The first all-electronic autopilot
The precision targeting made possible by the bombsights demanded a higher level of precision in maintaining a plane's course, attitude, altitude, and trim--far beyond what could be attained with a bombardier-pilot team or commercial autopilot.   

Some of the B-17s in the late 1930s came equipped with a Sperry A-3 commercial autopilot. The gyros in the A-3 sensed only simple angular displacement of the aircraft from the desired course. It used pneumatic hydraulic servo systems that were sluggish, and since there was no measure of velocity or acceleration, the system tended to overcompensate in rough air and thus oscillate.     

The Norden company developed an autopilot called the stabilized bombing approach equipment (SBAE), also based upon the earlier displacement only signal technology of commercial auto pilots. The Norden SBAE's mechanically sliding trolley-contact electric servos had simple dashpots or shock absorbers that produced negative clamping to eliminate oscillations, but this also showed response either to wind buffeting or to commands from the bombsight. The result was flight control no better than that of the Sperry A-3 commercial autopilot.    

For the new Sperry S-l bombsight, Frische invented the first all-electronic autopilot, the A-5. It was based on three dual element vacuum tube amplifiers, each corresponding to a different axis in the aircraft's control system: roll, pitch, and yaw. The tubes had been subjected to accelerated life testing, temperature cycling, and vibration to ensure unprecedented reliability.     

Each tube amplified the weak signal measured from the autopilot's own set of sensors on the high-speed induction gyros. More important, in addition to the displacement-error signal, the A-5 autopilot adjusted for the first and second time derivatives (the velocity and acceleration with which the aircraft departed from the base reference signal). The amplified signals controlled independent electro hydraulic servomechanisms, providing fast response for stabilizing the aircraft.

This resulted in a system that was critically damped, thus allowing for the aircraft's inertia, and was much more responsive than the electromechanical technology to wind gusts and command signals from the bombsight.     

Controls for the Sperry S-1 bombsight were electrically connected to the A-5 autopilot. Once the bombing run was begun, the pilot turned the aircraft over to the bombardier, who then flew the bomber by tracking the target through the bombsight. When the bombsight determined that the release point had been reached, it alerted the bombardier and dropped the bomb. The combination of the two Sperry mechanisms into one system led to unprecedented accuracy in targeting while in combat.     

The officials of the U.S. Army Air Corps were so impressed with the Sperry A-5 autopilot's performance that on June 17, 1941, the company was awarded a government contract authorizing the 186, 000 square-- meter plant in Great Neck, N. Y., for manufacturing the A-5 autopilot and S1 bombsight. The Air Corps also issued a Teletype message noting its decision "that in the future all production models of bombardment airplanes be equipped with the A-5 Automatic Pilot and have provisions permitting the installation of either the M-Series [Norden] Bombsight or the S-l Bombsight."

Rivalry and salesmanship
The Norden company was not pleased with Sperry's growing competition. As early as July 29, 1937, when the Air Force's Major General Andrews began encouraging Sperry to develop a bombsight for Air Force planes because Norden could not meet the demand, a conference was held at the Norden company in New York City between Navy, Air Corps, and Norden personnel.