pewterfish and I have just spent the past 20 or 30 minutes working out how fast the paper plane launcher I've got fires planes at, using an improvised timing system consisting of a white dot on the launcher and my computer monitor. CRTs make for nice accurate strobes when you use an image consisting of a horizontal white line on a black background.
We clocked the launcher at around 90 revs per second (for large values of around), which makes 5400 rpm ("Dude, that's hard disk speed!"). The launcher discs are about 5cm across, giving a circumfrence of 15.7cm or so. Times by 90 for cm/s, divide 100 for m/s, gives about 14 m/s rotational speed. This works out at 50.9 km/h, or about 32mph. Of course, the launcher isn't perfect, and a guesstimate for the actual launch speed is a respectable 16 miles per hour. This is using the stock supply of 2xAA. I've tried uprating it to 4xAA, but the motors shake enough to make it hard to hold (and I don't fancy the idea of a structual failure at stuipdly high revs).
'Course, we then spent the next ten minutes discussing ways of measuring the rpm more accurately, using a 555 timer, a bunch of white LEDs, a variable resistor, a PIC and a few seven-segment LED displays to show the strobe frequency.