Our experiences

A brief history

At the beginning, the observations were only for trivial curiosity. We were curious about "listen" this noise that some people said that we sometimes could receive by a radio tuned to Graves.

By the utility Spectran 2 (sw, open source) for radio-amateurs we saw the transit and heard the sound of meteors ehoes like a "golf-club" in the waterfall area.

– A Spectran 2 screenshot -


Once some mistakes were solved, we realized that we were detecting a huge numberes of echoes. So the first thought was to "measure" these echoes by recording the data, going beyond the mere listening.

Pointing to a continuous recording, we discarded the idea of a personal computer use (always on). So we decided to analyze the sound generated by the radio receiver with a microprocessor.

In the first version the circuit drew up the audio signal from the receiver simply analyzing it in amplitude.

- Data from a week: each circle represents a meteor echo, the size is proportional to the duration of the echo, each column represents an hour -

Recording all the echoes regardless of their origin let us to compute as meteoric echoes any transient event, not even meteoric: both those revealed by the radio receiver (lightning and electrostatic discharge) that those present on the power net (electric motors, neon lights etc.).

- Graphical output containing also not meteoric transients (lightning, little truck,
electric motors and neon lights). -


In the second version of Rambo, in order to limit this drawback, we have chosen a frequency to voltage converter to try to reveal those transients using a frequency threshold equal to or higher than 1000 Hz.

The purpose was to restrict the analysis to those transients characterized by a frequency component that is typical of the meteoric echoes, ie the one of the signals at 1000 Hz detected by the SSB sideband (in amplitude modulation AM) and obtained due to the 1 kHz tune shift upper the Graves carrier.


To further refine this technique, we had created a third version of Rambo in which, in place of the frequency-to-voltage converter integrated circuit, we used the MSGEQ7 integrated circuit. This IC is substantially an analog spectrum analyser that examines seven frequencies, one of which is precisely the 1000 Hz.


- Analisys with MSGEQ7 -

- The seven frequencies analised by the MSGEQ7 -

The rate between the amplitude of the signal at 1 KHz and the amplitude at other frequencies was is the new value examined by Rambo 3.

On board, RAMBO 3 had also a GPS receiver useful to assign time and date at each echo and a mini SD card, in which we recorded the datas in a csv file.


With the fourth version we introduced one more improvement, ever in order to avoid the electric impulses with give us “false positives”.
We have used a second radio, tuned to a different frequency of the Graves one. The audio output of this receiver will detect only transient, not meteoric generated. So, every time the second radio registers a peak, the data coming from the first receiver is discarded.


- Version with two receivers -

- Output data from Rambo 4.0 in a week -


With this version we have tried to improve this technique but the big difference in sensitivity between the two our radio equipments hasn’t led us to the expected results.


With the sixth version we have finally reached a different and final design. Now Arduino no longer runs only as an amplitude meter (V meter), but simultaneously also acts as a frequency counter.

Running both functions at the same time allows to perfectly discriminate between meteor and electrical transients. With this version we now have a system that does not suffer more 'of "false positives". In addition, we have improved the circuit that allows the measurement of the signal amplitude, a measure that jointed with the echo duration time provides us with the information about the meteor mass.

A further innovation of Rambo 6 is the choice of Arduino Yun, apparatus that also allows to send daily the data of the day by an email attachment. In this way the data can be read and processed anywhere with an internet connection.


The sound card created for Rambo 2 has been revised. Now is a printed circuit boards and so now the project has become easily reproducible.

In this card there are also a potentiometer for the audio level adjusting and seven LEDs to monitor the Arduino computation live.

By the combination between Arduino Yun, the sound card, a radio receiver and an antenna you can realized one or more observatory similar to Rambo to share the results everywhere.

- RAMBO hardware - (Receiver, Power supply, Rambo) -

- Microprocessor with group filters (1) -

- Microprocessor with group filters (2) -

- Receiver -

The antenna

The Rambo purpose is to measure events that occur in a well defined area of the sky at the 100 km altitude (share of ignition meteors) and along the line that connects the transmitter to the receiver (Dijon-Bologna).

- In red the gain of an antenna -

We had tried to use different antennas. But, what is the optimal antenna for this purpose?

Our experiments led us to choose a directional antenna: 20809 Tonna Yagi 9 elements and 13 dB of gain.


- Antenna Yagi Tonna 20809 -

In fact, a low directivity antenna affects a very large area of sky, too much higher than that in witch the phenomenon of meteor scatter occurs for ours purposes. In addition, its low directivity involves a low gain that penalizes the reception of low intensity echoes.

- From June 17 to 19 we have received with an omnidirectional antenna -

If the antenna is too much directive (Tonna Yagi 20817 to 11 elements and 15.3 dB of gain), we have the advantage of increasing the gain (though that advantage is limited by the radio and the Arduino process sensitivity). But it narrows the "field of view".

So, it could be possible that the area investigated by an antenna with very high directivity became smaller than the area in which the meteor scatter phenomenon occurs.

- On June 30 and July 1 we received with a very high directivity -

At the end of our experiments we think that an antenna with a good directivity (7-8 elements) is the best solution.

The "field of view" with this antenna is of the order of some tens of square degrees (about 20 °).


The Graves transmitter broadcasts with circular polarization, therefore the best antenna for receive could be a circular polarized antenna.

With linearly polarized receiving antenna we need to knon what is the most probable carrier polarization direction the reflection on to the ionization cylinder.

The reflected carrier will be perpendicular to the direction of the ionization cylinder. Therefore the question is: which is the most likely direction took by the meteors paths. More horizontal or more vertical?

Our experience is that the stronger field is receiving in vertical polarization.

- On July 2, receiving antenna with horizontal polarization -

So it follows, as indeed it as was expected, that meteors travel most likely with ground parallel paths.