Meteors by Radio

    The objective of this article is to share my own experience with anyone wishing to get started in the detection of Meteors by radio. In particular, my hope is that the information given here will enable a student to do a Science Fair or other special project that will lead to some measure of success rather than to disappointment. Further, such a project could become a relatively long-term hobby.

    Actually, before 'getting my feet wet' in detecting meteors by radio, I had the impression that it was just a matter of tuning a radio to some distant FM or TV station and presto -- I would hear meteors. But worse, I really did not know what a meteor sound would be like, and I really did not know what the listening conditions should be.

    Even though I have been working with radio and electronics for many years, I still found myself somewhat puzzled -- and even mistaken -- by the public-domain information that came my way. So, my hope is that my own experience with this project will enable the interested student to assemble a 'meteor by radio' observatory with a minimum of grief.

    First, What causes meteor sounds in a radio? Basically, the scenario is to imagine the meteor trail of ionized gas as a momentary reflecting surface, so that a distant radio transmitter's emission -- preferably from beyond the horizon -- briefly 'sees' a reflecting surface. The result is that the signal strength at the receiver suddenly increases. The increase might last only a fraction of a second or so.

    More specifically, imagine a situation in which the observer can just barely hear the distant radio (for TV) station; then, when a meteor trail is in the appropriate position, some of the transmitter's energy will be reflected from the meteor trail and reach the observer. If the meteor is traveling at exactly right angles to the great-circle line joining the observer to the transmitter, then the frequency of the reflected signal will be the same as the transmitted because the distance between the reflecting surface and the receiver is not changing noticeably. In other words, there is no Doppler shift. In this case, there is simply a constant time-delay (phase shift) between the directly received and the reflected signals. At most, the observer might hear a kind of echo effect.

Figure 1: Idealized diagram of a meteor event detectable by radio. The receiver detects two signals--one that has reached the receiver more or less directly following the curvature of the earth with no appreciable Doppler shifting; and one that has reached the receiver via the moving (relative to the receiver) reflective surface of the meteor's ionized trail and thus exhibiting appreciable Doppler shifting.

    However, if the meteor is traveling in such a way that a velocity component of its trajectory exists along the line joining the observer to the transmitter, then the wave reflected from the meteor tail is moving relative to the observer -- there is a Doppler shift! If this relative velocity is constant, then the Doppler shift will be constant and the observer will hear a steady tone for the duration of the event (usually less than a second). In other words, the beat-frequency is constant. If, on the other hand, the relative velocity is changing, then the observer will hear something that can be described by the word 'ping' -- something similar to striking a tuning-fork. (1) In other words, the beat-frequency is changing.

    This is probably enough theory -- how do we go about hearing these meteors?

    Of the several different techniques (2) for observing meteors by radio, yours truly is convinced that detection of a TV video carrier using an AM radio is the way to go for the entry-level observer. To make a long story short, the strategy is to have an AM radio tuned to some distant TV station's video carrier, such that the video carrier -- which produces a slight buzzing sound (this is the 60 Hz vertical synchronization pulse), usually changing somewhat in intensity due to changes in the average level of the video signal. Once we hear that buzz, then we expect the following: whenever some of that video carrier is reflected from a meteor's ionized trail towards us, then we expect to hear something like a 'ping'. If the observer wants to observe meteors at random times, then a TV station is an appropriate source of a signal because many TV stations are on almost 24 hours a day.

    So far so good, but what kind of equipment do I need? First, there must be at least one unused TV channel within your area (3), preferably in the low VHF range (channels 2 to 6) (4) because reflection efficiency falls off as frequency increases. Since we need a radio that is capable of AM detection and tunes accurately to the low VHF TV channels, it turns out that a type of radio popularly known as a 'scanner' is the most cost-effective. (5)

    Do I need a big antenna? No. In my installation, I have a dipole (6) antenna hanging inside my bedroom. This is a viable solution in my situation since the house is of non-metallic construction.

