By Don Hewes (EAA 32101) 12 Meadow Drive Newport News, VA 23606
_L HIS IS THE final installment of a three-part article on the subject dealing with the responses of tail-first airplanes to rain, snow, bugs or whatever. The term "Flight Behavior Change", or "FBC" for short, has been coined to refer to this type behavior. This behavior results primarily from changes in the aerodynamic characteristics of the lifting surfaces. This part covers several recommendations and suggestions for people who are designing, building or flying this type airplane.
Recommendations and Suggestions The keyword for conducting FBC tests, as in the case of any flight testing, is BE PREPARED. Initial flight testing of any homebuilt airplane, even if it is the 500th copy of a well proven design, should always be treated as a truly experimental flight test operation with all appropriate precautions taken. After all, what is on the sign that you had to place on the door? If you don't remember, it is
spelled E-X-P-E-R-I-M-E-N-T-A-L. It is suggested that the critical portions of the tests be delayed until all others are completed. However, you can perform the preliminary portions of the tests while doing the other normal tests but be sure that the lifting surfaces are dry and clean and avoid flying in the rain. Most of the following tests should be done using a mid-location for the CG to begin with and then repeated with the CG moved progressively forward and rearward. The test pilot may elect to eliminate some of the intermediate steps depending on the observed behavior from the previous tests. Although there are many tail-first airplanes currently flying which show no significant problems when they encounter rain, bugs or whatever, there are enough cases of FBC problems to indicate that caution should be exercised when flying an airplane of this type with unknown characteristics. Because of the many variables associated with current techniques used in constructing homebuilt fiberglass-foam airplanes, I believe that the only way a
builder of a tail-first airplane can be certain of the airplane's flight behavior with contaminated surfaces is to conduct flight tests designed specifically to evaluate this condition. I believe that such testing is in keeping with the builder's responsibility for the airworthiness of his airplane. Furthermore, it is a matter of the builder being fair not only with himself but his family and anyone who rides in or flies his airplane that this be done. This recommendation applies to any canard or tandem wing airplane whether it be the first copy of a new design or the umpteenth copy of a proven design. In making the decision to conduct the flight tests, it is recommended that the builder contact the designer to discuss the tests and obtain his advice. The builder should be familiar with conducting such experimental test flights and have some recent flight time. As is true for initial flight testing of any homebuilt airplane, if the builder does not fit the role of test pilot, he should find some qualified person to do the flying for him. It is suggested that the builder prepare a flight report of the tests and submit them to the designer. The report should cover the effects of contamination on at least the following items: 1) take-off distance, 2) rate of climb following take-off, 3) landing speed and distance, 4) stick force and trim control travel, and 5) maneuvers at approach speed. Other factors such as elevator deflections and angle of attack changes should also be included. The designer can then correlate this information with his own tests and those of the other builders as a way of isolating those factors which are most important in affecting the FBC. Included in the builder's report on an airplane with a significant FBC problem should be detailed measurements of the various critical parts of the airplane. Concerning the accuracy required to provide repeatable aerodynamic characteristics for these airplanes, it is suggested that designers 1) evaluate the need for greater accuracy, and 2) provide information on how to achieve the greater accuracy, if it is required, taking into account the wide spectrum of builder skill involved in the current
homebuilding movement. Incorporation of final contour templates and alignment jigs should be considered. In the following sections, a series of flight tests to evaluate FBC is suggested for use in lieu of any other specific information supplied by the designer. The actual procedures will need to be developed by the builder and pilot following these suggestions. Also, suggestions are made regarding pilot preparedness and possible cures for
severe FBC problems. SPORT AVIATION 61
