With its extension to heavy-duty trucks, onboard diagnostics (OBD) is becoming an even greater challenge than it has been for cars and light-duty trucks, a panel told attendees at an SAE 2014 Congress forum on “The Changing Face and Complexity of OBD.”
The heavy-duty trucks, primarily diesels, raise very high the subject of detecting misfire, which is primarily a fuel quality question, said Ford Technical Leader on Diesels and Aftertreatment, Michiel Van Nieuwstadt. The U.S. issue, he said, is cetane rating, whereas in China and India the top quality problem is sulfur content. Although, he added, U.S. “farm gas,” which isn’t regulated, is high-sulfur and a problem when used in on-road vehicles.
Misfire is not a simple “it does or it doesn’t,” as there are many intermittent signals along with inputs that can simulate misfire, and it varies according to everything from road surface to vehicle load. A company such as Cummins has a particular challenge, explained Jeffrey Potts, Chief Engineer on Diagnostics. Cummins does the OBD calibrations for its engines for all the manufacturers to whom it sells, and misfire is a very different calibration for a vehicle going into garbage truck service vs. long-haul operation. And as a result, there also may be specific service diagnostics and procedures for technicians, he said.
The heavy-duty truck makers will find misfire is very different compared with passenger cars and light-duty trucks, according to John F. Van Gilder, the OBD Technical Fellow at General Motors. He said they would be learning more about combustion in their diesels than they ever knew.
Preventing misfire codes
Gilder said that on its cars and light-duty trucks GM decided to spend more to try to eliminate misfire and considered that a more effective approach than simply detecting it and then paying to fix it. The primary steps were in ignition system upgrades, including a changeover to platinum-tipped plugs, then coil-on-plug modules and even two plugs per cylinder (which improves the combustion, not to cover a failure mode). Other steps included more uniform air distribution in the intake manifolds. Now he said, when GM looks at combustion patterns for upgraded engines, “they look good and other engines (not yet upgraded) look ratty.”
The area of OBD monitors and related system controls is evolving, said Harsha Nanjundaswamy, Manager of Light-Duty Diesel Development and Controls, FEV. He sees it moving from look-up tables to adaptive strategies operating in real time and application-specific. Now he said, there is an emerging focus on single component failures, but for the future there has to be more emphasis on separating system level interactions between components with minimum overlap.
The will be new sensors, such as for cylinder pressure and PM (particulate measurement for DPFs). But there’s just so far the industry should go with sensors. Instead FEV’s Nanjundaswamy said he saw the need for more model-based OBD, an “intelligent OBD,” which would have the potential to minimize sensor requirements. He gave the example of a “virtual SCR” for control of NOx.
Ford's Van Nieuwstadt agreed with the need for more modeling because “sensors are slow,” but said control is very different from diagnosis. “And we need modeling of a system that is out of range, not just a good one.”
At present, even such seemingly straightforward OBD requirements as cooling are becoming more difficult, Cummins’ Potts said. The manufacturer must deal with heat draw limits across multiple applications and such things as placement of sensors and other parts affect the calibration and validation.
This is leading to difficulty in developing monitors that can run in the restricted service of many trucks, and for which the thresholds for setting a code are different. With a passenger car or light-duty truck, a shop technician might be able to take a short drive under the right conditions to get all monitors to run. However, with a truck, that would require a technician with a special commercial license, as well as the ability, not always available, Potts observed.
Fixing it right the first time
The problem of getting the vehicle fixed right the first time is, of course, particularly important for a work vehicle, but developing the heavy-duty truck diagnostics has been especially difficult. Getting the OBD ready for demonstration took seven weeks in 2007, 36 weeks in 2010, and 72 weeks for last year, noted Potts. Further, Van Nieuwstadt observed, “we look for simple failures, not half failures,” which although he said delivers 90% for the effort, still leaves 10% to solve.
This increasing complexity would seem to indicate that the manufacturers should integrate OBD into the vehicle and powertrain engineering process. However, the panelists agreed that this is not easy, as many late developments and calibrations not only improve the product significantly, but would change the diagnostics. But trying to “build OBD on the side” also wouldn’t work, said Darren Gosbee, Navistar’s Engineering Director for Powertrain Calibration, who moderated the panel.
And as the emissions numbers come down, the challenges become greater, GM’s Van Gilder said. “Some of them are just too small to measure on OBD.”
This has led to questions about the best ways to direct the diagnostics, said Van Nieuwstadt. OBD monitoring is system-based, but more specific information is needed, so component diagnostics would be more helpful for service. In the case of an actuator or sensor, the diagnostic would have to provide real guidance, so the technician wouldn’t just be replacing with a new component. But it would have to recognize that one-end-of-a-circuit-to-the-other would be too labor-intensive.
CARB asks for lots of data
The control algorithms have to be practical for failure modes, causing no harm or the least amount, said GM’s Van Gilder. “A smart device may have to give up its default function to save a $400 electronic control unit.” He pointed out that the California Air Resources Board is asking for data on default functions, “and you may get only hours to answer.”
Further, he said, where a Tier 1 supplier has what it might like to maintain as proprietary software with an emissions-related component, as a result “there are no secrets between you and the OEM certifying the system.”
The entire industry will be facing new issues in 2017, when greenhouse gas emissions are a regulatory area, said Navistar’s Gosbee, adding that new OBD algorithms will be required for alternative fuels and their assuredly different misfire detection requirements.
The industry will have to be even more active in dealing with regulators in the writing of the rules, Van Gilder said. “CARB personnel change and the new people may have different interpretations.” CARB used to publish information on its interpretations but hasn’t done so in years, he told the session, which puts a premium on knowing how the new people in regulatory positions think. Further, he added, “we have the obligation to participate in the regulatory process to make sure the regulations are written to be as clear as it is possible.”