IARU Region 2 Monitoring System

Home Report Documents Tech Stuff Newsletters Contact the MS IARU Region 2


WHY  MEASURE  FREQUENCIES  ACCURATELY?


There are three good reasons for accurately measuring the radio frequency of an intruder signal.
  1. To be able to distinguish "your" intruder from other intruders, heard by other monitors, on nearby frequencies;

  2. To determine the technical characteristics of the intruding signal, characteristics such as frequency drift and frequency error or offset from a nominal frequency, which can often help in identifying the station;

  3. To simplify the work of the national telecommunications administration.
The first two reasons are what I call short-term, operational reasons, and they have immediate importance for the day-to-day operations of the Monitoring System. By themselves, they are good reasons, and I believe that you should give them your attention. In addition, both of them contribute to the longer-term goal, the third reason, of making the work easier for the national telecommunications administration. Anything that will help the national administration to identify and locate an intruder, and to get him out of the Amateur bands, is a Good Thing.

Spectrogram of Russian single-letter beacons on 40 m


For a good example of the importance of being able to separate signals which are on almost the same frequency, consider the spectrogram above which shows three Russian beacon signals on about 7039 kHz with spacings of 0.1 kHz. If it weren't for the fact that each of these beacons identifies itself with a unique letter in Morse code, it would be very difficult to tell them apart without using some measuring technique that provided better than 100 hertz frequency accuracy. The three beacons in this spectrogram, heard in eastern Canada on a late-winter evening, are reported to be :


Kaliningrad, Russia ("P") 7038.80 kHz
Arkhangelsk, Russia ("S") 7038.90
Moscow, Russia ("C") 7039.00

In this example, the receiver was in CW mode and tuned to "7039.00" kHz, so the beacon on that frequency is heard and displayed at 700 Hz on the spectrum analyser, while the other two appear at 500 and 600 Hz. The receiver bandwidth was set to 500 Hz (notice the band of background noise about 600 Hz wide) to prevent nearby strong signals from being displayed. At the time this recording was made, there was also a weak 4th signal on 7038.78 kHz, about 480 Hz on the display, which could not be identified (by the way, see how easy it is to measure frequencies using a spectrum analyser? All we need now is a calibrator to allow us to avoid our receiver's internal errors). 


Analog versus Digital Methods

There are two major factors in making accurate frequency measurements. The first is technique: using good methods that will produce reliable and repeatable results with minimum errors. The second factor is technology: having a frequency calibrator that can be easily adjusted to produce "markers" at precise intervals of frequency, as well as some means to compare the frequency of the intruder with the frequency of the nearest marker.

It does not matter whether your receiver has analog or digital tuning and display. It is important only that you know how to use your receiver's display to make an accurate estimate of the frequency.  Even though some modern digital receivers display the "frequency" with an apparent precision of 10 or even 1 Hz, it is certain that they are not truly accurate to this degree and the receiver should always be calibrated against an accurate, external frequency standard, such as station LOL, WWV, or CHU, before the display can be trusted. My own HF transceiver, a Yaesu FT-990, has a display precision of 0.01 kHz (10 Hz) and through experience I have found a normal temperature-dependent error of between -5 and -10 Hz, so if the receiver is tuned to "14010.00" kHz it is actually somewhere between 14009.990 and 14009.995 kHz. Your receiver could be similar. 


So how did I get such a small measurement error? I used an old analog receiver (Drake R4-C) whose dial could be read to about 1 kHz (if you were really careful!), an old Heathkit audio oscillator and an even older Heathkit oscilloscope, a frequency counter (Ramsey CT-70) that I had built from a kit (to read the frequency of the audio oscillator), and a home-built crystal-controlled 100 kHz frequency calibrator with a multivibrator designed to divide the 100 kHz harmonics down to 1 kHz. The audio oscillator was used to produce a "Lissajous" pattern on the 'scope' in order to measure the frequency difference (beat note) between the unknown signal in the receiver and the nearest 1 kHz harmonic from the calibrator. What was usually the biggest source of error in my measurement method? Probably reading the frequency of the audio oscillator using the counter. In theory these readings are easy to do but in practice it can be difficult as amplitude and phase changes in the intruder's signal, caused by propagation effects, affect the stability of the Lissajous pattern on the 'scope', in turn making it difficult to know when the audio oscillator has been set at the right frequency. 


How  Accurate?

How accurately do we need to measure radio frequencies? All MS Coordinators would be quite happy if all the monitors could reliably and consistently measure frequencies with an accuracy of 100 Hz. With a little care and the right equipment, 100 Hz accuracy is easy to achieve using either analog or digital equipment. Using the same techniques and a little more care, it is relatively easy to go one step further and achieve 10 hertz accuracy. 

But two Hz. If I can do it, then you can, too! And it doesn't matter whether your receiver is analog or digital, except that when using analog equipment you have to do a little more work to get there. Having said that, it should be recognized that it is sometimes impossible to achieve accuracies of better than 5 Hz on the HF bands simply because the propagation medium won't allow it. This is particularly true when the radio path between the intruder and your location is long and crosses a polar region, or even sometimes when it crosses a sunrise or sunset transition. In these cases, it is probable that the moving, changing ionosphere will cause a Doppler shift in the intruder's radio frequency. There are other considerations, too, and we will discuss them in a little while. 


Return to previous article

Go to the next article

Return to Techstuff index page

Home