ABSTRACT

T h e Gulfstream 11 bus i~~ess jct is a low wing aircraft of moderate sweep featuring a T-tai l arrallgcnicnt and twoaft- mounted turbofan engines. The paper out l i i~cs aerodynamic development of the configuration and colitrol system, high- lighting significant design clecisions ancl aerodynaniic char- acteristics with appropriate experimental data including force, pressure, and flow visualization rcsults. Aerodynamic design considerations included: wing optimization t o mee t the requiremenrs of high and low speed performance and a pitch-down stall break a t a l l flight conditions while provid- ing adequate fuel volume; nacelle-pylon-wing relationship for optimum drag, engine characteristics, and airplane bal- ance; and an empennage arrangement providing satisfactory stability and control a t a l l conceivable flight conditions. T h e Gulfstream I1 configuration provides strong aerodyhamic resistance to inadvertent secondary stall entry and more than adequate recovery capability.

THE GULFSTREAM I1 is a twin-jet swept wing corporate transport designed to provide high speed and long range ca- pability without sacrificing the airport performance, reli- abi l i ty , and other operational advantages of its predecessor, the turboprop Gulfstream I. T h e design originated when product improvement studies on the Gulfstream I indicated only a new configuration would realize the improvements in performance desired. On the basis of these studies, per- formance goals of coast to coast range, 0.80 Mach number cruise speed, and airport performance equal t o Gulfstream I were established. During preliminary design, the cruise speed requirement was raised to 0.83 Mach number to pro- vide intermediate range route times competitive with air- l ine schedules.

The Gulfstream 11, shown in Figs. 1 and 2 , features a low wing of moderate sweep, two aft-mounted engines, and a T- ta i l empennage arrangement. T h e engines are Rolls- Royce Spey RB 163-25-turbofans rated a t 11,4001b of take- off thrust, giving a minimum thrust to weight ratio of 0.4. Maximum take-off gross weight is 56,000 lb. A fuel capac- i ty of 21,50011, is sufficient for a theoretical (dry tanlts) range of 3400 nautical miles. Take-off and landing per- formance is illustrated in Fig. 3. T h e cabin is comparable in size t o the DC-3, typically accommodating 10-12 pas- sengers in a corporate interior.

CONFIGURATION ARRANGEMENT

In establishing the aircraft's general arrangement, sug- gestions from corporate pilots and the lessons and examples of a successful prcdecessor weighed heavily within the nor- m a l periormancc and geometrical restraints. 'The general

f stream

Aerodynamic Design

I . T. Waaland and E. J. Curtis Grumman Aircraft Engineering Corp.

Fig. 1 - Gulfstream I1 general arrangement

s ize, fuselage cross-section, and low wing location are ex- amples of "inherited" configuration details, whereas the wing planform and ,contours reflect the increased perform- ance spectrum.

Preliminary design of the wing was influenced by both cruise and low speed considerations. A wing sweep evalua- tion indicated a quarter chord sweep l imi t in the order of 25 deg for acceptable maximum l i f t capability and good stall characteristics without the cost and complcxity of lead- ing edge high lift devices. This degree of sweep appeared to be near opt imum for the 0.80 Mach number cruise re- quirement and satisfactory for the maximum range condi- tions. T h e critical performance requirement for the airplane is the long range cruise guarantee which'was established as a m i n i m u m 0.75 hlach number at40,OOO ft for 3200 nautical miles. This theoretical range is equivalent to New York to Los Angeles against a 90-knot headwind using the National

Business Aircraft Association ( N B A A ) jet range format. Para- met r ic analysis of this conditiori indicates dn optimum wing loading in the order of 55-60 psf and an optimurn sweep somewhat less than 20 deg. Adoderate variations in sweep and aspect ratio, however, arc not critical. Coricurrentwith wing :izing, preliminary evaluation of the engine location favored a fuselage mounting iihich precluded a low stabi- lizer location. Accordingly, a nose-down pitching break a t the stall due to a favorable stabilizer contribution was unlikely, and the aspect ratio of the wing was co~llpromised t o minimize wing pitch-up tendency.

T h e engine location is obviously not an "inherited" char- acteristic, its selection entailing extensive analysis and de- sign iteration. Aerodynamic analyses and wind tunnel in- vestigations indicated the wing and fuselage mountings offered approximately equal performance potential, although the sensitivity to interference effects appears gea te rewi th the wing installatiorl. Design studies revealed considerable problems in achieving a satisfactory compromise of the aero- dynamic, structural, and ground clearance requirements of the wing-mounted configuration. These problems were com- pounded by the large engines and a working environment that includes operation onunswept runways. T h e advantages of a cleaner wing (particularly with respect to flap geom- etry), less severe single-engine control requirement, and an improved cabin environment with respect to noise and visibility, which are inherent in an af t engine installation, were decisive when contrasted with the difficulties encoun- tered with the wing installation.

T h e aft engine location restricts placement of the hori- zontal ta i l but the decision to mount the stabilizer a t the vertical fin tip derives from many considerations. From an aerodynamic viewpoint, the T-tai l arrangement maximizes both the horizontal and vertical ta i l stability contributiors and minimizes empennage and trim drag penalties. Wind tunnel tests indicated undesirable characteristics with a mid tai l location, including nonlinear pitching and yawing mo- m e n t variations and a pitch-up a t the stall due toear ly wing wake influence. The effect of tail height a t very high an- gles of attack was not a decisive factor.

