I like David's constant pressure design the best for an inverted
system. I'm not sure why the design was dropped so quickly on the
list.
    I have been trying to find a way around Aubrey's objection to the
constant "high" pressure when the engine is at rest.
    The engine is at rest for quite sometime before the liquid cools
down and retracts reducing pressure. Is this enough time to get the
o-ring shift?
     How much does a 75F to 30F then back to 75F ambient pressurize
the system? When the temp falls liquid is suck up from the A-can. Then
the temp comes backup and pressurizes the system. Extremes happen like
this in Arizona where 75F day and a 30F night mid winter. Summer time
might have even higher extremes. Can this produce enough pressure to
get the o-ring shift.

But, to relieve pressure at rest.....
    If we put a spring in the liquid side of the P-can as in the
attached picture (badly drawn red spring). Then charge the air side
with the upper bleeder open, so we get the air and the spring in
equilibrium. Then when the system is at rest we have a relatively
non-pressurized system.
    ?Ideally? the spring provides equilibrium in a short throw. So
that when the system starts to expand it will leave the spring behind.
    At rest with the spring and air in equlibrimum and a ambient temp
cycle, as above, the pressure of the system should stay more constant
by allowing the system to grow without exceeding trigger point to
overflow.

--
Blake C. Lewis

The way I see it there are two major problems. As far as I can tell  there is
no way to get dissolved air and gas out of the coolant. Second, it  is expensive
to make as you need a large lathe. The whole idea with the Mazda  RX8 system
is to get the air out on a continuous basis.

BTW the cooling system in my RX8 appears to be far more robust than
in my RX7. From time to time the RX7 would get hotter than the  thermostat.
The RX8 never seems to do that.

Paul Lamar ...No rotor no motor.

If formation of air bubbles is a one time event as suggested by the
experiment with boiling water on the stove (sorry I forgot who did
that but it was very informative )  then, there could be a  bleed
line with  valve which would be opened only for the the initial
engine run to allow air to escape. The bleed line would not have to
be part of the floating piston can.   Jerry

It was Bill Schertz that said you had to boil the water.
After you boil the water in the block by letting the engine overheat
how do you know you have all the disolved gas out of the coolant?