Lockheed XF-104 Starfighter

Last revised December 5, 1999

The Lockheed F-104 Starfighter was the result of an attempt to reverse the trend towards ever-increasing weight and complexity in fighter aircraft. When it first appeared in the mid-1950s, it had a futuristic look about it, and its small wing area and needle-nose earned it the appelation of "missile with a man in it". The F-104 was the first operational interceptor capable of sustained speeds above Mach 2 and was the first aircraft ever to hold the World Speed and Altitude records simultaneously.

The Starfighter was destined to serve only briefly and in relatively small numbers with the air force of its country of origin. It was to be in the service with the air forces of other nations that the Starfighter was to achieve its reputation. The Starfighter won a large contract for NATO's next-generation multi-role fighter capable of delivering nuclear weapons, and was built in large numbers by a European consortium of aircraft manufacturers from Germany, Italy, Belgium, and the Netherlands, and Canada as well as the USA. The Starfighter became an important part of NATO's nuclear deterrent during the 1960s and 1970s, and served with the air forces of Denmark, the Netherlands, West Germany, Italy, Belgium, Greece, Turkey, Canada, Spain, and Norway. It was also built under license in Japan. The Starfighter also served with the air forces of Jordan, Taiwan, and Pakistan. Although the Starfighter has been superceded by later types in most of the air forces with which it was originally associated, the Starfighter still serves in fairly large numbers with the air forces of Greece, Taiwan, and Italy,

A total of 2580 of all Starfighter types were produced, making the aircraft one of the most important Western postwar military aircraft. However, in the mid-1960s when the Starfighter first entered service in significant numbers, the aircraft was involved in a large number of accidents. Because of the high accident rate, the Starfighter became a controversial aircraft and there were charges that the aircraft was an intrinsically flawed and dangerous design. However, in retrospect, the Starfighter was not intrinsically any more dangerous to fly than lots of other military aircraft of the day, and the high accident rate can be blamed more on inadequate and insufficient crew training rather than on any flaw with the basic design.

The Starfighter had its origin in a November 1952 unsolicited proposal by Lockheed's Clarence L. "Kelly" Johnson for a lightweight and relatively unsophisticated air-superiority fighter. Weight and complexity would be minimized in the pursuit of unmatched speed, altitude, and maneuverability. Johnson had visited Korea in December of 1951, and while there he had talked to fighter pilots then flying in combat over North Korea. He asked them what kind of fighter plane would be ideal. Their general consensus was that the trend toward ever-increasing weight and complexity had gotten completely out of hand, and they would gladly trade in their existing fighters for a lighter, less costly fighter with clearly superior speed, ceiling, climb rate, and maneuverability. Following his return to the USA, Johnson tried to convince Lockheed management that they should design a new type of fighter plane, one that was uncomplicated, lightweight, and inexpensive but one that would be able to outperform any other fighter in the world.

Even though the Air Force had no official requirement for such a fighter, Johnson was nevertheless authorized by Lockheed management to proceed with an initial private venture design.

In March of 1952, Johnson assembled a team of first-rate engineers at the famous Skunk Works at Burbank, California to work on the new fighter. The team started with the low-winged Project 227-0-6, which had a MiG-21 type center nose cone intake. This was followed by the mid-mounted delta-winged 227-0-11 with lateral intakes, a flush cockpit in a conical nose, and a high-set tail on top of a sharply-swept vertical fin. The 227-8-1 which appeared a month later reverted to a conventional low/mid-wing layout but weighed about 30,000 pounds. The Model 227-16-2 of a few weeks later was similarly configured but had a chin intake and weighed only 8000 pounds. It featured an ultrathin straight wing. This design was later scaled way up into the Project 227-13-1, which weighed about 50,000 pounds. At the same time, the team looked at the 227-15-3, which was a rocket-propelled aircraft looking a lot like a stretched Bell X-1. The 227-20-1 of October 1952 completed the Project 227 studies--it stretched the 227-13-1 concept even further but at only half the weight.

