Figure 1 - Republic P-47D-40 Thunderbolt

During World War II, the United Ending Harry Tope of Mt. Pleasant, Michigan. States introduced four new fighter de- signs produced by four separate man- TEST AIRCRAFT ufacturers: the Republic P-47 Thunder- bolt, the F4U/FG-1D Corsair, the Republic P-47D-40 Thunderbolt the F6F Hellcat, and the The aircraft is shown in Figure 1. Se- North American P-51 Mustang. rial number 45-49181. Powered by Pratt These fighter designs introduced in & Whitney R-2800-59 engine with Cur- the early 1940s have become legen- Argument tiss Electric four-bladed, constant dary. Many Army Air Corps, Navy and speed propeller. The aircraft and induc- Marine pilots became aces at their con- tion system had been modified by re- trols. Much has been written, and even By John M. Ellis III moval of the exhaust-driven tur- more said, concerning the exploits of Kal-Aero, Inc. bocharger with engine exhaust being these famous aircraft. Today, more than 5605 Portage Rd. routed through the former wastegate. 45 years since these aircraft entered Kalamazoo Municipal Airport The 100 gallon auxiliary fuel tank had combat squadrons, the facts and myths Kalamazoo, Ml 49002 been disconnected and only the main of their capabilities are blurred together fuel tank, 270 gallons, was usable. The as memories fade and the few remain- and test aircraft was equipped with one ing examples of these former warriors stores pylon mounted under each wing. are preserved by museums and avia- Christopher A. Wheal It is one of five P-47s believed to be in tion enthusiasts. Flight time on these Veda, Inc. flyable condition in the . aircraft is usually limited to public air 300 Exploration show exhibitions and, in deference to Lexington Park, MD 20653 Goodyear FG-1D Corsair the aging airframes and scarcity of en- The aircraft is shown in Figure 2. gines and parts (not to mention their Bureau number 92509. Powered by cost), they are normally operated well Photos Courtesy Kalamazoo Aviation Pratt & Whitney R-2800-8 engine with within their original design limits. History Museum and Harry Tope Hamilton Standard Hydromatic three- "Ending the Argument" documents bladed, constant speed propeller. With the results of a limited flight test pro- tually raised more questions than it the exception of modern communica- gram which was designed to compare, answered. tion and navigation equipment, the air- as far as possible on equal footing, the The P-47D Thunderbolt, FG-1D Cor- craft is unmodified from its original man- performance and flying qualities of sair and the F6F-5 Hellcat used in the ufactured condition. The Corsair also these four historic World War II fighter program were borrowed from the had one stores pylon under each wing. aircraft. Unfortunately, the title we Kalamazoo Aviation History Museum, It is one of approximately 25 Corsairs chose for the paper turned out to be the "", of Kalamazoo, Michigan. of different models which exist in flyable rather ambitious as the test program ac- The P-51D Mustang was on loan from condition in the United States. SPORT AVIATION 47 Grumman F6F-5 Hellcat Rolls Royce V-1650-9 Merlin engine at approximately 2,000 BHP. The varied The aircraft is shown in Figure 3. with Hamilton Standard Hydromatic dash numbers refer to minor differences Bureau number 79683. Powered by four-bladed, constant speed propel- such as accessory position and car- Pratt & Whitney R-2800-10 engine with ler. Major modifications include the re- buretor type. The Merlin engine in the moval of a 65 gallon fuel tank located Hamilton Standard Hydromatic three- P-51 was rated 1490 BHP. bladed, constant speed propeller. With behind the cockpit and installation of a the exception of modem communica- passenger jumpseat. Modern com- MISSION BACKGROUND tion and navigation equipment, the air- munication and navigation equipment Although all designed as fighters, craft is unmodified from its original man- had been fitted as with the other aircraft there were in fact considerable differ- ufactured condition. It is one of five in the test program. Approximately 125 ences in the ways the four aircraft were Hellcats in flyable condition known to flyable P-51 aircraft are believed to employed in combat. In addition, follow- exist. exist. ing a trend which has persisted to this day, they were all sooner or later used North American P-51D Mustang ENGINE PERFORMANCE in the air-to-ground role, though to The aircraft is shown in Figure 4. Se- The R-1800 engines in the test air- greatly different extents. rial number 45-11586. Powered by craft were virtually identical and all rated The two Army Air Corps fighters, the

