observation - Radio telescope targeting

Large radio telescope have pretty good pointing accuracy:

The second critical parameter is the beam width. Beam width depends highly on frequency:

beamwidth = wave length/dish diameter

When you use several dishes in an interferometer, you can increase their effective accuracy and decrease their collective beam width.

A bit more info on the mechanics of the Lovell telescope:

A control computer calculates the required drive rates to follow each radio source. The drive motors are servo-controlled, so there is a continuous check that the correct rate has been achieved. The position of the telescope is constantly monitored and fed back to control computer to ensure that the telescope is pointing correctly.

For good tracking the pointing accuracy should be about a twentieth of the resolution. Since the resolution is proportional to the wavelength being received (see below), it follows that the pointing accuracy is more critical at shorter wavelengths. The control computer is able to correct for pointing errors caused by the telescope bowl sagging under its own weight as it moves up and down. In this way the pointing errors can be kept to about 10 arcsec.

So, servo motors and presumably, calibration are what enables this accuracy.

Further reading: the story of Jodrell Bank, and a Radio Electronics article.