Introduction

There are several ways in which HF propagation can be studied using a fixed frequency SSB receiver (as opposed to a swept or multi-frequency arrangement). There are broadly two types of measurement:

• Time Domain (pulse and chirped radar, digital sequence types)
• Frequency Domain (Doppler carrier measurement)

The ZL2AFP STANAG 4285 Receiver is interested only in the header which accompanies each frame of data. The transmissions are PSK, at 2400 bps, with a sub-carrier frequency of 1800 Hz. The baseband signal spans 600 to 3000 Hz, but can be adequately received with a 2.4 kHz bandwidth receiver. Each block contains 256 bits, so the frame rate is 256/2400 = 9.375 frames/sec.

We are interested only in the 80-bit header which accompanies each frame. This contains about two and a half repeats of a 31-bit Pseudo-Random Binary Sequence (P-N sequence), which in the NATO application is used to align the data before demodulation, as a means of avoiding data errors induced by propagation effects. The P-N sequence has the interesting property that when cross-correlated (compared bit-wise with a copy of the known sequence), the output is zero at all times (relative to the frame timing) except the correct one, and is a maximum when the timing matches, allowing the receiver to locate the data to within 1/256 of the correct point (about 0.4 milliseconds).

A typical Correlation from a reasonably strong signal. Note the three peaks,
the strongest in the middle, and less than average noise between the peaks.

The ZL2AFP STANAG 4285 Receiver uses the header information to study the propagation effects. The program locates and makes measurements on this P-N sequence using a very sensitive cross-correlation technique. The program is able to measure both the timing and the frequency of the transmission, and plots both the timing (propagation delay) and frequency (Doppler shift and drift). Furthermore, it includes a clever display developed by Peter Martinez G3PLX, called a Scattergram, which displays timing and frequency effects in the one image.

Guided Tour

The program consists of several panes, some controls and a menu. These are shown in the picture below:

Correlation
The tall thin pane on the left shows the performance of the Cross-Correlator, and is called the Correlation pane. You can see that the blue line has several peaks (it is more obvious while watching the live program). These peaks are surrounded by a quiet patch (if the signal is strong), while the rest of the Correlation will be noisy. There will be three peaks, and the user chooses the centre one (as shown by the red line). There is one strong peak and two weaker peaks, because the P-N sequence is repeated about two and a half times. Ideally the user should click just below the largest peak, or if ground wave is visible, right on the peak.

Correlogram
The pane at the top is a Correlogram, which depicts changes in timing over the duration of reception. The vertical axis is time, as indicated, with an approximate distance equivalent in parentheses. You can see two faint groups of lines, spaced about 13 ms apart. This is because the span of the chart is greater than the 13 ms spacing between the P-N sequence correlation peaks. The lines of interest are the ones near 0 ms. If the signal is within ground wave range, there will be a strong and very clean line which ideally should be set to zero on the Correlogram. The Correlogram moves along in time, with time marked underneath. The speed and averaging of the chart can be adjusted.

Dopplergram
The pane below the Correlogram, with a dark background, is the Dopplergram, which shows frequency variations in the received signal. In this example, recorded at about sunrise, there are several frequency products visible. At the right of this pane, you see (as the lowest line) the E-layer response, and above that the F1 and F2 responses, followed by double-hop F1 and F2 responses, plus scatter. The double-hop returns are generally only seen at sunrise, and represent rays of the signal that have been returned from the F layer, bounced off the earth, then again returned by the F layer. These are characterised by double the Doppler shift of the main F layer returns, and are weaker due to loss, particularly on reflection from the earth. The height of the Dopplergram is ±5 Hz, and the Dopplergram moves along in time, with time marked underneath. The background level of this graph varies with audio level, signal strength and the effects of receiver AGC.

Scattergram
To the right of the Dopplergram is the Scattergram, which is a unique combination of frequency and timing, and slowly changes with time. As you can see, the vertical axis of the Scattergram exactly matches that of the Dopplergram, and shows (averaged) the exact same frequency effects seen on the latter. The horizontal axis of the Scattergram is propagation time. Zero at the left and about 16 ms on the right, so you can easily read off the additional time-of-flight and frequency shift of the various products with remarkable precision (0.25 ms and 0.04 Hz resolution).

In the example picture above, there is a whole line of products stretching up in frequency as far as +3 Hz, and +10 ms. For the first time, we can recognise and measure these individual products. Note how the individual dots are roughly in a straight line and some are very sharp. The increase in frequency is caused by Doppler shift, as at sunrise the apparent height of each layer moves closer to earth with increasing refractive index (ion density). Below the Scattergram is a small box, in which the frequency offset and delay are indicated when the mouse is hovered over a point on the Scattergram.

