Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site lanl.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!gamma!epsilon!zeta!sabre!bellcore!decvax!genrad!panda!talcott!harvard!seismo!cmcl2!lanl!jp
From: jp@lanl.ARPA
Newsgroups: net.physics
Subject: Re: Reality of Electric Field versus the Photon Heterodyne
Message-ID: <28098@lanl.ARPA>
Date: Wed, 10-Jul-85 10:58:11 EDT
Article-I.D.: lanl.28098
Posted: Wed Jul 10 10:58:11 1985
Date-Received: Sat, 13-Jul-85 12:02:26 EDT
References: <405@ttidcb.UUCP>
Organization: Los Alamos National Laboratory
Lines: 41

> 
> After all, radio is the same as light, but slowed down. Consider the
> superhet receiver. Using a local oscillator to beat the incoming carrier
> down to a lower frequency for filtering and amplification, it is well
> known to have a signal to noise ratio several times that of a power
> detecting diode. Nobody thinks photons with radio (or even television).
> 
I think the benefit of a superhet receiver over straight diode detection
is that the signal to noise ratio is improved by reducing the bandwidth
of the spectrum coming into the detector.  The usefulness of the superhet
is that it is easier to obtain a given bandwidth at lower frequencies.
For example, a 1 Khz bandwidth at 50 kHz requires a tuned circuit with a 
Q of 50 (readily obtainable with lumped circuit elements - coils and
capacitors) whereas to obtain the same Q directly at 50 MHz requires a
Q of 50000.  Such a Q is obtainable in a cavity with dimensions on the
order of a half wavelength ( several meters).  Hence the interest in
supperhet receivers when small bandwidths relative to the frequency are
of interest.

There is an interesting technique called coherent detection that is related
to the superhet business.  If you are doing an experiment where the 
result can be modulated (say by switching the polarization of the incoming
beam in a polarization experiment) you can dramatically improve the 
signal to noise ratio by adding all the data obtained when the polarization
is one way and subtracting it when it is the other (opposite).  This is
equivalent to multiplying the incoming data by + and -  1, depending on the
beam polarization, and averaging the result.  The multiplier is in fact 
a mixer, whether done with an electronic switch or in a computer, and the
if frequency is 0 Hz.  The bandwidth obtained is equal to the reciprocal
of the integration time.  I once did such an experiment with a 1 kHz
reversal frequency and an integration time of 24 Hours.  Let's see,
that is equivalent to a 1 kHz filter with a Q of 86,400,000.       

      

The advantage of doing this at 1 kHz instead of at 0.1 Hz, for example, is
that you also beat down the 1/f noise that arise from such things as
fluctuations in beam current, etc.


Jim Potter  jp@lanl.arpa