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).
Analysis of motor car acceleration, comparing theoretical calculations with experimental data from accelerometer tests.
| Identifier | ExFiles\Box 113\2\ scan0002 | |
| Date | 14th September 1912 | |
| 490 THE AUTOCAR, September 14th, 1912. The Acceleration of a Motor Car. and then to compare the curves given by experiment with figs. 3 and 4. Two methods are in use for the measurement of car acceleration: (1) By deduction from a space-time graph, (2) by the direct indications of an accelerometer. Of these two methods the former has the attraction that none of the measuring apparatus is carried on the moving vehicle: It is, moreover, a simple matter to measure the distance travelled at the end of equal time intervals; the slope of this space-time graph at any point measures the velocity, and a velocity-time curve can then be constructed. And the slope at any point of this velocity-time curve measures the acceleration. This process is, however, both long and inaccurate, particularly at the gear change points. The same data could have been treated equally well by an analytical process based on the calculus of finite differences; such a method has been used at Brooklands, but it suffers from the alternative disadvantages that, if a large number of terms in the infinite series (given by the method of the calculus) be taken, the process is exceedingly laborious, and that if few are taken the method fails in accuracy, particularly when the acceleration is changing rapidly. It is far more difficult to measure the acceleration of a car than of any other vehicle, owing to the discontinuity of motion in the gear changing. The use of an accelerometer avoids these lengthy and troublesome calculations, but has certain disadvantages peculiar to itself. It is open to the objection that the measuring instrument is carried on the moving vehicle, and is therefore subjected to a good deal of vibration, and possibly to zero-error should the floor of the car change appreciably in its upward or downward tilt relative to the road. Vibration gives little trouble in a suitably designed instrument, and is actually useful in that it helps to overcome any statical friction there may be in the mechanism. The difficulty due to change of tilt cannot, however, be avoided by any attention to the mechanism of the accelerometer. But its amount has been the subject of careful study on a variety of cars, and been found to be less than can be measured in practice; were it otherwise, it would be necessary to take the precaution of first blocking the car springs. Acceleration Curves Experimentally Obtained.—Acceleration tests were made with one of the recording accelerometers above mentioned on the car for which figs. 3 and 4 were predicted. The results are shown in figs. 5 and 6. The former was taken on Brooklands track, when the driver was instructed to accelerate as rapidly as possible, and the latter on the road on the way to Brooklands without the driver knowing that any test was in progress. In both cases the road was pretty nearly level, rising about 1 in 80 in the case of fig. 5, and falling about 1 in 100 in the case of fig. 6. There is a striking difference between the “test” and “non-test” starts, and the latter was, of course, much the more comfortable for the passengers, the rate of change of acceleration being far less. It is interesting to compare the prediction and the reality. On second and on top gears the predicted acceleration agrees very nearly with the actual figures, but on bottom gear the reality is some 2ft. per second per second, or 25 per cent. below the prediction. Now, the factors which all influence acceleration, particularly on bottom gear as the car starts from rest are: (a) The momentum stored in the rapidly rotating flywheel is suddenly liberated when the clutch is let in. (b) The carburetter is liable to choke with the rapid change in engine output. (c) During the engagement of the clutch the engine is exerting torque. (d) The road wheels are liable to slip. (e) The momentum given to the car is partly absorbed giving momentum to the rotating parts. As regards (a), the momentum stored in the flywheel of this car was 1½% of that stored in the car when the top gear was in. Before starting the driver may spin his engine and flywheel at a rate—say, 2,000 revolutions per minute — equal to that equivalent to running at 40 m.p.h. on top gear. And if this momentum be given to the car, it will produce in it a speed of 1½% of 40 m.p.h.—or 0.6 m.p.h. — corresponding to about 1ft. per second per second if the operation were completed in just one second. During the same time interval, whatever it may be, the engine is also exerting torque, and this (c) is a further reason why the starting acceleration should exceed the predicted value. It is difficult to assess the engine torque during engagement when the engine speed falls rapidly from 2,000 revolutions per minute to only a few hundreds. In the limiting case we may take the torque during this brief interval as equal to the maximum torque on that speed, and draw the dotted line TN shown in fig. 1. Cause (b) has probably—with (d)—much the most important influence on the acceleration—enough influence, in fact, to neutralise the factors tending to make the predicted acceleration higher than the true and to reduce the actual substantially below the prediction, for it is not to be expected that when the load on the engine is changing so rapidly as it does during the starting of the car that the carburetter will be able to supply fuel exactly proportional to the load, as it undoubtedly does in a good engine when running on the bench test from which the torque curves in fig. 1 are derived. In short, the carburetter is likely to choke and so cause loss of torque. This explanation is borne out by the fact that when grade climbing on bottom gear the carburetter has time to act, and it was found that the car in question was capable of climbing a grade steeper than 1 in 5, equivalent to an acceleration reading exceeding 6½ft. per second per second. Some carburetters have better starting qualities than others, and curves such as these bring out the differences. As regards (d), an initial starting acceleration of about 7 would, with, say, two-thirds of the car weight on the rear axle, correspond to a ratio of tractive effort to axle load of about Fig. 3. Fig. 4. Fig. 5. Fig. 6. Fig. 7.—Curve of ideal acceleration. Figs. 3 and 4 show predicted acceleration. Figs. 5 and 6 show the acceleration as realised. | ||