<... I guess you're joking ...>
Yeah.  Sort of.

<... After a bit of such testing you should have a pretty good certainty that your indicator works the way you want it to ...>
Standing stock still, no turbulence, at the [reduced?] power setting and attitude at which you conduct the test.  Perhaps a lot of fuss and bother could be avoided by approaching the problem from another angle - create a "pinhole leak" (touted as the limitation of pressure gauge), run the engine at high power for a little bit and see if you get a pressure drop, and then measure and see how much (if any) coolant you lost.

I'm still looking for some convincing (key word here) evidence that a coolant level indicator will detect a problem that a pressure gauge won't detect sooner.  Been reading lots and lots of ramblings, some ideas and various brain farts and a couple of pretty elegant solutions - all addressing a problem that may not exist.  I still haven't heard anyone describe an incident that actually happened where a coolant level indicator did or even could have detected a problem that actually happened that a pressure reading did not or could not detect at least as well.

The Cozy and Velocity forums periodically launch into incredibly lengthy threads proposing elaborate, usually expensive and often very elegant solutions to problems that don't exist.

P.V.O.R.T. is part of the fun of building and flying experimental airplanes .... Jim S.

Finn Lassen wrote:

I guess you're joking, but some ground and in the air (near airport) testing should be fairly easy.
Drain enough coolant so that the level is just below where you want your warning to start. Run the engine at various power levels and attitudes and notice the behavior of the indicator. Add a bit of coolant and repeat. After a bit of such testing you should have a pretty good certainty that your indicator works the way you want it to. In normal orientation (not plugs up), you should safely be able to run the engine with a pint and maybe even a quart of coolant missing. At a guess the danger level is when only 1/2 of the pump impeller is covered with water. You warning level should be significantly higher than that. Of  course if you do the testing while you have a real leak that's a different matter :)

Finn