Figure 5 - Engine Cooling Switches 48 JUNE 1990 Figure 6 - FG-1D Wing Lock

P-47 and P-51, are best known as long continuous settings. nose. In addition, sitting above or be- range bomber escorts. While their final were limited to "low blower" ranges hind a large straight wing creates a combat scenarios would have been which restricted the investigation to al- further blind area in a direction which practically indistinguishable from those titudes between the surface and 10,000 could well be critical in combat with the of the F6F and FG-1D (apart from the ft. MSL. Unfortunately, this limitation opposition coming up to meet you from opposition being Me. 109s and FW precluded an evaluation of the aircraft below. Rather than describe every fault 190s rather than Kates and Zeroes), an at typical "bomber escort" altitudes. All in detail, we shall pick out some of the important distinction lies in the fact that of the aircraft had permissible CG best and worst features. by the time battle was joined, the Thun- ranges of five and one half to eight in- P-47 - The massive size of the P-47 derbolt and Mustang pilots had already ches. The P-47 was tested at a CG is misleading. In fact, it had the most been in the air for several hours. Thus range very close to the forward limit, cramped and uncomfortable cockpit of the levels of noise, vibration and com- the F6F at one third aft, the P-51 at all four aircraft. The rudder pedals were fort in the cockpit environment, while exactly mid CG and the FG-1D at three adjustable for reach and the seat for never unimportant, assumed particular quarters aft. height (as were all the aircraft), but the significance in ensuring that the pilot ar- maximum available seat back to pedal rived comparatively fresh and ready for distance was only 41 inches which cor- the fight. A tired and cramped pilot COCKPIT EVALUATION responds to a 20th percentile leg length. whose head is buzzing from a noisy en- The cockpits were all essentially typ- Longer legged pilots were condemned gine is not likely to fight as effectively ical of 1940's vintage tailwheel aircraft, to suffer without thigh support, which as one who has spent the previous which is to say they had some major was uncomfortable enough during three hours or so in a comfortable seat ergonomic shortcomings. Important these short duration test flights and in a quiet cockpit. These considerations controls were often positioned with little would certainly have been hard to en- must be put in the context of the primi- regard to whether the pilot could see, dure on a long range bomber escort tive radios and indifferent sound attenu- reach or operate them without consider- mission. The tops of the pedals could ation qualities of 1940s helmets. able effort. Each of the aircraft posses- be folded down to allow the pilot to put sed fields of view which on the ground his feet on them but even that brought were limited chiefly by the precise ex- scant relief. TEST ENVELOPE AND LIMITATION tent of the area ahead blocked by the The control switches for the oil cooler Recognizing the age of the aircraft in engine and forward fuselage. The need flaps, intercooler flaps and canopy were the test program, it was agreed that for constant S-turns to avoid hitting clustered under the canopy rail at the structural loads would not exceed 6g. things while taxiing was part of the pilot's left shoulder (Figure 5), almost Military power was permitted for takeoff charm of flying these aircraft. In all impossible to see and very hard to iden- and climb to 10,000 ft. MSL consistent cases the field of view improved consid- tify.The cooler control switches in par- with manifold pressure limitations ap- erably from the moment the tail was ticular were arranged one on top of the propriate for operating on 100LL fuel. raised during the takeoff roll; but even other, operated in the same sense, and The non-availability of higher octane in flight the forward fuselage and engine were difficult to tell apart by feel. Inad- fuel reduced the maximum allowable continued to present a serious obstruc- vertently operating the canopy switch in manifold pressure by approximately tion to forward and downward vision flight would be immediately obvious but four inches for the three aircraft pow- and must have caused considerable dif- operating the wrong cooler switch ered by Pratt & Whitney R-2800 en- ficulties in shooting at large deflection would not and, if followed by a hectic gines. The Rolls Royce Merlin 1650 in angles or in air-to-ground aiming with a period at high power, could lead to en- the P-51 D was unaffected. For all other depressed sight-line when the target gine damage before being detected. tests, power was limited to maximum would tend to disappear under the The tailwheel lock control was on the SPORT AVIATION 49 >•

