From the Rolls-Royce experimental archive: a quarter of a million communications from Rolls-Royce, 1906 to 1960's. Documents from the Sir Henry Royce Memorial Foundation (SHRMF).
Investigation into the cause of Autovac petrol feed failure, specifically testing scenarios related to a leaking float.
| Identifier | ExFiles\Box 40\4\ Scan226 | |
| Date | 2nd February 1927 | |
| To CWB. from HS{Lord Ernest Hives - Chair}/AJL. AUTOVAC PETROL FEED. We have carried out tests to attempt to reproduce the conditions suggested in CWB10/GM30.12.26. as being the cause of Autovac failure. It has been found impossible to fill the Autovac float by the means suggested to such as extent that it will not function as a float. We have run a car for a month fitted with an Autovac float having a hole drilled in the bottom .150" in dia., the Autovac worked quite normally and gave no indication whatever that there was a leak in the float. If there is a leak in the bottom of a float, the amount of petrol leaking into the float is limited by the air trapped in the float above the petrol; as soon as the holes in the spindle inside the float are covered with petrol there is no outlet for the air trapped above the petrol and it is not possible to fill the float sufficiently to permit it working as a float. It must be approx. 75% filled with petrol before it will cease to function as a float. Suppose, as suggested, that it takes three working cycles to empty a leaking float - also suppose the engine is turned to allow the Autovac to function once and one third of the petrol is drawn from the float, but as this does not uncover the holes in the spindle none of the trapped air is drawn out, and consequently the petrol cannot rise higher in the float during the resting period than previously. We have carried out numerous tests but have not been able to fill a float to such an extent that it sinks in the petrol and ceases to function as a float. We have also put the question before the Autovac people and their reply confirms the results of our investigations. With reference to the report on CYK.10 enclosed with CWB9/GM16.2.27. Again we find it impossible to fill the float to such an extent by means of allowing the suction valve not to close, that it will not work. What happens in this case is that the air outside the float is sucked into the induction pipe and its place taken by petrol before all the air is sucked out of the float through the small hole in the spindle. We do not think that the self-draining float can be blamed for the trouble - we do not imagine from our investigations that the float when taken out of the tank was so full of petrol that it would not function; the trouble contd :- | ||