With an angle of 30 degrees at the half power points a good compromise is to fix it at 25 degrees elevation. Remember ... LEO's spend 70% of their time below 65%
elevation.
But there is no need to compromise for a LEO.
This is because 30% of the time they are below 10 degrees. So setting an antenna at 25 degrees is going to degrade 30% of all pass times by as much as 3 dB. There is no need to give up that 3 dB.
It is best to set the antenna to a tilt angle of 15 degrees so that the maximum lobe is still max at the horizon where the satellites are the fartherest away (3000km) and you need the gain the most. This makes the antenna pattern down only 3 dB down at 45 degrees where the satellite is 6 dB closer (1500km), so there is still net gain. The first null of such an antenna is then at 75 degrees, which a satellite will enter for less than 1% of all pass times, not worth worrying about.
If that 1% is that important, then switch to a rubber duck, since the signal at that point is 10 dB stronger than it was on the horizon, and again, nothing is lost.
In summary, Satellite stations need gain on the horizon, and do not need much gain above 30 degrees where the satelite is so much closer. Sacrificing gain on the horizon leads to a much much more significant loss (30%) in access times than sacrificing gain overhead (1% loss of access time).
That said, if your antenna cannot see the horizon, due to trees, or buildings, then by all means, don't waste gain there. In that case, a 3 dB smaller beam, tilted slightly higher will have plenty of signal, though you will have much less access times.
see the table: http://www.ew.usna.edu/~bruninga/rotator1.html
Bob WB4APR
More than 6 elements is not! recommended for fixed elevation installations, the horizontal beamwidth will be too sharp. Less elements = broader beamwidth ... (and lower gain) ... "horses for courses" Mike.
73 John la2qaa@amsat.org