There is a wide gulf between the calculated heat rejection of any particular system and the actual performance of that system in real life.

Just ask Mistral. Or anyone else who tried to cool a rotary. As in any car, the water pump is sufficient to cool the car in Death Valley on the hottest day of the year, at idle. So, when we spin up the engine to 5,500 RPM and the pump is spinning 7,000 RPM we can pump a good sized stream of water over a two story building. Good to know, but how about a radiator. Not quite as good an outcome. The radiator sharp edged tubes stick up into the manifold and defy fluid flow through them. Some racing radiators are filled with epoxy right to the tips of the tubes to provide a very smooth transition into the tubes and flow is profoundly improved.

 

So the formula says in part that more flow means more (Better)  cooling, and this is accurate. But when you see that the radiator is the biggest restriction in the coolant loop you might guess that a low pressure area could develop between the inside of the radiator and the water pump. And in every case it dose. So now you notice that all lower (suction side) radiator hoses have a big spring inside to prevent the hose from collapsing. So, the suction side of the pump can pull quite a low pressure on that hose, correct? Even with a 22 pound pressure cap and a really big free flowing radiator. There is never more than 14.7 pounds available to crush that hose, so the pressure inside the hose must be lower than that by a good margin. And if the pressure inside the hose is that low, what is the pressure inside the cool side radiator manifold? Notice in the olden days that the radiator died from the bottom tube ends (up and down radiators) seem to rot away and leaks at the bottom killed the radiator.

 

No, it was cavitation.

 

Also cavitation can kill a pump quickly. It eats away the pump vanes like acid. Notice that the top hoses are smaller than the bottom hoses?

 

Could drag increasing at the square of velocity be performing the function of a..........dare I say restrictor?  If the top hose is long enough it has that effect. Now in the lab and in the drawings all of that fluid is incompressible. And the surface of the tubes is uniformly exposed to the fluid, so we predict that at thus and so, a flow rate we expect rejection value X. But we seem to not achieve value X in actual practice. Because, while the fluid is very nearly incompressible the air bubbles in the fluid are easily compressed, and thus allow for volume changes, and then for both high and low pressure to exist in the same closed system.

 

Removing the radiator as the biggest restriction in the circuit just about eliminates suction side cavitation. So I installed a restrictor in the water outlet of 5/8" diameter. It is welded in place. It never changes. Been doing it since 1980. Have yet to overheat a rotary. Have never lost a water pump or radiator.

 

I would not use that small a restrictor for 5,500 RPM. Probably way too small. My engines were used between 7,500 and 9,600 RPM. This with the small crank drive pulley and the stock water pump pulley.

 

What I have deduced from this may be completely wrong. But, it does work for me. Or, perhaps my system is so overly large it is just tolerating my folly. 

 

Lynn E. Hanover

 

 

 

I saw Lynn’s coolant diagram with a restrictor plate in it  …  you guys with evaporator cores and 1” coolant hoses have a 1” restriction, this based on Mazda’s design of 1.5” inlet/outlet on the stock water pump and the stock design includes a thermostat.  With all of that as a background (never had a thermostat), I decided to try a restrictor plate  in my coolant system, using a 0.75” hole in a plate at the water pump outlet into my 1.5” radiator hoses.  I can say that it doesn't do any harm and may have actually provided about 5% improvement … more testing to follow.