The story behind the avenger’s replacement
By mid-1944 the U.S. Navy knew that it would take much longer than previously thought for the Allies to win WW II. The Japanese were still putting up very stiff resistance to the American demand for complete capitulation. Thus, it was expected that an invasion of Japan in late 1945 would be required to attain unconditional surrender.
The Grumman Avenger had taken quite a beating in the War’s early years, when four major torpedo problems made pilots go way too close to enemy ships to obtain hits at slow 120mph launch speeds. This was because, in 1926, the U.S. Navy stopped the development on torpedoes when they ran out of WW I R&D money. Development efforts had now made torpedoes that could be launched at up to 400mph and thus be released twice as far away from the guns of the target. The Navy knew that such a launch speed required at least twice the engine power of the Avenger. Grumman offered a twin-engine configuration called the XTB2F-1. With all of the new Navy requirements, however, this aircraft soon became too large for Essex-class carriers. It was canceled shortly after the Navy’s full-scale mock-up inspection in June 1944. Grumman then submitted five single-engine proposals:
1. G-70 (Grumman design number) with the R-2800
Pratt & Whitney engine rated at 2100hp and a dorsal gun turret like the Avenger’s.
2. G-70A with a Curtiss-Wright 2500hp R-3350.
3. G-70B with a 3000hp Pratt & Whitney R-4360 engine and a G.E. (General Electric) remote-controlled gun turret for rear protection.
4. G-70C with a 3000hp Curtiss-Wright R-3350 engine and G.E. I-20 jet engine, but without the turret.
5. G-70D had the higher-thrust Westinghouse 24C jet engine as a replacement for the I-20 in the G70C. With both engines running, the G-70D met all the Navy requirements: 394mph, rate of climb 4,880 ft./min., and takeoff distance of 204 feet with a 25mph wind over the deck!
Grumman and the Navy soon discovered that none of the production schedules of the R-3350 or R-4360 piston and I-20 or 24C jet engines would meet the Navy’s required deployment schedules. The Navy then asked Grumman, in January 1945, to prepare a sixth design with the 2300hp Pratt & Whitney R-2800-34W (now with water-Injection power increase) and the Westinghouse 19XB turbojet. This model had only a little less performance than the G-70D. Its speed at sea level would be 367mph, and takeoff distance was only 330 feet. The Navy contracted with Grumman to build three XTB3F-1s to this design. The first flight was made on December 23, 1946. The Navy put a stop-order on the program the next day because it had just given large contracts for the Douglas AD-1 Skyraider and the Martin AM-1 Mauler to take over the primary dive-bombing and torpedo roles.
The Cold War soon spurred an important requirement for a state-of-the-art, multi-seat Anti-Submarine Warfare (ASW) aircraft to replace the aging Grumman Avenger/Wildcat Hunter-Killer teams. On February 27, 1947, the Navy instructed Grumman to complete the second and third XTB3F-1s, without jet engines, to be redesigned as prototypes for this much improved ASW Hunter-Killer team.
Removal of the jet engine provided space for three equipment operators needed for these missions. These two XTB3F-1s prototypes were flown in November 1948 and January 1949. The first production aircraft, now called AF-2W, with the AN/APN-20A radar (Hunter version) and AF-2S (Killer version), carrying armament, flew one year later. A total of 386 aircraft were built. They began taking over the Grumman Avenger ASW role in November 1950, and they remained in service until November 1957. It came to be known as the Guardian
I was never a project pilot for either of the XTB3F or AF programs, but as senior engineering test pilot in November 1945, I was brought into the program many times to check development fixes and to fill in after the sad demise of three AF project test pilots! Pat Gallo had experienced a fatal low-altitude parachute jump from an F8F-2 Bearcat. Bill Cochran died in a USAF record-attempt takeoff in a KC-135 while flying as an FAI (Federale Aeronautique Internationale) witness. Mike Ritchie was unable to fly after making a very high-speed, 200-foot altitude parachute exit when the first XTB3F-1 crashed on Long Island. He landed on top of the crash and was hospitalized for many months before returning to an engineering ground assignment. The Pratt & Whitney engineer in the aft fuselage, performing an in-flight propeller vibration survey, died in the crash. Because of these depressing events, I logged 218 flights in the AF program in addition to my regular job of fighter testing.
Solving some interesting Guardian
After Pat Gallo’s death, I was to find out why the present
project test pilot, who had just arrived at Grumman, had determined that the AF-2W prototype was unstable longitudinally and would not maintain its trimmed altitude flying hands off. This would make a long, low-altitude cruising flight in turbulent air intolerable to the pilot. Before flight, I made my usual full-deflection test of the three flight controls and found that the stick motion fore and aft had an excessive amount of
friction—11 pounds. Stability in the air was usually measured below that force, so I asked the plane captain what the cable tension was in the elevator control. He stated 135 pounds. I decided that it would be pointless to make a flight with this much friction and asked him to
re-tension the cables to 25 pounds. The new test pilot stated emphatically that the Navy’s sacrosanct minimum requirement was 135 pounds and could not be reduced!
