Technical Articles

Engines and Heat by Kevin Cameron

The familiar motorcycle engine is powered by heat, so a great many of its most difficult problems concern the management of heat and the results of high temperatures.
    I like to construct word pictures for myself of what happens in engines, step by step. The reason I like to do this is that, unless I remind myself of what the underlying physical phenomena are, I can easily fall into accepting the validity of “the way it’s always been” without question.

    Let’s begin with the intake stroke, assuming that our engine of interest is operating at or near peak torque. This means its cylinder filling is at maximum, so its heat release will also be maximum. Let’s also assume our engine has been running for some time, so the temperatures of parts have stabilized.
    As the intake valve begins to open, some time before TDC, the piston is still decelerating, moving rather slowly. The exhaust valve is in the final process of closing, so at this point, the reflected low-pressure wave in the exhaust port arrives at the cylinder. It helps to pull exhaust product out of the small clearance volume above the piston (now at or very near TDC), and then it passes on through the intake valve, causing early motion of charge toward and into the cylinder.
    The metals walls of the intake port are hot, and as the intake flow begins, this flow becomes turbulent, which accelerates heat transfer from hot metal walls to intake gas. As the flow nears the cylinder, it flows over the very hot intake valve, where it collects still more heat.
    This heating of the intake charge both helps and hinders. It helps because this added heat evaporates any fuel being carried in the intake flow as droplets, or as a creeping liquid layer on the intake walls. Any fuel not evaporated by the moment of combustion passes through the engine unburned. Therefore a certain amount of intake heat, applied to evaporating fuel, is valuable.
    On the other hand, we know that as the temperature of a gas rises, its density falls – it expands. For maximum power, we need maximum charge density in the cylinder. Therefore any intake heating reduces power by reducing intake density. This is why air-cooled engines develop better power on the first lap of a race than they do on lap five – the temperatures of parts are lower at the start, but rise toward steady-state values as the engine runs, causing loss of intake density through intake heating. Intake heating beyond what is needed to evaporate fuel is known as intake superheat and is always undesirable.
    How much is the intake charge heated during intake? In an NACA paper from the late 1930s, this was measured in an air-cooled aircraft engine cylinder. Intake temperature rise in this engine was 75 degrees F! It’s hard to understand how this much temperature rise can occur as a gas flows only a few inches at high speed, but there are reasons for this. First, the metal temperature of the head of a hard-working air-cooled engine is very high. Second, the fast-moving gas flow is highly turbulent, constantly presenting a different part of the cool flow to the hot walls as the gas tumbles and swirls on its way. Turbulence promotes heat transfer.

    There is a second reason why this gain in intake temperature is harmful to engine performance; it pushes the combustion process towards detonation. Detonation is auto-ignition and supersonic combustion of the last parts of the charge to burn, after the spark has ignited the charge in the normal way. As the flame consumes the charge, the hot, expanding burned gases compress the remaining unburned part of the charge. This heats it. In the unburned gas, this heat drives so-called “pre-flame reactions”.  The hotter the unburned gas is, the faster these pre-flame reactions proceed. They are heat-driven chemical changes that break down fuel and air molecules, forming sensitive hydrogen-oxygen fragments called radicals. When the population of these radicals reaches a certain level, the gas of which they are a part ignites by itself, before the normal, spark-ignited flame front can reach them. Because these radicals are so chemically reactive, their combustion proceeds, not at the normal leisurely combustion rate of 50-150 feet per second, but at or above the local speed of sound. In hot gas, this speed is very high – thousands of feet per second. This shock wave, hitting the inside surfaces of the engine, makes the sounds we call ping or knock – detonation.
    The hotter the intake charge is, the faster pre-flame reactions are driven, and the closer the engine is to detonating rather than burning smoothly and normally. Therefore intake charge heating is harmful to performance – especially if it goes too far. Anything that heats the intake charge makes detonation more likely. If an engine detonates at 12:1 compression, we reduce the compression until we find a ratio at which detonation ceases, and we lose the extra torque that this high compression gives us. If we could reduce intake charge temperature enough, we could accomplish the same thing – stopping detonation – without reducing compression ratio or losing torque. This is why highly-developed race engines always take their intake air from the coolest available source, and exclude hot air from radiators or cooling fins through the use of such things as intake airboxes.

