As part of our celebration of R&WI's 50th anniversary this year, we are reprinting selected columns by perhaps our most popular writer, rotorcraft aerodynamicist Ray Prouty.
The term “fly-by-wire” leaves much to the imagination and therefore generates different reactions among pilots and designers facing the idea for the first time.
To establish a base for discussion, let’s first examine the older concept: “fly-by-iron.” Early helicopters — and some current small ones — give the pilot direct control of the main and tail rotors through a system of mechanical linkages consisting of push-pull tubes, bellcranks and cables connecting the cockpit controls to the appropriate rotor. Figure 22-1 shows how this works in the longitudinal cyclic system.
With these simple systems, the controls can be easily moved when the rotors are not turning. But when they are turning, the pilot must act against sizable dynamic and aerodynamic forces. Much early design effort went toward developing refinements such as bungees, Chinese weights and screw jacks to balance or reduce the pilot-control forces.
At some point, as helicopters got larger, the control forces with fly-by-iron systems became too large to be handled comfortably and helicopter designers followed the lead of airplane designers by using hydraulic power for the muscle — thus introducing the era of “fly-by-oil.”
Figure 22-2 shows a typical system — suitable for a medium-sized helicopter — in which the mechanical linkage from the cockpit are retained. But instead of going directly to the swashplate, they go to a servovalve on a hydraulic actuator. Moving the servovalve results in the actuator extending or retracting to move the swashplate. The actuator does all the hard work while intercepting loads coming from the rotor and reacting them against structure instead of the pilot’s hand.
Only a single hydraulic system is used, but it includes a bypass that opens if the system fails. This allows the pilot to control the aircraft by moving the body of the actuator, but with greatly increased forces. This is similar to the power steering of your car.
On most large helicopters, the forces generated by the main and tail rotors are too large for even a strong pilot to overcome without help; so at least one more hydraulic system is added for redundancy. Great care is taken to make each system completely independent so that no one failure will leave the pilot without control.
Even though it takes only a small amount of force to move the servovalve and hydraulic redundancy exist, both the FAA and the military insist on stout control systems between the cockpit and the actuators. This is because someone getting in or out of the cockpit might inadvertently put high loads on the sticks on pedals.
Another reason is that control systems occasionally get jammed by foreign objects. The system should be robust enough so that the pilot can bend a misplaced screwdriver or crush an errant flashing while doing whatever is necessary to maintain control.
The presence of a servocontrolled hydraulic system opens the door to giving the pilot some help by artificially improving stability using gyros or other devices. Although components producing mechanical inputs were originally employed, the current practice is to use devices emitting electrical signals. These signals can be sent to a stability and control augmentation system (SCAS) computer, and then to an electrically powered servovalve that moves the actuator independently of the pilot’s input.
Typical of these systems is the one used on the McDonnell Douglas AH-64A Apache. The servovalve controlled by the pilot can move the actuator from fully extended to fully contracted. However, the electrohydraulic valve controlled by the SCAS computer is limited in tis authority to 10% of full stroke on each side of the pilot’s commands. This prevents a short in a sensor or a glitch in the computer from producing a fullstroke actuator hardover. Even the effect of a 10% hardover is minimized by monitoring the SCAS and immediately nulling its command if it is detected doing something dumb.
Note that the hybrid system ahs the capability of accepting signal from any type of sensor, processing them in any manner, and then using them to improve the flying qualities. Fly-by-wire advocates sometimes claim this capability exclusively for their own systems.
One type of signal going to the computer is from the linear variable differential transducer (LVDT) attached to each cockpit control. These signals improve controllability by overriding the gyrostabilizing signals that would normally fight the pilot during maneuvers. The existence of these LVDTs leads to the possibility of true fly-by-wire.
As a matter of fact, if the Apache mechanical control system is severed in combat, a backup control system (BUCS) is automatically brought into play using the LVDT in the severed channel and giving it full authority over the appropriate actuator. By design, when the Apache is flying on BUCS, the aftected channel does not accept signals from the other sensors, so control is like a pure fly-by-iron system. Since in this emergency situation the pilot is relying on a single LVDT, wire, and computer or “single-thread system,” he is advised to quickly finish whatever he is doing and head home for repairs.
The Apache BUCS concept can be used as a basis for a full fly-by-wire system by not drawing in the push-pull tubes, bellcranks and cables found in the original design.
Now reliability becomes a major design consideration, since jeopardizing the aircraft with any single failure is not permissible. This is usually handled by using triply- or even quadruply-redundant independent systems with some sort of “voting,” so that any system which is out of step with its mates will be ignored.
The control system is said to have “fail-operational” capability; that is, its performance is not degraded by a single failure. Physical separation of the various elements is also important. These considerations are of special concern to designers of combat helicopters. There is obviously a weight saving achieved by eliminating the mechanical elements but the necessity of adding the redundant electronic elements at lea partially reduces the saving.
The recent thinking for such combat helicopters as the Army’s proposed light helicopter, the LHX, is that, for a variety of good reasons, a sidearm controller will replace the cyclic stick and perhaps the collective lever and pedals as well. Using a fly-by-iron system with a sidearm controller does not appear possible because of the detrimental effect any free play — or “slop” — would have in a system moved with small wrist motions instead of large arm-and-leg motions.
Sidearm controllers will probably have little or no motion, but be responsive to forces instead. For these reasons, with these devices, it is quite certain that fly-by-wire is the only way to go — or is it?
Fly-by-wire can do everything the designer wants. But it might also do something he doesn’t want. During a thunderstorm, the wires might act as antennas and generate transient electrical spikes in response to lightning strikes or even near misses, just as a radio produces static. Radio static might only be annoying, but control static could be downright dangerous. Lightning isn’t the only source of electromagnetic pulses. They can be deliberately generated by sophisticated devices in a combat situation.
To get around this, inventors have developed fly-by-light systems in which a coded series of light pulses traveling through an optical fiber carries the information fro the cockpit controls and other sensors to the computer and then to the actuators.