I suggested to him that the test pilot’s objective was to find ways to make an aircraft meet its requirements safely; then, if necessary, have the engineers obtain the needed changes to the Navy requirements. I had experienced similar problems several times during my prior five years as an engineering test pilot.
With the cable tension now at 25 pounds, the measured friction was only half a pound, so I took off for a 15-minute flight. I found that the AF-2W now had the longitudinal stability to make turbulent air cruising satisfactory. The engineering department persuaded the Navy to lower the cable tension to 25 pounds.
AF directional stability problems
A few months later, I was asked to evaluate the AF-2S Hunter prototype to find out why pedal forces went to zero when the rudder was pushed to full deflection in flight. During my preflight walk-around, I noticed that this prototype now had a very large dorsal fin attached to the main fin, which should have greatly increased rudder pedal forces at full deflections. I also noticed that the rudder had a large balance tab that was usually installed to reduce pedal forces. I assumed at first that the balance tab was the problem. I soon found that it and the increase in fin area were both culprits. In flight, the rudder forces went to zero at full deflection, so I landed and asked the mechanic to reduce the balance tab gear ratio with rudder deflection to half its present travel. On the next flight, the rudder forces decreased much less at full deflection, so I had the balance tab ratio reduced further to 25 percent motion, and that cured the rudder-overbalance condition.
I was also concerned that such a large increase in total fin area would greatly increase the pilot’s rudder deflection requirements for safe crosswind takeoffs and landings. This excessive fin area also made the aircraft weather-vane more easily into the wind and made it difficult to control on the ground during these maneuvers. I took off in a 10mph, 90-degree crosswind. It took almost full rudder pedal deflections for satisfactory directional control. When I landed in the same crosswind, I decided to ask why the big dorsal fin was installed. The project aerodynamicist said they thought that such an addition was necessary to cure the rudder force reversal problem. After I had the big dorsal fin removed and the original quarter-size dorsal fin reinstalled, the crosswind weather-vaning problem disappeared with satisfactory rudder-control inputs.
During my flights, I also noted that two long leading-edge spoilers had been installed on the wing. They caused a large amount of buffeting during takeoff and landing. The engineer told me that they were designed to correct excessive wing rolling in takeoff and landing condition stalls. I then inquired whether they had tested for any increase in stall speed since installing them. He answered no, so on my next flight, I measured stall speeds and asked the engineer to compare them to before the strips were installed. He acknowledged that the stall speed had increased by 7mph. This buffeting drag caused longer takeoff distances and an increase in energy that the cables needed to stop the aircraft during an arrested landing. They also gave the pilot too much buffeting, causing him to increase his approach speed, thinking his aircraft was flying much too close to the stall.
I suggested that because the right wing strip was 30 inches long and the left 36 inches, the right strip be eliminated and the left one shortened to 6 inches. My flight tests showed that the stall speed had decreased 7mph and the excessive buffeting had been reduced. The 6-inch remaining tripper strip still provided satisfactory warning without any wing rolling at the stall. I wish I could have taken credit for being smarter than the engineers. I couldn’t. Once again, my previous years of flight testing had provided me such insight.
Navy concerns in the AF-2W
In February 1949, the AF-2W Hunter prototype with the Westinghouse AN/APS-20 radome installed in place of the bomb bay was evaluated at the Patuxent Naval Air Test Center for carrier trials and general handling characteristics. It was found to have nose-over tendencies after catching the arresting wire. Its propeller diameter was reduced from 13 feet 6 inches to 12 feet 2 inches, which gave sufficient deck clearance during arrested landings. It also had the prototype hydraulic-powered flaperon/spoilers inboard of the ailerons to increase its rolling power during carrier approaches. During this evaluation, however, Navy test pilots had found it to have unacceptably low directional and lateral stability.
The Grumman project engineer had gone to Washington and told the brass that the aft fuselage structure wasn’t strong enough to add the amount of fin and rudder area necessary to cure the directional problem. He said it would require a total aft fuselage redesign. Washington replied the next day that if an increase in fin area couldn’t be made on the present structure, they would be forced to cancel the program! It was interesting to see how rapidly this ego-centered engineer changed his mind. The aft fuselage, as designed, was indeed strong enough to add a 43 percent increase in vertical fin area. He had already known that answer before he went to Washington! Grumman test pilots had not been satisfied with the enlarged fin and rudder fix, so two vertical fins were added at the mid-stabilizer span. It was now considered by Navy and Grumman test pilots to be satisfactory.
One other major problem had been fixed before our visit to Patuxent during the weekend of June 24 to 28, 1950: the elimination of many fuselage leaks that caused excessive carbon monoxide from the engine exhaust to enter the cockpit and aft cabin areas.
I was the test pilot with the new team that went to Patuxent to solve two problems that remained and hopefully to find acceptable fixes. The first was to improve the Guardian’s lack of lateral stability. I flew three flights a day for three days to test four engineering lateral stability fixes without solving the problem. After thinking about it during those flights, I suggested that we rig both ailerons 10 degrees down and install fixed tabs bent 25 degrees up to prevent air loads from pushing the ailerons back up. I hoped that the down-deflected ailerons would act like the wing had a greater dihedral angle, which hopefully should increase the lateral stability. Lateral stability was now deemed acceptable for reasons that still baffle me and the Grumman aerodynamics department! Sometimes intuition does provide answers.
