Manchester XPAG Tests Fuel and Tuning – Part 2: Choice of Fuel & Tuning Carburettors


In the previous article I suggested steps that could be taken to mitigate the problems a number of classic MG owners suffer from: Weak Running, where the engine stops in slow moving traffic, especially on hot days, and the Hot Restart problem where a hot engine cannot be restarted after stopping for 5 – 10 minutes. The Manchester XPAG tests identified a further problem with modern petrol: Slow Combustion. Particularly at normal road driving speeds and with high throttle settings, modern fuel appears to burn too slowly, overheating the valves, cylinder head and exhaust system. This in turn increases under bonnet temperature, making the Weak Running problem worse.

In this article I will discuss how both by the careful choice of petrol and by re-tuning the carburettors, the severity of the Slow Combustion problem can be reduced.

The implication of these results is that the normal additives used to enhance the octane rating are possibly to blame for Slow Combustion and fuels that use ethanol or other chemicals as an octane enhancer perform better.

Beware! Before using ethanol blended petrol in your classic car, read this article first.

Remember that our cars are all different and the severity of the problems experienced by owners varies immensely, even between the same models of car. The suggestions in these articles should be taken just as that, suggestions for people to try; they are not intended as solutions to be blindly adopted. While the specific details only apply to the 1 ¼” carburettors fitted to the XPAG engine, the general principles apply to all SU carburettors.

Slow Combustion

Previous articles described how poor mixing of the air and petrol in the carburettor can lead to a problem called Cyclic Variability. This is a condition where, in every cylinder, the timing of each combustion cycle can vary significantly cycle by cycle, leading to some cycles firing too early, or pinking, and others firing too late, resulting in very hot gases and unburned fuel leaving the cylinder. It is as though something is making huge adjustments to the ignition timing every time the engine fires.

Carbon monoxide (CO) in the exhaust gases is an indication of poor combustion. CO isproduced when there is insufficient oxygen present to burn the carbon to carbon dioxide (CO2). In an engine that is running rich, insufficient oxygen is inducted with the fuel and this will lead to high levels of CO in the exhaust gases. However, at Manchester both the mixture and timing of the engine were set to the optimum for each test. Therefore, the high levels of CO are a direct indicator of poor combustion caused by imperfect mixing of the air and fuel in the cylinder – one cause of Cyclic Variability.

The measurements of the SU suction piston heights at Manchester provided an indication of the scale of this problem, particularly when using full throttle settings below 3,000 rpm. In addition, there was a strong correlation between the degree of enrichment seen below 3,000 rpm on full throttle and the levels of CO in the exhaust gasses for the different fuels, confirming the link between the Slow Combustion problem and poor combustion.

Choice of petrol

At Manchester 9 different fuels were tested, along with a 20% kerosene mix and the use of a nebulizer to improve petrol atomisation and mixing. Each of these combinations produced different levels of CO. As the Slow Combustion problem appeared at its worst on full throttle between 2,000 and 3,000 rpm, the diagram shows the average CO levels in the exhaust gas for each fuel or combination for these tests. The lower thelevels of CO, the better that sample of fuel is burning. The grey bars show the special fuels we tested and the orange bars the ones containing ethanol.

These results show significant differences in the levels of CO between the different fuels. The top performing fuel is the Sunoco Optima 98 that was also rated highly as not suffering from the weak running problem.

Of the top 6 best performing fuels, three were ethanol blended, one an ethanol blended Super grade, came in 3rd. Remember Cleveland Discol that was introduced in 1928 and mentioned in the first article? This fuel contained alcohol (ethanol) and these findings support the manufacturer’s claim that it “contributed to a brilliant performance and better mileage because it keeps the engines cooler and cleaner”.

The following diagram shows the average relationship between unburned hydrocarbons in the exhaust gas against exhaust temperature for the 2,000 to 2,750 rpm range. As the engine was fully tuned for each test, high levels of unburned hydrocarbons reflect poor combustion. As might be expected, the less of the fuel that has burned the lower the gas temperatures and the less efficiently the engine is running.

It is interesting to note at the bottom right hand of the graph, the 95 Octane + foil and 95 Octane + 5:1 kero both burn better than the 95 Octane on its own. Adding kerosene certainly improves the combustion, suggesting any fears it will not burn are unfounded.

The three red tests at the top right of the graph show that increasing the volume of ethanol in the petrol both improves combustion and reduces exhaust temperatures as more hydrocarbons are replaced by oxygen molecules. However, before filling up your car with an ethanol blended fuel, read the warning later in this article.

