Systems

In Britain, the R.A.E. engaged Smiths Instruments to develop a new, all electrical autopilot under F. W. Meredith. The military version was known as the Mk 9 Autopilot, while a civilian version was made to B.O.A.C. specifications, and released as the Smiths Electronic Pilot, or S.E.P.1, in 1947.

Correction was divided into short-term, and long-term. Short-term, fast-acting, correction was driven by three electrically-powered gyros, for roll, pitch and yaw. Electric pick-offs gave an output proportional to the rate the situation was changing, hence the term “rate gyro”, and their minute signals were amplified by transductors*. Electric servo-motors reacted proportionally to the amplified output. In other words, the rate of correction was the same as the rate of disturbance. Long-term correction was a double-check, taking care of any residual lag or drift using pendulums or mercury switches to confirm that the gyro-platform was absolutely level, and a flux-valve to verify the offset compared with magnetic North.

In fact, the Mk 9/S.E.P.1 was the first successful use of the rate/rate principle, and it was active within 1/100th degree of actual movment on the gyro gimbals. The benefits for an aircraft were reduced stress on the airframe, smoother ride, and improved fuel consumption, because smaller movements created less aerodynamic drag.

*Transductors, or magnetic amplifiers, were developed in Germany during the war. During the late 1940s, they took over from valve amplifiers in heavy-duty or safety-critical applications, because they were solid state, with no moving parts, maintenance-free and capable of handling large currents. Their popularity eventually dwindled after the mass-production of transistors in the late 1950s.

For safety, the servo-motors were engaged using an electro-magnetic clutch which had to be held in place by a constant current. If there was a failure, power to the servos would be cut, and control would return to the pilot. In addition, the pilot could disengage the roll, pitch or yaw channels individually at any time.

The S.E.P. 1 could perform the usual requirements of holding the wings level, nose level, and the current heading. In addition, the pilot could make the aircraft turn, climb or descend while the autopilot was still engaged. This was similar to the way bomb aimers had been able to fly B-17s and B-29s from their Norden and Sperry bombsights during the war. With the S.E.P., the pilot had a little joystick, spring-loaded towards the centre. If the pilot held the joystick in any direction, little worm-drives jacked the entire gyro-platform, fooling the autopilot that it needed to make a correction. If he let go, it would maintain that rate of roll, which was designed for flying in holding patterns, or pitch, for climbing or descending at a particular angle.

On October 2, 1947, Flight magazine reported that engaging the auto-pilot was an imperceptible transition, compared with other systems.

"This", Flight wrote, "was 'George' with a difference.”