Flight Testing Before commencing these tests, the airplane should be equipped with accurate airspeed and rate of climb indicators and inflight calibrations of the airspeed installation should be made to insure reliable airspeed measurements within a couple MPH or better. An outside air temperature indicator should be available for airspeed corrections. Remember that these measurements need to be accurate so that reasonable comparisons can be made of other measured quantities obtained from various different airplanes using airspeed as the basis for comparison. If airspeed is in error, then the comparisons will not be reliable. Note that airspeed probes located in different positions of the airplane may produce significantly different readings because of local pressure differences (position error). An angle of attack indicator will be quite useful but is not an absolute necessity. A relatively simple vane mounted on a boom extending about a foot or two forward from the front surface near the tip can be used but it must be calibrated to account for flow upwash. To do this, fly the airplane at constant altitude for a series of speeds over the speed range and compare the vane reading with an inclinometer mounted in the cockpit. You will be dealing with only a fairly small angular range of about 15° and you should obtain each reading with an accuracy of about 1 A° or better. You can use a small electrical potentiometer attached to the vane and a meter hooked up in simple Wheatstone- bridge circuit to obtain the vane readings. Because the existence of a laminar boundary layer is the key to the FBC phenomenon, it is important to determine the amount of laminar flow that exists on your particular airplane for the various flight conditions. Therefore, it is recommended that the first test should be one to obtain a visual indication of the boundary layer flow conditions on the wing and tail surfaces. This can be done with either of two fairly simple techniques, the first is the one described in Reference 3 using sublimating chemicals. This has been used very successfully by NASA and is highly recommended. The second technique is similar but uses plain motor oil in place of the chemicals. It has been used extensively in wind tunnels but I have not had any personal experience with it. Also, I don't have any convenient reference for the steps involved, so a bit of experimenting will be involved. You should be able to find a couple quarts of old used motor oil that are heavily loaded with soot so that the oil is very black. Wipe this uniformly over the lifting surfaces to provide a thin layer of oil which will tend to migrate to the region just aft of the TRANSITION POINT. Be sure that the oil does not tend to form drops that linger otherwise you are creating the same effect as rain or bugs. You may have to adjust the viscosity of the oil by thinning with kerosene or thickening with heavier weight oil to get the proper effect. You will probably need someone to help observe the flow patterns which will undoubtedly change with the different flight conditions. You should expect to see flow transition point somewhere near the mid-chord position on the tail and probably further forward on the wing. If this test shows that the transition point is near the leading edge, then there is relatively little laminar flow and the subsequent tests will probably show only a mild FBC if any. On the other hand, if the boundary layer is laminar near or past the mid-chord position, it is possible the behavior will be much more
evident. It should be necessary to conduct this test for only one loading condition but at least three speeds from landing to cruise conditions should be covered. Remove the flow visualization material at end of test.
During this test, observe the position of the elevator
required to maintain straight and level flight for a given 62 JULY 1983
loading condition. Normally the elevator should be very
close to zero deflection or whatever was specified by the designer. If there is greater than a couple degrees from the desired setting, carefully inspect the airplane for improper rigging or some of the other factors covered in the previous sections. Following this test, conduct a series of baseline data tests to measure take-off speed and distance as well as the rates of climb at take-off power in the range from take-off speed to somewhat greater than that for maximum rate of climb using increments of about 5 mph or so. The measurements should be taken as you climb through the same altitude for each different speed so a series of saw-tooth climbs and descents will have to be made. If you wish, at the same time you can conduct idle-power tests which should also be made to measure the rates of descent for the same speed range. (You may have to be concerned with excessive cooling and heating cycles involved.) Be sure that you have a well stabilized climb or descent established before reaching the specific altitude for taking the measurement. You will probably need several hundred feet to do so. It takes practice . . . and a passenger to take the readings is very handy. A small hand held tape recorder is also useful in place of the passenger. Note touchdown speed and landing distance. The next step is to make a complete elevator vs. airspeed calibration (similar to that shown in Figures 9 and 10) for each of the loading conditions at power required for straight and level flight as part of the baseline data for the subsequent special tests. All that is required is a scale on the elevator control calibrated in terms of elevator position and on the trim control calibrated in terms of any convenient scale appropriate for the type of control handle used. If the trim control is a crank type, then you will need to keep track of the number of turns. The elevator scale should be accurate to within about one-quarter of a degree and should be read to at least one-half a degree. The trim system scale should be accurate to about 1 or 2 percent of the full travel. It is recommended that all these tests be flown at safe altitude of at least three thousand feet above
the local terrain.