In November of 1952 the team started a new series of studies known as Project 242. This time the emphasis was to be on a substantially smaller fighter. The 242-19-1 proposal had a miniscule mid-mounted wing with a vee-shaped windshield and a fuselage-mounted tailplane. Its estimated weight was only 9000 pounds empty. Apart from its tiny dimensions, it had the basic shape and layout of the design which was eventually to emerge as the F-104. The 242-23-1 which appeared at the end of 1952 was a scaled-up version of the earlier proposal and had a stabilator mounted low on an extended-chord vertical fin. The slightly-smaller 242-27-1 of February 1953 had the stabilator moved to the top of the vertical fin. This tail arrangement and the 242-27-1's unraked cheek intakes were then combined with the more elongated Project 246-1-1 which had evolved in the previous few weeks. This was finalized in the spring of 1953 as the Lockheed L-246 or Model 83, which had an empty weight of 12,000 pounds. The maximum takeoff weight weight was only 15,700 pounds, less than half the weight of some of the "Century Series" of fighters that were appearing at the time.

The L-246 design was based on an ultra-thin low-aspect ratio, symmetrical section trapezoidal wing with a thickness/chord ratio of only 3.36 percent and a quarter-chord sweepback of only 18 degrees. The thin, straight wing had the advantage of providing a high lift potential at low angles of attack and high speeds, but with a penalty of a high induced drag at large angles of attack and in high-g conditions. The leading edge radius was only 0.016 inches, sharp enough to require a felt covering strip during maintenance to protect ground personnel from injury. The wing had 10 degrees of negative dihedral in order to improve roll control during high-G maneuvers and to enhance stability at high speeds and high altitudes. Since the vertical fin was only slightly shorter than each wing mainplane, the anhedral was thought necessary to counter a marked roll tendency from rudder application.

The potential flutter problems that might arise from possible aeroelasticity in the thin wings were minimized by the short span, which gained some endplate and damping effects from the mounting of long jettisonable 170-US gallon tip tanks. The aircraft was supposed to be capable of supersonic performance even when these tanks were mounted. It was proposed that the wingtip tanks could be replaced by mounting rails for a Philco Sidewinder GAR-8 AAM.

This wing design was based on earlier Lockheed experience with the X-7 ramjet test vehicle as well as on the results with high-speed rocket flights over the western desert with various wing profiles. Test flights of the Douglas X-3 Stiletto experimental aircraft also played a role in the design of the wing. In order to recoup its losses on the X-3 program, the Air Force had insisted that Douglas deliver the aircraft plans to Lockheed.

Since the wing was so small, special techniques had to be devised to keep the landing speed at an acceptably low level. The Lockheed team employed boundary layer control to increase the amount of lift available at low speeds, thus decreasing the landing speed. This system operated by blowing compressed air from the engine over the trailing edge flaps, reducing turbulence in the boundary layer due to flow separation thus reducing the stalling speed. Full-span leading edge flaps were also fitted, which drooped in coordination with the trailing edge flaps during take-off, landing, and low-speed maneuvering.

The powered, all-flying horizontal tailplane was mounted atop the vertical fin. It moved as a unit and had no elevator. The tailplane was situated on top of the vertical fin in order to get it out of the turbulent air flowing over the wings and fuselage. It was hoped that this would help to improve lateral stability at high speeds.

The rocket-like fuselage was of a high fineness ratio (i.e., highly tapered toward the nose). All of the internal fuel was housed inside the fuselage, there being no room for fuel inside the thin wings. There was no space in the wings for the retracted landing gear either, so they had to be accommodated entirely inside the fuselage. The main oleos were mounted on skewed pivots that twisted during forward retraction so that the wheels lay flat in the lower belly.

The engine was to be the General Electric J79 engine, which was currently under development. It was an outgrowth of the J73 and was known at that time only as the J73-GE-X24A. The proposed J79 was to be capable of producing 9000 lb.s.t. dry and 15,000 lb.s.t. with afterburning. It was designed to be capable of Mach 2 performance. Since the advanced J79 would not be available for several years, the afterburning Wright J65-W-7 was selected as an interim propulsion system for the first few examples.