CONTROL SYSTEM

T h e selection of the Gulfstream I1 control system also involved considerable analysis and design iteration. Cus- tomerpreference favored retention of the Gulfstream I man- -

Fig. 2 - Gulfstream I1 flight photo

ual control systcm on the basis of reliability and ease of maintenance. T h e use of servo tabs or other forms of re- versible tabs, however, m e t strong opposition on twocounts:

1. Probability of t ab stall a t high angles of attacltwould require incorporation of a backup system for deep stall re- covery.

2. T h e increased possibility of encountering flutter prob- lems during flight test with prohibitive cost and t ime pen- alties.

T o retain t h e reliability of mechanical ly connecting the control column and control surface by means of cables and push rods, pilot assist was provided by means of a dual hy- draulic boost system employing tandem actuators. The principal drawback t o a fully powered system was the added complexity of a n artificial longitudinal feel system.

Primary control about a l l three axcs is provided by sealed, internally balanced, trailing edge surfaces. T h e longitu- dinal system utilizes full span elevators with pilot feel de- riving directly from the elevator hinge moments reduced by a factor of four through the boost system. During much of the configuration development, longitudinal trim was provided by a t r immable stabilizer. Control system analysis indicated appreciable stabilizer movement was encountered only at low speeds and that the use of a t r im surface of greater pitch power than the primary control surface intro- duces considerable complication and safety considerations. T h e rotation and landing holdoff benefits of the stabilizer a re retained in the design by mechanically gearing i t to the f lap drive while trim capability is provided by a n irreversible trailing edge t a b on t h e elevator, 1 /32 the size of the stabi- l izer . T h e tab is cab le connected to a trim wheel in the cockpit but is normally operated by an electr ic trim switch on the control wheel. Mach trim compensation of the tran- sonic tuck tendency is provided by an independent autopilot unit driving the tab through the trim motor. -

For la teral control, the relatively small ailerons are aug- mented by multipurpose, fully-powered spoilers. All three spoiler segments are automatically actuated t o full deflec- tion on the ground by a squat switch on the m a i n gear(with system armed and throttles retarded) to maximize decelera- t ion capability. T h e flight spoiler, which utilizes the two oitboard segments, doubles as a speed brake and lateral con- trol device. As a la teral control, the spoiler was geared to

Sea Level Standard Day Zero Wind

Take-Off (Flaps 20'1 /

Field 3 Landing Lengfh ,- f t Flaps 40'1

2

Gross Weight - 1000 Ib

Fig. 3 - Take-off and landing performance

I. 1'. WAALAND AND B. J. CURTIS

ini t ia te deflection at 20;'~ oi aileron trnvcl to avoid spoiler clcflections wit11 trirn or autopilot inputs. Pilot feel is again directly from the surface hinge ~nor i ie~ l t s and a boost ratio of tllree was used to pro\~ide lateral coiltrol forces similar t o t17c Gulfstream I . Laleral trim is obtained with a cable driven tall on the left a i l c ro~l .

The directional control system is fully powered in nonnal operation I ~ u t the mechanical connection to rlie rudder is similar to the aileron and elevator colltruls (infinite boost ratio). The directional control actuator is torque limited to prevent excessive rudder deflectiorls s t t!le higher air- speeds. I<u~lcler forces are obtained from a two slope spring with a nlnximum force of 100 Ib and directional t r imis ac - complished by shifting the feel spring. In manual reversion, due to loss of both hydraulic systems or both engines, the actuator pressure is eliminated and pedal forces r e f l e c ~ both the surface hinge moment and feel spring force. Dutch roll stability augmentation is provided for p:issenger comfort and miniinum pilot effort by means of a yaw damper connected in series to eliminate feedback to the rudder pedals. A damper authority of + 3 deg of rudder travel provides esscn- tially dead beat damping throughout the flight envelope.

CONFIGURATION DEVELOPMENT

The aerodynamic development program consisted of an- alytical and experimental studies. T h e wind tunnel program was largely conducted on two models: a 1/15-scale , sting- mounted high speed model , used for seven series of tests in the Cornell Aeronatutical Laboratory b-ft transonic tunnel; and, a l / lO-scale low speed model mounted either conven- tionally on centerline struts (complete model) or on a re- flection plane (half model). The low speed tests covered seven series in the Grumman facility, reflection plane tests t o 1 / 3 full scale Reynolds number a t the Cornell Aeronau- t ical Laboratory, and a c o n ~ p l e t e model series to 1 / 2 full scale Reynolds number a t the NASA Ames 12-ft pressure tun- nel. In addition to the standard force and moment meas- urements, these tests provided inlet performance, mass flow influence, coiltrol surface hinge moments , stabilizer loads and moments, pressure surveys, and flow visualization for speeds t o 1.0 Mach nurnbcr and angles of attack to 50 deg. T h e reflection plane tests were somewhat limited in use- fulness by cl~fficulties in "mirroring" the srabilizer without fouling but permitted rnensurement of forces with secondary airflow in the nacelles to simulate mass flow effects, and high angle of attack chnr:~ctcristics wirliout support inter- ference.

