You slam on the brakes and your vehicle comes to a halt. This sudden action causes a complex chain reaction if your vehicle is equipped with an anti-lock braking system (ABS). No matter what type of surface, the ABS must produce the optimum combination of stability, steerability and deceleration–a real challenge to engineers who must test the ABS under every conceivable driving condition.
ABS electronically decelerates your vehicle as quickly as possible without permitting the wheels to lock. Although an ABS generally provides better stopping distances than traditional brakes, its main purpose is to let you continue to steer under adverse conditions. If the front wheels lock, you no longer can safely control your vehicle.
By preventing the wheels from locking, ABS allows you to maneuver, even while applying constant brake-pedal pressure during a panic stop. This is accomplished by modulating the hydraulic pressure to the brakes based on feedback from ABS wheel speed sensors.
Testing Challenges
General Motors (GM) began installing ABS on the rear wheels of pickup trucks in 1988. Two years later, the company went to a full four-wheel ABS on Astro and Safari vans, and since then has been adding this feature to all of its truck models.
On the GM test track, the vehicle is put through a number of predefined maneuvers. One test is turning while braking; another consists of evasive maneuvers.
During testing, instrumentation has to acquire, manipulate and display a variety of analog inputs from vehicle sensors measuring force, pressure, rpm and temperature. As a result, test methods must be versatile enough to handle mixed analog signals and ensure compliance with strict performance requirements.
Measured Parameters
Test engineers in the GM Truck Engineering group perform two kinds of testing on an ABS: product assurance testing and development testing. In product assurance testing, vehicles are run on a matrix of different surfaces, speeds and pedal forces to verify that everything in the system is working correctly. In development testing, they focus on making improvements to the system.
Both kinds of tests use similar sensors and test setups to measure the variables. One variable is vehicle speed, which is measured using a fifth wheel (similar in appearance to a bicycle wheel) attached to the back of the truck.
Another variable, truck wheel speed, is acquired from the standard ABS sensors on the vehicle. Pressure sensors measure brake-system pressure out of the master cylinder and after the ABS.
Deceleration of the vehicle is measured with a decelerometer. The vehicle also is equipped with a yaw sensor for measuring rotational forces. A strain-gauge transducer measures how much force is applied to the brake pedal.
The steering angle is measured with a potentiometer encoder. Lateral acceleration of the vehicle, or turning, is measured with an accelerometer. Various internal and external temperatures also are acquired using thermocouples.
PC-Based System
For this type of data acquisition and analysis, GM engineers use an on-board 486 PC running Snap-Master for Windows software from HEM Data Corp. Signal conditioning and analog-to-digital (A/D) conversion is accomplished with PC plug-in cards made by Analog Devices.
Ahead of the A/D converter, one of the signal conditioning modules converts fifth-wheel rotations (frequency) to a voltage, which is converted to distance by the software. Strain gage and thermocouple signal conditioning takes place on two other boards. Figure 1 is a diagram of the ABS test system.
Before the current testing system was installed, test engineers used digital data collection methods, but without real-time analysis. The new PC-based test system offers many benefits over earlier data acquisition systems, including real-time graphic display of data, real-time analysis, a readily customizable human interface, and batch processing of recorded data.
Having immediate feedback in the form of a versatile real-time graphic display was the primary reason to change to the new system (Figure 2). On many occasions, this feature has made it possible to spot and fix a problem before an entire test is run, saving both time and money. For example, during one test, a data acquisition transducer cable came loose. By observing the screen display, the technician noticed the problem and corrected it without the need to graph and plot results before finding out that something was wrong.
Custom Instruments
A major benefit of the PC-based system is Snap-Master’s ability to create customized virtual instruments. These take the form of PC screen configurations that duplicate the capabilities of conventional test instruments.
A user defines an instrument by specifying sets of test variables and selecting display configurations from Snap-Master menus. This eliminates the need for multiple physical instruments.
The standard setup for GM’s Truck Engineering ABS testing includes six custom analog instruments. One displays up to 16 analog inputs as digital meters. Another is used to acquire and display, but not save, data similar to an oscilloscope. Another acquires, displays and saves analog data as continuous time plots. Two others perform pre- and post-test calibrations, and the sixth replays the data immediately after a test.