    When should I listen? According to the literature, as well as to my own observations, there are more 'pings' heard in the early morning than in the late afternoon. This statement is based on Figure 2, which is a plot of the number of hits at these two times of day.

Figure 2: Plot of the number of 'hits' as a function of day and time of day. The 10-minute listening sessions (around 8 AM and 5 PM local time) did not occur at exactly the same time from day to day; however, it is being assumed that a few minutes one way or the other should not introduce any major surprises. The number of 'hits' does not take into account differences in kinds of 'hits' -- it sufficed for my purposes to have identified an event as a 'ping' rather than anything else.

    What else do I need? Well, I'm glad you asked. An acquaintance of mine here in Halifax had been trying to do radio detection of meteors for a couple of years and was getting nowhere. In desperation, I guess, he contacted me and it soon became apparent that his radio was being 'overloaded' by several local FM stations -- he lives in a part of the city where his residence is in full view of the FM transmitter antennas just a couple of miles away. The selectivity (7) of his radio was simply not good enough, with the result that the FM stations were spilling over into what was a potentially useful TV channel. He had tried some filters between the antenna and radio but they were simply not selective enough to solve the problem. The solution was a band-pass filter with a Q (quality factor) (8) of at least 300 tuned to the TV channel of interest (channel 6). Success was immediate. The background noise-level in his receiver dropped to the point where he could begin hearing those elusive 'pings'.

    My interest in meteors began with this event. Once my acquaintance's detection problems were solved, I began to monitor meteors myself for a number of weeks, mainly to familiarize myself with the problems associated with getting started in this kind of activity, as well as becoming somewhat familiar with the kind of data that becomes available. I should mention that every once in a while, there would be a day or more during which the background noise was higher than usual, thereby no doubt masking some meteor 'hits'. The source of this additional background noise has not yet been identified. But even then, over a ten-minute period, there would usually be at least one fairly definite 'hit'.

    What can be done with the data? First, the observer could be interested in doing a survey experiment, in which observations are conducted, preferably on a daily basis, at some relatively fixed time(s) of the day. In my case, I found the early morning (around 8 AM local time--12:00 UTC) convenient, and listened for exactly 10 minutes. For comparison purposes, I did the same thing around 5 PM local time (21:00 UTC). The data is plotted as shown in Figure 2.

    Second, the observer may prefer to concentrate on the way that the 'pings' vary in frequency. This is of potential interest because it enables the observer to calculate the velocity-component of the meteor along the line joining the observer's site to the meteor. The calculation is done by estimating the frequency or pitch of the 'ping' and then applying the relationship:

    For example, a nominal 'ping' frequency of 500 Hz, and video carrier frequency of 83.25 MHz, gives a nominal radial velocity component calculated from (500/83.25 E+6) x 3 E+8 = 1800 m/s = 6500 km/h. The observer could then produce a plot showing the distribution of the number of 'pings' per unit time as a function of radial velocity component. Evidently, there is a challenge here to estimate the 'ping' frequency easily--perhaps a piano might be handy. (9)

    Conclusion: A procedure for detecting meteors by radio has been described, along with some results that are taken to be typical. The hardware is rather simple and could be the basis of a Science Fair project. Much additional information on meteors is available on the web from the American Meteor Society, easily found by your search engine (put the full name in quotation marks). (10)

    Acknowledgment: Many thanks to my colleague of many years, Mr. Robert Schultz, VE1IF, for helpful discussions and for providing a high-Q bandpass filter.

Footnotes:

(1) Actually, the observer is likely to hear a number of different sounds. Sometimes there will be relatively long 'pings'- -perhaps several seconds; then there might be tones that 'flutter'--this is quite probably due to the motion of aircraft producing constructive and destructive interference effects; etc. It takes about a week or so of daily listening to eventually distinguish between various sounds so that the short-lived 'pings' are readily recognized.

(2) There is a fair amount of information on the Web. You might try simply searching under "meteors".

(3) "Your area" is defined approximately as the circular horizon centered on your receiving site.

