Limitations of the chassis dynamometer
Like all devices intended to simulate a phenomenon, the chassis dynamometer has its limitations, which are easily overlooked. It cannot be too strongly emphasized that a vehicle, restrained by elastic ties and delivering or receiving power through contact with a rotating drum, is not dynamically identical with the same vehicle in its normal state as a free body traversing a fixed surface. During motion at constant speed, the differences are minimized, and largely arise from the absence of air flow and the limitation of simulated motion to a straight line.
Once acceleration and braking are involved, however, the vehicle motions in the two states are quite different. To give a simple example, a vehicle on the road is subjected to braking forces on all wheels, whether driven or not, and these give rise to a couple about the centre of gravity, which causes a transfer of load from the rear wheels to the front with consequent pitching of the body. On a chassis dynamometer, however, the braking force is applied only to the driven wheels, while the forward deceleration force acting through the centre of gravity is absent, and replaced by tension forces in the front-end vehicle restraints (remember: braking on a rolling road tends to throw the vehicle backwards).
It will be clear that the pattern of forces acting on the vehicle, and its consequent motions, are quite different in the two cases. These differences make it very dif- ficult to investigate vehicle ride and some aspects of NVH testing, on the chassis dynamometer. A particular area in which the simulation differs most fundamentally from the ‘real-life’ situation concerns all aspects of driveline oscillation, with its associated judder or ‘shuffle’.
The authors have had occasion to study this problem in some detail and, while their analysis is too extensive to be repeated here, one or two of their conclusions may be of interest