The software provides the capability to switch between the different instruments or to view a selected trigger channel and voltage with a single keystroke. Typical real-time displays include speed vs pressure, pressure vs deceleration, and speed vs lateral acceleration (Figure 2).
Each instrument can have up to eight different display screens, switchable with one keystroke. The number of data displays on a screen is limited only by the size required to view them. Up to 10 data channels can be included in each display.
Some of the more useful displays include speed vs brake pressure, another that shows pressure vs deceleration, and yet another that plots speed vs lateral acceleration and deceleration (yaw). Technicians can switch immediately among displays while the test is in progress or immediately afterward. These custom displays provide real-time results on the screen, avoiding the previous practice of downloading the data to a different system for later viewing and manipulation.
Math functions also are available to find statistical properties such as the minimum and maximum, to integrate and differentiate signals, and to digitally filter signals. More analysis can be performed as a post-test process. Sample acquired and derived waveforms are shown in Figure 3.
Custom Human Interface
One of the Snap-Master’s important features is its ability to link to other programs through an optional application program interface. This communication link between a Visual Basic (or other Windows programming language) program and Snap-Master is done through Dynamic Data Exchange (DDE), Microsoft’s standard method of enabling Windows applications to communicate with each other. Snap-Master has more than 250 DDE commands to allow it to communicate efficiently with other Windows applications.
The DDE feature is important because technicians in GM’s Truck Engineering group have become familiar with a certain user interface. Rather than having them learn Snap-Master directly, a virtual control panel interface, similar to what the technicians had been using, was created using Visual Basic. The control panel is shown in Figure 4.
From the control panel, technicians set up test parameters, descriptions of channels, calibration factors, test name, run number and instrument displays. The program sends these commands to Snap-Master using the DDE feature, which configures the data acquisition, manipulation and display functions.
By selecting the instrument file name associated with the type of test to be run, the user can move to a different test in a few key strokes. These tests are preconfigured by the test engineer and ensure that the technician runs the test according to plan.
To start the test, the technician presses the space bar on the keyboard or a user-specified trigger is used, such as the output from a sensor that detects brake-pedal movement. To stop the test, the technician can either press the space bar or use another trigger to end data collection.
Batch Plotting
After a full day of testing, numerous data files must be plotted, often in more than one format. Snap-Master allows the user to pre-select the files and define how they are plotted with a replay instrument configuration.
The batch-plotting operation can run unattended using the Visual Basic application program and DDE feature, which sequences Snap-Master through all the files to be plotted. This is a major convenience, because all runs taken in a day can be automatically plotted overnight and ready for viewing the next morning.
Summary
Improved productivity and better information are two of the biggest benefits of the new data acquisition system implemented by GM. More tests can now be run in the same amount of time since time wasted in rerunning tests has been eliminated. The findings from the ABS tests, because they are displayed in combinations of variables most meaningful to system developers, help engineers make more informed decisions.
SIDEBAR
In-Vehicle Network Data Acquisition
Besides the external sensor data, the GM ABS test system simultaneously acquires data from the vehicle’s electronic control network. The in-vehicle network has several purposes:
o It is the way that various on-board controllers talk to one another and share information.
o It is used by development engineers to test the controllers and various subsystems, such as the ABS, to determine the vehicle’s performance.
o It is used by manufacturing to verify quality as the vehicle rolls off the production line.
o It is used by service to diagnose a vehicle problem.
A separate virtual instrument acquires data from the vehicle’s electronic control unit (in-vehicle network), and additional sources such as the hydraulic system and motion transducers. Each signal is traced on a single display and analyzed with a separate algorithm.
Snap-Master’s analysis and real-time plotting capabilities make it easy to compare the test sensor data to that collected from the in-vehicle network. It is important to be able to compare these data sources because the analog sensor data generally represents what is truly occurring in the real world, which provides a reality check on the in-vehicle network. Figure 3 is a comparison of the in-vehicle network and the analog sensor data.
Copyright 1995 Nelson Publishing Inc.
January 1995
|
|