The engine was to be fed by a set of lateral air intakes, one on each side of the fuselage just ahead of the wing leading edge. Fixed half-cones were to be incorporated in the lateral air intakes to reduce the speed of the air entering the engine. These half-cones were supposed to reduce Mach 2 airflow to about Mach 0.7 at the engine's face.

A downward-firing ejection seat system was selected, since it was feared that conventional upward ejection would be highly dangerous if not impossible at the high speeds at which the CL-246 would be operating.

On October 31, 1952, Johnson presented the CL-246 proposal to Lockheed management. They were enthusiastic, and gave him the go-ahead to present it to the Air Force. Even though the USAF did not have a standing requirement for such a fighter, the USAF thought sufficiently highly of the general idea that they issued a General Operational Requirement on December 12, 1952 for a lightweight air-superiority fighter to replace the North American F-100 in the Tactical Air Command beginning in 1956. However, in order to be completely fair, the USAF had to request competitive bids for the project from the aviation industry.

In response to the request for proposals, Republic submitted its Model AP-55, based on its XF-91 Thunderceptor, but with a solid rounded nose and NASA-developed flush-type engine air intakes. North American submitted its Model NA-212, which was an advanced version of the Super Sabre which eventually emerged as the F-107. Northrop submitted its Model N-102 Fang, a proposal for a J79-powered aircraft fed by a ventral, bifurcated air intake.

Lockheed's head start was just too much for the competitors to overcome, and in January of 1953, Lockheed's proposal was selected. On March 12, 1953 a letter contract for two prototypes was issued under Weapon System 303A (WS-303A). The designation XF-104 was assigned. Lockheed assigned the aircraft the company designation of Model 083-92-01.

Under the guidance of Clarence R. "Kelly" Johnson and project engineer Bill Ralston, the project rapidly moved ahead. The mockup was inspected on April 30, 1953, and at that time it was decided to substitute a single General Electric Vulcan Gatling-type cannon (then under development and known as the T-171) in place of the two 30-mm cannon originally proposed. The T-171 (later to be designated M61) cannon was to be mounted on the left side of the fuselage and was projected to be capable of firing up to 6000 rounds per minute. The cannon was 72 inches long and weighed about 300 pounds. It was to be fed by a 725-round drum of ammunition. The cannon was to be integrated with a Type K-19 fire-control system and incorporated an AN/APG-34 radar and a computing gunsight.

The first prototypes were to be powered by a non-afterburning Wright J65 turbojet (license-built Armstrong Siddeley Sapphire), but production aircraft were to be powered by a single afterburning Wright J65. The J65 would serve as the interim powerplant until the more advanced J79 could be ready.

Construction of the first prototype XF-104 (53-7786) began in the summer of 1953 at Lockheed's Burbank, California factory. This aircraft initially was powered by a non-afterburning Buick-built Wright J65-B-3 turbojet. Construction of the second prototype (53-7787)--the armament test bed--began in the autumn of 1953, but work on this aircraft proceeded at a slower pace in case revisions were needed. The air intakes of the two XF-104s were of fixed geometry without presence of half-cones, since the J65-powered aircraft was incapable of Mach-2 performance. The air intakes were similar to those of the F-94C, being mounted slightly proud of the fuselage, with an inner splitter plate for the boundary layer bleed.

The first XF-104 (53-7786) was ready in early 1954, and was trucked out to Edwards AFB in high secrecy during the night of February 24-25. Veteran Lockheed test pilot A. W. "Tony" LeVier was to do the initial testing. Taxiing runs began on February 27, 1954. On February 28, 1954, the XF-104 made an scheduled short hop of about five feet off the ground during a high speed taxiing run. Its first official flight took place on March 4, 1954. During that flight, the landing gear would not retract. After a low-speed flight of about 20 minutes, Tony LeVier landed. Some adjustments were made, and LeVier took off again, but the landing gear still would not retract. The problem turned out to be low pressure in the hydraulic system, which was fairly easy to correct. However, inclement weather kept the XF-104 on the ground until March 26, when flights three and four were carried out with the landing gear retracting adequately.