Unfortunately, these findings show that not all Super Grade fuels performed as well. The Super grade with ethanol produced 1.55 ppm CO while a different brand of Super grade without ethanol produced 4.16 ppm CO – more than double.

One question most people will ask is “did you see any difference in power output between the different brands?” The answer is yes. The average full throttle power output of the test XPAG between 2,000 and 2,750 rpm was 25.2 BHP; the difference in power output between the worst performing petrol and best was 1 BHP. While measurable, this difference is small and would not be noticed during normal road driving. However, this measurement wasobtained for an engine that was fully retuned for each fuel and rev setting. Different fuels showed different levels of enrichment due to the Slow Combustion problem. When an engine is running rich, it produces less power. Hence, when used on the road and where the engine is not being continually re-tuned, as in the tests, these fuels will produce less power than the tests predict.

The fuels that gave the maximum average power output were Sunoco Optima 98 and the Branded 95 Octane (Batch 2) with nebulizer.One surprising result is that adding 20% kerosene to the 95 Octane (Batch 2) reduced the CO emissions by nearly 50%. However, it also gave the lowest power output. It is not clear why adding kerosene should significantly improve burning and reduce the effect of Slow Combustion while, at the same time, reducing power output. The power reduction is not caused by the kerosene failing to burn properly, the level of unburned hydrocarbons in the exhaust (141 ppm) was slightly less than that of petrol on its own (149 ppm), suggesting peoples’ worries about the kerosene not burning are unfounded.

This test was run with a very high concentration of kerosene (1part kerosene to 5 parts petrol or 20%). Adding kerosene to standard fuel is perhaps something worth trying. Best to start with lower concentrations, e.g. 5% kerosene (1part kerosene in 20 parts petrol) and increase it if it appears to improve matters. Owners of high compression engines need to take care as adding kerosene also reduces the octane rating and could cause pinking. This is discussed more fully in the next article.

These results demonstrated differences in the performance of different brands and grades of petrol in the XPAG engine. High ethanol content fuels appear to perform best, but they bring with them another set of problems. Trying different fuels to find one that best suits your car, is not easy. Look at the difference between the two batches of the same brand of 95 Octane petrol, both bought within days of each other in Manchester at a filling station close to the University. The one without ethanol burned worse producing 50% more CO but with an exhaust gas temperature 50oC lower than the ethanol blended batch. Also, remember the composition of the same brand and grade of petrol will vary across the country.

However, for normal driving, remember the advice that was given in the last article; to avoid the Slow Combustion problem, do not open the throttle fully below 3,000 rpm. If you wish to accelerate, select a lower gear first. Conversely, if you are cruising in 3rd or 2nd gear and your revs are above 3,000 rpm you should change UP to a higher gear.


A number of people have asked about the nebulizer and could it be used in a road going car? Unfortunately, the answer is no, it would not be practical.

The nebulizer consisted of a special nickel foil fitted between the carburettor and inlet manifold. The holes in this foil were around 8 micrometres in diameter, the same size as the dropletsproduced by afuel injectionsystem. As the inducted mixture passed through the foil, the droplets of petrol were forced to break up with the resulting turbulence improving the mixing.

Two things were special about this foil. Firstly, it was very thin and secondly, it had a 70% Free Area, i.e. the holes occupied 70% of the area of the foil, or the material of the foil only reduced the overall area of the inlet by 30%. As a result, the air flowing into the engine was virtually unaffected by this foil. A wire mesh would not have worked in the same way as they typically have a Free Area of only 30% – 40%. Using a mesh would have reduced performance by effectively reducing the diameter of the carburettor from 1 ¼” to under 1”.

Unless the air and petrol entering the engine can be filtered to remove any particles greater than 8 micrometres in diameter, such a foil would soon block and choke the engine. The reason it cannot be used for normal road driving.

Ethanol Blended Petrol

There have been numerous articles on the dangers of ethanol blended petrol. I will only discuss two aspects here, enleanment and water absorption.


Ethanol contains chemically bound oxygen molecules which is probably why it burns better in the XPAG. The oxygen is contained with the carbon and hydrogen molecules rather than having to rely on the turbulent mixing of the petrol and air. However, to offset this, a richer mixture is needed than with an unblended petrol.

