While servos have been in use as far back as the steam age, their impact on rc models has been a recent one. During this blog, we'll trace the evolution of servos to the high tech products of today.
First, let us define what a servo is, and does. A servo is a small modular actuator developed by the radio control model industry for the remote operation of everything from model boat rudders and car steering linkages to model airplane flaps and robotics. Servo motors were first used in 1908 by the inventor and professor Elmer Sperry to power gyroscopes on ships. In 1916, Lawrence Sperry, an aviation inventor, filed a US patent application for the use of servo motors to direct fins on aerial torpedos. By the end of World War I, servo motors were becoming a common term among America's community of electrical engineers. General Electric began design efforts in 1922 on an electronic Selsyn-servo for use in directing naval guns. Within three years, GE engineers perfected an electronic servo using proportional control with feedback for stability, principles of modern servos. During the early 1930's, both Westinghouse and General Electric began producing strip chart recorders that used electronic servo motors for driving the pen mechanism.
However, it wasn't until the 1940's that serious attempts to both develop and fly radio control models came to fruition. The early radio transmitters used a 27 Mhz carrier wave system, which operated by vacuum valves with massive 120 volt dry batteries plus a 2 volt lead acid heater battery. It was often necessary for two people to carry the transmitter from the vehicle to the flight area, where an 8' 6" antenna was erected. Receivers of the era used gas filled valves, which worked on a voltage range of from 22 to 45 volts. An additional mini two volt accumulator was used to power the valve heater. Some of these units were actually built from dismantling old car batteries. The control surfaces of model aircraft at the time were relatively small with the aircraft mostly free flight models designed with a lot of wing dihedral to keep them stable. The radio control system could not fly the rc plane, but merely alter its path. The 1950's brought about development of small electric servos, which could be carried aboard rc model aircraft. While better than the earlier pre-wound elastic bands, these servos only offered limited control of rc plane surfaces. By the late 1950's, pulse proportional systems came into use. These units utilized a mechanical or electronic mark/space pulsing circuit in the transmitter. The receiver switched a spring centered electric motor, backwards and forwards in the model, using a fast pulse rate. This produced a crude, but proportional control of the rudder. Another pulse rate system, called the galloping ghost (due to the noise it made when gliding to a landing) was a much slower pulse rate system, which kicked the elevator of the model upward as the transmitted pulse rate slowed. Though a relatively crude system, the galloping ghost offered effective control of both rudder and elevator control surfaces. At the end of the 1950's, transmitted audio tones could work several control surfaces on rc planes.
By the early 1960's transistors were becoming more common, replacing the tube and electric motor driving control surfaces. In both tube and early transistor sets the rc model control surfaces were usually operated by an electromagnetic escapement controlling the stored energy in a rubber-band loop, allowing simple rudder control and other functions such as motor speed and kick-up elevator. The first inexpensive proportional systems did not use servos, instead employing a bidirectional motor with a proportional pulse train consisting of two tones, pulse width modulated. A more sophisticated and unique proportional system was developed by Hershel Toomin of Electrosolids Corporation called the Space Control. This benchmark system used two tones, pulse width and rate modulated to drive four fully proportional servos. The Space Control system was widely imitated but was priced out of reach for most rc modelers at the time of its introduction. The Space Control system was followed by a feedback proportional control arrangement, which offered proportional control of two control surfaces. This system used a fast variable mark/space transmission, which was smoothed out to a voltage swing at the receiver. Analogue servos were designed with a feedback potentiometer to follow the voltage swing. The receivers could also produce an additional voltage swing by detecting a change in the rate of the mark/space, which produced a proportional output for the second servo. While this system was highly successful, analogue servo systems had their share of problems. Attempting to coordinate more than two servos proved difficult as the control of one servo tended to effect the position of other servos. There was also a lag or delay of the servos getting up to speed as the transmitter control stick was moved, with the servos overshooting the command position, then quickly returning to the correct position.
The next major step in the evolution of servos (also in the early 1960's) is the digital proportional radio control system. Designed by Doug Spreng and Don Mathes, NASA engineers, to control space satellites, the digital control system revolutionized servo applications in rc models. They had developed a servo which would sit at the center position of its swing with a repetitive impulse of 1.5 milliseconds. The digital system produced quick, precise and accurate proportional control without the time delay or over-swing of analogue type servos. Also, digital servos, due to their speed advantage, were able to effectively control several other servos, unlike analogue devices. This was accomplished by using the transmitter to send out control pulses for each servo in sequence, similar to the firing order of spark plugs on an engine. Spreng and Mathes developed a technique by which repetitive time frames of servo pulses could be electronically adjusted, narrowing the gaps between transmitted pulses, while increasing pulses transmitted to the receiver. A twenty millisecond frame rate used by Spreng and Mathes could control up to eight servos. Since the pulses were generated with separate timing circuits, there was no interaction of servo positions, as with analogue systems. They also developed a signal pulse receiver that could count signal pulses and direct them to the appropriate servo. These digital pulse timings are still in use today.
Servos of todays market are a lot more capable units than those of just a few years ago. For example, high-voltage servos have recently come into use for the rc modeler. High-voltage (HV) servos are run off of an unregulated lithium-ploymer (LiPo) battery pack. This negates the need for a voltage regulator, which may be a source of breakdowns. LiPo packs offer the advantage of a more consistent voltage level throughout discharge as compared to nickel-cadium (Ni-Cd) packs, where voltage drops off continually from the start. LiPo packs also provide better servo performance during the entire rc model flight. The primary negative to high-voltage servos is the receivers they're coupled to might not accept the higher voltage limits of the servo. Conversely, some non high-voltage servos may outperform HV servos. Due to the popularity of HV servos and LiPo packs, a number of radio manufacturers are now making their receivers compatible with HV servos. Though digital servos offer several advantages to analogue units, a number of rc radio companies offer servos which allow you to reprogram their microprocessors. These servos allow the modeler to change the travel direction, servo speed, neutral point and end points of the units. The primary advantage of programming the servos themselves is that the rc operator will need less equipment inside the aircraft, saving weight. If you fly an rc plane with dual flaps, one servo needs to rotate clockwise, while the other counterclockwise. By using a servo programmer, each digital servo could be programmed with the proper rotation, deadband width, neutral points and end points. The two servos could then use a single receiver channel without any computer radio mixing involved. Even the most recent digital servos are capable of sending pulse adjustments of at least 300 times per second, six times faster than the latest analogue units, the only disadvantage of the digital units being greater power consumption associated with the speed of the servos.