Simanaitis Says
On cars, old, new and future; science & technology; vintage airplanes, computer flight simulation of them; Sherlockiana; our English language; travel; and other stuff
VW’S DIESEL SCAM
THERE IS technical interest in Volkwagen’s scamming of government and customers. According to Bloomberg Business, almost 500,000 VW and Audi diesels, beginning in Model Year 2009, are involved in the recall. Their “clean diesel” powerplants are fitted with software capable of turning emissions controls on or off, depending upon whether the car is being emissions-tested or used in ordinary driving.
Like the guy marveling about a thermos bottle keeping hot things hot and cold things cold, “How does it know?”
The background: Unlike spark-ignited counterparts, diesel engines ignite their fuel through the heat of compression. Diesel combustion temperatures are a bit higher than gasoline’s. And, as a consequence, a diesel has greater emissions of nitrogen oxides, NOX. Diesel combustion also generates particulate matter, nanoparticles of soot that are especially invidious to the human respiratory system.
Diesel design and calibrations balance control of particulates and NOX. Hotter combustion reduces particulates, but causes more NOX. Less heat generates less NOX, but more soot.
Emissions testing is performed on lab dynamometers (think treadmills for cars); this, to make the testing as accurate and reproducible as possible. Specific test procedures are followed, from which emissions certification as well as fuel economy data are derived.
A laboratory chassis dynamometer is an automotive treadmill. The driver follows a test cycle shown on the screen. Image from “54.5 Mpg and the Misinformed.”
Our Federal government has two basic test cycles, the EPA City test (evolved from a real-world drive around downtown Los Angeles) and the EPA Highway test (based on a trip in the environs of Ann Arbor, Michigan). Governments at state (and some local) levels have other procedures of periodic inspection.
VW got in trouble when researchers noted discrepancies between emissions-tested data and those derived from actual operation. NOX, in particular, was exhibiting as much as a 40-fold increase in the real world.
As I noted in “Turbo Foot” earlier at this website, “… these clearly delineated City and Highway Cycles offer engineers opportunity to optimize for the tests. There’s nothing deceitful in this; it’s sound engineering practice.”
However, to devise means of differentiating between tests and the real world isn’t optimization, it’s evasion.
The basic EPA test cycles. Image from “Turbo Foot.”
So, how do I conjecture it was done? Were it only the EPA Cycles, it would be straightforward to program engine-control software that recognizes the first several seconds of each cycle’s operation and acts accordingly. The EPA City begins with a short idle followed by acceleration and some dithering around 20 mph; the EPA Highway has an immediate acceleration to around 35 mph.
State inspections vary, but a diligent and painstaking engineer could identify common electronic signatures that could trigger the NOX control into action. Absent these, engine calibrations could be optimized for performance, not emissions. This wouldn’t be costly to finesse: Once an electronic system is in place, software programming is cheap.
There could be an even more elementary scam: Many inspections begin with electronic querying of a car’s Onboard Diagnostic system. It would be easy to gin up a cross-linking of the car’s emissions controls and its OBD receptacle: Plug in/NOX control on, plug out/NOX control off again.
Back in 2007, Casper’s Electronics, of Mundelein, Illinois, got caught for marketing its O2 Sim. This gizmo tricked the engine into seeing bogus oxygen-sensor data monitoring its operation. Casper’s was fined more than $74,000 by the U.S. Environmental Protection Agency and Department of Justice for selling 44,000 of these devices.
In VW’s case, the U.S. Clean Air Act of 1970 and its revisions have provision for a maximum penalty of $37,500 per car, more than $18 billion with the 482,000 VWs and Audis involved. Hitherto, the largest automaker fine was less than a tenth of this, Toyota’s $1.2 billion, for its floormats potentially interfering with pedal function.
VW’s outright deceit is rather more serious, with implications in selling, buying or registering the affected cars, even after recall. There is also the possibility of VW executives risking criminal charges. ds
© Dennis Simanaitis, SimanaitisSays.com, 2015
Simanaitis Says 
This is no laughable fraternity prank. It is a willful flouting of our laws. The engineers who developed the defeat technology should be stripped of whatever professional certifications they have.
VW should be assessed the full $18 Billion in penalties! And that is before the fully justified class action lawsuits kick in.
As they say in racing, “If you ain’t cheating, you ain’t trying!”
Great article Dennis.
Many thanks, SK.
Back around the time heavy truck diesels first got extensive electronic controls (Tier 3?), some engine manufacturer(s?) tried this too. Similar trick – when on the road they were set for higher NOx to keep the smoke down and reduce fuel consumption. They got caught, and besides some nominal fines they had to reflash all available engines to be compliant. Supposedly, all trucks on the road with those engines were to be reflashed, but I don’t remember what was done to enforce that. It was understood that the ones on the road would probably not be reflashed voluntarily, since they ran better (lower fuel consumption) as-is from the standpoint of the operator. Might be worth looking up if you have nothing better to do.
Since 1998 the truck engine industry (primarily Caterpillar, and Navistar in partnership with MAN) has been on the EPA short list. It was just two months ago that Navistar was fined $300M for earlier emissions violations.
In 2010 world renowned Caterpillar, with all it’s wherewithal, abandoned this market because of their emissions problems and fines.
From the 60’s these truck engine issues were exacerbated by the Mack “Maxidyne” engines that were coupled with 5 and 6 speed transmissions. The wide splits – compared to typical 9,10, or 13 speeds – required considerable over-fueling for good performance, but resulted in excessive black smoke after a shift. Those engines are long gone – for the last 10 years Mack, now a Volvo subsidiary, has used Volvo engines, or the occasional Cummins. Today’s HD class 8 truck is ~$25,000 (20%) more expensive because of emissions regulations, but this is offset by a ~50% mpg improvement vs. 20 years ago.
Interesting that it seems the real issue is with the “miracle” 4-cyl diesels that, unlike all others, don’t need urea injection/SCR aftertreatment to meet emission standards. Of course, as it turns out, they don’t meet emission standards in normal use. They probably *do* need urea and catalysts like everybody else, or will (when properly calibrated for emissions) run horribly. Shades of the CVCC & thermal reactors vs catalyst game back in the 1970s. As one story in R&T pointed out, the VWs seemed to be too good to be true, and it turns out they were.
Assume all testing is done on a chassis dyno, as with the Miata in the photo here. Could the engineer use wheel sensors (for ABS etc) to detect that the non-driven wheels are not moving, and change the engine calibrations for legal emissions levels for that situation? (Sorry, but I haven’t figured out to make this work with awd!) But thanks for a good article, as I haven’t heard any other speculation as to how they did this.