The tests at Manchester used three ethanol blended fuels, two from the UK and E10 purchased in France. The good news is that none of these fuels showed significant degrees of enleanment. However, if you are planning to use E10, the advice is to enrichen the mixture by 1 – 2 flats on the jet adjusting nuts and keep an eye on the plug colour to ensure the mixture is correct.

Water Absorption

Separate tests have shown this poses a serious threat.

Ethanol blended fuel attracts and absorbs moisture from the atmosphere and once a certain concentration of water is reached, the ethanol/ water mix will phase separate from the fuel. Under normal circumstances, the risk of this is relatively small. However, storing your car forlong periods with its tank filled with an ethanol blended fuel poses a greater risk.

What is a more serious threat is that the fuel systems of classic MGs are far from waterproof. The filler cap is at the top of the fuel tank so rain can easily get in. The float chambers have ticklers on them that will also allow water in.

Even if only one droplet of rain gets into ethanol blended petrol, it will absorb ethanol from the petrol and sink to the bottom of the tank or float chamber. This droplet of water/ethanol mix is highly corrosive and will probably remain in the same place, corroding the metal, especially if the car is not being used.

I am sure you can imagine how shocked I was when I opened the storage container used to hold a test sample of steel and a section of float chamber in a water/ethanol mix for around 6 months. Yes, they are still in there hidden by the rusty water mix!

The before and after pictures below show a piece of mild steel (left) and a section of a float chamber right that have been stored in water that had previously been mixed with E10. The degree of corrosion of both the steel and aluminium is extreme and it is hard to believe these are pictures of the same pieces of metal taken only 6 months apart.

While these are test samples, the problem shows itself in real life. This picture shows similar corrosion inside one of my float chambers, probably caused by water getting in through the tickler pin when driving in the wet. When I first looked into the float chamber, there was what looked like a worm, sitting underneath the petrol at where the corrosion is. This was almost certainly a small quantity of water.

For the past 6 years, I believed I had only been using an ethanol free, super grade petrol. This corrosion in my float chamber shows that, unbeknown to me, at some time, I must have filled up with an ethanol blended petrol.

Once a small quantity of water settles at the bottom of the petrol tank or float chambers, it will sit there permanently absorbing ethanol from the petrol and corroding the metal.

Unfortunately, the additives sold to protect fuel systems against ethanol have no effect; whilst they will mix with the petrol, they will not mix with the ethanol/water mixture.

The only practical solution to avoid this problem is to use a petrol known to be ethanol free, such as Sunoco Optima 98, or to drain the petrol tank and float chambers once a year and allow them to dry to ensure no water / ethanol mix remains. Perhaps it is also time to think about slosh coating your petrol tank, with an ethanol resistant coating, as this may help protect it.

Tuning the Carburettor

When race tuning an engine, the aim is to get the correct mixture containing as much liquid petrol as possible, into the cylinder to maximise the power output. Typically, engines are run on full throttle, high revs when they are raced. In stark contrast, full power, high revs is rarely used when driving on public roads. When driving on the road an XPAG will mostly run between 2,000 and 3,000 rpm on part throttle. Unfortunately, these are the conditions where the Slow Combustion problem is at its worst.

The steps taken to race tune an engine, such as matching the manifolds, gas flowing the cylinder head, etc., all reduce the turbulence and mixing of air and fuel flowing into the cylinder, making the Cyclic Variability and Slow Combustion problem worse. Ironically, Cyclic Variability can also reduce power output and it is possible, that when running on modern fuel, a mildly race tuned engine produces less power at road driving speeds than an un-tuned one would.

This effect is clearly shown in the tests using the nebulizer. The nebulizer was fitted between the carburettor and inlet manifold where it forced the inducted petrol to break up into small droplets approximately 8 microns in size, around the size produced by fuel injection systems. It did this at the expense of restricting the airflow into the engine, something to be avoided in race tuned engines.

A comparison of the power output using the same fuel and same level of tune showed the nebulizer IMPROVED the power output by 1% below 3,000 rpm. Above 3,000 rpm where Slow Combustion is not a problem, the restricted airflow reduced the power output.

This again points to the Slow Combustion problem being the result of Cyclic Variability caused by poor atomisation and dispersion of the petrol in the inlet manifold and cylinder. It also shows that, for road use, improving atomisation and dispersion can give a power gain, even at the expense of restricting the airflow into the engine.

The next sections suggest how the carburettor can be tuned to improve fuel atomisation by setting the petrol height in the jet and by ensuring the correct suction piston/spring combination is used.

