Satgen490  Antennas and Azimuth 3  by GM4IHJ  (BID SGEN490) 1998-08-15

Once you have marked out a permanent "north heading mark" on your
fence or the ground near your antenna mast (Sgen488 and 489 refer).
You can easily align and set up a fixed antenna on any azimuth you
chose. However, it is less easy to calibrate an antenna and rotor
combination , so that it can be turned accurately to any azimuth.

The problem lies with the simple first order servos fitted in almost
all rotators sold to radio amateurs. Bigger , more expensive rotators,
rarely means greater pointing accuracy. You simply pay more for
greater strength to handle bigger antennas. Which can result in paying
more, to do less, when say, a big narrow beam UHF or S band antenna
array , is positioned by a servo which cannot meet the accuracy
required by the narrow beam of the satellite tracking antenna.

You can check how your rotator will perform in the following way:-
1. Check its accuracy running in to north from 90 east
2. Repeat the check by running in to north from 270 west
You may be surprised at the difference
3. Set up for the best compromise you can get between these two tests.
Then 
4. Repeat tests 1 and 2 above running in to south not north.
Again you may have to compromise. Taking note of the possibility that
say - much of your satellite work will feature one direction or the
other depending on where your favourite DX is located and whether you
generally take north going or south going orbits.
5. Next check your antenna rotator combination using increasingly
smaller steps eg run in from 45deg, 30 degs,20 degs,10 degs etc. to
discover at which size of step the stiction (static friction) in your
rotator system begins to produce irregular results . You will then
know the smallest step shifts your system will take whilst retaining
some degree of accuracy.

Remember most azimuth rotators offset a potentiometer , producing a
voltage proportional to the offset, to drive the motor. Small offsets
simply cannot overcome the inherent stiction in the drive gears. Too
little stiction is even worse producing overshooting beyond the
designated azimuth. Other rotators use a switch which feeds power to a
motor whilst you watch a voltmeter display where, volts are
proportional to degrees of bearing. These systems are only as accurate
as the initial setting up of the resistance chain feeding the
potentiometer producing the variable voltage. So follow the makers
setting up instructions to the letter, do not skip this. Used properly
this second type can be generally more accurate than the first type.
But it has the disadvantage that you must keep the switch pressed and
watch the votmeter when you need all the hands and eyes you have to
follow signal doppler and carry out a QSO. 

Computer control of your rotator gets round some of this sensory
overload. But it is expensive, requires accurate keplerian elements
and can give jumpy tracking if the computed shifts are only just big
enough to overcome stiction.