The advances in \\rind tunnel testing during the last dcc- ade were found particularly useful in developing the con- figuration and eliminating the usual trouble spots. The abil- i ty to obtain on-line measurements of overall forces and monlents, individual strain gage reaclings, and a large num- ber of local pressures, simultaneously \\,it11 direct visual and/or film recordiilg of flow r c p r c s e ~ l t : ~ t i ~ ~ n s , greatly re- ducecl thc l ime and "art" rccluired to ol~tai l l practical, ef- ficient aerod y n a n ~ i c sliapes. The increased flexibility of

the tuiinel as a development tool permittccl improved co- operation with lofting engineers and on-tlle-spot resolution of internal volume and external shapc conflicts such as wheel fairings and wing-fuselage intersection. Further, the use of numerical corltrol lofting and machining techniques permitted compatibility and agreement between the wind tunnel models and the airplarle to an unprecedented degree.

WING DESIGN - Primary development effort was devoted to establishing wing geometry. During preliminary con- figuratior~ sizing i t became evident that fuel capacity would be a critical factor in meet ing the range guarantee. Also, t h e conflicting maxirnuln cruise speed requirement was raised to O.E3 Mach number to insure scheduled airline per- formance a t intermediate ranges. Research effort was in- itiated t o develop the airfoil geometry recluired for maxi- m u m sweep benefit from the selected planform, to increase

. t h e degree of freedom in selecting thickness and fuel vol- u m e . T h e interference problem a t the wing-body juncture wa$ treated by modification of the airfoil shape and thick- ness over the inner third of the wing span. An inverse c a m - ber line modification was calculated using Kuchemann's theory. T h e restoration of inboard loading and an increase in upper surface velocities to achieve a favorable isobar dis- tribution was accomplished by a combination of incidence, thickness, and thickness distribution. T h e treated wing pro- vides a relatively large thickness a t the root (13%) and the improvement in fuel volume permitted some latitude in selecting outer panel thickness variation.

T h e basic airfoils for t h e m a i n area of the wing are similar to those of the A 6 -A at tack aircraft and utilize NASA six series thiclaess distributions combined with an in-house mean line. T h e forward c a m b m of these airfoils is magnified by in- creasing the leading edge radius tangent t o the basic upper surface contour, the leading edge radius factor varying from 1.4 a t the 1 / 3 span location to 4.7 a t the tip. In develop- ing the \sing contours, information derived from the Eng- lish literature was of invaluable assistance in attaining an efficient configuration in a reasonable t ime.

High speed wind tunnel tests confirmed the favorable sweep benefits of the treated wing and also demonstrated t h e insensitivity in upper surface flow development to mod- ifications of lower surface leading edge geometry. T h e l im- iting condition in increasing leading edge radius for ben- ef icial low speed characteristics was the deterioration in negative lift buffet boundary due to cusping behind the lcad- ing edge. The performance of the configuration was better than expccted due in part to a favorable compression tend- ency ahead of the shock, illustrated by the pressure distri- bution of Fig. 4 . A significailt influcncc of the nacelle 6n wing loading and shock wave developmcnt was also notcd. T h e compression influcncc of the nacclle has a beneficial drag efrect but adversely affccts sllock-induced separation on the outboard portion of the wing. Also ttie overlap of t h e naccl lc and wing creates a region of accelerated flotv that intensifies wit11 iricrcnsing blacll nunlber, increasing the trailsonic tuck tendency. The wing pressures in tllc viclnity of the nacel le illustrate this acceleration (Fig. 5). It is in-

GULFSTIGAM I1 AERODIWAMlC DESIGN 1153

tercsting that no separation was observed in t l~ i s region, the flow changing tlirectly from normal rccovery to a trailing cdgc shock< contli tion.

On the basis of the wind tunnel results, the maxirrium speed prediction for the aircraft was raisecl to 0.15 Mach number. A bulfet boundary coinmensurate \,:it11 this speed capability was attained by incorporating a row of co-rotat- ing vortex generators on the outer \iring panel. in designing the vortex generators, a papcr by Pearcey (1)" permitted es- ta'blishmcnt of an optimum configuration on the initial a t - tenlpt. The inl lue~ice of the vortex generators on the buffet onset boundary is sllown in Figs. G and 7 . Botl~ the vortex generators and a bsundary layer fence contoured to the up- per surface streariilines manifest no measurable cruise drag penalty

In developing the wing contours, cognizance was taken of the aircraft's low speed requirements by tailoring th; leading edge raclius to preclude leading edge separation and by considerntion of local loading. The maximllm lift c a - pability of th: wing was not critical due to the low wing loading (large wing area recluired for long range cruise and fuel volume recluirements), the high thrust loading, and ad- vances in braking capability. The high lift configuration, which consists of a one piece, single-slotted Fowler flap of

'Numbers in parentheses designate References a t end of paper.