Petrol height in the jet

The petrol height in the jet is controlled by the weight of the float and by the setting of the forks in the float chamber. Normally, this is adjusted by inserting a rod between the lid face and the inside curve of the hinged lever and bending the lever. When set correctly the needle valve should be just closed when the forks meet the rod. With the H type float chamber use a 5/16 inch rod; with the HS type with a hinged nylon float use a 1/8 in rod.

There is a problem. This time not caused by modern petrol! With an HS type of float chamber and the correct fork setting, a float weighing 28gm to 30gm is needed to achieve the correct fuel level of 3/16” below the jet bridge (approximately half way between 1/8” and 5/16” recommended in the factory handbook for the TC). The original brass floats only weigh 24gm which gives a petrol level below the jet bridge of 3/8” (as normally specified for SU carburettors). Modern brass floats can weigh as little as 22gm and the plastic stay-up floats 20gm. (Note: All floats made by Burlen are to the original drawing specification of 20-24 grams).

A lower petrol level in the jet has a negative effect on fuel atomisation and dispersion. The partial vacuum in the choke has to both raise the level of the fuel to the top of the jet and provide the energy to atomise it as it is sucked out of the jet. A lower level of petrol in the jet, requires more force to raise it to the top of the jet, reducing the energy that is available to break it into droplets.

Although no specific tests were run at Manchester, the first set of tests run by the students in 2013 used the standard weight floats and fork settings, which would have given a petrol height in the jet about 3/8” to 1/2” below the jet bridge. The second set of tests were runwith standard fork settings but using heavier floats to give a fuel height of 3/16” below the jet bridge. The latter tests showed an average increase in power output around 5% for three different fuels, however, this figure should be taken with caution as these tests were run many months apart and other factors may have influenced the measurements.

A photograph of the carburettor taken during the first set of tests shows the (artificially coloured red) petrol leaving the jet as a stream rather than a dispersed mist. Something not seen in the second set of tests.

Unfortunately, it is not possible to achieve the fuel level recommended by MG in the jet using the lighter floats. While it is possible to bend the forks up to raise the fuel level in the jet or down to lower it, care must be taken not to bend them up too far or the float will foul the chamber lid and cause flooding. Should you adjust the forks, it is worth inserting a pencil through the bolt hole in the lid and the centre of the float to check it does not foul the pivot support (highlighted on the picture by the red circle) before the needle valve shuts off.

Be careful to position the float chamber with connecting arm at right angles to the carburettor body. While it is tempting to move the float chamber closer to the engine to make it easier to remove the float, this both exposes it to more heat and lowers the fuel level in the jet.

To achieve the MG recommended setting, I carefully add solder to brass floats to increase their weight to 28gm – 30gm using digital kitchen scales to measure the weight. As stay-up floats are solid, it may be possible to put self-tapping screws into their base to increase their weight.

It is also very important that the float chambers are open to atmospheric. It is easy to think the purpose of pipes fitted to the top of the float chambers is to remove any petrol that overflows. This is not strictly true, they also act as breather pipes allowing the air pressure in the float chamber to remain at atmospheric. Care must be taken to ensure that these pipes are not blocked and the correct stepped washer is fitted between the lid of the float chamber and the boss on the pipe.

Suction Piston weight/spring force

Early TCs had fixed weight suction pistons of 8.5oz (240gm). Later cars 4oz (110gm) pistons and either a “red” spring which gives downward force of 4.5oz when the suction piston is closed or a “light blue” spring which gives a downwards force of 2.5oz. These springs are fitted on top of the suction piston to increase its effective closed weight to either 8.5oz or 6.5oz.

If it has not worn off, springs can be identified by a coloured band on one end.

The advantage of using the “light blue” spring is that it reduces the choking effect of the carburettor allowing more air to flow through it at high loads/ revs and is one modification favoured when race tuning engines. The disadvantage for road use is that it reduces the velocity of the air flow through the choke, which in turn reduces fuel atomisation and dispersion. Either a fixed weight piston or a light piston with the red spring fitted is suggested for road use.

It is understood that some remanufactured carburettors are fitted with “light blue” springs, so if you have new carburettors on your car, it may be worth removing the suction chambers and checking.

Effect of piston weight on mixture

When I was researching the effect of different springs on mixture, I found very little information on the internet; the purpose of this section is to clarify this point. NOTE: It is worth having read the article on Carburettors before continuing.

