lower pressures are towards the top

things come out right), I attack the wing airfoil

near the fuselage, and sometimes the fuselage shape itself, until the two get along. The "tufts" in Figure 6 are well-behaved behind the trailing edge. Figure 7 is another pressure distribution plot. You can see in this pressure diagram that the modified wing root indeed slows the air down on the side of the fuselage. It shows up as the red, yellow, and green colors being pulled forward on the fuselage towards the top of the wing. This raises the pressure of the air on the aft-facing fuselage surfaces behind the wing. I have colored this plot so that anything which is dark blue shows higher than ambient pressure, pushing on

the airplane. You can see that most of the tail cone is colored dark blue. I ran the fuselage through this analysis with and without the wing. On this airplane, the fuselage with the wing attached had lower drag than the fuselage by itself. This is one way of verifying that the design has been optimized. A detailed knowledge of pressures on an airplane also let the designer know how much lift occurs at each spanwise station along the wing. This allows the wing shape to be refined, and its stall pattern altered so that a gentle stall will occur. On this forwardswept wing, the lift falls off rapidly as you get out towards the wingtips. I had to increase the taper ratio (bigger root chord and smaller tip chord), change the camber in the airfoils

SPANWISE LIFT COEFFICIENTS - FLAPS UP STALL

detail that I've ever had to do on any of the airplanes I've worked on. All the changes you see in Figure 3 between versions 18 and 20 are solely due to this fuselage/wing

the air will take a circuitous path along the fuselage ahead of and behind the wing. Being the founder of the Atilla the Hun school of airplane design (we rape, pillage, loot, and do whatever else is necessary to make

I 800 -T1.500 ~p

1 200 -r 0.900 - 0.600 -: 0 300

:

0 000

———-

0.000

50.000

100.000

150.000

200.

BUTTLINE

FIGURE 8 SPORT AVIATION 33

air and throw it much harder at the ground. This produces a kind of drag called induced drag. It is the price the airplane pays for throwing air at the ground.

FIGURE 9

at the root and tip, and also increase the wing incidence at the tip in order to get a perfect span loading for climb, and a high maximum lift. You see, it's easy to get a gentle stall: just put 10 degrees of washout in the wing, and it will definitely stall root first and protect the tips! The reason you probably don't want to do this is that your wing will then start to stall at the root while the outboard part of the wing is barely doing any lifting at all. The end result is a much higher stall speed for the airplane. Conversely, if you make the whole wing stall all at the same time, you'll get the lowest possible stall speed, but when the sucker lets go it'll let go all at once, and it will help to have a rodeo rating on your pilot's license to recover from it. A forward-swept wing helps to make the wing root stall first. I found that I could load up the wing outboard a bit more than for a straight wing, and still have nice stalling behavior. Figure 8 shows what the lift coefficients are at the flaps-up stall speed, at each location down the wing. This information came from the detailed three-dimensional pressure data. REDUCING THE CRIME RATE AT LOW SPEEDS

Skin friction really takes its toll at high speeds. At low speeds the airplane is struggling just to stay in the air. At high speed a lot of air is passing over the wings. Since the wing has a lot of air to work with, it only has to throw it very gently at the ground. As you remember, the ground gets upset by this and pushes the airplane away - the effect we call lift. At slow speeds not much air is passing over the wings. To make up the difference, the wings have to take this smaller mass of

There are two ways to reduce this kind of drag. The first and by far the most powerful is to make the wing wider by adding span. An airplane flying at low speed is like a bicycle moving slowly in soft sand. Both are stuggling not to sink. A bicycle with a narrow racing tire is going to struggle more than one with a wide balloon tire. Increasing the span of the wing is the same as increasing the width of the tire. They both work better at low speeds. If you just don't care about drag at low speeds, you can make the span as short as you like. A good example is the F-104. Since there is a structural price to be paid for long, skinny wings, eventually you reach a practical limit to wingspan. So the next step is to make the lift distribution from wingtip to wingtip look like the top half of an ellipse. The success of this approach depends upon having the fuselage carry over the lift from the wings, making as much lift as the wing buried inside it would have made. Otherwise, you get a big "lift hole" in the middle, which screws up the induced drag. Looking back at the pressure diagram in Figure 4, another thing to notice is that at the juncture with the wing, the sides and top of the fuselage are colored red, indicating low pressure. This proves that the lift is being carried from the wing up and across the top of the fuselage, making the fuselage lift as much as the wing inside it would have. This is the lift carryover we spoke of. It took quite a bit of work to accomplish this, but it should pay off in climb performance. The combination of large span and good span loading should help the airplane meet our design goal of good performance at altitude, because the induced drag varies with indicated airspeeds, which are lower at high altitudes. All the intersections on Forte were refined using these kinds of pressure studies. The engine cowling and windshield shapes were modified many times until their pressures matched some mystical ideal I carry around in my imagination. OTHER COMPUTER TOOLS