% Chord

Fig. 4 - Wing pressures (outboard)

I

20 40 60 80 100 "A Chord

Fig. 5 - Wing pressures (inboaril)

3070 chord, performed accorclirig to expectations and primary low speed development effort was clcvoted to stall characrer- istics. Stall iniriotion on the basic wing occurrcd a t mid- span but spread rapidly t o the t ip , particularly at large flap deflections. T h e addition of an upper surface fence a t about 1/3-span provided a strong pitcb down a t the s tal l , without sacrificing maximum l i f t , and also afforded an aclequatc rnargin between initial and tip stall . Inlet pressure meas- urements, however, indicated the possibility of walte flow ingestion and significant compressor face distortions during power-on s tal l for this configuration. A compromise be- tween stall and inlet requiremeilts was achieved wit11 a mid- span fence location. For this configuration, tunnel results indicate stall initiation a t the fence with a gradual spread inboard (Figs. 8A-8C), trailing edge separation in the aileron area 6-8 deg after iliitial s ta l l , and full wing separatioil 2-3 deg further. This configuration provides a 2-3 deg margin between ini t ia l stall and onset of inlet distortion at full en- gine mass flow, and a 5 deg margin with idle power.

' Mach No. - M

Fig. 6 - Effect of voftex generators on buffet boundary

50000 Ibs GW

Mach No. - M

Fig. 7 - Effect of vortex generators on buffet boundary

Fig. 8A - \Ving stall progression - a = 15 Jeg

I. T. WAALAND AND E. J. CUR'I'IS

NACELLE DlJSIGN - The sensitivity of nacel le drag pen- al ty to locatioli and orientatiori was established cnrly in the clevelopmcnt program, but difficulty in minimizing nacelle intcrference al.ose. from conflicts with nonaerodynamic fac- tors. An overlap ol ~ l a c e l l e ancl wing, for example , was un- avoidable within the geometrical and weight and balance restraints of tlie configuration, yielding a strons sensitivity t o fore and aft location. Early coniiguration studies called for an off-the-shelf engine pod with integral cascade-type t h r u s ~ reversers, but modifications reducing the inlet length and reshaping the i ~ i l c t cowl for t l ~ c Gulfstream I1 applica- tion were dictated by drag considerations. When the cas- cade-type reversers Ifere el iminated, due t o cruise drag and net thrust penalties, in favor of target type reversers, full design freedom for the nacelle shape was realized.

Reduction of the wing-nacelle interference to a p ~ a c t i c a l minimum involved a most aft nacel le location within the limits of balance arid structural considerations, a minimum inlet length consistent with crosswind requirements, and op- t imizat ion of nacel le incidence. T h e influence of nacel le pi tch orientation derives from both inlet alignment with the loca l flow conditions and modification of the flow channel shape between the nacelle and wing. T h e opt imum drag orientation also provides full pressure recovery a t the engine compressor face a t a l l flight conditions. In the developed configuration, the sensitivity to further aft movement has been reduced to a n iriignificant l eve l and vertical location is not a drag consideration within the limits investigated.

An additional interference regiori is the channel formed by the nacelle-pylon-fuselage intersections. T h e engine

mounting geometry and engine-pylon interface precluded symmetry in the pylon-nacelle relationship, aggravating the

j~ylon lower surface channel geometry. Significant relief, ' particularly a t high &Inch numbers, was achieved by lolling the engine and pot1 relative to the fuselage, opening the

'! channel , and eliminating acute corners as sho\vn in Fig. 9 . In optimizing the nacelle geometry, careful attention to small geometry details was iound necessary to minimize iriterference or compressibility effects and avoid separation. T h e relationship between cowl curvature and pylon leading edge shape and location, or the establishment of the pyloll trailing edge geometry, are examples of fruitful areas of refinement. T h e resulting nacelle drag penalty at cruise conditions is shown in Fig. 10 and includes an adverse na- ce l le influence on span loading. Note that a beneficial in- terference effect is in~roduced a t the higher Mach numbers.

T h e disadvantages associated with an overwing nacelle location have been stressed, but there are also benefits to be derived. Because of its relationship to a rather large turn- ing vane , the inlet operates a t low angles of attack with no measurable pressure. recovery penalty throughout the n o m ~ a l flight envelope. Also, the problem of wing wake ingestion a t s ta l l , buffet onset or with wing spoilers extended is avoided and the nacel le wake irlfluence on the ta i l a t very high an- gles of a t tack is considerably minimized.

Detailed wind tunnel studies were made of the static and dynamic compressor face pressure characteristics over a wide range of angle of a t tack, sideslip, and speed. Negli- gible distortions were noted up to sideslip angles of 15 deg and.the crosswind investigarion, which indicates the l ee en- gine is c r i t i ca l , justifies clearance for 40-knot crosswinds. 1

Fig. 9 - Nacelle roll effect

Fig. 8B - Wing stall progression - a = 1 8 d e g Cruise Cond Increment from Model *

+.002 - -,-y--"- , 3 ,- . p

4

~ -.- .*. ', -y. -e,:; ..