    This raises another point. We assume that cooling air should cool the exhaust side of the head first, because that maximizes the temperature difference between coolant (air) and the hot object (the cylinder head). Think again. Cooling air passes through the exhaust-side fins and in doing so becomes hot. Then it flows on, to the intake side of the cylinder head, and this hot cooling air heats, rather than cools, the intake side of the head. This heating causes excess intake heating and therefore earns us some power loss.
    It is for this reason that modern liquid-cooled cylinder heads route the cooling water to the intake side first, then allow it to flow across the head to cool the exhaust side. This keeps the intake side cooler and thereby slightly boosts performance. Any for air-cooled engines? The last of the great air-cooled radial aircraft engines, the P & W R-4360, was also reverse-cooled, with the intake sides of its 28 cylinders facing the cool air, which then flowed around the cylinder to cool the exhaust side afterward.

    Back to the intake process. As the intake air enters the combustion chamber, it encounters the glowing, red-hot exhaust valve. This heats the intake charge even more. In the case of supercharged aircraft engines, with their large-diameter valves, exhaust valve temperature acted as an upper limit on engine performance. Push the compression ratio or the supercharger boost too high and the hot exhaust valve either pushed the charge to detonate or ignited the charge before the spark plug could do so (preignition). In either case, the engine was soon wrecked because preignition also leads to destructive detonation.
    In smaller engines, the distance from the hot center of the exhaust valve, out to its rim of contact with the cooler valve seat is short enough to keep the valve temperature reasonable. In larger engines, the valves finally had to be internally cooled by partially filling a head-and-stem cavity with sodium. At operating temperature, the sodium became liquid and, sloshing back and forth between head and stem, carried excess heat out of the valve head, to the cooler stem. There, it could be conducted away into the valve guide. This in itself was a problem, because it caused oil in the guide clearance to coke, then seize the valve in the guide. During the early 1950s, Norton had to run a small exhaust-valve cooling radiator on its Manx racers, and BSA had to use a new, high-tech gas turbine alloy (Nimonic-80) in the exhaust valves of its Gold Star engines to prevent valve cupping from high-temperature creep. The bigger the valve, the worse the temperature problem it tends to have. This is especially so if the tuner has decided to use narrow exhaust valve seats – the narrower the seat, the hotter the valve will run.
    Back to the intake process again. Not only does the fresh charge get heated by the hot exhaust valve, but also by the hot metal of the piston, combustion chamber, and cylinder wall. The better the engine cools, the lower will be the temperature of these parts, and so the cooler will be the final temperature of the intake charge. This in turn makes detonation less likely, allowing a higher compression ratio to be safely used, etc. If, on the other hand the engine is poorly cooled (like the rear cylinders of Harley-Davidson engines, or any cylinders which receive little cooling air), all its parts will run hotter,will heat the intake charge more, and will thereby be driven closer to detonating. You the tuner will then have to respond by reducing compression ratio, retarding spark timing, and/or using super-octane fuel to make the engine run without knocking itself to pieces. In air-cooled engines, cooling is a major key to performance; the cooler the engine can be made to run, the higher the compression ratio it will tolerate without detonating, and so the greater the torque it can safely make.
    Once the intake process is finished, compression begins. The act of compressing a gas adds energy to it, and this energy appears in the form of a higher temperature. Therefore, the higher your engine’s compression ratio is made, the more the intake charge will be heated during compression. This is why a higher compression ratio pushes engines closer to detonation.