One of the most tedious jobs you'll face as you design your own homebuilt is lofting the fuselage. "Lofting" refers to the task of generating cross-sectional drawings of your fuselage every few inches down its length. The goal of the side and top views is to enclose the engine and people in a pleasing

shape. The engine needs clearances to handle engine starts and shutdowns, when the engine often shakes on its mounts. I found a "gotcha" in the fact that the spark plug wires angle outboard from the tops of the cylinders. Extra clearance is needed to accommodate these dumb wires. Check a couple of planes in your local maintenance shop, and you'll see what I mean. North American's P-51 Mustang was, to the best of my knowledge, the first airplane whose fuselage was constructed using mathematical equations. The advantage of using equations is that once you have them, you can find any dimension anywhere along the fuselage's length quite conveniently. It's also easier to ship one page of numbers than a mountain of drawings! Another advantage is that mathematical curves won't have flat spots or ridges in them. These discontinuities are easy to construct using French curves, since sometimes curves which look pretty smooth to the human eye will have pressure distributions which are anything but smooth. The reason is that pressures don't respond to surface slopes (first derivatives), or surface curvatures (second derivatives), but to the rate of change of curvature (third derivatives). If you don't understand that, don't worry - you don't need to. The point is that the human eye isn't calibrated to see third derivatives, because this wouldn't help us to catch or grow food, or to escape from predators trying to eat us! It's better, then, to trust the equations. Roy Liming, who was Head of Engineering Loft Mathematics at North American Aviation, wrote a book about his system (see references). Based upon this simple system, Peter Garrison wrote a very useful computer program, called FLOFT, for designing fuselages or other similar shapes (see Sport Aviation, September 1990). The basis for the system can be understood by studying Figure 9. This shows the cross-section (front view) of a hypothetical fuselage station. On it you can see a large square divided into four smaller squares. In order to define a fuselage cross-section, you need to know the following details: the top of the fuselage the bottom of the fuselage the width of the fuselage the vertical location of the maximum width point the roundness of the top corner the roundness of the bottom corner Since the fuselage is symmetrical left to right, you only need to draw one side. I'm showing you both sides in Figure 9, which also shows the widest point of the fuselage halfway between the top and bottom. Of

course, the maximum width might not be

5 LINES DEFINE MAX WIDTH

halfway from the bottom to the top, and usually isn't. You can see that I scaled Figure 9 so that any diagonal line from the center to a corner has a length of 1. What's missing

is a system for specifying the "roundness" of the fuselage corners. Liming's system uses K factors.

7 LINES DERNE TOP LINE 4 LINES DEFINE BOTTOM LINE FIGURE 10 34 FEBRUARY 1991

The mysterious K factor is simply a percentage of the length of this diagonal line. Looking at Figure 9, I've drawn a circle in red. The K factor for a circle is .707 (square root of 2 divided by 2, for you math buffs). So for a spinner, you set the fuselage height equal to its width, make the maximum width

occur halfway from top to bottom, and set

the K factor at .707. This gives you a round spinner by pulling the curve 70.7% of the

way towards the corner. If you set the K factor to 1, you'd get a square corner. In effect, it would pull the "curve" 100% into the corner. Usually you want something squarer than a circle, but less square than a square. K factors of .80 and .90 are plotted on Figure 9 to show you what curves pulled 80% and 90% of the way to the corner look like. The shape at any fuselage station is completely determined then, from just 6 numbers. How do you get these numbers, then ? Figure 10 shows the answer. In Liming's method, only a few points are needed along the fuselage. Look at the bottom of the fuselage in Figure 10. Only four points were defined: 1) at the bottom of the spinner, 2) below the pilot's seat, 3) behind the back seats, and 4) at the aft end of the fuselage. I've marked points 2 and 3 with tick marks, so you can see them. At each of these points, a tangent line is attached. A tangent line is like a ruler which you can rock on the surface. It shows the local angle of the fuselage as it passes through each point. The four tangent lines intersect each other, forming three corners. This series of intersecting lines sure isn't pretty! What's missing is a way of creating mathematically smooth curves from these straight lines. This is done by using K factors to pull curves into the corners formed by the intersecting lines. For the fuselage bottom, the K factors I used are .7316 for the front of the cowling, and .79 for the other two comers. FLOFT has a built-in editor, which allows you to change the slopes of the tangent lines, move the control points, or change the

K factors until the shapes please your eye.