9, 0 -

,- A

A CD 2.--a* *

& a. . "d * a- -.002 -

1

Build-Up Test - C = 0.2 I

I \\ .-.om , I I I 0.70 0 75 0.80 0.85 0.90

...-,- ---* A d 2 kc.-, .L>+L-- Mach No. -M

Fig. 8C - Wing stall progression - a = 23 deg Fig. 1 0 - Nacelle drag increment

C;IJLFSTP&Ahl I1 AERODYNAMIC DESIGN 3.155

Rclativcly largf static distortions coupled wit11 a high level of tlyrlarnic fluc~uations were measured, however, a t angles of a t tack Ibcyond wing stall , particularljr a t n ~ n s i n i u m en- gine mass flow. lleak to peak total pressure fluctuations in the order of 6':'~ were noted for take-off power conditions as compared t o a rnaximu~n of 3% at idle poiier. Smoke stud- ies were usclul in confirming the \\ling separation-inlet in- gcstion r e l a t i o l ~ s h i ~ ~ and the timing and intensity of the dis- tortions were motlified by changes in wing stalling behavior and maximizing nacelle vertical location. T h c static dis- tortions at a n y co!idition are well within the engine manu- facturers specification but the adverse influence of the su- perimposed dynamics on engine stall margin could not be predicted. Rolls-Royce interest and assistance on the prob- lem included tests in both its sea level bed and altitude chamber to ascertain the effect of combining dynamics with ltnown static distortion levels. The results indicate a reduc- tion in crlgine stall margin but siall frce operation of the e n - gine is anticipated on the basis of these tests even during stall penetrations to high angles of attack.

DEEP STALL - A detailed analysis of the contributing fac- tors and physics of the deep stall problem is unnecessary and redundant considering the attention it has received in the literature (Rcfs. 2 and 3 a le excz!lent and thorough exam- ples). The worldwide deep stall a larm was sounded during the early dcvclopment studies of the Gulfstream 11. I-Iigh angle of a t tack wind tunnel tests and analog computer mo- tion studies, which were accordingly incorporated into the dcvelopmel~t program, provided insight into the basicmech- anisms and an appreciation of the influence of configuration

variables. The cariclid nature of industry-wide communica- tion on the deep stall also contributed significantly to its understanding and the ability to evaluate the adequecy and safety of the Gulfstrcam I1 design.

T h e high angle of a t tack investigations on thcGulfstream I1 indicate:

1. Stable t r im conditions exist up to 45 deg angle of a t - tack.

2. T h e elevator deflection required to trim t o the pri- mary stall a t most forward center o i gravity is sufficient to trim a deep stall a t t h e aft center of gravity.

3 . Recovery from deep stall is immediate upon forward stick motion, that is, more than adequate nose-down e le - vator control is avai lable (Figs. 11 and 12).

In addition t o adequate control effectiveness at a l l an- gles of a t t ack , the control column is mechanical ly con- nected to the elevator such that no auxiliary devices are re- quired t o insure airplane response. The acceptability of the Gulfstream I1 high angle of a t tack characteristics and the absence of a deep stall influence on configuration sizing and arrangement is attributed to the mitigating influence of the nacelle-wing overlap on nacelle contribution. Configura- tion buildup studiesreveal the adverse nacel le influence on tai l pitching moment contribution above 30deg angle of a t - tack (Fig. 13) is not unduly severe and no appreciable effect on elevator or stabilizer effectiveness is evidenced (Fig. 14). Recoverability of the aircraft from any stall condition is a prerequisite to configuration acceptability but a strong re- sistance to deep stall entry and the attainment of good pri- mary stall behavior a r e equally important. A designrequire-

Fig. 11- - High angle of a t tack moment characteristics Fig. 1 3 - Nacel le effect on tail contribution

Angle of Attack - Deg

10 20 30 40 50 0

is = O

A C, -.2

CG @ 38% Clean Confiq

-.6-.

0 10 20 30 40 50

Angle of Attack -Deg

Fig. 1 2 - High angle of attack trim elevator deflections

Full Nose Down ISe+ 13 ) -.3

I I I I I

Cm8, Per Deg Nacelles Off

-.01-

0 10 20 30 40 50 -.4 Angle of Attack - Deg Nacelles Off (WBVH-WB)

Angle of Attack - Deg.

Fig. 14 - Nacelle effect on control power

I. T. WARLAND AND E. J'. CUIXTIS

melit on Grummnn aircraft is a nosc-do\\rn moment break a t tlie stall ant1 t11c configuration ilcvelopment eiforts de- voted to this cnd were describccl c:~rlicr. Asnolecl, the low spcccl wind tunnel results indicntc ail increase in stability at the stall and n margin of 6-S:!cg angle of attaclc i~e tween initial stall and lip scparatiorl. l ' h e a m l o g studies using these data indicate deep stall entries for the Gulfstream I1 will require near-masirnt~m attailiablc control deflections or elevator deflections \\re11 in excess of those for primary stall coupled \\ritil tligli approach rates or strong gusts.

During the coursc of tlie high angle of a t tack studies, some observations allti i tems of ~ c r i c r a l interest were rioted. A large variation in Reynolds number ( 1 t o 7 million) pro- vided anticipatcci eifecls on lift and pitching moment var- iations in the vicinity of primary stall hut virtually noeffect a t angles of attack above 2Sc!eg. hlach number variations from 0.13 t o O . C O also had a proriou~lced influence on pri- mary stall angle of attack without sigilificantly altering the secondary stall behavior. Reflection plane tests to high an- gles of attack intlicarcd slightly less tail contribution to sta- bility below wing stall , attributed to difficulties in obtain- ing proper reflection at an appreciable distance from the support point without fouling, but equivalent to better tail contribution at the higher angles of a t tack. It is not known whether this is due to the absence of support interference or alteration of fuselage vortex influence. The deterioration in elevator or stabilizer effectiveness with increasing angle of attack was also somewliat less for the reflection plane rilodel suggesting uriaccounted, adverse suppxt effects ]nay remain in the complete model results. A complete model evaluation of nose lcngth sho\\rcd no change in ta i l contribu- tion between the standard fuselage ancl cutting back to a hemispherical nose ahead of the wing body juncture.

