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