Volkswagen - How were they caught?

[maxparrish]

One might, in a generous moment, suppose that for once a government agency was actually doing routine policing - you know, the mega billions EPA and its six figure scientists and labs caught a major violator due to their bullet proof system of regulation.

Naw...you knew that couldn't be right.

Turns out it happened by accident. A little self-funded engine lab at West Virginia University got a little grant in 2012 to research and write a paper on how well American Diesels perform compared to German engines. The little group of five busied themselves with the exciting prospect of being the first to road test engines - a journal article for their CVs.

They flew out to California to be near the big State labs,

rented a BMW X5 sport-utility with a diesel engine from a car dealer there, and two Volkswagens—a Jetta and Passat—through a website that allows people to rent out their vehicles to strangers.

They installed emission-monitoring equipment on the vehicles. The devices piped exhaust into a series of measurement systems that could gauge how many pollutants the engines were spewing at different speeds. But unlike they would have done for standard tests for emissions, they didn’t run the cars on a stationary test bed. They took them out and drove them hundreds of miles—measuring pollutants along the way.

http://www.wsj.com/articles/volkswa...sed-by-routine-university-research-1443023854

They were perplexed. BMW emissions were as expected, but not Volkswagen. After numerous retests they accepted the obvious...Volkswagen must have put a defeat device or code into its control systems. The researchers did not even know how it was done, just that it was the only explanation.

What is curious is that this has been happening since 2009 and NO ONE in the EPA bothers to run verification checks on manufacturers? What do they do in the labs, rubber stamp official paperwork and take off early for golf?

 

[Derec]

Go to this link instead and click on the article. Google is a defeat device for the paywall.

The EPA (and various EU regulatory agencies) should have caught the scam, that's for sure. But WVU (WV is VW backwards. Coincidence?) is a public university.

Extra points (of Godwin variety) to WSJ for using a photo of Hitler, excited as a schoolboy over the model for the original VW Bug.

By the way, the original name was KdF (Kraft durch Freude) Wagen or Strength through Joy car. No wonder it didn't stick.

 

[bilby]

One might, in a generous moment, suppose that for once a government agency was actually doing routine policing - you know, the mega billions EPA and its six figure scientists and labs caught a major violator due to their bullet proof system of regulation.

Naw...you knew that couldn't be right.

Turns out it happened by accident. A little self-funded engine lab at West Virginia University got a little grant in 2012 to research and write a paper on how well American Diesels perform compared to German engines. The little group of five busied themselves with the exciting prospect of being the first to road test engines - a journal article for their CVs.

They flew out to California to be near the big State labs,

rented a BMW X5 sport-utility with a diesel engine from a car dealer there, and two Volkswagens—a Jetta and Passat—through a website that allows people to rent out their vehicles to strangers.

They installed emission-monitoring equipment on the vehicles. The devices piped exhaust into a series of measurement systems that could gauge how many pollutants the engines were spewing at different speeds. But unlike they would have done for standard tests for emissions, they didn’t run the cars on a stationary test bed. They took them out and drove them hundreds of miles—measuring pollutants along the way.

http://www.wsj.com/articles/volkswa...sed-by-routine-university-research-1443023854

They were perplexed. BMW emissions were as expected, but not Volkswagen. After numerous retests they accepted the obvious...Volkswagen must have put a defeat device or code into its control systems. The researchers did not even know how it was done, just that it was the only explanation.

What is curious is that this has been happening since 2009 and NO ONE in the EPA bothers to run verification checks on manufacturers? What do they do in the labs, rubber stamp official paperwork and take off early for golf?

What they do in the labs is test the cars.

What the EPA could not reasonably have been expected to know is that VW had explicitly written software to detect lab testing, as opposed to road testing, and to respond by adjusting the engine characteristics to burn cleanly, but less powerfully, when the car was tested in such conditions.

The VW cheat determined (from the lack of air-pressure and steering changes) when the car was running, with the wheels turning, but was neither moving through the air, not steering around obstacles; and concluded from these facts that it was on a 'rolling road' in a laboratory, and not on a real road.

