The Weak Signal Propagation Reporter (WSPR), to give it its proper title, was designed as a way of getting real-time propagation information by using low-power digital signals. A transmitter radiates a short digital signal using a power of a watt or two, then listens to see if any similar signals are coming in. If they are, it decodes them and puts the information on a website. If anyone hears you, they do the same. Then the process repeats itself, all controlled by a PC. The signals you receive are listed by the controlling program and you can see them on your PC at any time. If you want to see if anyone has received you, then you have to go on the Web and find out. It's a sort of automatic DX machine and can be left running, unattended, for many hours at a time. What you do with the information is up to you. At first, it's fascinating to see just how many stations, world-wide, are receiving your pipsqueak transmissions, then more serious applications may occur to you. Some amateurs leave their WSPR stations running for months at a time and log all the results so they can see how propagation varies hour by hour and day by day. Others use them to catch very short duration DX openings that you will never hear by just listening casually. Yet others use them for comparing aerials - verticals vs. horizontals, loops vs dipoles, and so on. It's fun to see just how low a power you can get away with to be heard on 80m in the USA, etc, etc. Over the last few years I've done all of these things; my current project is to leave it running during daylight hours on 10m to see how often that band is open outside Europe even in these low-sunspot days. The answer is - most days, something I would never have suspected just by casually tuning over the band now and again.
One of the rather nice things about it is that you don't need QRO or huge aerials. In fact, the use of powers over 5w is actively discouraged, because that just jams the other signals in the rather narrow band allocated to WSPR. However, that power is the actual radiated power, so you can put 10w into a short vertical whip that's only 10% efficient and get away with it. It's sometimes a bit difficult to turn down the power so low in the normal 100w rig and anyway most amateurs don't want to commit their main rigs to WSPR for long periods so several specially-designed QRP rigs have appeared on the market. What I use is my Yaesu FT-817, a little portable rig that doesn't see much use except when I go out in the summer. It's maximum CW power is only 4w and it can be turned down to less than 1w quite easily. For powers less than 1w I use a small attenuator that can get me down to 1mW into the aerial. Don't ask me how you measure such small powers - that's an article in its own right! So those with small aerials and low-powered rigs are not at a disadvantage; WSPR is something you can play with no matter how modest your setup.
So what do you have to do? Well, read up on it first, - G4ILO's pages are a good start - http://www.g4ilo.com/wspr.html. Then you need a PC (or laptop), of course, on which you can install the WSPR program, downloaded either from G4ILO or direct from the WSPR homepage (see http://www.physics.princeton.edu/pulsar/K1JT/wspr.html). The bit where you have to do some work is the interface between your PC and your rig. To transmit, your PC puts out an audio tone, around 1500 Hz, generated by the WSPR program which has to be sent to the rig as a modulating tone; the rig operates in USB SSB mode. A very crude way of doing this is to use the audio to drive a small loudspeaker which is placed near the mike on the rig, but a much better method is to input it to the rig via its external audio drive auxiliary pin (usually found on a DIN socket at the back). That way you can accurately control the audio level and hence power output. The rig, of course, also has to be switched to "transmit" at the same time but the WSPR program controls Tx/Rx switching via one of the COM ports on the PC. For the reverse process, delivering received signals to the PC, receiver audio is taken to the PC audio board. There is usually a pin on the auxiliary DIN giving a fixed-level audio signal so you don't have to twiddle the audio gain control. The WSPR program takes care of all the rest. Details of how to connect up the various interface systems are given on G4ILO's site. I recommend using an interface that completely isolates the PC from the rig, by using transformers or opto-isolators, to prevent hum and other spurious modes appearing. It's quite easy to make your own, but commercial versions are advertised in "Radcom".
There are two other matters requiring attention, the first being the frequency stability of your rig. WSPR is a narrow-band system and requires that your rig can hold a frequency within about 2 Hz for 2 minutes. Modern rigs can do this fairly easily even at 10m if they're allowed to warm up for an hour or so, but if you decide to go for 6m or higher it's worth checking. Most VHF/UHF rigs cannot do it at 2m and above and require specialised frequency control, which is why you will not hear much about WSPR at VHF. The second is accurate timing. WSPR transmits for two minutes at a time and the start of each transmission must be exactly at an even minute of time, within 2 seconds. This is absolute time, locked to GMT, and no PC I know of can maintain time this accurately for more than a day or so. You can get over this by using the PC's "Internet Time" feature (right-click on the bottom-right time block). Alternatively, there are several "Internet Time" programs (free!) to be found on the Web - I use one called "Dimension 4" that gives more facilities than the PC built-in one.
Having done all this what can you get? If you consider a WSPR "contact" as a QSO (ARRL don't!) then you can do a "WAC" in about ten minutes on 30m, and DXCC in a couple of days. Or leave it running all night on 80m and see how many W's you've "worked" by morning. You can leave it running on different bands each day and see exactly when each band opens and closes to different areas. I've found it quite interesting to run 10w to an 8ft mobile whip on 160m and see how far it gets round the UK. I was a bit shattered to find one night that it was being heard in DL, SM and OK - the radiated power could not have been more than 100 mW.
Doing aerial comparisons it's a bit boring to have to sit there manually changing over aerials every few minutes so I built a little automatic aerial changeover switch that switches between two aerials on alternate transmissions. That way I could couple up two aerials, each with its own ATU, and see how they differed in the DX they captured. In particular, I spent some months earlier this year comparing a vertical quarter-wave against a horizontal half-wave on 80m and found the horizontal won every time, at all distances. In fact, much of the real DX (US West Coast, JA, etc) was only workable on the dipole. So much for the much-vaunted "low-angle" radiation of verticals! There are endless experiments of this type you can do - is that new aerial really any better than your old one? Does it work better in a slightly different location? Would a longer "long wire" make any difference? And so on.
Have a go - and have fun.