Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!think.com!mintaka!gaak!jpexg
From: jpexg@gaak.lcs.mit.edu (John Purbrick)
Newsgroups: comp.robotics
Subject: Re: Getting narrow-beam range data
Message-ID: <1991Jun9.051512.29194@mintaka.lcs.mit.edu>
Date: 9 Jun 91 05:15:12 GMT
References: <1291@sousa.ltn.dec.com>
Sender: news@mintaka.lcs.mit.edu
Organization: MIT Laboratory for Computer Science
Lines: 30

>In article <160412.13057@timbuk.cray.com>, 
>	kilian@cray.com (Alan Kilian) writes...
>> I can understand that the beam divergence could be proportional to the
>> transducer width but I don't see how a larger transducer can produce a
>> narrower beam. Is this really true?
>> Also how does an array of transducers produce a narrow beam?

Alright, here's a rough & ready explanation:

If you have a single transducer, at some distance it is essentially a point
source. Sound will radiate and be returned in all directions, hence your 
transducer is geometrically omnidirectional, with only the characteristics
of the transducer to give you limited directionality.

If you have a large transducer, or a lot of small ones in parallel, you 
generate a nearly flat wavefront advancing into space (like those physics
problems involving flat capacitors "big enough that we can ignore edge 
effects"). 

Essentially, if the transducer is large compared with the distance to the
target, you will get the desired parallel wavefronts; if not, you get
waves on a circular pattern. 

A team at the University of Nottingham, England, invented a phased-array 
sonar system; once the target was located they would "steer" their beam
across its surface by firing the transducers at the right times so that at the
desired point on the target all the beams would be in phase, while at all other
points the beams would cancel at least partly.

John Purbrick