Tune Indicator
On the far right is a Tune Indicator, which has a coarser scale and greater range than the Dopplergram and Scattergram. This allows you to tune in the station more easily. You need to adjust the tuning to within 1 Hz. If your receiver does not have 1 Hz steps, you can use the two + and - buttons between the Scattergram and the Tune Indicator. Use these to centre the lowest product on the Dopplergram. These buttons will also centre the Scattergam, but it changes more slowly. Several nearby frequencies may show stable tuning, but the correct spot will also show clear peaks on the Correlation pane.

Choice of Receiver and Antenna
It goes without saying that the receiver must be very stable, preferably 0.1 ppm or better, ideally with step resolution 1 Hz. It is best to use a good quality conventional Communications Receiver, at least at first.

With care, you can achieve good results with an SDR receiver, but since they often have clock slip, drift and limited accuracy, this is a field for experts. Not all receiving software is suitable. The V3 and other better SDRs work well with SDR# on a fast computer if adjusted carefully. You will need to use a third-party audio piping program to capture the audio.

You might think it a good idea to use a remote Internet receiver, such as a KiwiSDR, but experience has shown that signals received via the Internet have excesssive clock offset, combined with random timing variation. This is not the fault of the receiver, just a natural effect of the way data is routed across the Internet. The results will probably be unusable, so stick with a conventional synthesised receiver.

Achieving best results will strongly depend on the antenna used. If the Scattergram is particularly noisy, or the Correlogram or Correlation are weak, try a different antenna. Conventional dipoles, active Miniwhips and small Active Loops work well. Some larger loops, Loop on Ground and untuned verticals are less effective for a variety of complex reasons.

Other Controls

The Menu line at the top of the program allows the user to set various options. These are:

File/Autosave Correlogram, Dopplergram, Scattergram
Allows these products to be saved automatically. They will be placed in an appropriately named sub-folder. The Scattergrams are saved at a rate of one every 30 seconds, and are 53k BMP files. As each autosave is enabled, a checkbox appears to the left of the drop-down.

Audio Input/Select Input
Allows you to select the appropriate audio input from the receiver. This can be done while the software is operating. An error will be generated if the selected device is busy or missing, or if there is no appropriate device.

Hints:
• If the Correlation peaks are in an inconvenient place (top or bottom), reselect the Audio Input, as this restarts the Sound Device and will randomly change the Correlation timing.
• If you perform file operations on the computer while the program is running, there is a high risk that the timing will slip inadvertently.

Tools/Correlator Averaging
Sets how many frames the Correlation is averaged before display on the Correlogram. This directly affects the speed of the Correlogram. It has no effect on the Correlation display or any other products.

Tools/Display Brightness
Allows you to individually set the brightness of three products, the Correlogram, Dopplergram and Scattergram. First adjust the master Gain control so the Correlation (blue line at the left) shows peaks about 50% high, then adjust the others to suit.

Tools/Scattergram black b/g
Does what it says! Changes the Scattergram to white on a black background.

Tools/NATO, Tools/Amateur
These are alternative settings. NATO stations use 2400 bps data and 1800 Hz sub-carrier, while because Amateur equipment cannot effectively transmit such wide signals, the AMATEUR option is reduced to 2000 bps data and 1500 Hz sub-carrier. ZL2AFP offers a matching transmit program which transmits the AMATEUR version with a simple ASCII payload (call sign and locator).

Clock Offset
This slider control adjusts for differences between the transmitted 2400 Hz clock and that of the receiver sound device. If the Correlogram rises, and the Correlation peak moves up relative to the red line, move the slider to the right a little, and vice versa. Check again after a few minutes, and if necessary make another fine adjustment. Unfortunately it is not possible (even with GPS timing) to lock directly to the incoming signal.

Hint:
• If the received Clock offset is excessive (> ±50), perhaps because of a poor quality sound device, you can manually tweak the first line in the STANAG.SET file (created by the program). If you mess up, delete the file and restart the program. Once you have the correct value, don't touch the Clock Offset control, or you'll go back inside the ±50 range. Divide the ppm number you want to use by one million before changing the value. For example, for +50 ppm, use the value 0.00005; for -60 ppm, use -0.00006.

Gain
This is the Master Audio Gain control. Start with the level at 1. When tuning in, increase this so you can easily see the peaks, but once set, reduce the level so the Scattergram has very little noise (remember, the Scattergram changes slowly, so be patient). You may only just see the peaks on the Correlation. If the levels of other panes are unsuitable, use the Menu/Tools/Display Brightness dialog.

Tuning a Station

Start the program, tune in a STANAG 4285 signal (in the South Pacific try 4588.6 or 9910.6 kHz), and select the appropriate Audio Input device so that you can see noise on the Correlation. Adjust the Gain control so you can see peaks (if you aren't yet correctly tuned, advance the Gain so the Correlation pane noise is about 1/3 scale).

