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 failure of the 'dumb bell' link G.83302 for the Phantom III front brakes.
| Identifier | ExFiles\Box 92\4\ scan0352 | |
| Date | 23th August 1937 | |
| By.{R.W. Bailey - Chief Engineer} from Rm{William Robotham - Chief Engineer}/AFM.{Anthony F. Martindale} c. Sg.{Arthur F. Sidgreaves - MD} c. Hs.{Lord Ernest Hives - Chair} c. Da.{Bernard Day - Chassis Design} c. MX.{John H Maddocks - Chief Proving Officer} c. RHC.{R. H. Coverley - Production Engineer} c. Ed.{J. L. Edwards} 306 PHANTOM III. FRONT BRAKES. Failure of 'dumb bell' link G.83302 The stress in the above link is very low. The normal maximum load in it is, say, 300 lbs. which gives a tensile stress of 9,500 lbs/sq.in., and possible superimposed stresses of 15,000 lbs/sq.in. due to torsion and the same due to bending, the bedding of the ball affecting the latter stresses. The greatest principal stress is unlikely to exceed 30,000 lbs/sq.in. under any circumstances, and nickel steel should not fall below 110,000 lbs/sq.in. In proof of the above, we have fitted a car with soft aluminium 'dumb bell' links and have failed to break them in spite of abuses. The aluminium should fail at 75,000 lbs/sq.in. i.e. it has less than 2/3 the strength of the piece used on Production. There is a possibility, if the bell crank lever G.84118/9 is wrongly machined, or if the brakes are incorrectly set up, of the dumb bell striking the side of the bell crank in the neck of the former, every time the wheels go on to re-bound. The force causing the knock is the pull-off spring load, equivalent to 12 or 15 lbs. at the dumb bell end. While this force is quite insufficient to break the link outright, on rough roads the link might receive 500 such impacts per mile, which in 10,000 miles might fatigue the metal and cause failure. By setting up the brakes on a car incorrectly, i.e. by lengthening the cross rods, removing the stops, and winding the brake adjustment to the fully worn position, we have obtained this foul and have succeeded in breaking two cast iron links on the bumper drums with 3/4" cams, by making the car tramp. We cannot break aluminium or mild steel links. It therefore seems impossible that this foul should cause failure directly on a customer's car. We cannot break an aluminium link by making the servo judder. The geometry error which causes these impacts will be cured when Wraith type wheel movements are introduced next August. continued | ||