There is something quite un-nerving and frustrating, not to mention embarrassing, when an older, beautiful looking Proper Motor Car stumbles and behaves more like a kangaroo when negotiating a traffic island or traffic lights. When it happens, such behaviour can last for twenty seconds or longer and may even result in the car failing to proceed for a while. Fortunately, in the context of this article, the problem is self-rectifying and a call to the vehicle recovery company is unnecessary. The irritation is more common on hotter days and when the car has been driven fairly hard for a period of time such that engine temperature is high. It is more prevalent in some cars than others and the absence of undertrays and bonnet flaps will not help the situation. An accumulation of heat in the region of the Autovac, situated as it does, high and to the rear of the engine compartment, is likely to initiate the problem and it is here that the impediment to normal, uninhibited fuel movement and the occurrence of fuel vaporisation is going to be centred.

Most owners have attempted to remedy the problem by enveloping the Autovac with either aluminium foil or a combination of foil and heat insulation material. Such a measure is an unsightly affair and is probably not a satisfactory solution. Others have by-passed the Autovac completely, having installed an electric pump. Moreover, a design fault in the layout of the engine compartment exacerbates the problem in that the exhaust manifold and downpipe are in close proximity to the Autovac and its associated fuel lines. The modern constituency of petrol makes matters worse still in that it is significantly more volatile than the very old equivalent and probably produced in this way to suit modern engines. The temporary stops at traffic lights or queuing to negotiate an island renders the air in the engine bay very hot indeed, especially on a hot day, and under normal engine idling speeds the radiator fan in our older cars is simply incapable of controlling the situation satisfactorily, particularly in view of the fact that the Autovac region is something of a ventilatory cul de sac.

Autovac Lid, left to right: air vent, induction pipe, petrol inflow

The Physical Properties of Petrol.

Let us examine the physical properties of petrol. The petrochemical industry does not produce the stuff we nozzle into our tanks as a pure-bred substance. Rather, it is a blend of perhaps 100 or more distillates of hydrocarbons of different carbon chain lengths. The smaller the component, the more volatile it is and vice versa and their boiling points will range from 20-200degrees Celsius. Such a blend allows the carburettor and combustion chambers to do their jobs effectively and since it is the ratio of air to vaporised fuel in the combustion chambers that really matters, it is important, for example, that during a cold start the smaller hydrocarbon fractions vaporise quickly to enable ignition. The function of a choking carburettor is to ensure that there are more of the more volatile components available to maintain a reasonably optimum air to vapour ratio. Conversely, with a very hot engine, the increased fuel temperature induces a large increase in vapour pressure as the less volatile hydrocarbon components are able to leave their liquid state more readily. In consequence, the rising vapour pressure will begin to exclude some of the essential air, and therefore oxygen, from the chamber mixture. This will result in a loss of engine power because effectively the chambers now have a lean mixture in that the fuel vapour part is dominated by the more volatile fractions. Furthermore, lean mixtures burn slowly on account of the wider separation of the fuel molecules and in turn this gives rise to an increase in coolant temperature caused by a more prolonged burning period in each cylinder. It also results in overheating of exhaust valves because the combustion process continues as the valves open. For this reason, we are all encouraged to keep a regular eye on the temperature gauge and to take appropriate steps accordingly.

The petrol supplied to car owners in the early to middle part of the last century had a much narrower band of components based on their volatility. In the more recent decades newer, much shorter length and newer, much longer length hydrocarbons have been introduced. The old fuel blend equates more to the fuel type used in racing cars where the more volatile components are excluded to prevent vapour lock caused by very high temperature engine systems. In effect therefore the owners of some vintage cars have become the victims of a relatively incompatible fuel blend. This is not to say that, with the old type of fuel, cars never suffered the problem of vapour lock but rather that its prevalence in the last few decades has become more noticeable.

Getting to the Root of the Matter.

Some years ago, Steve Lovatt of Ristes Motors in Nottingham felt urged to investigate the exact cause of this problem and to devise a means of overcoming it. He had in his workshops a 1931 Phantom 2 in the process of a total rebuild. It was a low mileage chassis, two owners since new, having spent most if not all of its life in USA and with a well-known history of vapour-lock.