Very slowly tune the station in, looking for the sudden arrival of stable peaks in the Correlation, and then adjust the tuning precisely so that in addition the Tune Indicator is centred. If the signal is weak, this may be noisy, but should not hunt up and down. Make you final fine tuning to centre the lowest line on the Dopplergram. You can do this using the + and - buttons between the Scattergram and the Tune Indicator. All known STANAG stations transmit on a multiple of 100 Hz, so there's a hint for you if your receiver calibration is accurate to a few Hz.

Click just below the largest of the three peaks, as shown in the picture above. This will place the red line and select the zero time point for the Correlogram and Scattergram. The peaks might be anywhere on the Correlation pane. Even though the peaks may not be tall, they should be stable, while all around is just noise. Finally, adjust the Gain so there is enough signal to see on the Corellogram, but not enough to make the Scattergram noisy. About one third the width of the pane for the tallest peak should be plenty.

As time progresses, watch for a slope in the Correlogram, which will indicate clock drift. Correct this in small steps as outlined above, using the Clock offset control, also making peak click adjustments as necessary. Once you have the clock drift corrected, reception for several hours without drift should be possible.

Note:
• The frame timing of some stations arbitrarily changes phase occasionally, which causes the Correlogram to slip. We think this happens when traffic starts or stops. The same will happen if you perform file operations while recording the frame timing can also slip. So if you are recording Scattergrams to make a movie, keep a beady eye on the Correlation!

Hint:
• Before you make the movie, review all the frames in the Scattergram folder with Explorer set to show large icons. Weed out (delete) any frames that are noisy or have timing slip.

Making Movies

The most interesting and revealing products generated by the program are the Scattergrams. Here is an annotated example:

The closest dot to the left, centred vertically, is the E-layer return. If the station is within ground wave range, there would be a further dot, very close by, even further left. Because there is generally very little Doppler on the E-layer return, the dot is small and concise. You can see a faint trail to the right of this dot, which is believed to be caused by an intermodulation in the software when the signal is strong.

The second and third dots are F-layer returns, and are slightly fuzzy because the layer is very active. This picture was captured at sunrise, when the F layer signal is not only strong, but the apparent height of the F layer is moving towards the earth at high speed, causing positive Doppler shift of about 1 - 2 Hz. 1 Hz shift at 4.5 MHz represents a relative movement of about 235 km/hr.

On HF frequencies, often at night the F layers merge, and then separate into two layers during the day. What happens on any particular day depends on overall solar activity.

The fourth and fifth dots have twice the Doppler shift of the previous two, because they are the result of a signal refraction off the F layer, then a bounce off the earth and a second F layer refraction. The final fuzzy dot is caused by scatter from points not on the straight line between stations.

It is very obvious that all these dots are in a line. As the F layer settles well after sunrise, the dots generally line up horizontally, as there is then little Doppler shift. In the evening, the line of dots bends over in a downward direction. Sometimes the line of dots waves around during the day, indicating disturbance in the upper ionosphere.

Watching the Scattergram in real time is fascinating. It is an averaged thing, so every time it slowly repaints, it enhances static dots and tends to suppress noise. However, you can also make movies of the captured Scattergrams, and compress time by a factor of about 600:1.

The software used to make .AVI movies is included in the archive (Image_to_AVI_maker.exe). Move it into the Scattergram folder, (it will only work correctly here), run the program, select 'Add all BMP files in folder' to select all the Scattergram frames (having already deleted any dud or shifted images), then click Make AVI. Once the movie has been made, a dialog will appear asking for a file name. Enter a name and press Save. (Don't save in a different folder - the program will only work correctly in the current folder). Then select the video compression option - choose 'Cinepak Codec by Radius'.

Test play the results with Media Player. If necessary, close the program, weed out further dud frames and try again. AVI playing is not universal, and the files are large. You can make more widely recognised compressed .MP4 movies from your AVIs with a program called AVIDEMUX. Set the Output to 'MPEG4AVC(x264)' and the output format to 'MP4 muxer'.

Transmitting

Clearly there aren't STANAG 4285 stations everywhere you might need to study propagation from, nor are they always on the air when you need them. Therefore, it's a good idea to transmit your own signal, or have a friend transmit it for you. For use with Amateur transceivers, it is best to use the slightly narrower AMATEUR option in the program, and there is a simple program that will generate this signal. This is 'TX_data_CRC.exe'.It transmits the standard STANAG 4285 frames, but instead of an encrypted payload, it carries a simple ASCII text message and CRC check. Typically it sends the user callsign and locator.

The new ZL2AFP STANAG 4285 Receiver is not capable of decoding or displaying the payload message. However, the older program ZL2AFP PSKSounder has this capability (but cannot display Scattergrams). Both receiver and transmitter are in this archive. Provided both receive programs are operating with the Amateur settings, you can use both receivers at the same time.