I recently read a 1960s report, made by an Air National Guard unit, on their experience with the air-cooled Pratt & Whitney 28-cylinder R-4360 radial engine. Much of the report was devoted to the problem of the exhaust valve-seat rings coming loose in the aluminum heads.
As it happens, this has also been a problem with some of the four-stroke motocross engines.
In the aircraft case, it was explained that the seat rings (nearly 3 inches in diameter!) were installed at 0.005–0.007-inch interference by heating the head to 500 degrees F and cooling the seat ring with dry ice. During testing with an instrumented cylinder, it was found that if the engine were started and then rapidly warmed up, the seat ring would quickly heat to 600 degrees while the temperature of the head—a much more massive part—rose much more slowly. This large temperature difference caused the seat to expand enough to force the aluminum head material around it to yield. The result was a loose exhaust seat.
Likewise, it was found that in-flight engine shutdowns or even steep power reductions could have the same result. The rule that came out of this sounds rather like the old admonition, “Moderation in everything,” for the best way to keep exhaust seats firmly in place was gradual warm-up and gradual cool-down.
Maybe the laws of physics have changed in 40 years, or maybe none of this is relevant to water-cooled four-stroke MX engines. But it’s something to think about.