A Liquid-Cooling Motorhead Myth

[Hugh Janus]

Kevin Cameron (Robert Martin/)

This is my great favorite because it shows that the same mind can hold two completely opposite views of the same reality and see no conflict between them.

Reality No. 1: The radio tells us “Record cold continues, with area temperatures in the single digits and a wind chill factor of 20 degrees below zero. People spending time outdoors are advised to dress extra warmly.”

“Reality” No. 2: “I took the thermostat out of my bike’s cooling system and it overheated. See, the coolant is going through the engine so fast now that it doesn’t have enough time to pick up the heat from the engine.”

Wind chill factor is a measure of how much faster a warm object, such a person’s body, loses heat in cold weather as a result of wind. We all accept this as true because we’ve all lived it—when it’s cold and the wind blows, we need warmer clothing than if it was equally cold but windless. Yet many people continue to believe that coolant, pumped extra rapidly through an engine’s cooling system, will pick up less heat, not more. Yet we all know that the faster the “coolant”—wind—moves past us, the faster it cools us.

Which reality is true? No. 1 or No. 2? If it’s No. 2 that’s true, then the whole development of liquid-cooling systems has gone in the wrong direction.

Early liquid-cooled bike engines surrounded their cylinders and heads with lots of water in big water jackets and had no pump at all. Good examples are the early water-cooled Bultaco TSS roadrace singles of the later 1960s. They circulated their water only by convection: Hotter water expands, becoming less dense, and therefore the hot water slowly rose out of the engine through a really large hose to the radiator where it was cooled, contracting slightly and becoming more dense, falling back to the engine through a second hose at the bottom of the radiator. This system was given the grand name of “thermosiphon,” and it worked OK until engines made more power. Then the thermosiphon’s very slow circulation rate allowed formation of steam pockets, which in turn overheated engines.

The answer, engineers decided, was to move the water through the engine fast enough to scour away potential steam pockets, so they added a pump to pep up the circulation. There were still problems so next they began to make the coolant passages slimmer, so the coolant had to move through them faster. In fact, coolant passages in Formula 1 engines were eventually made so small that Honda’s V-10 F1 engine’s entire cooling system contained only 2.1 liters (roughly two quarts), circulating at very high velocity through tiny coolant passages. Such cooling systems have proved extremely effective.

How can that be? If spending more time in the engine and in the radiator are the key to best cooling, why did the very slow circulation through big water jackets fail to keep up with engine power increases?

The answer has to do with the difference between slow-moving and fast-moving flows. When flow moves slowly it tends to move in layers that don’t mix with each other. We feel this as the cooling that happens too quickly after we’ve settled into nice hot bath water. Yet stirring the water makes us feel the warmth again. What has happened is that the layer of water next our skin has given us its heat and by doing so has cooled, making us feel less warm. When we stir the water, we bring water from distant layers—which are still hot—into contact with our skin. Ah! Best of all for those who enjoy a hot bath is a Jacuzzi, which rapidly circulates its hot water, constantly scouring away the cooled layers of water next to our skin, and replacing it with nice hot water from deeper in the flow.

In an engine with a large water jacket and slow-moving coolant, the layer of water next to the hot cylinder and head surfaces is quickly heated—possibly enough to boil it—but water in more distant layers remains cool.

When we stirred the bath water to bring distant hot water to make us feel warm again, we created turbulence—random swirling, mixing motion. As fluid flow accelerates, a point is reached at which movement in layers—so-called “laminar flow”—is replaced by turbulent flow. In engine cooling systems, turbulent flow improves cooling by constantly bringing cooler fluid from all parts of the flow into direct contact with the hot surfaces we want to cool.

Heat flows from a hotter object to a cooler one in direct proportion to the temperature difference between them, so turbulent flow improves cooling by constantly bringing all parts of the flow (not just a thin layer) into contact with hot surfaces. This is why making coolant passages smaller to force coolant to become more turbulent has been so successful in improving cooling. The higher the coolant flow velocity is made the more turbulent it becomes.

OK, but how can removing the thermostat from a liquid-cooling system sometimes lead to the opposite—overheating? One of the most critical variables for centrifugal pumps is pressure on the intake side. As flow increases, which it does when you remove the restriction of the thermostat, intake side pressure, sometimes called “suction head,” falls. If it falls far enough, the pump can cavitate—pull the fluid apart to form cavities in the process known as cavitation (the same cavitation that can occur in suspension dampers if they are not pressurized). Cavitation causes pump output to fall—perhaps enough to cause overheating.

In my own experience with water-cooled race engines I have found that a little internal smoothing on the intake side of the pump—especially if there is a right-angle bend there—can pull coolant temperature down by as much as 5-10 degrees. That suggests that some cavitation was taking place there, provoked by flow over sharp edges.

Source

[Pedro]

I like this, but it's way over simplified and not accounting for pressure.