Figure 3 - Grumman F6F-5 Hellcat

cockpit floor outboard of the seat at- the pilot could exert while seated and any size pilot. The seat was inclined aft tachment point and required an awk- necessitating his unstrapping or getting and the pedals were installed higher ward stretch to reach; but on the posi- help from a plane captain to manually than in the other aircraft which gave the tive side, the incorporation of an au- lock the in the spread position. feeling of sitting in a high class sports tomatic tailwheel lock which operates Doing so would delay a mission or car. An enhancing feature was the fully whenever the is lowered perhaps induce the launch of an aircraft automatic operation of the coolant door is an enhancing feature and one which with the wing hydraulically but not and oil cooler shutters which required could well have been copied in other mechanically locked. Other Corsair no attention in flight other than monitor- designs. owners indicated that the wing locking ing the appropriate temperatures. FG-1D - Anyone familiar with the mechanism on this aircraft was not un- Overall, the four aircraft were ranked cockpit of the SNJ/AT-6 family would typical which suggests that the several approximately the same for each of the feel instantly at home in the bottomless documented cases of Corsair wings three categories of pilot comfort, ease pit of the Corsair. Anything dropped and folding in flight could have been caused of operation of cockpit controls, and not secured to the pilot's person is prob- by this problem. field of view in the air and on the ground: ably gone for good, or at least until after F6F - Largest and roomiest cockpit of P-51, F6F, FG-1D and P-47. the flight, given the natural tendency of all. Adjustable seat and rudder pedals inanimate objects to wedge themselves gave a maximum seatback to pedal dis- TAXI into corners from which they cannot be tance of 47 inches, enough to accom- dislodged even by determined applica- modate a 98th percentile leg reach. The The P-47, F6F and FG-1D all had tail- tions of negative Gs. only control difficult to reach was the wheels which were locked for takeoff All cockpit controls are in the pilot's emergency landing gear extension and landing or free castering for taxiing. normal view and within easy reach ex- lever which required the pilot to unlock Directional wheel control was by diffe- cept for the wing fold lock control at the his harness to operate it. rential braking. The P-51 had a tail- rear of the left cockpit shelf (Figure 6). P-51 - The most comfortable cockpit wheel linked to the rudder pedals which It was difficult to reach and very stiff in of all. Seat and rudder pedal adjust- was steerable through 6 degrees either operation requiring more strength than ments were of a range to fit practically side of neutral. The steering mecha- nism could be disengaged by pushing the stick full forward which allowed the tailwheel to become fully caster- ing for tight turns or maneuvering in TABLE I confined spaces. All aircraft were easy to taxi but taxiing in a crosswind with TAKEOFF AND MILITARY POWER CLIMB the three fully castering tailwheels re- p. quired riding the downwind brake, in- P-47D FG-1D F6F-5 S1D creasing wear and possibly leading to Takeoff Weight 1 1 ,535 1 1 ,055 10,681 8,900 Runway Temperature 54°F 77°F 62°F 62°F brake overheating. Once again the P-51 Liftoff Speed 110mph/96kts 85kts 90kts 115mph/100Ms came out on top. Takeoff Distance 1500ft 1200ft 1400ft 1500ft Climb Speed 155/mph/135kts 135 Ms 130 Ms 175mph/152kts TAKEOFF HANDLING Takeoff for all four aircraft was made 50 JUNE 1990 P-47

