New dyno helps squelch NVH problems

Temperature, speed, pressure, and vibration can all cause noise, vibration, and harshness (NVH) problems in wheel-end components. Rotors, pads, pins, calipers, bearings, and the like change dimensions as they wear, producing squeals, moans, jutters, rattles—and unhappy customers.

Using traditional brake dynamometers and chassis-roll dynamometers, you can control most of these parameters when performing NVH tests, but vibration is still a problem. About the best you can do to reproduce vibration is to use a realistic road texture on a chassis dyno.

The PERDC will be located at Ford's Essex Engine Plant in Windsor, ON.

At Robert Bosch Corp., we developed a system that reproduces the environment experienced by wheel-end components. Our Road Load Dynamometer (RLD) simulates

• wheel RPM,
• inertia,
• line pressure (and therefore component temperature),
• acceleration in the vertical (z) and fore/aft (x) directions, and
• the mechanical impedances attached to the wheel-end components (and therefore the stress and strain environments in those components).

The RLD consists of an inertia dynamometer, a two-axis hydraulic-actuation system, a fixture to couple these systems, and a sophisticated, distributed control system. The dynamometer produces RPM, inertia due to vehicle gross motion, pressure, and, if needed, environmental temperature. The two-axis hydraulic-shaker system induces the accelerations at the hub and, therefore, the movements of the various parts and the inertia-induced forces on them. The fixture provides the static and dynamic loading caused by the vibrational inertia of the vehicle. Combined, these systems also work to reproduce the stresses, strains, and component and assembly resonances that could affect the wheel end components in field use.

The system uses a brake NVH dynamometer from Link Engineering (Plymouth, MI; http://www.linkengineering.com). The main shaft of this dynamometer connects to an output flange embedded in the wall of a large environmental chamber. The chamber includes a T-slot bedplate centered on the dynamometer main shaft.

Two hydraulic cylinders from MTS Systems (Eden Prairie, MN; http://www.mtssystems.com) provide the vibration. The vertical actuator supplies a 50-kN force and a 150-mm stroke (11 kip/6 in.). In addition to supplying dynamic vertical inputs, the actuator supports the vehicle corner weight. A 24-kN/150 mm (5.5 kip/6 in.) actuator supplies horizontal input.

The fixture has five components:

• An actuation frame anchors the hydraulic cylinders. It is modular and movable, so we can reposition it with respect to the dynamometer shaft. This allows us to simulate two-wheel drive and four-wheel drive vehicles.
• The vehicle frame attaches the wheel-end components to the baseplate. It provides a mechanical impedance similar to the one that the wheel-end components see on a real vehicle.
• The weight frame holds weights that simulate the weight of the drive train and chassis.
• The vehicle-frame support supports the vehicle frame and absorbs braking
• torque forces on the frame.
• Adapters connect the test subjects to the test equipment—one at the dyno and one at the actuation frame.

The system controller, a Component RPC III (cRPC) from MTS, feeds modified real-time field data to the "second layer" controllers of each device in the system. It also monitors data generated by the UUT and modifies its inputs, taking into account the capabilities and impedances of the control systems and the UUT.

This is possible because the algorithm the cRPC uses is iterative. The controller outputs a signal, and then it monitors the results, computes modifications to the transfer function, and finally, creates a new output file.

The second layer controllers include the dynamometer control system that controls speeds and line pressures. We modified this system with hardware and software that accept analog signals from the cRPC controller. MTS 407 controllers control the hydraulic systems. They also accept analog signals from the cRPC controller instead of using their internal function generators.

We built an interconnected array of safety mechanisms into this control system to protect personnel and equipment. The system monitors everything from hydraulic fluid temperature to frame x-direction acceleration. If one alarm goes off, the whole system either goes to rest or shuts down (depending on the severity of the fault).

The system requires two data-collection systems. The field data-collection system collects data from a vehicle operating on the road, a test track, or on a four-poster simulator. The lab data-collection system records test data from capacitive-coupling probes and high-speed imaging systems to record the behavior of the unit under test. The bottom line is that this system can accommodate most on-road and off-road data for passenger cars and light trucks.

Because we use cRPC instead of the full version of RPC, our system can control only four channels, but our system cost only about half of a full-blown RPC III capable of controlling six channels. Fortunately, the accelerations are not all independent and we can achieve our goal by controlling only four channels: wheel RPM during motoring only, brake line pressure during stopping only, z acceleration, and x acceleration.

Using the RLD, we have been able to simulate the operating environment of wheel-end components in our lab. This simulation is at least as accurate as a road test, and using the RLD eliminates variability due to weather, route, driver, lane changes, and traffic. And because it is programmable, we can use it to simulate a variety of vehicles and an infinite number of test routes.

Edmonds, David E., "Road Load Dynamometer: Combining Dynamometery with Multi-Axis Vibration Testing," Paper 2003-01-1638, 2003 SAE NVH Conference (May 6–8, 2003; Traverse City, MI); SAE International, Troy, MI. http://www.sae.org.