The next step is to repeat the last step taking data for coordinated banked turns of 30 and 60 degrees. Then the following step is to perform a series of banked coordinated turns at the normal approach and touchdown speeds using slow and then rapid control inputs. Do this with power to maintain essentially constant altitude and then with normal reduced power for landing. Start with shallow turns and observe any increased tendency to pitch down or up. Carefully note any additional aileron/rudder deflections required to correct for possible roll/yaw tendencies caused by partial flow separation on only one side of the tail or wing. If a pitch break occurs, hold controls steady if possible to observe the motions of the airplane. Note rates of descent. Attempt to recover by pulling further aft on the control. This may result in an aggravated FBC with the airspeed increasing significantly. In this case, you may need to PUSH THE CONTROL FORWARD as you would in the case of the stall of a conventional airplane. .'Remember that the tail is in a stalled condition with the elevator deflected downward. Raising the elevator will allow the tail to become unstalled.) Note altitude lost in recovering. Increase the bank angle in small increments and observe any further tendencies for FBC, roll or yaw. Note the elevator deflections and airspeed at which they occur. Bank angles in excess of the limits normally observed need not be reached. CAUTION: The objective of the following steps is to fly the airplane with an artificially induced turbulent boundary layer on portions of the tail so as to simulate the effects of contamination. These steps pose some addi-
the last step checking to see that the elevator deflections and airplane flight behavior are still within acceptable limits. If the incremental changes experienced with the two different lengths of tape are insignificant, then you can proceed repeating the tests with the tape applied full span. Otherwise, proceed with smaller increments until the results indicate that you should proceed no further. Repeat these tests with the CG moved to the mid-point
ELEVATOR DEFLECTION DE6.
V UAX.
•0
100
110
and then the forward locations. If the full-aft limit for the
CG was not tested previously, it should be done also unless prior tests indicate otherwise. Be aware that these tests will probably be more critical so proceed with caution.
140
INDICATEDAinSPfCO, KNOTS
20 18
FIG. 9 -FLIGHT TEST DATA FOR VARIE2E.
16
min trim
TRANSITION
OFREE D FIXED
14
tional hazard and should not be performed by anyone who is not fully qualified and prepared to handle the airplane under conditions requiring emergency actions. When you are ready to start the final series of tests, carefully clean off the first 3 inches of the WING and TAIL leading edges using a cleaning agent to remove any traces of dust or oil so that a strip of masking tape will stay firmly attached in flight. Apply a double thickness strip (about .008 to .010 in.) of'/»to 3A inch masking tape about 2 inches from the leading edge, both top and bottom, starting inboard and extending out to 1A the span of each tail panel. Be sure that the tape used has very good adhesive qualities and press it firmly onto the surface. It is necessary that the tape be applied in short sections of about 12 inches each so as to eliminate the possibility of one loose end causing the whole length to peel off. Tape is applied inboard only for the first flight to minimize the possibility of large rolling moments caused by unsymmetrical separation, and to approach the most severe condition (full span trip) in a careful manner. With the tape in place, conduct a series of high speed taxi runs checking to see that there is sufficient elevator power to lift off within the first