Characteristic changes in the pressure altitude and vertical acceleration traces of the FDR recording indicate that lift-off occurred 51 seconds after brake release at an airspeed of about 167 KIAS. Following lift-off, the airspeed continued to increase for a further two seconds until a peak airspeed of 172 KIAS was attained. The aircraft crossed the departure end of the runway six seconds after lift-off, at about 170 KIAS. Thereafter, the airspeed continued to decrease until a stall occurred.

It proved impossible to determine an altitude profile of the flight from the pressure altitude trace of the FDR because of static pressure errors associated with the occurrence of the stall. However, eyewitness observations and the radar controller's observations of the radar Mode C readout suggest the aircraft gained a maximum altitude of 125 feet. The Mode C readout as observed by the radar controller did not change from the 500 feet asl readout that was indicating at the commencement of the take-off roll. The readout indicates in 100-foot increments, thus it is possible that the aircraft could have climbed a maximum of 100 feet above the start of take-off roll altitude (125 feet above the runway departure end) before the altitude readout would have changed to 600 feet. Eyewitness observations were consistent with a maximum altitude gain of 125 feet, although in all likelihood the altitude gain was less than that.

The vertical acceleration trace proved to be unsuitable for estimating a flight path; nevertheless, the rapid fluctuations in acceleration values immediately after take-off indicate the aircraft was in a stalled condition. Further evidence of this condition are the extreme oscillations of the pressure altitude trace which are the result of the rapid pressure changes experienced in the stall regime. The fluctuations in both vertical acceleration and pressure altitude values were in marked contrast to those of previous take-offs. The alteration in heading which commenced about five seconds after lift-off was not inconsistent with control difficulties experienced during stall onset and is typical of swept wing aircraft accidents where aircraft stall was a factor.

Because of the unreliable time sequence associated with the vertical acceleration trace of the FDR, it was not possible to determine precisely when the stall occurred. Nevertheless, when viewed together, the vertical acceleration, airspeed, and heading traces indicate that the aircraft was in a stalled condition within 10 seconds of lift-off.

Early rotation would normally result in aircraft lift-off at about 161 KIAS, if the crew used a pitch angle of eight degrees while on the runway. Analysis indicates that the aircraft lifted off at about 167 KIAS, six knots above the predicted speed. However, since the actual pitch history of the take-off and brief night is not known, it is not possible to conclude with certainty that the lift-off speed was abnormal.

The performance of the aircraft during the take-off was compared closely with the theoretical performance data provided by Douglas Aircraft Co. A normal DC-8-63 at the calculated weight of the accident aircraft and under the existing environmental conditions should accelerate to liftoff in about 47 seconds, using 6,700 feet of runway. After lift-off, the aircraft should climb and accelerate while transitioning to the climb configuration.

The performance of the aircraft during the take-off was below that predicted. Although acceleration corresponded well with that expected to rotation, lift-off occurred four seconds later than predicted assuming normal take-off reference speeds were used. Over 1,000 feet of additional runway were used. Nevertheless, sufficient flying speed was achieved, and the aircraft lifted off well before the end of the runway. The later than normal lift-off should not have had any adverse effect on the remainder of the take-off.

The performance of the aircraft after lift-off was significantly below predicted values. The evidence is conclusive that, following lift-off, both the climb rate and acceleration were well below normal. Although an initial climb was established, it was maintained for less than 10 seconds, and no more than 125 feet was gained during this brief climb. Similarly, the aircraft continued to accelerate for only two seconds following lift-off. Thereafter, the aircraft began to decelerate until the stall occurred.

Based on the data provided by Douglas Aircraft Co. and from the DC-8-63 Aircraft Flight Manual, the 1G stall speed, at the weight calculated by the Board and for the configuration of the accident aircraft, is 148 knots. As determined from the FDR, the aircraft stalled within 10 seconds of lift-off. Airspeed during this 10-second period varied between a peak of 172 knots, which was achieved two seconds after lift-off, and a low of 163 knots, which was the recorded airspeed 10 seconds after lift-off. Thus, the aircraft stalled at an airspeed between 15 and 24 knots above the predicted stall speed. Application of the estimated error bounds of the FDR airspeed trace results in a stall speed range between 10 and 29 knots above the predicted stall speed. It should, however, he noted that the recorded airspeed during the take-off roll agreed closely with that predicted by the Douglas Aircraft Co., evidence that the recorded airspeed values were substantially correct.

Further analysis was conducted to determine the theoretical lift and drag penalties necessary to result in the observed differences between predicted performance and the actual performance of the aircraft during the accident take-off.

Theoretical analysis demonstrated that the performance of the aircraft after lift-off was indicative of a significantly decreased value in coefficient of lift and a significantly increased value in coefficient of drag. During the brief climb which followed lift-off, the aircraft decelerated. Assuming an altitude gain of 125 feet, the coefficient of drag value necessary to produce the deceleration was calculated to be 0.267, well above the normal coefficient of drag value of 0.13 provided by Douglas Aircraft Co. for the conditions and aircraft configuration during take-off. The calculated coefficient of drag was about 100 per cent higher than the normal value.

An altitude gain after lift-off of less than 125 feet would require an even higher value in coefficient of drag to produce the observed deceleration. The calculated coefficient of drag values that corresponded to altitude gains of 100 feet and 70 feet were 0.281 and 0.29 respectively.

Although the recorded airspeed could have been subject to a maximum error of five knots, any error would have been constant, and thus would have no effect on the validity of the deceleration used to calculate the coefficient of drag.

The increase in both lift-off speed and stall speed is indicative of decreased lift-producing capability of the wing (i.e., coefficient of lift ). The calculated decrease in CL maximum necessary to account for the magnitude of the increase in stall speed was at least 0.38. According to data provided by Douglas Aircraft Co., this corresponds to about a 22 per cent decrease in maximum coefficient of lift.

The conclusions of the computer simulations conducted by UDRI agreed closely with this analysis. Their solution of the aircraft's equations of motion determined that an approximate 30 per cent loss in coefficient of lift had occurred accompanied by at least a 100 per cent increase in coefficient of drag.