The XF-104's original yaw damper was ineffective, allowing the nose to wander left and right. This problem was corrected by revising the rudder-centering device.

The XF-104 could not exceed the speed of sound in level flight when powered by the nonafterburning J65-B-3 turbojet. However, Mach 1 could be easily exceeded during a slight descent, and the transition to supersonic speed was quite smooth.

In July of 1954, the J65-B-3 non-afterburning engine was replaced by the long-awaited afterburning J65-W-7 turbojet rated at 7800 lb.s.t. dry and 10,200 lb.s.t. with afterburner. In that same month, 17 more service test aircraft were ordered. They were also to be powered by the J65-W-7.

With the afterburning engine installed, the performance of the XF-104 was markedly improved. Maximum level speed was Mach 1.49 at 41,000 feet, and an altitude of 55,000 feet could be attained in a zoom climb. Mach 1.6 could be attained in a dive.

The second prototype (53-7787) flew on October 5, 1954. It was fitted with the afterburning J65 from the start. Since it was to be the armament test bed, it was fitted with the 20-mm Vulcan cannon and was equipped with an AN/ASG-14T-1 fire control system. Initial aerial firing tests with the Vulcan cannon were successful, but on December 17, there was an explosion during a firing burst, and the J65 engine started to run rough. Test pilot Tony LeVier immediately shut down his engine and glided back to make a successful dead-stick landing at Rogers Dry Lake. An investigation later showed that one of the 20-mm cannon rounds had exploded in the breech, blowing the bolt out the rear of the gun and into the forward fuselage fuel cell. Jet fuel gushed into the gun bay, and leaked out of the gun bay door joints and into the left engine air intake. The engine immediately flooded with fuel, choking it to death. Tony LeVier was lucky to be alive.

XF-104 number one achieved a top speed of Mach 1.79 at 60,000 feet on March 15, 1955. Lockheed test pilot J. Ray Goudey was at the controls. This was the highest speed achieved by either of the XF-104 prototypes.

The second prototype (53-7787) was lost on April 14, 1955 when test pilot Herman R. "Fish" Salmon was forced to eject during gun-firing trials at 50,000 feet. The gun malfunctioned during a test firing, and severe vibrations began to build up which knocked loose the ejection hatch on the belly of the plane. Cabin pressure was immediately lost, and Salmon's pressure suit pumped up and covered his face so that he could not see. Recalling Tony LeVier's harrowing experience with the exploding cannon shell the previous December, Salmon believed that the same thing had happened to him and that he had no option but to eject. This he did. He later found out that he could have saved 53-7787 by simply bringing it down to a lower altitude and waiting for his pressure suit to deflate.

With the loss of the armament testbed, Lockheed engineers were forced to find an alternative. Armament trials were continued on a modified Lockheed F-94C Starfire.

The first XF-104 was accepted by the USAF in November of 1955. XF-104 number 1 was lost in a crash on July 11, 1957, when it developed an uncontrollable tail flutter while flying chase for F-104A flight tests. The entire tail group was ripped from the airframe, and Lockheed test pilot Bill Park was forced to eject.

Consequently, no XF-104 prototype survives today.

Serials of Lockheed XF-104 Starfighter

53-7786/7787 	Lockheed XF-104 Starfighter 
			c/n 083-1001/1002 


  1. The World's Great Interceptor Aircraft, Gallery Books, 1989.

  2. Lockheed F-104 Starfighter, Steve Pace, Motorbooks International, 1992.

  3. Lockheed Aircraft Since 1913, Rene J. Francillon, Naval Institute Press, 1987.

  4. The American Fighter, Enzo Angelucci and Peter Bowers, Orion, 1987.

  5. The World's Fighting Planes, William Green, Doubleday 1968.

  6. American Combat Planes, Ray Wagner, Third Enlarged Edition, Doubleday, 1982.

  7. Lockheed F-104 Starfighter, John Fricker and Paul Jackson, Wings of Fame, Vol 2, Aerospace Publishing Ltd, 1996.