The average person gets humbled pretty quickly trying to hover a helicopter for the first time. People who try to fly Remote Control helicopters have the same problem. Control is usually lost within a couple of seconds. This is a direct result of not having positive dynamic stability: the helicopter won't just sit there if you don't move the controls. Instead, constant control input is required in order to maintain a constant position and height above the ground.
Part of the problem is that the cyclic is not a position control. You don't move the cyclic 1 inch to the right to move the helicopter 1 inch to the right. The cyclic also isn't a simple rate control. You don't move the cyclic 1 inch to the right to move to the right at 1 inch per second. The cyclic is an acceleration control. You move the cyclic 1/4 inch to the right if you want to slowly accelerate to the right. You move the cyclic 1 inch to the right if you want to accelerate extremely rapidly to the right. The problem is that usually in hovering flight we don't want to deal with accelerations, we just want to deal with absolute position over the ground.
If the cyclic were a position control, the pilot could look out the window, decide he wanted to be 1 inch to the right, move the cyclic 1 inch to the right and be done with it. The actual process the pilot needs to go through is more like this:
All this aside, the basic technique to cyclic control is fairly easy. If you have read the previous section on Attitude Flying you understand how the pilot uses the horizon to maintain a specific airspeed. This same mechanism is used in a hover. The pilot has to learn what pitch attitude will give him zero airspeed flight, and then simply maintains that attitude. If the helicopter drifts forward or backward, slight attitude changes are made to gain and lose forward or rearward speed until the desired position over the ground is achieved.
Besides pitch attitude, the pilot needs to also control the helicopter laterally with the cyclic control. He does this by banking the aircraft. In forward flight, banking the aircraft causes the helicopter to turn. In hovering flight, banking the aircraft simply translates the helicopter laterally. By maintaining a no-bank attitude, the pilot will maintain a zero drift rate.
Most collective controls also effect engine power output. This is because the change in pitch angle of the main rotor almost always requires a change in engine power to maintain a constant main rotor RPM. There are three basic configurations:
The most basic is when there is not a connection. In this case, the pilot is required to roll the throttle on and off while making collective pitch adjustments. This takes quite a bit of practice to become proficient.
The next step up is a "correlated" collective in which the collective is mechanically connected to the engine throttle such that moving the collective also moves the throttle. Correlators can be very effective if the engine power output can be accurately matched to collective position. Some correlators are not well matched, or are misrigged, and can end up being worse than no correlator at all, because the pilot still ends up having to make throttle adjustments, but has to compensate for the amount of throttle that the correlator is adjusting.
The final way that collective and engine can be connected is with a governor. This is a device which actively tries to maintain rotor RPM at a preset value. A governor would typically sense the change in main rotor RPM caused by a collective pitch adjustment, and would increase or decrease throttle as required to maintain desired RPM. Some governors sense collective movement and start adding or removing fuel right away in anticipation of the effect the movement will have on RPM. This is called a "compensator" and is typical in turbine engines which are slow to accelerate and decelerate (and therefore would experience large RPM fluctuations without the compensator).
All this talk about the collective being connected to the engine is because the throttle is typically mounted on the end of the collective control (but is sometimes on the overhead panel on machines with governors). You can think of the two controls together as the "power" control, or you can think of them seperately. Which makes more sense depends on the type of helicopter you are flying, and what maneuver you are trying to perform.
Most single main rotor helicopters require the pilot to manipulate the pedals during torque changes. Failure to do so will result in yawing the aircraft. Exceptions are helicopters such as the MDHC Notar helicopters which will automatically increase anti-torque force by the nature of the way the anti-torque force is developed, and also helicopters with Stability Augmentation Systems which incorporate a yaw damper, which is typically a gyroscopic driven device which automatically changes the tail rotor pitch in response to the aircraft yawing (it's a simple auto-pilot).
During hovering flight, the pedals are a rate device. Pushing on the left or right pedal a certain amount will cause the helicopter to yaw right or left at a particular rate. The more you push the pedal, the faster the helicopter will yaw.
On a calm day, the pedals will hardly move in a hover except to counter torque changes, or to yaw the aircraft on purpose. On a windy day however, the pedals will be in constant motion as tail rotor thrust varies due to wind and main rotor downwash effects.
On most helicopters, the tail rotor is not mounted at the vertical CG. Changes in tail rotor thrust will cause a right or left rolling tendency depending on which pedal is being pushed, and whether the tail rotor is mounted above or below the CG. The rolling tendency has to be countered by the pilot moving the cyclic control. This cross coupling adds to the pilot's workload.