When I was ten years old, I heard about how a family friend, who was (and is, to this day) an Alaska bush pilot, had been compelled to fill his tanks with suspect fuel in order to fly out of some remote place. In the flight that ensued, the poor fuel detonated, causing damage that put one cylinder out of commission, but plane and pilot reached Fairbanks safely.
The other day I attended an aerobatic contest and walked into the display trailer of an aircraft engine rebuilder. On the wall was a display of failed components. In the display was a piston, largely destroyed by detonation. It was marked, "Failed owing to pre-ignition."
Will this saga never end? Now, I don't expect pilots to care as much for engines as for flying -- that's a given. I also do not expect rebuilders to be engineers. On the other hand, since the late Harry Ricardo discovered the difference between detonation and pre-ignition almost seventy-five years ago, it seems reasonable that the information might have dribbled down to the users by now. I think a person lost in the wilds of Alaska, knowing his fuel is suspect, needs to know the fundamentals about how engines respond when fuel octane rating is low. Use the minimum possible power, listen for knock, run rich to reduce peak combustion temperature. Know also that an unexplained rise in cylinder head temperature is a very good indicator that light, inaudible detonation is at work. Know also what detonation can do in the way of engine damage, and how quickly this damage can be done.
Therefore, for review, here is some of what Sir Harry found out just after World War I.
Normal combustion in a spark ignition, gasoline-fueled engine is not an explosion - it is smooth, progressive burning -- like that seen in a forest fire or grass fire. Random turbulence in the fuel-air charge causes the flame front to be extremely wrinkled and distorted, and it is because the flame is thus burning over so large a surface area that the cylinder charge can be burned as quickly as it is. In completely still mixture, the so-called laminar flame speed of normal gasolines is remarkably low -- of the order of a few inches per second. Thus, charge turbulence is a necessity if the charge is to be burned in a reasonable time. In engines operating normally, therefore, the flame speed is something between, say fifty and two hundred and fifty feet per second.
In an actual explosion, the combustion reaction propagates at or beyond the speed of sound in the reacting medium. This ranges from some thousands of feet per second upward to a few miles per second. The shock wave that results from sonic or supersonic combustion is a very steep, severe one; on one side is whatever the initial pressure may be, and on the other the pressure may be of the order of thousands of pounds per square inch -- or much more, in the case of solid explosives.
When engines run knock-free, combustion is a smooth, progressive process, and the rate of pressure rise inside the cylinder is of the order of tens of pounds per square inch per crank degree of rotation. This kind of pressure rise rate is mechanically safe and tolerable.
We all know that engines do not always run so smoothly. Lugging at low RPM, under heavy load, a peculiar knocking sound can sometimes be heard, coming from the engine -- especially when running on poor fuels. When such knocking is allowed to continue, characteristic engine damage results; first, overheating, followed by erosion of some metal surfaces within the combustion chamber. In heavy detonation -- as in an aircraft engine forced to continue running under heavy throttle on poor fuel, the eventual result is collapse of the piston ring lands and, eventually, complete break-up of the piston itself.
At first, in the WW I period, this was blamed on pre-ignition, which is ignition of the charge before the intended timing, as a result of the presence of something in the combustion chamber hot enough to ignite the charge. This could be glowing carbon deposits, an overheated exhaust valve, or the incandescent center- or ground-wire of a sparkplug of too hot a heat range for existing conditions. Based on a belief in pre-ignition as a cause of the engine damage observed, changes were made, but even in the proven absence of carbon deposits or overheated metal parts, the phenomenon of knocking combustion was still observed.
Ricardo used a high-speed engine indicator -- a device that graphs cylinder pressure as a function of crank degrees -- to find out what was happening. He discovered that the sudden pressure rise associated with the knocking did not occur until long after the ignition spark had set the charge burning -- not, indeed, until the final stages of combustion had been reached. This was clearly something quite distinct from pre-ignition.
Next, he discovered that there are great differences in the tendency of fuels to knock in this way. A supply of fuel made up from Indonesian crude oil was found to have outstanding resistance to this kind of knock. Other variables also affected to liklihood of knock. Knock became more likely as inlet temperature, engine temperature, compression ratio, throttle opening, or spark advance were increased.
In time, it was found that detonation -- as the phenomenon was called -- was a chemical process that depended upon the exposure of the remaining unburned charge to heat over time. As the piston compressed the charge, and the progressive burning of the charge after the spark further heated and compressed the unburned part of the charge, changes occurred in it that could ultimately result in detonation.
Now it is understood that heat and the passage of time cause so-called "pre-flame reactions" in the unburned charge. Molecular collisions set the fuel molecules vibrating, and those molecular species that are most fragile -- long straight chains -- are actually broken up. The result is a population of touchy fragments called free radicals, chiefly OH, increasing with time. When this population reaches some necessary level, that portion of the charge can auto-ignite and burn at sonic or supersonic speed. It is the impact of the resulting shock waves on the engine's parts that produces the metallic ping or knock that is heard. To sum up, we can say that the treatment of the unburned charge causes changes in its nature, possibly resulting in auto-ignition.
Detonation and pre-ignition are often interrelated, so that their effects are mixed. If hot carbon deposits do produce pre-ignition, the heating of the parts resulting from early ignition will certainly lead to detonating combustion and its familiar pattern of damage. Likewise, light detonation causes an increase in heat flow from the combustion gas to the engine's parts. This is because the parts are normally protected to a great degree from heat by the existence of a sluggish boundary layer of gas, close to and adhering to parts surfaces, acting as insulation. The violent disturbance of detonation of even a tiny portion of the charge scours away this boundary layer, causing the increased heat flow. On an instrumented engine, as detonation begins, there is seen a fall in exhaust gas temperature, for heat that would have left the cylinder in the exhaust gas is now instead being transferred to the piston and combustion chamber. The immediate effect of this can be overheating of the spark plug electrodes -- up to the point that pre-ignition begins as well.