(4) The video-carrier frequencies are: 55.25, 61.25, 67.25, 77.25, 83.25 MHz for Channels 2 to 6 respectively. In my situation, channel 6 was the most appropriate.

(5) In my installation, I have a deluxe scanner called the ICOM R-7000, which in fact had been acquired for some other experiments. This radio is rather expensive (more than $1,000). A perfectly suitable alternate is the recently introduced Radio Shack model PRO-60 (my acquaintance is using one), priced at around $300. In general, the radio must tune to at least one uncluttered TV video carrier and it must do this in AM (Amplitude Modulated) mode.

(6) The Radio Shack antenna #42-2385 or 'rabbit ears' type #15-1827, along with a balun #15-1140 and adapter (from F type to BNC type) is all that is needed.

(7) Equivalent to the band-pass characteristics.

(8) Such a filter is in the form of a metal can long enough to contain a 1/4 wave resonant line. Construction details can be found in any recent issue of The Amateur Radio Handbook, published by ARRL, etc. The unit I supplied was a modified surplus VHF 'can' filter as used by telephone companies in their mobile radio installations; new ones cost several hundred dollars. It could also be mentioned that a prospective observer could be living in a part of the country where there are simply no usable 'blank' TV channels. In this case, perhaps a combination of high-Q band-pass and band-reject filters might bring success. At the moment, I have no further recommendations in this kind of situation. Contact your local amateur radio store for suggestions on sources of filters.

(9) The frequency or pitch of the 'ping' could be obtained electronically by means of a frequency-to-voltage converter or by a computer-controlled frequency counting software. It should also be mentioned that a 'ping' is usually not just one well-defined frequency, because the velocity component of the meteor trail is not likely to be traveling at a constant velocity relative to the observer. However, most 'pings' do exhibit some 'dominant' or 'average' frequency or pitch.

(10) Additional sites: steyaert@vvs.innet.be for the Radio Meter Observation Bulletins, and 72632.1427@compuserve.com.

SSB Mode. This is Single-Side-Band, which is usually not found on typical scanners that are capable of tuning to the video carriers of TV stations in the low VHF range; i.e., channels 2 to 6. The two significant advantages over AM are: First, the bandwidth is about half that of AM (all else being equal) so that the background noise is lower by a factor of two (corresponding to about 3 db gain in signal-to-noise ratio); Second, the original, non-Doppler-shifted carrier is not needed--the SSB modes supplies its own carrier which then produces a "beat" or heterodyne product with the Doppler-shifted signal. This feature is handy if the TV station is so far away that the direct signal from it is too weak to hear.

This mode has been tested at my location recently using the SSB mode on my ICOM R-7000 and the results are quite noticeable -- the background noise drops by about 3db on the S-meter and now more meteors, plus other "strange" things can be heard.

Are there any complications? Yes, there is one, but it is significant only for the observer who wishes to measure the line-of-sight velocity component of the meteor. The problem is associated with the frequency-stability of the SSB local oscillator in the radio. If this oscillator drifts, say by 100 Hz, then the error in the velocity component will correspond to this 100 Hz offset. Even with the ICOM R-7000, there is usually an offset of about 100 Hz (sometimes below and sometimes above the desired frequency) when I first turn on the radio, and then the offset will drift around as time goes on. The offset produces a beat--in this case, 100 Hz--which can be reduced to zero (or almost) by tuning the radio slightly one way or the other relative to the exact frequency. Evidently, this is bad news for the observer trying to measure the meteor's line-of-sight velocity component unless some way can be found of stabilizing the radio's local oscillator so that it drifts by no more than, say, 5 or 10 Hz.

For the observer who is simply counting meteors, the oscillator drift can be compensated by simply tuning the receiver frequency whenever it is needed -- surely a small price to pay for significant improvement in counting meteors.

