probably a reed switch.
That's what I thought, but the resistance is always high (I've measured it) and there's just not enough room on the impeller or axel for a magnet. I've also tried holding a compass next to the probe, there's no deviation of the needle, so there can't be any magnets involved.
you could always use an ultrasonic transmitter/receiver pair
I've not seen this done anywhere, and it sounds like a major project in it's own right. I'm looking for a lightweight and reasonably cheap and quick solution.
thanks