Recovery studies on the analog computer indicated re- covery can be accomplished under a l l conditions (and to the limits of a large tolerance band on the aerodynamic data) by returning the control column to its original trim position. Minimum altitude loss during recovery ( in the order of 1000 ft from a fully stabilized deep stall trim condition) is achieved by restraining nose ovcr pitch rate to a reasonable level. The addition of power is beneficial as it adds a large force vector normal to the flight path, assisting the pullout.

HANDLING QU2-\I,ITIES - Excellence in handling quali- ties is a Grumman reputation that is fully maintained by the Gulfstream 11 design. Handling qualities specifications are currently the subject of much research, and evaluation of the design du"rng its development involved examination of a nuinber of existing and proposeil criteria in addition to the usual in-house standards. The greatest influence ofhan- dling rlualitics duriilg configuration c le~e lopment arose from lateral-directional co~lsiclerations, particularly the dutch roll mocle. The principal design parameters influencing the dutch roll arc the static lateral stability (dihedral ef iect) , damping in yaw ai1~1 roll, and product of inertia ( I ). xz

Incrcascs in vertical tail area and tail arm wcrc lncorpor- atc0 during devc lo j~mci~ t to incrc,lsc yaw d a m p i i ~ g and a t - tain the dcsircd lcvcl of directional stability. T h e aft en-

g ine , T- ta i l arrangement incurs an adverse priilcipel axis inclination but cognizi~nce of this factor early in the pro- grnnl permitted significant reduction in the product of in- er t ia as the design evolved. In developing the configura- t ion, restriction of tlic geometric dihedral was attempted to offset the i~ lc rease in dihedrai effect with increasiiig lift , common t o swept wing aircraft. Limitations impcscd by the fuel system, and by flap trailing edge clearance in an ex- trc.me landing situation, restricted the geometric dihedral to a minimum of 3deg . Wing bending i n normal flight in- creases the lateral stability equivalent to an additional 1.5 deg of wing dihedral. Stability calculations with the esti- mated characteristics indicate acceptable dutch roll damp- i n s throughout the flight envelope but less than satisfactory levels a t very high altitudes and possibly objectionable work- load in approach and landing. Stability augmentation, in the form of a series yaw damper, was incorporated into the configuration to insure the desired absence of extraneous pilot effort during approach and landing as \brell as satisfactory passenger comfort in high altitude turbulence. Loss of the yaw damper a t higher altitudes is not critical.

Configuration criteria based on static longitudinal sta- bility and c o ~ ~ t r o l considerations also provide satisfactory longitudiilal handling qualities under al l flight conditions. In terms of dynamic stability, damping of the short period and phugoid oscillations satisfy the appropriate military spe- cification. New criteria suggested for longitudinal handling qualities, taking into account aircraft response and control- labi l i ty , include Birhle's Control Anticipation Parameter (CAP) (4) and Shomber and Gersten's bulls-eye approach(5). T h e Gulfstream I1 CAP ranges 20-40 deg /sec /g , indicating satisfactory precision control capability, particularly in the approach and landing conditions. On the basis of theShom- ber and Gersten approach, satisfactory precision control is predicted for cruise conditions up to approximately 30,000 ft and acceptable characteristics at the higher altitudes. For approach and landing, the characteristics are centered in the appropriate bulls-eye. Statically, the wind tunnel data evidence linear stability and control characteristics to stall or buffet onset conditions while flying quality calculations indicate speed stability levels well within FAA maximum and minimum requirements and satisfactory maneuvering gradients without recourse t o springs or bobweights.

FLIGHT RESULTS

Flight testing is ideally a confirmation process but almost always includes a sharc of the configuration development effort. T h e Gulfstream I1 flight testing t o d a t e lias beenvcry encouraging, sho\\ring close agreement with most predictions and geilcrally better performallce tlian est imated, but changes i n the configuration have been made in the flying qualities area t o improve lateral control and reduce trim changes due to power anil flaps. Revisioris were alsorequired with respect t o stall cl~aracteristics and to el iminate an ob- jcctionable buffct a t full flap deflection. T h e flap buffet

GULFS?'LEl\i\/I I1 A ~ O D Y N A h ~ l I C DESIGN

was mild at approach flaps but severe \vi t l~ lai~tling flaps. Flight tuft sluilies (Fig. 15A) corll'irmed the suspected separ- ation in tile vicinity of the flap tracks as well as fully :it- tachecl flow tlirough the slot. A turning vane configuration developed carlier on the low speed tunnel modcl was in- corporatecl, rcilucing ttie separation and associated buffet to an i n s i g ~ ~ i ricaiit level (Fig. 150).

Stability levels closely correspond to wind tunnel meas- urements and the predicted dutch roll behavior a t high al t i - tucles has been confirmed as a c c e p ~ a b l e (yaw damper off). For the approach and landing conditions, d a ~ n p i ~ ~ g appears better than predicted and is colisidered satisfactory with the. damper inoperative. Buffet boundary deterlnination indi- cates buffet orlset lift coefficients slightly tiigl-~er than esti- mated , confirming the generous buffet margirls of the de- sign.

