There are two types of gyros we use for model helicopters. Rate damping gyros and rate demand gyros.

Standard non-heading hold gyros are rate damping. What this means is that the gyro will always try to counter movement. Even when you give a rudder command a damping gyro will try to counter it. Your saving grace is that you have more "stick" control than the gyro has damping control. Before heading-hold, the trick technique was to "over-drive" the available control throw. You did this by maximizing servo wheel throw with no regard to whether or not you were binding the pushrod. This allowed you to turn up the gyro gain to get good performance from the gyro. In flight you didn't have to worry about binding up the controls because the gyro was damping, or using up, some of the available control throw. You had a good compromise of tail holding, lots of gyro gain, for backwards flight and stuff and still had max throw for a good piro rate. But even with the best piezo gyros and the fastest servos it was still a compromise and during hard 3D the tail would get away from you.

Then came Heading-Hold gyros. These are rate demand gyros. The premise upon which this gyro operates is the continual force-vector calculations performed in the microprocessor. This calculation is an angular velocity. Every time the gyro moves it moves a certain direction, the angle of movement, at a certain speed, the velocity of the movement. What the gyro does is read this information from the piezo sensor. It will then immediately feed to the servo any and all available throw required to produce the opposite angular velocity to bring the gyro back to the center position. Hence, rate demand. If the movement from the gyro is gentle then the microprocessor will demand gentle corrections. If your flying in a viscious cross wind then the microprocessor will have to demand more aggressive servo action to counter the movement. The higher end the gyro, the more controls you have over the characteristics of this demand. This is a nut-shell explanation as there are alot more factors involved.

Now then, in the real world, the calculation is not done every now and then. I'm not sure the rate of the calculations. But I think its several hundred times a second and is why the helicopter seems to hold its heading. Once you establish "center" the gyro will perform the calculations, and command the servo to maintain that center, or heading.

However, there are some things that can affect the calculations, If the processor is not a very capable one then you will get rounding errors. The higher quality processors can make the neccessary calculations, at the neccesary speed, to a higher number of significant digits. Rounding the 10th significant digit has much less impact as opposed to rounding the 4th significant digit. This is because the 10th digit number is much more accurate of a measurement than a 4th digit number. Again, I'm not sure the number of digits calculated in the different gyros. What I'm getting at is that after several hundred calculations over several seconds, the error produced from rounding will be much, much greater in the lower quality processor. You'll see this in flight. Its one of the reasons a heading-hold gyro will drift after a while if you don't touch the stick. The caluclations to bring the gyro back to center will slowly degrade and the heading position will slowly drift off. If your constanly throwing the sticks around (flying 3D), of course you'll never see the calculation drift as the heading is never constant for any length of time. But, if your an FAI pilot where you do alot of precision hovering manuvers, you'll notice the difference right away between a high quality and low quality gyro.

Another problem is temperature change. Its no secret that electronics perform differently at different temperatures. The temperature of the components will affect the angular velocity calculation. It'll affect both the speed and ability of the processor and the piezo sensor. Futaba has found a neat way to make temperature much less of a factor in their AVCS (heading hold) gyros than other brands. First off is their processer. Its a very high quality and very temperature stable over a wide range. The biggest advantage though is the piezo sensor. It is mounted on a micro machined "vibrating ring". I'm not going to explain the physics of the vibrating ring because I'm 1) not sure I fully understand it and 2) not sure I could do a good job of it. But I think they are using vibration phasing to effectively cancel out temperature as a factor.

Back to rate demand. Up until now I have talked about gyros in a hovering situation. When you give a command to the gyro in flight, The gyro ceases to function in the classic sense as a gyro.... per se'. At this point it interprets the rudder stick movement as a rate demand. It will feed command to the servo to get the angular velocity you are asking for. Little stick movements generate little angular velocity demand while large stick deflections generate large angular velocity demand. This demand is a function of the ATVs you have set-up. This is why alot of people have a hard time understanding that the rudder stick has no control over servo throw limits in the tradinonal sense. In the higher end gyros, the rudder stick is a rate demand variable trimmer. Servo throw limits are set-up "In the gyro" itself. This is also why you should not have any rudder trim or subtrim in the transmitter. Any rudder command away from center will be interpreted as a rate demand and the gyro will drift around and not hold a heading. Sometimes though a couple of clicks of trim in the morning on a not so temperature stable gyro will "recenter" the gyro to allow it to hold. Usually by the end of the flight or a few minutes in the sun will require you take the trim back out.

Because the microprocessor constanly sees the information from the piezo sensor. It is very easy to set-up the core programming to interpret the information and make the gyro perform in a rate damping sense. That is why most HHold gyros also allow you set-it up as non-heading-hold gyro.

Mike