Steve carried out many, many checks over a period of days if not weeks in an attempt to get to the root of the problem. He even installed a temporary, replacement Autovac to no avail. What he did discover however was that when the problem manifest itself, a quick inspection of the interior of the Autovac revealed an almost empty lower chamber and he wondered where all the fuel had disappeared to. After much scratching of the head and a consumption of some 40 gallons of petrol over many days, he detected the cause. As part of his experiments, he decided to rig up a fuel line system using transparent plastic pipe between the tank and Autovac and on to the carburettor in order that he might be able to see the petrol in motion and any appearance of vapour or air pockets therein. On taking the car for a spin with the front floorboards removed, a colleague in the driving position and Steve alongside him, the fuel starvation problem developed. His observation was that the fuel was travelling in a reverse direction along the plastic pipe between the Autovac and the fuel tank. The fuel was followed by lots of vapour-laden air. This explained the absence of fuel in the Autovac and in turn the lack of fuel to the carburettor. His observations also indicated that it was a high probability that the fuel line from tank to Autovac would end up with a total absence of fuel.

Autovac Lid, top to bottom: air vent, induction pipe, petrol inflow

The Dynamics of the Autovac.

So, why does fuel commence making its way back to the tank? First of all, we must understand the workings of the Autovac. Comprising an inner (upper) and an outer (lower) chamber, the device relies upon a partial vacuum created by the inlet manifold when the engine is running. For this function, it is easy to identify the pipe (the induction pipe) leading from the lid of the Autovac to a union close to the rear inlet port of the manifold. Inside the upper chamber is a float, suspended from a spring-loaded toggle-arm which, when the float is elevated, closes the valve to the induction pipe and opens an air vent to the lower chamber. When the float descends to the lower part of its chamber, the toggle arm opens the induction valve and closes the air vent. The lower or outer chamber is simply a fuel reservoir with an air pocket above it kept at normal atmospheric pressure by the air vent. As the action of the induction pipe renders a partial vacuum in the upper chamber, the now relatively higher atmospheric pressure in the lower chamber closes a non-return or drop valve in it such that once the fuel has entered the latter it cannot re-enter the upper chamber by backflux.

When the Autovac is working normally, the fall in pressure in the upper chamber causes fuel to be sucked up against gravity from the petrol tank and thence by dribbling into the upper chamber. Importantly, this maintains the reservoir of fuel in the lower chamber and a rising level in the upper chamber until the float and toggle arm close off the induction valve, thus stemming further flow of fuel from the tank. This cycle of events repeats itself until either the engine is stopped or one runs out of fuel.

I would love to see a display Autovac made of clear plastic, but using the valves, float, toggle arm etc from a discarded device and married up to a fuel tank and engine. Only then would we be in a position to witness the true function of such a wonderful invention. Is there an engineer in our membership out there willing to devote some time to such a project and to present the finished article to The Hunt House for educational purposes?

The Vomiting Capabilities of the Autovac.

The rising temperature in the upper layers of the engine bay, particularly on the exhaust manifold side, causes excessive heating of the Autovac. The air in the inner or upper chamber then expands, incoming petrol from the fuel tank, getting hotter by the second, evaporates rapidly and their combined pressure pushes and keeps the float at the bottom of its chamber. Whereas the pressure of air above the reservoir of fuel in the lower or outer chamber is normally at atmospheric, thanks to the air vent on the lid of the Autovac, we now have a problem. With the air vent closed off, the drop valve open on account of the rising pressure in the upper chamber, the expanding air above the fuel reservoir, coupled with the rapidly evaporating petrol beneath it, the drop valve is overwhelmed and the hot air and now, technically 'boiling' fuel enter the upper chamber in an uncontrolled manner. In turn this adds to the rising pressure there and accelerates the reversal of the direction of petrol in the fuel line to the tank. In the meantime, with the float jumping around in the lower part of its chamber, the induction valve is still open. This provides the only other escape route for the vapours and what liquid state fuel remains in the lower chamber is finding its way over to the carburettor until starvation prevails.

The Autovac is quite a sophisticated piece of engineering and since most of us have never examined its innards we are relatively unaware of its dynamics. The events surrounding a bout of fuel lock must be the equivalent of retching and vomiting when our stomachs become upset but without the violent muscle contractions.

Thus, vapour lock will lead inevitably to a very lean mixture and consequential reduction in engine power and much spluttering. If the lock is severe and prolonged, a complete deprivation of fuel to the carburettor and a temporary failure to proceed will be the likely outcome until the Autovac and its fuel supply have cooled down significantly.

Incidentally, it is only as a result of writing this article that I have been more acutely aware of the exact location of the vacuum pipe's joint with the manifold. It is very close to cylinder No. 6. The flow of air/fuel vapour mixture from the central common inlet port (that delivering the mixture from the carburettor) means that the vacuum pipe will add more vapour to Cylinder No. 6's allocated ration. Could this be the reason why No. 6 spark plug is the one that tends to be more sooty than the other five, I wonder?