LATEML CONTROL - The lateral control effectiveness was suspecte~l to be less than optimum prior to flight but satisfied existi~ig and proposed criteria. In flight evaluation, it was iminediately evident that higher roll rates and lower force gradients than suggested by these criteria are consid- ered necessary for satisfactory "jet" performa~lce and con- trol harmony. The principal objection to roll performance

Fig. 15A - Flap tuft study - Basic configuration - 40deg

flap

Fig. 15B - F l ~ p tuft study - Turning vanes installed - 40deg flap

rclated to sensitivity ncar neutral and a nonlinear variation with wheel deflection; lnaxir~lunl roll rate capability was satisfactory. T h e illcreased sensitivity through neutral was achieved by revision of the spoiler-aileron mixing, inclutl- ing the use o l spoilers out of neutral, while the desired force levels were attained by increasing aileron boost ratio to 5.0. T h e measured roll performance for the revised config- uration, shown in Fig. 16, is considered optimum for the ap- proach configuration and satisfactory a t a l l other flight con- ditions,

TRIM CHANGES - During preliminary design of theGulf- stream 11, engine exhaust influence on horizontal ta i l flow was estimated as relatively minor and jet flow si~llulation was not included in the wind tunnel program. The location of the engine pod introduces a sizable direct thrust moment , however, and the engine nozzle was designed to orient the thrust vector closer to the center of gravity and el iminate trim changes with power. Measurement of engine mass flow influence during the wind tunnel program indicated changes in wing loading and a small nose-up moment but the mag- nitude of the change did not warrant revision of the nozzle design. Inflight measurements, however, indicated a sig- nificant nose-up moment with power application. A limited pressure survey confirmed the wind tunnel measurements for the wing but suggests flow field changes a t the ta i l are larger than estimated. It was decided toe l imina te the trim changes by reorientation of the thrust vector, the revision being a c - complished by redesign of nacelle and tai l pipe transition components without change t o the thrust reverser assembly. With the modified configuration, stick-free applications of po\trer yield insignificant changes in attitude.

Another t r im change considered during the design was that associated with flap retraction or extension. T h e gear- ing of the stabilizer t o the flap drive, discussed in the con- trol system development, was mecharlically accomplished by means of a s imple, direct linkage. Design conditions for the gearing were the minimum drag stabilizer position (zero incidence) with the flaps retracted and -4deg incidence (ap- propriate with desired take-off performance) a t the take-off flap setting (20 deg). This gearing yielded -5.5 deg of in- cidence with the 40 deg landing flap and permitted trim tab sizing on the basis of c lean requirements. Flap trim change estimates for this gearing were not completely satisfactory

Wheel Force

v

Wheel Deflection - Deg

Fig. 16 - Roll performance

in that push forces were indicated for flap extension bctwecn clean and approach (zero to 20cle.g). T h e forces were con-

$ sidered a c c c p ~ a l ~ l e , however, iri that they satisfied the FAA rcgulatioi~s and in view of the t ime associated witti the flap extension (15 sec). Iistirnated changes for extension from approach to landing flaps were small and in the pull direc- tion.

Flight testing confirmed the order of magnitude and d i - rection of the flap trim changes but alleviation of the trim c1i;tnge occurring witli flap retractioll after take.-off was rec- ommended. The flight results also showed the desired ro- tation and take-off performance could be achieved with less incidence at the take-off flap setting. As the undesirable yortion of the trim change occurred between zero and 10 deg of flap and was vcry satisfactory over the rest of the range, the dcsired effcct was accomplisliecl by introducing a-dwell range in the flap-stabilizer gearing relationship. Thedwell . unit, designed by Aiiiesearch, idles the stabilizer during 2870 of the drive travel, without hysteresis, as shown in tile gear- ing comparison of Fig. 17. The revised gearing provides small flap trim changes under al l conditions and sufficient stabilizer incidence at take-off and landing to satisfy t r im- mabil i ty and take-off performance requirements.

STALL, CHARACTERISTICS - T h e stall characteris- tics of the Gulfstream I1 a re best described as docile: a very mi ld , easily controllable lateral twitch or \vallow si- mulraneous with buffet onset, the buffet intensiry increasing

I. T. WAALAND AND E. J. CUIITIS

with increasing penetration. No g break is experienced and both la teral and longi~udinal control remain exccl lcnt . In- ilight tuft studies and the longitudinal control charncteris- \

i tics evidence close correlation with tunnel results (Fig. 18) but the increased stability evidenced a t stall initiation is riot discerned by the pilot. T h e absence of the d e s l ~ e d g break is associated with the gradual spread of the stall on the wing and the corresponding lack of do\vnwash change at the ta i l . The stall characteristics are satisfactoryin h e m - selves but d o not preclude stall penetrations to the point of secondary stall pitcliup. Inboard leading edge trippcrstrips, to promote a more widespread stall break, were investigated briefly but caused only pre-stall buffet or no discernible change. Removal o f the wi'ng fence was evaluated but e l im- ination of t h e stall initiation anchor point yielded an objec- tionable roll off. Rather than pursue a lengthy flight test rescarch effort, and in view of the excellent primary stall behavior, it was decided to mechanically l imi t the estent of s ta l l penetration.