Since we don't trust our eyes, FLOFT also

displays a graph showing surface curvatures

calculated mathematically. When you start, this graph usually looks like a profile of the

Rocky Mountains. But by using the editor to

make small changes, you can end up with

smooth variations in curvature, which Mother

it rounder towards the aft end of the fuselage (I used .775). To specify how the K factors change from nose to tail, you create a drawing just like those used for the top, bottom and sides, using the tangent lines and K factors. Instead of a dimension in inches, however, the dimensions are just the K factors. The variations in K factors along the body

Nature should applaud. The curvature plots for Forte are shown in Figure 11. The K factors themselves also change along the length of the fuselage. You start with .707 to get a round spinner shape, then

can be displayed by FLOFT. You smooth these curves using the editor, just like any

the pilots feet, to get a flatter floor, then make

view. I lofted the fuselage, canopy, and spin-

go to something more square (I used .90) by

other curve. FLOFT will then show you a slice anywhere along the body, even slices at an angle. It will also show you a perspective

FIGURE 12 SPORT AVIATION 35

where the fuselage is flatter, and smaller where the fuselage shape changes more rapidly. The desktop computer really has revolutionized the speed and accuracy of the airplane business! MacBravo allows you to write your own commands. Using this ability, it reads in the shapes I generate from the QuickBasic programs. Then it can shade the airplane, so that I can look at it from any angle. I can even move the sun around and put the shadows where I want them. Figures 13 and 14 show the MacBravo shaded views of Forte photographed off my Macintosh screen. OTHER TOOLS

The last item which has really helped me is a terrific book by L. Pazmany called Landing Gear Design for Light Aircraft. Mr. Pazmany has collected just about everything you ever needed to know about landing gear geometries, stresses, tires and brakes into one handy book. It is a real work of art. THE END FIGURE 13

ner as separate bodies. That way I could change the canopy shape without messing up the rest of the fuselage. FLOFT will combine these to show you a composite view. Figure 12 shows a perspective view with cuts every 4 inches down Forte's fuselage, photographed off the computer screen. FLOFT is a marvelous tool which has saved me hundreds of hours. Mr. Garrison has recently been offering it to homebuilders. Check the classified ad section of Sport Aviation for his advertisements. For making 2-dimensional drawings, I use a computer program called Vellum. It is available for both Macintosh and IBM-compatible computers. The first time I used it I was ready

to quit my job and become a salesman for Ashlar, who produced it! It is the only computer drafting program that actually makes it faster to draw by computer rather than by hand. Since Vellum can only do 2 dimensions, I use another Macintosh program called MacBravo to do 3-dimensional drawings. I wrote my own programs, using Microsoft's Quick-

Basic, which reads the files produced by FLOFT, and creates the wireframe models needed to do the pressure analyses. These programs find the intersections of all the parts, divides the airplane up into little panels, and creates the files I need for the pressure analysis. It even makes the panels bigger

From the letters and phone calls I've gotten, I know that many of you have found this series of articles useful. My goal was to put a lot of technical tools out there, with the hope that future homebuilt designs might be safer and better thought out. I also hope to force some of the kit manufacturers to be a bit more honest with us. The sad truth is that some published performance numbers are dishonest. Also, some kits offered to you are for unstable airplanes which could never pass the muster if they had to meet the FAA requirements. You now have the means to plug a homebuilt's numbers into a computer, and check out their stability and performance for yourself. I don't want this kind of information to end here. I've been trying to find a structural engineer to take over, and publish spreadsheets which you can use to size the structure for your airplanes. If you fall into this category, please do it! Yes, it is a lot of work, it doesn't pay one cent, and you'll be inundated with letters. It is still worth it! We are living in a time when the factories will no longer produce new light airplane designs and maybe not even old designs much longer. The hope for the future lies here with the EAA and homebuilt designs. The best gift you can give aviation is to pass on your knowledge to other people who also love airplanes. I also want to see a similar series for flighttesting your homebuilt. We need to find a professional test pilot willing to write the articles and spreadsheets for this part of the job. I you qualify, please go for it! Finally, I'd like to thank Jack and Golda Cox for their patience and encouragement. Without them, I would have given up months

ago. I hope they eventually forgive me for having massacred so many deadlines during the past year. (You're forgiven, John, but you still owe us a Chopin concert. - Editor) Now get out there and design some new airplanes!

REFERENCES Liming,

Roy

A.,

Practical

Analytic

Geometry with Applications to Aircraft,

The Macmillan Company, New York, 1944.

Pazmany, L., Landing Gear Design for Light Aircraft, Pazmany Aircraft Corp., Box

FIGURE 14 36 FEBRUARY 1991

80051, San Diego, CA 92138, May 1986.