FG-1D 8 - F6F-5 P-51D

FIGURE 7 CLIMB PERFORMANCE 2 -

30 60 90 120 150 180 210 240 270 300 330 TIME FROM BRAKE RELEASE - SECONDS

with the flaps up, setting field baromet- sentative because rough running at ference between the R-2800 powered ric manifold pressure while holding the 4,000 ft. required the power to be re- airplanes, the P-51 had a decided ad- brakes and increasing to military power duced to METO which smoothed out vantage both in acceleration and top immediately after brake release. The the engine operation but naturally re- speed. Both the P-47 and the FG-1D tail was raised to the takeoff attitude as duced the rate of climb somewhat. The were handicapped by underwing pylons soon as elevator control was available. Hellcat came out on top with a time of but it is doubtful if the results were sig- A positive rotation was made for liftoff four minutes fifteen seconds from brake nificantly affected. and the gear was retracted as soon as release to 10,000 ft. but there was less a positive rate of climb was established. than thirty seconds between it and the STALL CHARACTERISTICS Takeoff conditions are listed in Table I. FG-1D, while the Mustang would clearly Normal and accelerated stall charac- All the aircraft required some preset have benefited from the missing few in- teristics were evaluated for all the air- right rudder trim for takeoff. The P-47 ches of manifold pressure. The handl- craft in the cruise and landing configura- called for 5 degrees and the P-51 and ing of all the test aircraft in the climb tions with power at idle, cruise and FG-1D required 6 degrees while the was good and all control forces could METO at 10,000 ft. MSL. Normal stalls, F6F needed full rudder trim. For all but be trimmed out with the exception of power on or power off, were conducted the Hellcat, the preset takeoff trims pro- the rudder forces in the F6F Hellcat in straight flight using a slow decelera- vided good control of the initial swing which needed continuous right rudder tion (less than one kt./sec.). Acceler- with application of military power and for this and all other high power opera- ated stalls were conducted in a constant yet enabled the foot forces to be com- tions at fairly low speed. The forces 3g turn allowing the airspeed to decay pletely trimmed out during the climb. were estimated at 80 to 100 pounds and at less than one kt./sec. The results are The Hellcat, however, required addi- were high enough to be tiring even dur- presented in Table II. tional right rudder during the takeoff roll ing the few minutes in the climb to Aerodynamic warning of the normal and climbout when the foot forces were 10,000 ft. A prolonged climb to higher light and easy to apply. stalls in both configurations ranged from altitude would have been very tiring and scant to nonexistent. The P-47 was the In all other respects the aircraft were the probability of a pilot relaxing the best with some light airframe buffet oc- easy to fly and trim, with minimal trim right rudder force and incurring a loss curring about five kts. above the stall. changes from gear retraction and initial of climb performance would be high. acceleration. The P-51 had virtually no warning, the Corsair and Hellcat a couple of knots LEVEL ACCELERATIONS apiece. Secondary cues such as de- CLIMB PERFORMANCE The level acceleration tests were creasing effectiveness and in- AND HANDLING conducted at 10,000 ft. MSL using creasing longitudinal stick forces were The results of test day military power METO power. For all tests the OAT was noticeable in all except the FG-1D in climbs to 10,000 ft. MSL for the four 50 degrees F. The results are pre- which the stick forces tended to lighten aircraft are presented in Figure 7. The sented in Figure 8. It can be seen that, before the stall. Normal stalls for all air- data for the Mustang is not truly repre- while there was not a great deal of dif- craft in all configurations, power on or

