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).
Design and operational issues of petrol pumps.
| Identifier | ExFiles\Box 149\4\ scan0090 | |
| Date | 3rd November 1936 | |
| Rm{William Robotham - Chief Engineer} / X1263 Sgr and Hs.{Lord Ernest Hives - Chair} from E.{Mr Elliott - Chief Engineer} E.3/HP.3.11.36.. Petrol Pumps. Since dictating E.1/HP.2.11.36, we have been doing some intensive work on the petrol pump question, and there has also been a discussion between Mr Skinner of S.U. and Roy{Sir Henry Royce}, Rm{William Robotham - Chief Engineer}/Dixon, and E.{Mr Elliott - Chief Engineer} Skinner has had some experience with running a double pump scheme, and he strongly recommends running both pumps all the time, as he is afraid that the diaphragm in an inert state will deteriorate. Failing this the non-working pump should at least be always charged with petrol, which is a provision we had arranged for by putting the changeover tap on the suction side of the pumps. Skinner, and also our own Experimental Department now object to the use of a tap on the induction side owing to its great liability to leak air into the system, and cause the pumps to work violently, noisily, and ineffectively. A tap, however, on the delivery side is not of very much value in safe-guarding the pumps, neither does it prevent air leakage from a faulty pump upsetting the working pump, also if we follow Skinner's recommendation it would not be required at all. It is felt also that a tap if used would have to be made operative from the facia board, which is a complicated undertaking. Furthermore Skinner is of the opinion that half of the Phantom III pump will not be adequate for the Bentley engine, as the pump would be working too hard, and this is certainly true with the change to the G.3 diaphragm, as we at present find this new material is 40% down on pumping capacity. The longer solenoid and stronger spring cannot be used yet as it is at present an unknown quantity, and also is not obtainable if we could use it. Summarising the position therefore we think now | ||