Very early in the development of internal combustion engines, it was realised that different fuels produced knock (a very damaging form of combustion which can destroy an engine, and sounded to anyone listening like someone knocking on the engine block with a hammer) which at wildly different conditions than others. At that time, no standard test engines existed to rate the resistance of various fuels against this behaviour – and any standard rating system was therefore impossible. Since knock got progressively worse as the boost pressure of the supercharged engine was raised, pressure grew from engine makers on the fuel refineries to discover more, no very reliable system of knock detection other than the “careful listening ear” existed. Therefore once the best possible compression ratio or supercharger boost had been established – the engine developers were loathe to accept any variations in fuel type or quality, lest the dreaded knocking and its associated piles of scrapped pistons and ruined cylinder heads return.
Eventually in the USA this resulted in the Co-Operative Fuel Research Committee being formed in the 1921, and the Waukesha company was tasked with the design and manufacture of a standard test engine which would be used by all firms for testing new fuels. Thus in 1928 was born the first “CFR” Test engine (Co-operative Fuels Research). It operated by moving the whole cylinder head up and down relative to the piston crown with an easily adjustable mechanism. Which was regarded at the time as a far better solution than previous methods which included taking the cylinder head off and installing a thicker head gasket to decrease the compression ratio. The man mostly responsible was the Chief Research Engineer, Arthur W. Pope. Thousands of engines were built and sent all over the world in the following years. It was a tremendous commercial sucess.
A much more repeatable knock detector was incorporated, the “bouncing-pin” type, which was literally a metal pin which when propelled upwards by the violent “ping” from heavy knock would break an electrical contact spring-leaf and produce a signal.
This basic system was used until well after the end of the Second World War. Although for military use, it was developed very much further for highly boosted aviation engines in fighter aircraft, where the variable compression system was found to produce very poor results.
In Britain Henry Tizard, Pye and Ricardo collaborated in experiments on engine “knocking”, and Tizard, (having already discovered the excellent anti-knock properties of Toluene), proposed that a scale of anti-knock resistance be formed, based on Toluene as a reference standard. It was known (predictably) as “The Toluene Number”. For reasons which are not (yet) entirely clear to the author, this scale did not find widespread accepted use, and an opportunity to rate fuels sensibly was lost. (Some interesting discussion on the politics of how this may have come about are discussed in this very worthwhile paper:
“Early Liquid Fuels and the Controversial Octane Number Tests” by E. L. Marshall of the Newcomen Society
Oddly a paper appeared in 2016, by the Society of Automotive Engineers, which appeared to propose almost exactly what Tizard had wanted, possibly oblivious to being about 96 years late.
“If the scale in the antiknock rating system was based on toluene and n-heptane (Toluene Reference Fuel, TRF), rather than PRF, a single “Toluene Number”, TN, the volume percent of toluene in the toluene/n-heptane mixture, can describe anti-knock behavior of real gasolines reasonably well at different conditions, unlike RON or MON”
“Knock Prediction Using a Simple Model for Ignition Delay”, SAE 2016-01-0702, Gautam Kalghatgi et al.
Having ignored Tizards work with Pye at Ricardo labs on the south coast of England in Shoreham, in the 1920`s various other proposals were made to standardise the resistance of a fuel to knock, at the time Iso-Octane was selected as a fuel which nobody could forsee being surpassed for its anti-knock value (although in fact Tolune, as proposed by Tizard, had a considerably higher anti-knock resistance than even Iso-Octane). At the time Iso-Octane was only produced in tiny quantities at great expense in research laboratories, a fuel which began to knock at the same time as Iso-Octane was declared to have 100% the rating of Iso-Octane. And therefore such a fuel was called 100 Octane. For the zero rating, it was found that Heptane knocked in any engine, and therefore a fuel with the same knock limit as Heptane is known as Zero Octane. The behaviour in between is calibrated to mixtures of the two fuels, it is not strictly speaking a linear scale for this reason, as different mixtures do not produce an exactly proportional change in knock rating, but it was better than nothing at all.
