I've just got back from a jolly down Kent way so only just picked up on this, and it's getting interesting.

Amazing what you can get aint it?
For what it's worth, these are my thoughts based on what I have found out and from listening to others:
My reason for doing tests all stem from the fact that the single, most expensive item to go wrong is the head, and anything that can help to give an early warning has got to be the way forward, and that is the reason I purchased a TM-2.
I originally fitted it in the position as recommended (at the time of purchase - since altered) which was on the passenger side of the head - 2nd bolt back. After studying the system and speaking with others, I realised that the most informative place to get a reading was where the coolant leaves the head (the hottest coolant). The absolute ideal would be a sensor to replace the Mazda temperature gauge sensor at the front of the head but I couldn't find one substantial enough. Next best place would be an in-line sensor at the top hose stub, but I discounted this as being intrusive - I didn't feel happy about creating two more potential leak spots into the system.
I settled for the top hose metal stub, and this works very well, albeit a bit of a fiddle to fix.
Haydns current recommended position is a far simpler one, but you have to bear in mind that it only simulates the coolant temperature (nearest coolant is 50mm away). Having said that, it does give a good indication of what is happening.
Mike - I now have data and graphs of 4 Bongos with the sensor fitted to the top hose stub and they are all generally as shown below.
Maybe this sort of visual display could be a possible answer (if it could be arranged).
- Red and Blue dotted lines are stat max lift and closed.
- Main graph (blue) are the TM-2 readings
So as long as the ongoing TM-2 readings are somewhere between the Red and Blue stat positions, then the system is in control.
