More about signals


Arecibo has interchangeable receivers. The one used by SETI@home is called ALFA (Arecibo L-band Feed Array). It has 7 beams in a hexagonal pattern. The sensitivity of each beam is roughly a 2-D Gaussian function.

This picture shows the sensitivity pattern of the seven beams:

Each beam has two antennas, sensitive to radio waves vibrating in orthogonal directions. An antenna's direction of sensitivity is called its "polarization". There are 14 antennas. Currently, the SETI@home front end records and processes the data from each antenna separately.

Telescope motion

Although Arebico is a fixed dish, it can effectively be "pointed" at a sky location by moving the receiver.

Except for brief re-observation periods, SETI@home doesn't control the pointing of the telescope. Other experiments do so, and SETI@home records data from wherever they're pointing. There are four main modes of operation:

SETI@home's analysis reflects the mode of operation. For example, if the telescope is tracking, we don't look for Gaussians.

Time and frequency

Natural radio sources, like stars, are spread out in both frequency and time. So SETI@home looks for signals whose power is concentrated in a narrow range of frequency, time, or both.

The basic tool for this is the Fast Fourier Transform (FFT), which takes a time window of data and computes its power in a set of frequency bins. The number of bins is the number of samples in the window. This is called the "FFT length". Shorter FFT lengths give better time resolution; longer FFT lengths give better frequency resolution.

SETI@home's analysis uses FFT lengths ranging from 8 to 128K samples. For each FFT length, it divides the 107-second workunit into windows of that length, and computes the FFT of each window. This produces a matrix of power as a function of time and frequency. The program then looks for patterns in this matrix, as described below.


With radio waves, as with sound, relative movement between transmitter and receiver changes the received frequency. For example, if the transmitter is moving toward the receiver, the received frequency will be higher. This is called Doppler shift.

If the relative movement is changing - i.e. if the transmitter or receiver are in accelerated reference frames - the received frequency will change over time. Over short time periods the change will be approximately linear. The rate of this change is called the chirp rate of the signal. The Arecibo telescope is accelerated by the Earth's rotation and its orbit around the sun. Similarly, an ET transmitter located on a planet would be accelerated.

The SETI@home client takes chirp into account. There are two possible ways of doing this:

Coherent integration provides greater sensitivity because the power of a drifting signal is not smeared across adjacent bins. Of course, it uses far more CPU time because the analysis must be done separately for each chirp rate of interest. The number of chirp rates used by SETI@home varies with FFT length; for the longest FFT we check 5,409 different chirp rate.

Signal types

For each combination of FFT length and chirp rate, the SETI@home analysis detects several types of signals: Originally SETI@home detected only spikes and Gaussians; pulses and triplets were added in 2004, and autocorrelations in 2011.

Signals have various attributes:

Signal types have additional parameters; for example, Gaussians have a goodness-of-fit parameter.

Signals are assigned "scores" based on their various parameters. The score is an estimate of the likelihood of the signal occurring in noise.

Further reading

More info about signals (somewhat out of date) is here.

Next: RFI removal.

©2018 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.