SSB Hardware. Perhaps the simplest way of implementing SSB, short of acquiring an ICOM R-7000 or equivalent, is to build or otherwise acquire a Converter to convert the TV video carrier frequency to a frequency that is compatible with a communications type of receiver already fitted with SSB capability. Thus, a video carrier on 83.250 MHz could be converted to 28.0 MHz (called the Intermediate Frequency (IF) of the Converter). This 28 MHz is then fed to the antenna terminals of a typical HF (High Frequency) receiver that is already capable of SSB reception. It is then simply a matter of tuning the HF receiver to 28.0 MHz--perhaps a little below or above to compensate for lack of absolutely precise frequency-control--and you are in business. Frequency Converters are discussed in great detail in the Amateur Radio Handbook, for example.

On the other hand, one could try getting inside of the non-SSB scanner and access its IF, and connect this to the HF receiver--such a solution, however, calls for a steady hand and awesome good fortune because modern scanners are rather tightly packed inside, and usually there is no schematic available.

FM mode. This mode could be a simple and good approach if the observer wishes to simply detect meteors and is not trying to measure their line-of-sight velocity components. The strategy is to do the observing sufficiently far from FM transmitting antennas and/or install a sufficiently good band-pass filter between the receiving antenna and the radio, so that the strong FM signals adjacent to an empty FM channel do not enter this empty channel. One could try driving to a relatively radio-quiet spot far from a city and using the car radio. What will I hear? The "empty FM channel" is a relative concept; the channel may be "empty" in the observer's area or region but occupied somewhere outside this region. Hence, the signal from that distant station will momentarily reflect from the meteor -- for a fraction of a second or so -- and be heard. Any Doppler shift will probably not be noticed because of the relatively wide band-width of FM receivers. If the distant station is being detected in the background, then the meteor-reflected version will momentarily enhance the signal; the limiting action in the FM receiver will probably prevent any heterodyne effect between the non-shifted and the shifted signals. No doubt clever techniques could be used with the FM approach to extract any heterodynes, but such procedures are beyond the scope of this article. A word of caution. The band-pass filter mentioned above, if needed, in all probability MUST be high Q. That is, its pass-band must be quite narrow relative to the center frequency of the filter. Although I have not personally tried the FM mode, it seems to me that a filter with a Q of at least 100 would normally be needed, especially in or near any urban area. Hence, this rules out trying to make a filter using domestic-grade coaxial cable or twin-lead transmission line. The usual solution is to acquire or build a "can" filter, consisting of a metal pipe perhaps 20 cm in diameter and an inner pipe a few cm in diameter, both forming a coaxial structure that is 1/4 wavelength long, closed at one end and tunable to the frequency of the target FM signal. The construction details are outside the scope of this article. A "far-out" solution to the band-pass filter problem -- which I have not explored-- is to build or otherwise acquire a filter that uses crystals rather than coaxial structures. These days, crystals are easily available at amateur-level prices for any frequency in the FM band, and perhaps even a single crystal might suffice. The concept is always discussed in the Amateur Radio Handbooks published b y ARRL. Such a filter could be rather lossy but its compactness might justify the decrease in receiver sensitivity.

CW Mode. The letters stand for "Continuous Wave." This mode is usually associated with the use of Morse code (or equivalent) for communication. It is quite similar to the SSB mode discussed above, and is usually available on typical communications receivers. Basically, there is a BFO (Beat Frequency Oscillator) in the radio which produces a steady ac (alternating current) voltage at a frequency differing by a few hundred hertz, for example, from the incoming signal frequency. The two are usually combined in the IF (Intermediate Frequency) section of the receiver (usually centered on 455 kHz), producing a beat frequency of a few hundred hertz, for example. In typical receivers, the BFO can be adjusted by the user to produce a beat anywhere from 'zero' Hz to several kHz. If a beat frequency of 300 Hz, for example, is produced, then the band-width of the IF section can be reduced according, offering a corresponding improvement in SNR. As in the SSB mode, this mode does not require a detectable direct waves from the transmitter. As soon as a sufficiently strong and Doppler-shifted reflected wave arrives at the receiving antenna, then a beat frequency is produced and the resulting 'ping' is heard.

Back to main page

Updated Nov 1, 1999, by Ian McCarthy.