The other problem, caused by the radome of the APS-20, was that the AF-2W pitched up with right rudder deflections and down with left rudder deflections. Herb Crawford, chief of flight test, noticed that the Douglas AD-1 Skyraider had one elevator tab bent up 20 degrees and the other elevator tab bent down. He asked one of the Navy test pilots why they were so installed. The pilot told Herb that it cured a problem similar to ours. We installed tabs bent accordingly and cured our problem (with many silent thanks to the Douglas Aerodynamics department)
On Sunday morning, Navy test pilot Cmdr. Joe Reese evaluated all of our new fixes and pronounced them acceptable. That evening, I flew our team up to Washington with Cmdr. Reese in our Grumman Goose amphibian. At the Navy Bureau of Aeronautics on Monday morning, he reported his acceptance of our fixes to the Guardian project desk. What a long weekend!
Solving the AF’s weak tailwheel problem
Early AF squadrons had trouble with breaking tailwheels during carrier operations when approaching the deck and flaring out to catch the arresting wire. Trying to fix the problem, Grumman engineers had progressed from a stronger single tailwheel to a double tailwheel and then a triple tailwheel! None solved the problem. When I first noticed this amazing tailwheel configuration on our flight-test line, I asked flight test chief Herb Crawford what the problem was. He told me the long and incredible story. I suggested that I fly that configuration because I wondered whether the AF program was having a similar problem that we had previously cured on the XF9F-2 Panther. The Panther smashed its tailpipe many times during its first carrier landings tests by Patuxent test pilots. I fixed it by simply limiting its up-elevator travel from 30 to 15 degrees so it did not
have so much excess elevator-control power available during carrier-landing flare-outs.
I proposed some Guardian restricted up-elevator deflection landing tests with a nose-heavy center of gravity, which we found to be critical to this problem in the Panther. I made 12 simulated carrier landings with incremental reductions of the up-elevator travel by slowly reducing a chain length attached from the stick grip to a hook on the instrument panel during practice carrier landings. I determined that 11 degrees was sufficient up-elevator for a satisfactory flare-out during all carrier landings and center-of-gravity conditions. A single tailwheel would always hit the deck for proper hook snatching, but now with much less violence. Patuxent test pilots approved the 11-degree elevator-up limit, and it was put into production.
The Navy had the fix-implementation duty for all service aircraft, but they omitted spreading the fix information. They forgot to send it to all three Naval Overhaul and Repair (O&R) stations. Three months later, Grant Hedrick, chief of structures, told me that Navy squadrons were again breaking tailwheels, and my fix had been no good.
I asked the Grumman service department to make inquiries throughout the Navy, and we learned that only Guardians that had come out of one of the three O&R stations had their elevators re-rigged to 30 degrees up. I pleased to be out of Grant Hedrick’s dog house! All aircraft that had been through the O&R process now had their elevators re-rigged to 11 degrees up, and no more tailwheel bashings occurred.
ASW in one aircraft
I do not remember whether the Navy or Grumman suggested the possibility of putting all the ASW roles and missions of both aircraft into one aircraft, but it was a great idea. It gave all ASW carriers a 100 percent increase in capability. Thus, with the lighter and more powerful wing-mounted AN/APS-31, the AF-2S was selected to be the AF-3S for both Hunter and Killer roles. An additional piece of new ASW equipment, the Magnetic Anomaly Detector, which could locate submarines deep under the water, was also added to the aft fuselage. It could be extended 12 feet for best operations or retracted for landing. With the 4,000-pound armament weight and drag of this aircraft, its mission range was now considered too short, and two 150-gallon drop tanks were required to obtain the desired range. Forty AF-3S Hunter-Killer aircraft were ordered.
Now with the AF-3S Hunter-Killer capability in one aircraft, the ASW competition that followed provided the Navy with 42 proposals from 18 manufacturers! Grumman won with the S2F-1 Tracker twin-engine design. They were so successful in operational capability that the Navy ordered 1,169 between December 1952 and December 1957.
Arrogance ends the Grumman ASW monopoly
After building 11,489 Anti-Submarine Warfare Avengers, Guardians and its follow-on S2F twin-engine Trackers, Grumman lost the next Navy ASW carrier aircraft competition on August 1, 1969, to the Lockheed S-3A Viking. Although I was no longer director of business development, I was sent by the Grumman president to Washington to find out why. A former Navy friend allowed me to read the decision paperwork. It stated, “The Navy expected Grumman to win after their continuous 28 years of supplying the Navy with excellent ASW aircraft.
Because of Grumman’s pompous attitude expressed in their very skimpy proposal, they were rated 25th in a five-company competition!” I went home and read our proposal. It was a “Send us the money and we will make you the next ASW aircraft”—a real ego-trip dud by Grumman.
This team from engineering was never allowed to do a preliminary design again.