In the 1930`s people developing highly supercharged aero-engines, began to realise that all was not well with the Octane scale. In April 1938, Biermann in the USA at the NACA laboratories in Langley wrote:
“The data…show that for any one fuel, there is a definite relationship between the limiting conditions of inlet-air temperature and density at any compression ratio. This relationship is dependant on the combustion gas temperature and density relationship that causes knock….it is concluded that aircraft engine fuels cannot be satisfactorally rated by any single factor, such as octane number, highest useful compression ratio, or allowable boost pressure. The fuels should be rated by a curve that expresses the limitations of the fule over a variety of engine conditions”
NACA Technical Note#647
Just over a year later, in August 1939, Fritz Seeber – a fuels researcher at the German Aviation Research Institute near Berlin, published his findings, which not only corroborated Biermann`s findings – but propose a solution, and demonstrated it on a new single cylinder test engine, operating on the “DVL Supercharge Method”. Seeber wrote:
“The knock rating of aviation fuels by the conventonal octane scale in the CFR engine is defective and proceeds from misleading premises…Owing to the far-reaching adaptability of the test conditions of the supercharge method to the conditions of flight operation, a diagrammatic presentation of the knock characteristics is made possible, which is of great significance for the further development of high-output aircraft engines and their fuels”
“Neue Verfahren der Kraftstuffprüfung” – Fritz Seeber, Luftfahrtforschung, Vol 16 no.8 pp431-437
This German test engine, did not alter the cylinder geometry, but only the boost pressure, boost temperature and most critically, also mapped the response from Lambda 0.7 to 1.3. This allowed production of a 3D contour map of fuel behaviour showing maximum possible BMEP, vs Lambda vs charge temperatures. It was decades ahead of civilian transportation fuels research work of the time.
NACA had Seebers paper translated to English and send his work to their own engineers across America. It was not until 1942 when the American ASTM test methods for aviation fuels were officially updated in the USA to plot knock-limited BMEP vs Lambda. Britain had been slightly further ahead of the USA in understanding how radically differently fuels behaved at different Lambda values, and the wide variety of fuels used in aircraft engines made it clear than no standard curve to the left or right of Stoichiometric ideal could predict the knock behaviour without a full test. This resulted in several batches of 100 Octane fuel urgently needed by the RAF against the Luftwaffe having to be “doped” upon arrival in the UK, because the batch flaw had not been picked up by the American test method which tested at only one air/fuel ratio. The British test ascertained the fuel had inadequate knock-resistance at fuel-rich mixtures.
Thus it was that during the Second World War, high performance engine makers virtually abandoned the entire Octane rating system, and relegated it to simply identifying different grades. The actual determination of what a grade DID in an engine was carefully mapped across all major engine variables, and the maximum possible supercharge boost at any given condition was given for each engine type – not by any blanket interpretation of the Octane scale as far as determining possible engine performance.
Seeber`s DVL Supercharge test engines not only did away with varying the compression ratio (which cannot really be done sensibly as the squish characteristics of the chamber are totally altered, dramatically altering its burn characteristics) – but recommended that a single cylinder version of each service engine be used instead of a standard test engine. A single “0”, was inserted into the designations, the Messerschmitt 109E engine was the Daimler-Benz DB601. The DVL test engine for this type was known as the DB6001. Its cylinder head an exact copy of one cylinder from the V12 in use.
Given the irrefutable evidence that the Octane scale is at best an extremely poor approximation of what a fuel may do in an engine, why is it still used today ?
The Octane rating system has not changed appreciably since its inception nearly a century ago, it is still given at fixed ignition advance, fixed charge temperature and fixed air-fuel ratio.
The answer is likely a mixture of different reasons have allowed this scale to continue, when it appears well past its sell-by date,
1) Seebers research in Germany, being probably the first to realise how BMEP changes radically with Lambda, was really due to the German use of Aromatic fuels, which as can be seen from the second graph, are FAR more temperature sensitive than Paraffinic fuels which were used by the Allies. Thus Germany realised this first, the highly Aromatic fuels were used my necessity in German in WW2 aircraft due to the lack of access to crude oils. They could not afford to simply slosh pure Iso-Octane into their fuels to improve knock limits. Modern fuels today do not appear as temperature sensitive as these Aromatic fuels
2) Knock sensors, in WW2 the fighter aircraft did not fly with knock sensors, but had to use maximum possible boost. Therefore the testing done had to be acutely accurate to allow a mass produced engine to run on the limit line without any method of detecting knock in actual use. Therefore a fuel had to be characterised to extremely fine limits.
3) Modern engine mapping, allows engines to be tuned very accurately to the limits of the fuel being used in-car. Modern military aircraft do not use piston engines, and modern civilian aircraft use piston engines running far below knock limit.
Despite these reasons allowing us to “get away” with still using the Octane scale today, it may be time to ask if fuel makers should be sending out better test-data, which would allow small engine tuners to more easily predict what behaviour their engines may exhibit before testing actually begins. The significant differences between fuels in terms of their lean or rich performance may give the engine designer a head-start if he has a firm direction for the application of the engine if it is to run dominantly rich-or-lean.
A fitting conclusion is to list the final paragraph of Seeber`s 1939 paper.
“Chief advantages are as follows:
1) The knock tendency of fuels is tested direct in the actual engine cylinders.
2) The fuels are tested under conditions most closely approaching actual service.
3) The rating by several values permits conclusions for the development of 100-Octane engines and fuels.”