PERFORMANCE - On the basis of t h e testing accomplished t o d a t e , the performance of the airplane is better than esti- mated and the equal of our hopes. T h e estimated specific range has been exccedcd by approximately 770 throughout the speed, weight, arid altitude range of the aircraft. A repre- sentative specific range - Machnurnber curve, presented as Fig. 19, indicates the guaranteed New York-Los Angeles c a - pability ( 9 0 knot headwinds, NBAA je t range format) can be achieved a t 0. SO Mach number. T h e rrlaxirnurn speed capability is well above the guaranteed 0.83 Mach number a t 30,000 f t as well as t h e la ter 0.85 Mach number predic-

40" Landing " I

tion. T h e maxiniuin cruise Mach number a t 30,000 ft De9 20" T.O. 8 Approach is in excess of 0.86 and a heavyweight 0.85 Mach number

0 capability is indicated for altitudes in excess of 40,000 ft.

Original Gearing T h e expected single and twin engine c l imb performance has been demonstrated repeatedly. Twin engine climb

i,, Revised Deg l-L5?? -2 a t in near 15 minutes maximum from weight brake to release 40,000 with ft has a recorded been accomplished rate of

0 0 20 40 60 80 100

c l imb a t 40,000 ft: in excess of 1500 fpm. Preliminary take- off and landing checks indicate the estimated airport per-

Flap Drive Travel, % formance will also be achieved.

Fig. 17 - Flap-stabilizer gearing

Clean, Idle Power CG @ 28%

Wheel Pull, 40 LG

/

8 e 1

De9 -10 Buffet Onset - 5

8 12 16 20

Angle of Attack - Deg

Fig. 18 - Control characteristics i n stalls

.201 Flt Test Data at

I Cruise ~"a ran tee Specific.l2 at W/ 8 = 243000 Range

Mach No. - M

Fig. 19 - Specific range

1. M. H. Pearcey ancl C. M. Stuart, "Methods of Bound- ary-Layer Control for Postponing and Allcviating Buffeting ancl Othcr Effects of Shock-Induced Separation." S.M .F. Fund Paper No FF-22 Prescntcd a t IAS National Sunlmer Meeting, Los Angeles, June 1959.

2. R. S. Shevell ancl R. D. Schaufele, "Aerodynamic De- sign Features of the DC-9." Paper No. 65-738 presented a t AIAA/RUeS /JSASS Aircraft Design ancl Technology Meet- ing, Los Angeles, November 1965.

3. E. J. Ray and R. T . Taylor, "Effect of Configuration Variables on the Subsoilic Longitudinal Stability Character- istics of a High-Tail Transport Configuration." NASATech- nical Meinorandum X-1165, October 1965.

4. W. Bihrle, Jr. " A Handling Quality Theory for Pre- cise Flight Path Control" AAFDL-TR- 65-198. June 1966.

5. H. A. Shomber and W, M. Gersten, "Longitudinal Handling Qualities Cri ter ia and Evaluation. " Paper No. 65-780 presented a t AIAA/RAeS/JSASS Aircraft Design and Technology Meeting, Los ~ n g e l e s , November 1965.

Novembe I- 1 970

G . A . C . CAD ---.

NASA/G I I bIST 1sIE IGHT AIdD LALANCE SUMP.1.4RY ------ -,-. --.---

Denlo - Prod, - @ T y p i c a l G I I PI-oduct ion Oper, Weight 35600 1 bs.

e Remove: Non-essent i a l F u r - n i sh i ng i n Demo -804 0

e Renlove: G I 1 Wing and c o n t a i n e d c t l s , e tc , -7300 -7300

o Acld: Whitco!i;b scw 770 sq, f t . r e f . area and c o n t a i n e d

c t l s o e tc . (Note: Both wings as.sLrine prod. design,

denlo l acks L.E, s l a t s ) +7550 4-7970

: Ref ined .Ti;ld. fus . 1 i nes , radonie and "coke" + 558 4-338

e /.love a f t s e c t i o n ? a f t o i s ta , 456, a f t 54", i n c l u d i n g

engines and t a i l s and

- Add 54" s t r u c t u r a l 'Is1 ug" - + 500

-Bending, t o r s i o n and shear m a t e r i a l f o r added t a i l

l eng th , + 126

-Con t ro ls , w i res , d u c t i n g i nc rease l e n g t h + 20

Add - A i r Condit, and p r e s s u r i z a t i o n r e v i s i o n s f o r

inc reased cab in l e n g t h and V L 0

-Soundproof ing, t r i m , f l o o r cover , e tc . f o r l onger c a b i n ; . add (2) seats. +1 90

GI Add - F i n L.E. Sweep 4 change and o t h e r misc. m ino r

r e v i s i o n s t o A/c . + 50

e Unknowns +310 - B Opera t ing Weight; i n c l o (3 ) crew 36800

e Passengers and Baggage (12) 0

Ze ro Fuel Gross Weight 36800

Fuel - Usable - Wings Box Beam (Est. ax) ";* 22200

- A f t Lower Fuselage st) 1800

-Wing L. E. Glove ax 2000) 1700

Ranip Weight

Maxi lnu~n Fuel Vol unle Summary Wing Box Beam 22200 lbs . Wln!3 Leading

Ed e Glove 2000 A f t ?ow Fus. 1800

To ta 1 12%-Gm0