quarter of the runway. Use the mid CG loading condition for these tests. If it is determined that a satisfactory liftoff can be made without reaching excessive airspeeds, apply a small amount of the boundary layer flow visualization material used previously so as to check to see that the tape is tripping the flow. Place it in a location where you can easily see it, and then proceed with the take-off. Carefully note and record takeoff distance and airspeed. If the flow appears not to be tripped, land and add another layer of tape. Otherwise, proceed with repeating the previous rate of climb, rate of descent and elevator vs. airspeed tests for comparison with the initial data. Make a careful note of the minimum trim speed of the airplane for the landing condition. Then repeat the coordinated-turn maneuvers. Note any unusual characteristics that may be associated with the tripped flow condition. If any unsatisfactory or unsafe effects are noted, terminate the testing immediately. When landing, make only gradual turns and maintain an approach speed about 1.3 times the noted minimum trim speed. Note and record touchdown speed and landing distance. Compare the data and check to see that the elevator deflections required with the tripped flow condition are not excessive. There should be only very small changes from the original data of 1 or 2 degrees at the most. If the results are judged to be acceptable, add more
tape span wise and extend it to the Vz-span location. Repeat
ELEVATOR DEFLECTION
6
12 10
deq
60
80
100
120
140
160
180
INDICATED AIRSPEED, V ( . knots FIG. 10 -FLIGHT
TEST DATA FOP LONG-EZ
Cfi£F. 2)
Having performed all of these steps, you should have fairly well defined the operating characteristics of your airplane FOR THE MOST SEVERE CASE OF RAIN OR BUGS. If you were unable to complete the full series, then you know that some operational limitations should be placed on the airplane for conditions where the airplane might become contaminated. Be sure to remove the masking tape from the surfaces after a few days of testing to avoid possible damage to your finish due to the "curing" of the adhesive over a longer period of time.
Pilot Preparedness
If you are flying a tail-first airplane that has not been thoroughly tested for its FBC, you should be aware that you may encounter a FBC problem unexpectedly and you should be prepared to take the proper corrective action immediately. Do not attempt a flight if rain or snow are threatening or if the field is heavily infested with flying bugs. If possible, load the airplane so that the CG is in the mid to aft portion of design range.
Inspect the lifting surfaces and remove any surface
contamination that could cause flow turbulence. Inspect
the elevator travel for proper down travel limits. The travel stop should be positive with no "spongy" tendencies. Do not take-off from a field with long grass or weeds which can strike the leading edge. On take-off, check to see that excessive speed is not required to reach liftoff.
Abort the take-off if in doubt. If there is sudden power loss, avoid abrupt elevator inputs and banked turns.
SPORT AVIATION 63
If rain, drizzle, mist or bugs are encountered during flight, avoid slow flight and steeply banked turns. Make a SHALLOW STRAIGHT-IN approach with airspeed 10 to 15 knots higher than normal. Be prepared to ADD POWER if the plane suddenly starts to pitchdown and pick up speed. Avoid making abrupt aft stick inputs and allow speed to increase significantly before attempting to apply any aft stick. Allow for a longer than normal landing runout. In the event of a go around, remember that the airplane has higher drag and lower lift than normal and will not climb as rapidly as normal. Expect a shallow climbout and avoid any abrupt turns.