Suction piston weight affects the mixture in two opposing ways:

* The lighter the piston, the higher it floats for a given volume of air passing through the carburettor. This both reduces the choking effect of piston and increases the size of the annulus between the needle and jet…… making the mixture richer.

* However, the pressure difference between the choke and atmospheric is reduced, decreasing the force that is pushing the petrol out of the jet…… making the mixture weaker.

The following graph shows the relative effects on mixture between the fixed weight piston, a lighter piston with the blue spring and a lighter piston with the red spring for the 1 ¼” SU Carburettor

Both the blue and red springs change the mixture profile with increasing revs/load in the same way as changing the needle would. However, fitting a blue spring also makes the baseline mixture weaker.

Over the normal range of piston heights used for driving on the road, the ES needle can be used in all three cases. Carburettors fitted with a blue spring can be made richer by screwing out the jet adjusting nut which will move the whole blue curve upwards. Similarly, for those fitted with red springs, screwing the jet adjusting nut in, will weaken the mixture.

One problem with the blue spring is that when using the ES needle, the mixture becomes weaker as the engine revs and load increase. A more conservative approach is for the mixture to become richer as the extra petrol helps keep the valves cooler, helping to prevent damage to the engine. Using a red spring produces a more conservative mixture curve than with the blue spring.

1250cc MG TF 1½ inch carburettors

At this point it is worth mentioning that all original MG TFs were fitted with 1½” carburettors with “light blue springs”, rather than the 1¼” carburettors that were fitted to earlier T-Types. The findings from Manchester support the comments of the critics, who at the time said the bigger carburettors were not needed. Unfortunately, the engineers were overruled by the marketing people.

With these bigger carburettors, as the suction piston rises, the aperture is more like a slit than the taller, squarer opening would be in the 1¼” carburettors for the same throttle setting. With the 1½” carburettors, it is probably better to keep the blue springs to produce a squarer aperture and better mixing.


This article has discussed the Slow Burning problem which appears to be due to high levels of Cyclic Variability caused by poor atomisation and mixing of the petrol in the carburettor. It appears as though one possible cause is due to the chemicals added to petrol to boost its octane rating. It also suggests that steps to race tune an engine could make matters worse for road use.

The Manchester tests showed that with the exception of the specialist Sunoco Optima 98 petrol, the best performing, commercially available fuels are ones that used ethanol to boost the octane rating. This is not surprising as ethanol contains chemically bound oxygen, improving the oxygen/carbon/hydrogen mixing in the cylinder. Unfortunately, the dangers of any water in the fuel system when using ethanol blended fuels is very clear.

As Cyclic Variability causes an engine to run slightly rough, owners should choose the fuel on which the engine runs most smoothly; particularly on full throttle below 3,000 rpm. An indication of lower levels of cyclic variability.

Although Sunoco Optima 98 is around twice the price of pump fuel, its low volatility below 50oC, improved running characteristics, guarantee that it does not contain ethanol and long storage lifetime make it a fuel of choice for low mileage vehicles. It

can be ordered direct from the Anglo American Oil Company via their web shop or by telephone on 01929 551557. Be aware, the law limits the amount of petrol that can be stored in a garage, or anywhere within six metres of a dwelling to 30 litres.

Adding around 5% – 10% kerosene to pump fuel reduces its volatility below 50oC and, it is not clear why, also reduces the degree of the Slow Burning problem. Care still needs to be taken to try to buy ethanol free petrol and owners of high compression engines must watch out for pinking as kerosene reduces octane rating.

If you live in the UK, remember you can legally add kerosene to petrol for cars produced before 1956, but you will need to apply to HM Customs and Excise for a Concession. Write to:

Mr John Loughney, Excise, Stamps and Money Businesses
HM Revenue & Customs
3rd Floor West
Ralli Quays
3 Stanley Street
M60 9LA

…….requesting a “General Licence to mix hydrocarbon oils under Regulation 43 of the Hydrocarbon Oil Regulations 1973 (SI 1973/1311)” giving your name, address, model and dates of production of the model of your vehicle.

Finally, the article discusses how the SU Carburettors can be tuned to improve fuel atomisation and dispersion by adjusting the petrol height in the jet and by choice of suction piston spring.

Ultimately, reducing the degree of the Slow Burning problem will lower exhaust temperatures, helping to protect the engine from burned valves and damage to the cylinder head. Lower exhaust temperatures will also help keep the under-bonnet area cooler reducing the severity of the Weak Running problem.

Paul Ireland