SPORT AVIATION 51 240-

FIGURE 8 LEVEL ACCELERATION 10.000 FT MSL METO POWER

100' i i i I 40 60 80 100 120 140 160 180 TIME - SECONDS

power off, were defined by an uncom- poising motions and the ever present TURN PERFORMANCE manded nose down pitch accompanied likelihood of departure. in most cases by a wing drop, the direc- The P-51 provided virtually no warn- Sustained turn performance was tion and severity of which were strongly ing whatsoever of the accelerated stall evaluated at 10,000 ft. MSL at METO influenced by any departure from ba- other than a very slight trembling felt power. While again not reflecting the lanced flight conditions. In all normal through the stick immediately before true performance to be achieved at stalls except those in the Corsair with departure which was not obvious to an combat power settings, we believe that power on, it was possible to control the alert pilot with nothing else to concern the playing field was level enough to wing drop by coarse use of the rudder. him and would have been almost cer- allow reasonably valid comparisons be- The behavior during accelerated tain to be missed during the heat of tween aircraft. Instantaneous turn per- stalls is of more interest from the com- combat. At the stall, the aircraft invari- formance was evaluated in wind-up bat pilot's point of view and it was here ably departed with complete loss of turns from 10,000 ft. The results are that the most significant differences control, on occasion achieving 270 de- presented in Figures 9 and 10. were found. The most predictable and grees of roll before recovery could be The results are interesting in that, controllable were the Hellcat and the P- effected. The departure was accom- below about 130 kts., the sustained and 47. Both could be flown at will into the panied by a violent aileron snatch which instantaneous turn performances for pre-stall buffet, which at no time was was strong enough to take the stick out each aircraft were virtually the same, heavy enough to present problems with of the hand of an unprepared pilot. though separated by about 1g between tracking and held at maximum usable However, the two-handed pull necessi- the best and the worst. As expected, lift coefficient with ease. Any significant tated by the high maneuvering stick the Hellcat out turned the other three by sideslip became noticeable as a pro- forces would probably have reduced the a fairly conclusive margin. There was gressive wing heaviness which was the chances of that actually happening in too much scatter in the results to make pilot's cue to make a rudder correction, combat. Altitude lost in the departure a confident distinction among the rest, but in the absence of gross mishandling and recovery was about 500 ft. but it looks as though the P-47 came off there was little tendency to depart. From these results, it can easily be worst with the P-51 and FG-1D ahead The Corsair generated considerable seen that the P-47 and Hellcat were the but not by much. airframe buffet five knots prior to stall in most "user-friendly" of the four, with the Another interesting feature is that for a constant 3g turn but at the stall there Corsair next and perhaps more suited all the aircraft the corner speeds were was a pronounced g-break and a rapid to experienced pilots. The P-51's lack remarkably close to the maximum level right wing drop regardless of the direc- of stall warning and consistent depar- flight speed, implying a very rapid tion of the turn. With care, the wing drop ture constituted what these days would energy loss when turning at the struc- could be prevented with some anticipa- be classified as a Part I deficiency. In- tural limit. The Hellcat was in light air- tory left rudder but the aircraft then be- deed, as we discovered, there are plen- frame buffet at 6g at Vmax, while the came very unpredictable at the stall with tiful accounts of Mustangs spinning out P-47 exhibited light buffet at 4.8g and moderate to severe bucking and por- of combat engagements. moderate at 5.2g. The FG-1D and the 52 JUNE 1990 P-51D were buffet free up to 6g at their TABLE II maximum level flight speeds. Of course, the maximum level flight speeds would STALL CHARACTERISTICS have been higher at combat power set- tings but so would the structural limits: P-47D FG-1D F6F-5 P-51D C P1 Vwarn 107kts 90 kts 75 kts 89 kts the Corsair, for example, had an opera- CP'Vstall 97kts 85 kts 68 kts 87 kts tional limit of 7.5g. CP1 Height Loss 200 ft 250ft 200ft 250ft C P2 Vwarn 90kts 78 kts 68 kts 85 kts MANEUVERING STABILITY C P2 Vstall 87kts 76 kts 62 kts 83 kts The maneuvering stability as indi- C P2 Height Loss 50ft 100ft 50ft 100ft 1 cated by the stick force per g in wind-up LP Vwarn 91 kts 73 kts 65 kts 76 kts turns at Vmax is presented in Figure L P1 Vstall 87kts 69 kts 60 kts 74 kts LP' Height Loss 200ft 150ft 100ft 150ft 11. Probably the most conspicuous fea- L P2 Vwarn 74 kts 57 kts 58 kts 72 kts ture is the very high stick forces dis- L P2 Vstall 70 kts 55 kts 52 kts 70 kts played by the Mustang, averaging over L P2 Height Loss 50ft 50ft 50ft 50ft 20 Ibs. per g. In fact, the force gauge Accel Vwarn 1 26 kts 103 kts 100 kts 122 kts only read to 60 Ibs. so that the forces Accel Vstall 109 kts 98 kts 95 kts 122 kts for 4 and 5g were estimated. But Accel Height Loss 100ft 150ft 150ft 500ft throughout the test program, there was

1 no doubt at all that the Mustang was a NOTES : C - Cruise Configuration P - Power Off Accel - 3g Turn, Decaying two-handed airplane in which pro- L - Landing Configuration P2 - Power On Airspeed longed hard maneuvering was ex- tremely tiring. By contrast, the FG-1D with an average of only 5 Ibs. per g, and a local gradient between 4 and 5g (a great deal less) was almost too light. The P-47's 7.5 Ibs. per g was just about ideal while the Hellcat's 12.5 Ibs. per g was barely acceptable and verging on excessive.

STATIC LATERAL- DIRECTIONAL STABILITY < 2- The aircraft all exhibited positive sta- I- tic directional stability and positive di- hedral effect. Steady heading sideslips in cruise and land configurations re- SO too 150 250 AIRSPEED - WAS vealed nothing out of the ordinary beyond the fact that the rudder forces in both the Hellcat and Corsair were ex- FIGURE 9 tremely high. Full rudder sideslips gen- erally required 50-60% of available aile- SUSTAINED TURN PERFORMANCE ron deflection in cruise at 180-190 kts. 10.000 FT MSL METO POWER and 20-50% aileron in the landing con- figuration. Sideforces were highest for the P-47 which required 45 degrees of bank to the left at 190 kts. and 20 de- grees at 120 kts. Rather surprising in view of its deep fuselage and large keel STRUCTURAL LIMIT area, the Hellcat did not require particu- larly large bank angles, but this may well have resulted more from limited rudder control power and strong direc- tional stability restricting the amount of sideslip that could be generated. All the aircraft showed marked asym- metries between left and right sideslips, especially with regard to the bank ang- les which were as a rule substantially less for slips to the right than to the left. Unfortunately, steady heading sideslip

