In-vehicle data acquisition is nothing new. For many years, public safety and the operating expenses of large trucks have been simultaneously addressed by the tachograph, a paper-disk-based chart recorder installed in the cab of a truck. Road speed, engine rpm, and distance traveled are recorded vs. time. Analysis of the recording can reveal over-revving, downhill coasting out of gear, harsh acceleration and braking, speeding, and long periods of driving without a break.
Modern versions of the tachograph, such as the VDO FM100 Fleet Management System, use an on-board computer to log relevant vehicle data for up to 650 miles. Data is downloaded to a PC via a secure code key that identifies the individual driver. A battery in the computer protects against data loss in the event of a power failure, and an off-line Windows-based program provides detailed analysis of each trip.
The attraction of televised motorsports also is receiving a boost thanks to data acquisition. NASCAR race cars now carry telemetry equipment and in-car TV cameras. RACEf/x™ equipment developed by Sportvision in partnership with Pi Research provides statistics related to braking, lap position, speed, and engine rpm. This information is gathered from a combination of global positioning system (GPS) data and in-car electronics and cameras, processed, and inserted into the live television coverage.
Unfortunately, as exciting as it is for race fans, adding electronic equipment to a race car is not straightforward. During the Global Crossing race at Watkins Glen International held Aug. 12, 2001, the telemetry box in Robby Gordon’s car dramatically failed. It was later determined that the lithium batteries powering the GPS system had been damaged by high vibration, causing them to catch fire. Gordon was leading the field at the time, but the smoke and flames forced him to retire from the race.
Of course, you may not be trying to instrument a race car. Even so, in-vehicle data acquisition implies many more constraints than does capturing signal information in a lab setting. According to B & B Technology’s President Timothy Brooks, three key attributes of in-vehicle data acquisition systems are durability, power consumption, and connectivity.
“Extreme temperature and vibration conditions are our most challenging environmental issues. To reduce size and power, we have increased the channel count of our modules from eight to 32. And,” he continued, “using mass termination connectors standardized for all vehicles has helped us support the need to move the data acquisition system from vehicle to vehicle.”
Dewetron’s President Grant Smith also listed durability, power consumption, and connectivity, but emphasized the wide range of signal types that confront test engineers today. He described the many plug-in modules his company provides to accommodate strain gages, accelerometers, voltages from microvolts to kilovolts, frequency or tachometer inputs, position, and temperature.
The significance of a wide dynamic range was explained by HBM’s director of marketing, Robert Davis. With 20-b resolution, there is no need to select the measuring range in advance of a test. Even if the measurement signal occupies only a part of the maximum range, it still can be evaluated accurately. In practice, this means more reliable data collection despite inevitable human error.
The availability of the right modules in the right size is central to IOtech’s approach to in-vehicle data acquisition. “Portable systems have to provide a total measurement solution since it is difficult to mix instruments from different suppliers in a space-limited in-vehicle environment,” commented Tom DeSantis, the company’s president. “Plug-in card systems can be bulky and overkill for many applications. Chassis/card-based approaches often are too large for smaller systems and too small for larger systems.”
In addition to the usual types of signal conditioning, many in-vehicle data acquisition systems also offer CAN and J1850 modules. Signals such as temperature, pressure, and flow that appear on these buses in normal vehicle operation are processed by the modules so they appear to the data acquisition system just like any other channel of sensor information (Figure 1). Instead of a current clamp or voltage probe, these modules connect to the relevant signals via the vehicle’s on-board computer port.
To PC…
PC-based data acquisition systems are a good choice for many applications. They offer flexibility, and a large range of hardware and software is available. But they are not suitable in all cases. National Instruments’ (NI) data acquisition product manager, Brent Boecking, said that NI has used custom ASICs to address traditional PC data acquisition problems.
A memory access and interrupt control ASIC handles data transfers over the PCI bus while allowing the host processor to continue data analysis. This means that while data is being acquired, it can be analyzed on the fly. Each test is guaranteed to be successful because you know immediately if the data is faulty in some way. Before a large amount of test time has been wasted, you can reconnect probes or make other adjustments to ensure data integrity.
Timing and triggering are other areas where custom hardware has made an improvement. The data acquisition system timing controller (DAQ STC) ASIC provides all of the timing and triggering lines within a data acquisition device as well as the capability to route these signals as required. The real-time system integration bus (RTSI) is the common interconnection interface in which the timing and triggering signals are routed.
An example of a commercial data acquisition system based on NI products is the DRIVE system developed by Roush Industries. This vehicle dynamics test system gathers such information as acceleration, pedal pressure, steering effort, and steering angle from a race car. The LabVIEW-based VeDyan program, also developed by Roush, analyzes the collected data.
There’s no question that a notebook PC is well suited to applications requiring portable computing, but its compact form factor sacrifices the capability of accepting plug-in cards. A Measurement Computing designer, Buzz Mormann, described his company’s solution to this problem: “The PC-CARD product line provides a broad assortment of data acquisition cards for use in the PCMCIA slot of a laptop or notebook computer. The cards accommodate analog and digital I/O, counters, and communications and can measure up to 16 channels of analog inputs at speeds to 330 kHz.”
