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).
The performance issues of an experimental exhaust heated throttle, specifically concerning poor fuel distribution in a cold engine.
| Identifier | ExFiles\Box 50\3\ Scan195 | |
| Date | 20th September 1921 | |
| To R.{Sir Henry Royce} from Hs.{Lord Ernest Hives - Chair} c. to CJ. c. to E.{Mr Elliott - Chief Engineer} c. to Da.{Bernard Day - Chassis Design} c. to BY.{R.W. Bailey - Chief Engineer} c. to EP.{G. Eric Platford - Chief Quality Engineer} c. to EWT. for R.R. Inc. X4305 Hs{Lord Ernest Hives - Chair}5/LG20.9.21. EXHAUST HEATED THROTTLE. X.4305. We are not able to obtain such good results from the exhaust heated throttle made to LeC.1418. as we can from the experimental exhaust heated throttle which was produced by fitting a copper jacket to a present standard 40/50 throttle. The fault we have with the exhaust heated throttle is that the contour of the throttle has been altered so that the distribution when the engine is cold, is bad. Under/condition we collect three times as much liquid petrol from 4, 5 and 6 cylinders than we do from 1, 2 and 3. With the "Dorris" scheme for collecting petrol, this throttle is quite impossible with 20% paraffin mixture, cold engine. The "Dorris" scheme improves it tremendously, but it is not as good as the experimental throttle with the Dorris scheme. We have not yet been able to appreciate any advantage in using the 1¾" pipe. With our experimental exhaust heated throttle fitted to a standard 2" induction pipe and fitted with the Dorris scheme, we are able to start up a cold engine, open up immediately to 500 R.P.M. full throttle and maintain maximum torque. We have been able to do this on two sets which we have made. We have tried fitting pegs in the exhaust heated throttle casing in order to correct the distribution similar to what proved successful on the 40/50. So far, on this throttle, we have not been able to achieve the same result. We are continuing our tests to try and correct the distribution with this contd:- | ||