30 100 190 250 data was incomplete for the FG-1D and AIRSPEED - WAS the P-51. This precluded further investi- gation of a peculiarity which came to light during rudder-only turns. The Cor- FIGURE 10 sair's only response to left rudder in INSTANTANEOUS TURN PERFORMANCE either configuration was to drop its nose, suggesting weak or non-existent 10,000 FT MSL METO POWER dihedral effect with right sideslip. All the aircraft except the P-47 pos- SPORT AVIATION 53 ability of the aircraft to run out of a losing fight. Starting in level flight at 10,000 ft. TABLE V MSL with METO power from typical pat- DIVING ACCELERATION tern speeds, the dives were entered via (30° DIVE, 10,000' - 5,000' PULL UP - LEVEL (a 4,000') a minus 1g pushover to a 30 degree angle. Recovery was initiated at 5,000 P-47D FG-1D F6F-5 P-51D ft. The results are presented in Table V. START SPEED 110kts 100kts 100 kts 120 kts From the performance aspect, the re- MAX SPEED 350 kls 348 kts 315 kts 350 kts sults are slightly confused by the fact TIME 23 sec 32 sec 28 sec 25 sec that the starting speeds were not the same for all aircraft. The 10 and 20 kt. starting advantages that the P-47 and P-51 had over the Corsair and Hellcat TABLE VI probably influenced the results, though it is nonetheless significant that the P- LEVEL 180° HEADING CHANGE (METO) (START 220 kts (<> 10,000' MSL) 47 still outran the P-51 despite being 10 kts. down on it at the beginning. The P-47D FG-1D F6F-5 P-51D F6F apparently being four seconds 9.7 sec 8 5 sec 99 sec 10.0 sec quicker than the Corsair, despite start- ing from the same speed and ending 33 kts. slower, suggests that there may be some scatter in the results. The flying qualities encountered dur- sessed at least moderate degrees of to 25-27% for the Mustang, 28-33% for ing the dives were of more interest. In adverse aileron yaw, the worst being the Corsair, and 26-38% for the Hellcat, the Hellcat, the out-of-trim stick and rud- the Hellcat in which the yaw was quite making large changes of direction der forces built up faster than they could marked. Countering it during aggres- under g a fairly ponderous business. No be trimmed out and quickly became sive maneuvering was hard work and wonder that improving roll rate was a substantial. The P-51 experienced quite tiring in the Hellcat and the Corsair major goal for fighter designers of the moderate increases in both longitudinal because of their very high rudder period when gaining a few degrees per and directional control forces which forces. The P-51 also required continu- second could endow a substantial ad- could easily be trimmed out. In the Cor- ous rudder coordination but the forces vantage over the opposition. Roll rates sair, the longitudinal stick forces were were low enough for it not to be a prob- in the landing configuration were very light and easily manageable but the rud- lem. Aileron yaw was negligible in the similar for all the aircraft, time to roll der forces were high and required re- P-47 which made accurate tracking through 90 degrees being about 2.3-2.4 trimming. The P-47 showed moderate considerably easier than in the other seconds with an average roll rate of increases in both stick and rudder aircraft. about 38 degrees per second. forces which were easily controllable without retrimming but could also easily be trimmed out if desired. ROLL PERFORMANCE DYNAMIC STABILITY Roll performance was evaluated As expected, the dynamic modes SIMULATED MISSION TASKS using full aileron deflection in rolls to were typically well behaved. Damping AND AGILITY TESTING the right and left at 10,000 ft. MSL at ratios were high resulting in short period 200 kts. for 1g rolls through 360 de- This was, so to speak, the "proof of oscillation which were either deadbeat the pudding" part of the test program. grees and 220 kts. for rolls at 3g through (P-47 and FG-1D) or nearly so (F6F and The aircraft were subjected to a couple 180 degrees. The results are presented P-51 and a single overshoot). Dutch roll of basic agility tests (suggested by the in Tables III and IV. modes were also well damped, the Cor- reports on agility testing conducted at While the 1g roll rates are not particu- sair being the worst with three over- Edwards recently) consisting of larly impressive by modern standards, shoots, the F6F with one, and the P-51 evaluating the ease with which a target they were nonetheless quite respecta- and P-47 deadbeat. The phugoids were of 3g could quickly be captured and ble for their day. It is significant that all likewise well damped and not obtrusive held for 3-5 seconds, and the time taken the aircraft, with the exception once in normal flying. again of the P-47, showed substantial for a 180 degree heading reversal. Then simulated air-to-air and air-to- reductions in roll performance when ground tracking tasks were performed rolling under g. Eventually the P-47 lost DIVING ACCELERATIONS with the results recorded on a video 11 % of its roll performance in loaded Dives were performed both to quan- rolls to the right (though nothing to the tify the trim changes to be expected with camera. Agility Tests - The g capture and left), but this was negligible compared speed variations and to evaluate the hold was easiest in the P-47 which was predictable and accurate. The Hellcat bobbled a bit and overshot the target by 0.2g while both the P-51 and the FG-1 D TABLE III overshot by 0.5g before returning to a steady state. The results of the heading TIME TO ROLL 360° reversal test, performed from 220 kts. (10,000 ft MSL, 200 KIAS) straight and level at 10,000 ft. with METO power, are presented in Table P-47D FG-1D F6F-5 P-51D VI and are somewhat inconclusive, RIGHT 4.9 sec (74°/sec) 4.5sec(81°/sec) 4.6 sec (78°/sec) 4.8 sec (75°/sec) though apparently showing the FG-1D LEFT 5.9sec(61°/sec) 4.9 sec (73°/sec) 5.9sec(61°/sec) 5.1 sec(71°/sec) as the winner with 8.5 seconds for the 180 degrees. With 0.3 seconds cover- 54 JUNE 1990 ^