Curing Severe FBC Problems
Finally, we will address the question of what to do to the airplane if it demonstrates UNACCEPTABLE behavior. The first thing to do is review all information supplied by the designer (owner's manual, newsletters, etc.) and then contact him if you do not find an obvious source of the problem. If this is an original design, review the earlier discussions in this article. Check for misalignment. Check the tail for insufficient incidence and the wing for excessive incidence. (Both will cause excessive down elevator settings.) Shim the surfaces or make new attachment fittings. Carefully check the airfoil shapes using external templates to see that they conform to the design airfoils. Unless the designer can provide specific data for the desired contours of the airfoil, you will have to develop them from the normal construction templates. In this case, you will have to make allowances for the added thicknesses of fiberglass, resin and surface filler. An alternate method to making templates from the drawings is to make a series of exact half-templates (upper and lower) of the actual surface contour. These are then compared with a drawing of the desired airfoil. A simple procedure is to mask the chordwise section of the upper surface to be checked with a narrow strip of Saran Wrap or wax paper and then lay down a thin strip of "Bondo" plastic body filler about Vz inch wide. Before the filler hardens, press in a piece of plywood or hardboard previously rough cut to the approximate contour of the upper surface. The piece should extend below the upper surface at the leading and trailing edges. After the plastic hardens, carefully mark on the template the position of the trailing edge and a reference mark placed on the leading edge of the surface. Repeat the process for the bottom surface and then match the two parts of the template using the leading and trailing edge marks on the two parts. This process is quite easy and fairly quick but requires rework of the template whenever the surface is reworked. Some contour errors can be corrected by filling with microballoon slurry and refinishing. Others may require
building a complete new surface. If possible, the CG forward limit could be set further aft so as to avoid the higher loading of the tail. A possible redesign solution is an aileron reflex mechanism such as the system developed for the Quickie airplanes. Also, a small moveable horizontal surface located at the aft end of the fuselage to provide an inflight adjustable pitch trim moment could be installed in some
cases. Perhaps the most drastic but most satisfactory solution would be to build a new tail surface with an airfoil less susceptible to the flow separation problem. Unfortunately, there is relatively little data available on which to base a decision for selecting an alternate airfoil section. However, some of the latest airfoil computer design 64 JULY 1983
techniques now offer some hope for obtaining a suitable section. These steps represent major redesign effort and should not be taken unless all else fails, and then should be taken only after consultation with the designer. Of course, any modifications to the airplane should be checked with the FAA and thoroughly flight tested for all flight conditions.
Closing Comments We have specifically aimed this article at the problem of a more or less symmetrical stalling phenomenon in which there are no significant rolling or yawing moments present. However, it is quite possible that an unsymmetrical condition of the airplane exists so that the stall itself will be unsymmetrical. The pilot, therefore, should also think in terms of the potential for a ROLLOFF or SLEWING behavior associated with the FBC. This article has been presented to the reader for the purpose of exchanging information. Because of the highly experimental nature of both the information of this article and the flight activity associated with homebuilt airplanes, the author cannot assume responsibility for actions taken as the result of using this information, the suggestions or the recommendations presented herein.
Acknowledgements I would like to thank all of the following persons for their help in obtaining some of the information and data presented herein, and some of them for the advise and comments pertaining to this presentation: Dr. Bruce Holmes (Flight Research Engineer), Joseph L. Johnson (Head, Dynamic Stability Branch — full scale wind tunnel), Long Yip (Wind Tunnel Research Engineer) and Dan Somers (Airfoil Research Engineer), all of NASA's Langley Research Center. Also John Roncz (Airfoil Designer), 1510 E. Colfax Ave., South Bend, IN; Burt Rutan (President, Rutan Aircraft Factory); Gene Sheehan (President, Quickie Aircraft Co.); Bob Walters (Dragonfly Designer) and Rex Taylor (Viking Aircraft Ltd.). This acknowledgement should not necessarily be construed as representing an endorsement on the part of any one of the individuals or organizations mentioned.
References 1. Yip, Long P. and Coy, Paul F.: Wind-Tunnel Investigation of a Full-Scale Canard-Configured General Aviation Aircraft. 13th ICAS Congress/AIAA Aircraft Systems and Technology Conference, Seattle, Washington, Aug. 22-27, 1982. ICAS Paper Number 82-6.8.2. 2. Holmes, Dr. B. J., Obara, C. J.: Observations and Implications of Natural Flow on Practical Airplane Surfaces. 13th ICAS Congress/AIAA Aircraft Systems and Technology Conference, Seattle, Washington.
Aug. 22-27, 1982. ICAS Paper Number 82-5.1.1. 3. Holmes, Dr. B. J., Croom, C. C., Obara, C. J.: Sublimating Chemical Method for Detecting Laminar Boundary-Layer Transition. Handout available from
Dr. Holmes, Mail Stop 286, Langley Research Center, Hampton, VA 23665.
