This is the most basic slow speed mode.  It's just the same Morse code that you've probably heard all over the amateur bands, only sent much slower.   While conventional CW contacts might take place at 5 to 40 words per minute, QRSS is typically sent at much slower speeds.   Instead of specifying the speed in words per minute, it is common to describe the time that each dit takes in seconds.  For example, QRSS1 would mean "Morse with one second dits", and QRSS3 would be "Morse with three second dits". 

To convert back and forth, you can use the formula WPM = 1.2 / DIT, where WPM is the words per minute, and DIT is the time each dit takes in seconds (you some times see this formula with the time spent in milliseconds, which looks like WPM = 1200 / Dit).   If we plug this in, we see that one second dits translates to 1.2 words per minute, and three second dits translates to 0.4WPM.

By now, you are asking yourself: well, why bother?  Don't I want to send Morse as fast as possible?   The answer is of course that we can decode very slow morse signals at much lower signal levels than Morse code sent at conventional speeds.    You could just take my word for it and move on if you like, but understanding why is actually a pretty good piece of knowledge to acquire, so if you want to at least get a basic grasp of what's going on here, read on!

Bandwidth, Signal and Noise

First of all, technically CW stands for "continuous wave".  But a truly continuous wave doesn't turn on and off, so there is no way that it can carry any information.  It's just a steady carrier.  In the context of amateur radio, CW is normally meant to be "on/off keying of a continuous wave".   When we say CW, we will use it in this sense.  

In theory, a perfect carrier is a bandwidth of zero.  When you key it on and off however, you generate what are called "sidebands".   These sidebands differ from the original frequency, and cause the signal to occupy greater bandwidth.   Forgive the math for a moment, but let's say that the carrier is some function of time C(t) = sin(k1 * t).   If we have a keying transform K(t) that takes on values of zero when the transmitter is sending a space, and a one when the transmitter is sending a dot or a dash, then the output signal O(t) = C(t) * K(t).  Let's say that K(t) is also a sine wave, sin(k2 * t), so what happens?   Well, if you stayed awake in trigonometry class, you know that the sin(a)*sin(b) = 1/2(cos(a-b) - cos(a+b)).   What does this mean?  It means that when we modulate a carrier with a carrier of a different frequency, we get two output carriers, each carrying half the amplitude of the original, one at the sum and one at the difference of the two frequencies.  These are what we call sidebands.

So, if our keying waveform was a 10hz cosine, we'd generate two sidebands, one at the carrier frequency minus 10Hz, and the other at the carrier frequency plus 10Hz.   We would then say that the bandwidth of the signal is 20Hz (the difference between the highest and the lowest sideband).

But in our case, our keying waveform isn't a sine wave, it's a square wave.   But we actually can use the same approach to figure out what the the bandwidth of the resulting signal is.   What we will do is approximate the square wave as the sum of a bunch of sine waves.     If you know what a Fourier series is, you know that a square wave can be represented as an infinite sum of sine waves at higher and higher multiples of the square wave frequency.  For instample, if we look at sin(x), we have the following:

If, for some reason, we plotted the curve sin(x) + sin(3*x)/3, we'd get this...