Figure 4 - North American P-510 Mustang

ing the other three aircraft, no clear con- as a gunsight. The video tape shows, Mustang. The very high longitudinal clusions can be drawn. among other things, just how limited the stick forces required both hands on the Air-to-Air Tracking - The starting field of view was from these aircraft. stick and made delicate lateral correc- conditions for the air-to-air tracking task Accurate tracking was easiest in the tions difficult. Also, the complete ab- were 210 kts. at 10,000 ft. MSL, straight Corsair, at least partly because the light sence of any stall warning resulted in a and level. The target profile consisted longitudinal stick forces did not interfere totally unpredictable departure at about of a 3g turn to the left building to 4gs with lateral aiming corrections. The P- 4.5g during a longitudinal tracking cor- followed by a hard reversal into a 4g 47 and Hellcat were next with the order rection. The characteristics encoun- right turn. The video recorder was dictated by the increasing longitudinal tered during this test strongly supported mounted on the glare shield in each air- stick forces and difficulty of coordinating the conclusions reached in the stall craft in approximately the same position aileron inputs with rudder. Finally, the evaluation. Tracking accuracy was not

100 -i

P51-D FIGURE 11 80 - STICK FORCE PER G

60 -

F6F-5

40-

20 -

1 T 2 3 NORMAL ACCELERATION - G

SPORT AVIATION 55 The automatic control of the oil and TABLE IV coolant radiator flaps in the P-51 re- lieved the pilot of a great deal of fussy, TIME TO ROLL 180° (10,000 ft MSL, 220 KIAS, 3g) secondary workload and enabled him to direct more of his attention to matters P-47D FG-1D F6F-5 P-51D of major importance. It is surprising that LEFT 2.7sec(66c /sec) 3.1 sec(58°/sec) 3.7 sec (48c/sec) 3.3 sec (55°/sec) more effort was not given to developing -11% -28% -38% -27% similar systems for other aircraft of the RIGHT 3.0 sec (61C /sec) 3.7 sec (49°/sec) 4.0 sec (45°/sec) 3.4 sec (53°/sec) time. No Change -33% -26% -25% CONCLUSIONS The objective of the test program as the whole story however. The forward POWER MANAGEMENT stated, albeit rather tongue-in-cheek, field of view, while not good in any of was to decide which of these four air- the aircraft, was particularly bad in the The primary power management craft was the "best" U. S. fighter in Thunderbolt because, as well as being controls consisting of throttle, mixture World War II. The answer is, "it de- long, the nose was so wide. The Corsair and propeller controls were located in pends." For general, all-around comfort, scored well in this respect because its quadrants to the pilot's lower left side. field of view and ease of operation, the forward fuselage was narrow but still The Hellcat and Corsair used identical Mustang was a hands down winner. It accommodated the R-2800. hardware but the P-47 and P-51 were also scored high in performance, was Air-to-Ground Tracking - The air-to- different from them and from each well suited to long range escort mis- ground tracking task consisted of a 90 other. Secondary controls such as cowl sions and would do well intercepting degree roll into a 30 degree dive start- flaps, oil cooler and intercooler controls and defending against non-maneuver- ing from 200 kts. at 5,000 ft. MSL allow- varied substantially among the four air- ing targets. However, its extraordinarily ing the aircraft to accelerate in the dive. craft. high maneuvering stick forces, totally Recovery was a 3.5g right rolling pullout Throttle and Propeller Response - inadequate stall warning and vicious to a 90 degree heading change initiated Manifold pressure response to throttle departure characteristics make it quite at 2,500 ft. movement was instantaneous in all air- unsuited to the ACM environment. It is The flying qualitites were essentially craft, though no automatic manifold a tribute to the adaptability of the pilots unchanged from those seen in the div- pressure limiters were fitted and it was who flew them that Mustangs scored so ing acceleration though the smaller entirely the pilot's responsiblity to avoid many kills against the opposition. speed range reduced the magnitude of overboosting. Propeller