Thats what we call servo creep. First lets make sure everything is mechanically correct. Plug the rudder servo straight into the RX rudder port, bypassing the gyro all together. Now check that the servo arm (horn) is 90º to the tail boom as Marius illustrated without using any trim. You can unscrew the arm and replace it as close to 90º as you can. That is then mechanically set up.
Now the hard part. With a computer radio you can dial out that creeping with rudder subtrim in the menu, but yours isn't a computer radio??? So i will need to do some research to figure this out. I'll get back to you.
- - - Updated - - -
Looks like it should be ok. Just try to spool it up and adjust if necessary with the rudder trim tab on the front of your radio. Heres some light reading for you.
Gyro creep is a problem that generally occurs in cheaper heading hold gyros, although it can occur in more expensive models, especially if the transmitter is not set up properly. It is characterized by the tail rotor pitch servo creeping slowly to one end of its travel while the helicopter is left motionless on the bench, or the helicopter rotating slowly when in the air, even though no rudder inputs are made.
The cause is one of two things:
The gyro may sense that it is actually rotating and try to apply a corrective action, even though it is actually stationary. It may think it is actually rotating because the sensor has poor temperature compensation (the sensors warm up in use), or possibly the helicopter was moved during the gyro's initialization phase or the gyro's calibration has been otherwise disturbed.
The signal the gyro receives from the transmitter is not what it expects when the rudder control is left centered. This may be due to gyro misconfiguration, or due to trim, subtrim or revolution mix being used on the transmitter.
With cheaper gyros, such as the Align RCE500, the only available corrective action is to wait for the gyro to warm up to operating temperature, and then apply sufficient trim or subtrim to stop the gyro creeping. More expensive gyros usually read the rudder neutral signal when the power is first applied, and power cycling should cure the problem.
Note:
If the nose only creeps in the air, then chances are the gyro is subject to too much vibration, or the mechanical tail setup is not good enough. This is true especially if the tail holds well in rate mode, but not in heading hold mode.
If the helicopter spins rapidly, the the gyro sense is wrong.
I thing this is you
If the tail servo appears to 'stick' to one end of travel or the other only when the helicopter is turned by hand or rudder control is used when the helicopter is on the bench, then your heading hold gyro is working correctly.