Mid to late series 20/25s fuel tap mechanism with associated pipework; induction pipe to inlet manifold, vertical reserve and main fuel pipes, left and right respectively below tap

The Solution to the Problem..

For Steve Lovatt, from his experience in the Phantom II, the evidence was compelling, but what was his solution? He concluded that insulation of the Autovac and the use of heat reflective foil around it was a poor option. The reversed flow from the Autovac to the tank had to be stemmed and the introduction of a non-return valve in the feed pipe was a very easy means of achieving success. It worked and the Phantom he was in the process of restoring was later driven by the owner and Steve from Seattle to Pebble Beach without a hitch. The vapour lock problem for this car was now confined to history.

Based on his findings, Steve designed his own one-way valve system and began manufacturing them. The item is in brass, is a two-part, hexagonal union into which fits a circular valve which has to be cut to fit the interior and also an aluminium washer, suitably filed down to suit the hexagonal cross-section of the two halves. It measures about 40mm in length and requires two 3/8ths Whitworth spanners to secure it and make a petrol-tight fit. Its reference number is RMcarb1 and costs £47.36 plus VAT.

Brass, non-return valve (RM) in situ, nearside looking forwards

Easy to Fit.

Where should it be positioned? This is straightforward. Certainly in the small horsepower Rolls-Royces one will find a brass union between the main fuel pipe from the petrol tank and the ascending pipe to the Autovac. It is located just to the nearside of the clutch housing, directly beneath the firewall and to the rear of the starter switch unit. Simply by removing the front footboard, the union is more than visible. In the 20hp and early 20/25s there is but a single fuel pipe from the tank to the Autovac. If the chassis has a fuel reserve option, controlled by the fuel tap on the cab side of the firewall, there will be two such pipes, each with its own union. It is best to be sure which pipe is the reserve and which is the main; it is into the latter that the non-return valve must be installed. In my 1932 20/25, GAU 15, it is the one to the outside of the pair; in the photograph submitted, it is the one on the right of the two vertical fuel lines. To be sure, start the engine on the main fuel supply and keep the engine speed fairly high. After a few minutes, unscrew the union just beneath the circular tap system directly beneath the Autovac. As the pressure seal is broken, you will not see petrol gushing from the pipe since fuel will have descended the pipe back towards the tank and any in the pipe above the union will be either still ascending up to the fuel input of the Autovac lid or be held in the pipe by the induction pressure from the inlet manifold. However, the union will be wet. Try the same process on the other union with the fuel tap left on 'On' and you should find that the union innards will be bone dry.

Using the two 3/8ths spanners, the old union can be undone. Since there is a strong likelihood that this union has never been undone since the chassis was built, the process will be helped by using some WD40 penetrating oil or an equivalent. The operation is best performed with a low level of fuel in the tank, otherwise fuel will flow freely from the unfastened union. It will be necessary to increase the gap between the two halves of the original union very carefully and since the ascending limb is quite fixed, the pipe to the fuel tank is more flexible in this respect. There is a need for careful manipulation and alignment of the two pipes until the non-return valve fits. Mechanically, the new part will only fit one way, but it is best checked beforehand by blowing air into one end and then the other. One will immediately ascertain that the one-way flow has to be upwards. Once the threads of the old union are locating on the new part, the nuts should be carefully tightened up using both spanners, remembering that there are now four nuts in total to contend with. Once secure, the petrol level in the tank should be raised in order that any leakage of fuel can be monitored and if seepage is noticed, further tightening will be essential. After a run out in the car, check for seepage again.

The proof of the pudding will be evident during your next long outing, preferably on a hot day and during some hard driving in varying traffic conditions with numerous traffic islands and traffic lights to negotiate. The real test, I suppose would be a trip to The Alps and a long mountain climb to 7-8,00 ft (2,500m) followed by an over-run on reaching the summit and a descent.

Added 09/10/2010
The installation of the non-return valve has worked for me regardlesss of the weather conditions here in the UK. However, we do not compete with the extreme temperatures of Australia, I have to say.
Since writing the article some years ago, a number of people have discussed the issue with me. For some cars there is an additional factor. If the car hits a very bumpy part of the road, including some speed humps and railway crossings and the car pitches and rolls in response, there may develop within a few seconds a temporary fuel disturbance either in the autovac or in the carburettor, sufficient to cause 'kangarooing' in the car's engine performance. This will last for up to 20 seconds after the pitching and rolling of the chassis has subsided. If this happens, it is best to ease off on the accelerator pedal under the problem is rectified. Alternatively, you could always stop and turn the engine off for a minute until the petrol levels have been restored.