response for In a turning fight, the FG-1D emerged the out-of-trim stick and rudder forces. the Hamilton Standard propellers was with a slight advantage over its rivals. The major interest was to determine quick and positive though showing a Light and comfortable stick forces, good how the flying qualities affected tracking small tendency for the rpm to vary in performance, adequate stall warning ability. response to airspeed and manifold and docile behavior at the stall made it The P-47 was the best of the group. pressure changes. The Curtiss electric the "weapon of choice" among those During the 125 kt. acceleration in the propeller on the Thunderbolt was tested. dive, the stick and rudder forces re- noticeably slower to respond and The Hellcat, while possessing almost mained light and easily manageable lagged by as much as 2-3 seconds for a 1g advantage over the other aircraft without retrimming. The aircraft's crisp large rpm changes such as that be- at any given speed, was handicapped and deadbeat control responses, with tween METO (2550 rpm) and cruise by heavy stick forces which interfered negligible Dutch roll excitation, made (2100 rpm). It was also sluggish to re- with accurate lateral tracking correc- accurate tracking particularly easy spond to manifold pressure changes tions, very heavy rudder forces which though once again the extremely poor and had noticeable tendency to over- made coordination difficult, and an ex- field of view was obvious and would speed in dives though not to any treme reluctance to turn right at low have made depressed sight-line aiming hazardous extent. speed and high power. Despite these difficult or impossible. All the radial-engine aircraft de- shortcomings, it is worth remembering The Corsair was next best. The lon- pended entirely on the pilot to monitor that the Hellcat holds the air-to-air kill gitudinal control forces remained light engine conditions and adjust the con- ratio record of 19:1. and pleasant during the 100 kt. acceler- trols accordingly. In steady flight, it was In the air-to-air environment, the P-47 ation but the rudder forces became relatively easy to establish the correct did all that was asked of it, handling high, though not so high as to make cowl and cooler flap settings but engine nicely but unspectacularly; however, it retrimming mandatory. Control re- temperatures were sensitive to small was principally handicapped by its se- sponses were crisp and virtually dead- changes in manifold pressure, rpm and verely restricted field of view. In the air- beat but aggressive lateral corrections airspeed and in rapidly changing flight to-ground role, however, the light stick required continuous rudder coordina- conditions, especially at high power. forces, almost complete absence of ad- tion which was hampered by the high Management of the secondary controls verse yaw, and crisp, deadbeat tracking pedal forces. was a demanding task, especially when responses overcame that drawback The Mustang was similar to the Cor- coupled with confusing or awkwardly and made it particularly suitable for the sair in roll but the longitudinal stick placed control switches. The general mission. The other major weakness of forces were noticeably heavier though consensus would generally be to over- the Thunderbolt was the poorly de- rudder forces were lower and coordi- cool an engine rather than overheat it signed and extremely uncomfortable nated lateral corrections were slightly and that the best thing to do would be cockpit which would undoubtedly de- easier. to worry about the engine until after the grade pilot performance on a long mis- The longitudinal stick forces in the fight. Unfortunately, cowl and cooler sion. Hellcat increased approximately 20 Ibs. flaps are quite serious drag producers In closing, we would like to express during the 100 kt. acceleration and the and have significant adverse effects on our sincere thanks to Harry Tope and rudder forces became high enough to performance. Getting the answer wrong the Kalamazoo Aviation History be a considerable hindrance to accu- either way could have serious conse- Museum for making available the air- rate tracking. quences. craft used for this test program. 56 JUNE 1990