Drift scan

A drift scan is the simplest observation there is: point the antenna at a fixed altitude and azimuth, usually due south on the meridian, and let the sky move. The Earth turns 360.9856° per sidereal day, so a source on the celestial equator drifts across your beam at about a quarter of a degree per minute. Record spectra continuously and the Galactic plane rises out of the noise as it comes through, then fades again.

It needs no motor, no tracking and no pointing model, which is why the $150 receiver and the first-light session both use it.

How long a source stays in the beam

The time a point on the sky takes to cross your half-power beamwidth is the beamwidth divided by the drift rate. Away from the equator the sky moves more slowly in angle (it’s travelling on a smaller circle), so divide by cos δ.

Worked example: a 1.8 m dish at 1420 MHz has a half-power beamwidth of about 8.2°. At declination 0°, a point source takes 8.2 × 239.3 ≈ 1,960 s, about 33 minutes to cross it. A horn with a 20 to 25 degree beam keeps a given patch in view for well over an hour. That time is your integration budget per pointing, so check it against what the radiometer equation says you need.

Planning a session

1. Decide which galactic longitudes you want.
2. Use the Galactic-plane transit planner to find when they cross your meridian and how high they get. Below about 20° altitude you pick up more ground noise and RFI.
3. Set the antenna to that altitude on the meridian before the first one arrives, start recording, and leave it.
4. Take a reference spectrum away from the plane, or use frequency switching, so you can remove the bandpass.

The sky repeats four minutes earlier each night, so a drift scan you liked can be run again tomorrow at nearly the same time.