Data Translation also supports data acquisition via notebook and laptop PCs but favors the USB port for connection to separate modules. Tim Ludy, the company’s data acquisition product marketing manager, noted, “Our USB modules have the same performance as found in PCI and ISA boards with eight to 16 channels, high speed, and digital I/O. The USB plugs into any PC, including a battery-powered laptop unit. Any size instrument can simply be plugged into the back of a PC.”
Several manufacturers commented that standard PC hardware wasn’t sufficiently robust for many in-vehicle applications. For example, B & B Technologies’ Mr. Brooks said, “Our in-vehicle solutions are based on standard PC-based tools, and the real differentiator for our products is the software. By moving to flash drive architectures, we provide more protection from the elements and a more stable, conducive environment in which to gather data.”
In another example, IOtech’s DBK70 Vehicle Interface Module uses mechanically robust PC-Card memory instead of a hard disk. The DBK70 extracts signal information from J1850, J1939, ISO-9141, and CAN vehicle buses. The unit can be combined with the company’s LogBook/360 product and signal conditioning to provide stand-alone data acquisition and storage without a PC.
Or Not to PC…
You can avoid PC-related problems simply by not using one. Yokogawa has developed a series of MobileCorders with universal inputs and large amounts of memory. Basically dataloggers rather than high-speed data acquisition instruments, they include a floppy disk as standard, augmented by an optional 100-MB zip disk or 160-MB flash memory card.
In one application, the MobileCorder logged performance of yet another kind of vehicle, an electric forklift. Voltages and currents in the battery and inverter circuitry and device temperatures were measured as the forklift maneuvered over different kinds of pavement and with changing loads. Channel-to-channel isolation allowed thermocouples to be mounted directly on the parts to be measured which could be at different potentials.
Dewetron provides a degree of future-proofing in its data acquisition products by using a plug-in industrial PC. This means that the user can upgrade system performance by installing a more powerful processor card. Similarly, HBM uses a plug-in ML70 Computing Module to process the measurement channels in the MGCplus System. Both companies’ products also interface to PCs for further data analysis and report generation.
On the other hand, if you anticipate acceleration up to 100g, few if any conventional data acquisition products are appropriate. The compact data acquisition (C-DAQ) system from Gould Instruments can handle it. Capable of capturing up to 96 channels of analog data, the recorder provides CAN bus compatibility, GPS information, and telemetry within a 5² cubic package. Power dissipation is 10 W maximum, and data is stored on removable compact flash or PCMCIA cards.
For more information:
www.rsleads.com/207ee-236
Not Your Grandfather’s Signal Morphology
The controller area network (CAN) was developed by Robert Bosch GmbH, Germany, in 1986, but took a few years to be implemented in production cars. CAN has evolved into a so-called class C bus capable of supporting communications rates up to 1 Mb/s and is used for real-time control applications.
The J1850 bus specification officially was adopted by the Society of Automotive Engineers (SAE) for medium-rate so-called class B applications in 1994. J1850 and CAN buses are widely used today although they are not the only ones found in vehicles. New buses are under development for future safety-critical applications such as drive-by-wire.
These two buses and Ethernet are broadly described as carrier sense multiple access (CSMA) networks. All nodes comprising a network listen to the bus, and if nothing is being transmitted during the monitoring time, one or more nodes may begin to transmit their messages. These actions correspond to the carrier sense and multiple access parts of the description. The way in which each bus responds when a collision is detected distinguishes the three network protocols.
Ethernet nodes stop transmitting when they sense a collision. The nodes wait a random amount of time and then start again. After waiting, chances are that one node’s transmission will sufficiently precede another’s attempt that the bus contention will be resolved. J1850 and CAN buses arbitrate collisions more quickly through message prioritization. In both networks, the idea of dominant and passive states is used to transparently settle bus contentions.
CAN nodes use active pull-down devices, the high state provided by a weak resistive pull-up. Each node has a unique identifier value that is transmitted first and indicates its priority level. The lowest numerical value of the identifier corresponds to the highest priority. Nodes listen to the bus as they transmit. If a node transmits a high state but hears a low state, it stops transmitting because a higher priority node wants the bus. The last node still transmitting has the highest priority, and its message has not been corrupted by the arbitration process.
More than 2,000 unique identifiers are available in standard CAN networks with 11-b identifier fields, while more than 5,000,000 are provided by the 29-b field of extended CAN networks. Because there is no bus master, nodes broadcast their messages throughout the network. Based on the message identifier value, individual listening nodes decide if they will take any action and what it should be.
J1850 networks can use either a 41.6-kb/s pulse width modulated (PWM) two-wire differential scheme or a 10.4-kb/s variable pulse width (VPW) asynchronous approach. In the popular VPW implementation used by General Motors and Chrysler, the concept of a logical 1 or 0 corresponds to a longer or shorter period of time as well as to whether the voltage level is high or low.
Because the bus is actively pulled high and only weakly pulled low, the high voltage level is the dominant state. Arbitration operates in a way similar to the CAN bus: a simultaneously broadcasting node stops transmitting once it recognizes that a passive bit has been overwritten by a dominant bit from a higher priority node. The J1850 bus suits non-real-time communications of information between nodes.
Return to EE Home Page
Published by EE-Evaluation Engineering
All contents © 2002 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.
July 2002