Foundations of Amateur Radio

For some time now I've been discussing the potential of weak signal propagation and its ability to create a live map from the data that your own station transmits. There are several systems in place that show a map of where and when your station was heard in the past little while. Using 200 milliwatts, I've been transmitting a WSPR or Weak Signal Propagation Reporter beacon on 10m for the past few weeks.

At the moment, the furthest away my beacon has been heard is 13.612 km away. That's an 0.2 Watt signal heard on the other side of the planet, on 10m. As distance goes, it's a third of the way around the globe. I must point out that there's no way of knowing if this signal travelled the short path or the long path.

If you've heard those terms, short and long path but were wondering what they mean, here's how it works. If I get on my bike at my QTH in Perth in VK6 and peddle East until I hit Sydney, I'll have crossed Australia, taken about 184 hours and travelled about 3.746 km. That's the short path. If I head West instead and start swimming, visit Cape Town, Buenos Aires and Auckland along the way, I'll have travelled much further, still made it to Sydney, but taken the long path.

Radio waves can do the same. Depending on propagation, a signal might take either the shortest route, or go in the opposite direction and take the longest route along the great circle between two stations.

I'm mentioning this because WSPR doesn't tell you if it's one or the other and if you're using a vertical, it could be either. Even directional antennas might receive a signal from unexpected directions.

Using one of the mapping tools, wspr.live, I extracted all the sightings of my callsign and all the reports that I'd made from my receiver. It shows that my newest transmitter has now been heard by 11 stations across three continents.

Those numbers are just the beginning. I wanted to see on the map where these stations were, so, during the week I built a proof of concept world map that I used to visually show the four character Maidenhead grid squares that my station was heard in. I also had a look to see which stations I'd heard over the years and where they were. In all, 771 different stations are in my log, either as a receiver or a transmitter.

N4WQH heard me on 40m, 18.832 km away when I was using 5 Watts. My station has heard, or has been heard across 331 different grid squares. There's reports across some remote parts of Australia, Japan, India, South Africa, Europe, the United States, several across the Pacific and even a few in Antarctica.

I wondered how many of the world's grid squares have actually been activated and which station was heard the furthest and how much power was used. Those numbers will have to wait for another day. I initially started using wspr.live which has a neat way of allowing you to embed an SQL query as part of the URL to download the output.

I was getting some interesting results, so I thought, rather than hammer this lovely resource with my questions, I should download the raw data instead. So I did. Well, I am. Still. It's big.

As of today, there's 166 files, taking up 60 GB of compressed data, with over 3.5 billion reports.

The first spot in that data goes to N8FQ who heard WB3ANQ on Monday, the 17th of March, 2008 across 911 km on the 30m band transmitting with 28 dBm, or about 630 mW, reporting a signal to noise ratio of 1 dB.

Using preliminary data to get started I mapped all the activated squares, each shown as a red box and saw that my entire map was red. At that point I figured that either I've got a bug in my code, or something else is going on.

To give some context before I share what I found, a Maidenhead locator consists of a combination of letters and numbers. For four letter grid squares, there's a grand total of 32.400 different combinations, running from AA00 to RR99. They're 2 degrees wide and 1 degree high and their width depends on where on the planet they are. At the equator it's about 222 km wide and 111 km tall, at the North and South pole, it's 0 km wide. If you travel between two squares, you might have to move a meter, or the entire width of a grid square.

Among the report, I found stations who had activated more than one square. That's fair enough, you can move your station and start making noise where ever you like. I found stations with activations across more than a thousand different squares. Before I start pointing the finger, I will mention that if you attach a WSPR beacon to an aircraft, or a balloon, you can legitimately activate plenty of squares.

When you set-up a WSPR transmitter, you're required to manually enter the locator and mistakes happen. There's plenty of records with invalid Maidenhead locators, typically shown instead is a callsign. Then there are stations that pick desirable locators. This manual entry is also true for the power level and even the callsign, so I'm not outing these stations here, since it's entirely possible that the callsign shown doesn't actually relate to the transmitter or the licensed amateur.

What does all this mean?

It means that the information in the WSPR database cannot be trusted. I suspect it also means that the data used to lodge FT8 contacts across the planet can probably also not be trusted. It means that any propagation data you're deriving is likely contaminated by misreporting, deliberately or not.

As a community, if we want to use this for actual measurements, we'll have to figure out how to make this a trusted resource, because the information that WSPR can bring to propagation is in my opinion extremely valuable.

I would love to hear your thoughts on how we might fix it.

I'm Onno VK6FLAB