    Finally, the spark jumps the plug gap and combustion begins. A swirling, shredding, mixing flame spreads away from the plug electrodes, driven by turbulence. Now we come to the question of flame speed. If the combustion chamber contains vigorous turbulent air motion, flame propagation will be rapid. Typically, for every degree of spark timing before TDC, combustion continues by a like about after TDC. This means that if your engine runs best with ignition timing at 32 deg. BTDC, its combustion process occupies roughly 32 + 32 = 64 crank degrees, centered on TDC. This in turn means that the slower your combustion is, the longer your piston and cylinder head are exposed to high temperature. This makes them run hotter, so they heat the next intake charge that much more, making detonation more likely. The longer combustion takes, the longer the last part of the unburned charge is exposed to heat, and the more likely detonation becomes for that reason also. Another reason favoring fast combustion is that mixture burned near TDC will generate higher pressure than will mixture burned when the piston is already well advanced into its expansion/power stroke.
For these reasons, performance is better in engines with shorter combustion periods – that is, engines which make best power with smaller amounts of ignition advance. Unfortunately, tradition resists this idea. Even the language encourages them. Would you rather be thought of as advanced or retarded? Tuners know that more spark advance will often boost acceleration, so this makes running more advance seem desirable – the more you can use, the better. This leads to a general desire to run as much advance as possible, and the result is that some tuners are not upset when their engines need 38 degrees for best torque, or even 40 or more degrees. So much the worse for them!
    If steps are taken to increase the turbulence of the intake charge, flame speed will rise, and the chamber will burn to peak pressure earlier. This requires the spark to be retarded so that peak pressure is again reached just as the piston is about to begin its rapid descent on the power stroke – at about 13-15 degrees ATDC. Likewise, if multiple spark plugs are used, the maximum distance of flame travel will be reduced – resulting again in being able to use less ignition timing, and often, in making more power. In one example – the Harley-powered BOTT special called “Lucifer’s Hammer”, the highest power resulted from a combination that gave best torque with an ignition timing of only 23 degrees BTDC. If you can build an engine that needs less ignition timing for best torque, it will in general make more power and run cooler than  combinations that require longer timing.
    Another way to add mixture turbulence is to raise intake velocity. Older tuners relied on the “hog ‘em out” school of porting, and engines inherited from such tuners often have huge ports. The current trend in engine porting is to use the smallest possible valves and ports. The resulting higher intake velocity puts more gas motion energy into the chamber, and this speeds flame travel.
    During actual combustion, the burning of the fuel adds a large temperature rise to whatever the gas temperature was at the moment before ignition. That initial gas temperature is the result of;
a) the intake air temperature
b) the temperature gain of intake gas during the intake process
c) the temperature gain of intake gas during compression
         Combustion is added on top of these other temperature rises, so the temperature during combustion is very high, and pushes heat into piston and cylinder head rapidly. Obviously, the higher any of these temperature gains becomes, the more heat is driven into head and piston. The two most important factors are the compression ratio and the degree of cylinder filling. If the engine breathes poorly, cylinder head temperature will be lower than if it breathes well.
    As the crank turns, the piston begins its descent on the power stroke. As the piston drops, the gas above it is allowed to expand, so its pressure drops rapidly. By the time the piston has reached half-stroke, more than 80% of the power theoretically available in the combustion gas has been transferred to the piston. This is because near TDC, pressure is very high, while at 78-80 degrees ATDC, it is a small fraction (like 1/8) of its peak pressure. The second half of the power stroke produces little power, which is why opening the exhaust valve at 60 or more degrees BBDC doesn’t hurt performance.
    This also explains why the cylinder head runs much hotter than does the cylinder. The head is exposed to peak gas temperature at about 15 degrees ATDC, but at this point, very little of the cylinder wall is exposed to the hot combustion gas. As the piston descends, exposing more and more cylinder wall, the gas temperature drops rapidly.

    When an engine detonates – even slightly – its temperature begins to rise. This is because normally, internal surfaces are to a degree insulated by the layer of gas nearest them. Being mainly under the influence of the nearby metal wall, this gas, called the boundary layer, cannot move rapidly. Like the stagnant gas between the double glazing of a storm window, this boundary layer slows heat flow. But when deto begins, the shock waves of its sonic-speed combustion blast the boundary layer away, accelerating heat flow into head and piston. The hotter the internal surfaces of an engine become, the more the intake charge is heated, leading to detonation. This is a vicious circle that can rapidly destroy engines. Even light, intermittent deto causes a rise in engine temperature.
    Most people are surprised to find that, when an engine begins to detonate, its exhaust gas temperature (EGT) falls. This is because, with more combustion heat being transferred to head and piston, there is less left over in the exhaust gas, so EGT falls.

    Now for another surprise. The higher compression is raised, the cooler the exhaust valve and exhaust pipe will run. We naturally assume that if we raise compression, thereby raising engine performance, we must also raise EGT. The reverse is true. Raising compression ratio does increase combustion flame temperature somewhat, but by expanding the burned gases more before the exhaust valve opens, it cools them more, thereby lowering EGT. Exhaust valves are heated by combustion gases during combustion, but they are heated even more by the high-speed, turbulent flow of exhaust gases past them after the valve opens. Because this is so, exhaust valve temperature is more closely tied to EGT than to peak combustion flame temperature. Harley tuner Don Tilley commented recently that at 11.5 compression, almost the entire header pipe of his Buell Harley engine became red-hot on the dyno. When he raised compression to over 13, only the first six inches of the pipe became red-hot. At the same time, this increase in compression ratio increased cylinder head temperature, because that is determined more by peak flame temperature.

It is a little-known fact that about half of the heat that flows into an engine’s cylinder head is picked up from the exhaust port. Why? Again, the reason is that the high-speed, highly turbulent exhaust flow is more efficient at transferring heat to cylinder head metal than is the slower, less turbulent flow in the combustion chamber during combustion. It is for this reasons that rational engines are designed with;
a) exhaust ports that are as short as possible, to minimise heat transfer surface area
b) steel stubs that extend as far as possible into the head, touching it as little as possible (preferably surrounded by an air gap)
Short exhaust ports obviously reduce the area of head metal exposed to hot exhaust gas. Lining as much as possible of the exhaust port with a steel stub, backed by an insulating air gap, further reduces the heat transferred to the head. In more recent times, tuners have the insides of their exhaust ports coated with insulating material such as zirconium oxide.

In addition, engineers such as Jim Feuling have shown repeatedly that most engines do not need exhaust ports and valves as large as they have in stock form. Smaller ports of higher angle and more streamlined section flow just as well or better than stock size, and their smaller surface area further cools off high head temperatures. Again and again, better, smaller exhaust ports have reduced air-cooled engine head temperature by amounts large enough to permit safe use of higher compression ratio.

How are pistons cooled? The answer is, by any means possible. In air-cooled engines, piston-to-wall clearances should be as small as possible, so that piston heat can easily pass into the cooler wall. Full-skirted pistons may be better in this respect than are the modern box-skirted variety.  In some cases, three rather than the trendy two piston rings must be used – again, to promote heat transfer through that second compression ring. The less surface area of piston exposed to combustion heat, the cooler the piston will run, and that means, if at all possible to use the closest thing to a perfectly flat-topped piston. It is better to build compression by filling the head (provided that valve flow is not thereby impeded) than by building a steeple on the piston. The more surface area the piston has, the hotter it will run.

It is also valuable to place oil jets in the crankcase, aimed to direct oil onto the underside of the piston crown(s). Always remember that air-cooled engines need all the help they can get from oil cooling.

What about thermally insulating coatings applied to piston, combustion chamber, and valve faces? This is an area for experiment, because the results are divided. Some find the coatings help with specific problems, like piston or valve distress. Others find that, because the coating surfaces run hotter than uncoated metal surfaces would otherwise be, they lead to detonation. Try it and see.
Finally, how do air-cooled engines get rid of heat? This depends entirely on how much air passes through the fin space per unit time, and how much heat that air picks up in the process. In traditional air-cooled motorcycle engines, the cylinders just stick up like pieces of firewood, and whatever air enters the fairing and blows over them is all they get. Most of this air travels around the outside of cylinder fins, not through them, because this is the path of least resistance. Fins on cylinder heads are often blocked by rocker- or cam-boxes and pushrod or camchain tunnels. Cast-in cooling holes are often obstructed by casting flash, reducing their size and thereby cutting cooling airflow to the critical area between the valves. The result of all these limitations is high head and cylinder temperatures – unnecessarily high.

In air-cooled aircraft engines, finned cylinders are covered by baffles that force cooling air to pass only through the fin spaces. Air supplied to the fins is picked up in a region of high pressure, and hot air emerging from the fins is led to a region of lower pressure. It is this pressure difference that rapidly drives cooling air through the cooling fins. On motorcycles, the usual excuses given for not employing baffled cooling are that;

a) it won’t work
b) removing the baffles for service will take too long

These were exactly the excuses I heard for years, given for not employing intake airboxes on motorcycles. Yet every new bike sold now has just such an airbox, and gains significant power from its use. Just now at Daytona I saw more than one air-cooled motorcycle engine blow and blow again because it effectively had no cooling other than accidental. This is not necessary – air-cooling can work and work well, if anyone wants to apply established techniques to make it work.