This is not something that a reasonable testing authority could have expected; The only reason that VW got caught was because some amateur testers who didn't have the funding for proper lab tests did their resting on real roads (which is more prone to confounding factors than testing in a proper lab).

If the EPA did their testing on real roads in the great outdoors, the test results would vary due to weather and terrain conditions, and would not produce comparable results for vehicles tested on different days. Unfortunately for VW, the guys from WVU were prepared to tolerate this data degradation in order to save a few bucks.

But don't let reality get in the way of your ongoing spittle-flecked tirade against the very concept of regulation.

 

[Nexus]

What are the chances the engineers and executives responsible will most likely not be held accountable?

 

[bilby]

What are the chances the engineers and executives responsible will most likely not be held accountable?

The CEO has already resigned. It remains to be seen whether that protects him from any further penalties or punishments.

 

[Nexus]

I knew he resigned and others probably will as well. What I meant was be held accountable for the criminal fraud.

 

[bilby]

I knew he resigned and others probably will as well. What I meant was be held accountable for the criminal fraud.

I suspect that one of his motives for resigning would be to reduce the chance of that happening; The 'boy's club' may be able to convince themselves that being forced to resign is punishment enough, and of course the new CEO might have to accept a large fine for his corporation, but can deflect any suggestion of personal liability for actions taken before he was in the top job.

It's all about covering one's arse at that level in big business.

 

[DBT]

There is talk that the European Commission plans to ban petrol and diesel cars by 2050, at least in capital cities.

 

[angelo]

VW usually runs near last for reliability tests as well. They've had engine woes, transmission woes and other reliability issues. Yet people still buy them as they're [along with Toyota] the top selling car brand in many markets.

Luckily for them many people keep their cars for 2-3 years, and it's the poor sucker who buys it second hand who then starts to fork out big money to have them repaired.

 

[barbos]

Clearly Hitler is responsible for this.....again.

 

[angelo]

How many people have bought one these lemons thinking they're doing their bit for the environment! Wonder what they're thinking now!

 

[Achwienichtig]

What they do in the labs is test the cars.

What the EPA could not reasonably have been expected to know is that VW had explicitly written software to detect lab testing, as opposed to road testing, and to respond by adjusting the engine characteristics to burn cleanly, but less powerfully, when the car was tested in such conditions.

The VW cheat determined (from the lack of air-pressure and steering changes) when the car was running, with the wheels turning, but was neither moving through the air, not steering around obstacles; and concluded from these facts that it was on a 'rolling road' in a laboratory, and not on a real road.

This is not something that a reasonable testing authority could have expected; The only reason that VW got caught was because some amateur testers who didn't have the funding for proper lab tests did their resting on real roads (which is more prone to confounding factors than testing in a proper lab).

If the EPA did their testing on real roads in the great outdoors, the test results would vary due to weather and terrain conditions, and would not produce comparable results for vehicles tested on different days. Unfortunately for VW, the guys from WVU were prepared to tolerate this data degradation in order to save a few bucks.

But don't let reality get in the way of your ongoing spittle-flecked tirade against the very concept of regulation.

I'm not so sure about this. Most modern cars, not just these VW cars, have settings that detect laboratory testing conditions.

https://www.washingtonpost.com/news...ehind-how-volkswagen-tricked-emissions-tests/

Using a special engine setting for vehicle tests isn’t all that unusual, according to Consumer Reports. Most new vehicles do something similar because otherwise vehicles might interpret some of the testing procedures, like traction issues from being on rollers, as dangerous. But the problem here is that the EPA says the carmaker used its testing mode in an inappropriate attempt to beat the system.

So the software for detecting laboratory settings has been there for some time. The issue is more likely that the EPA functions on the assumption that cars are machines, but the reality is that cars are turning more and more into computers. Is there anybody in the EPA who even considered that computerization might affect testing?

Is it unreasonable to fault the EPA for not finding this cheat? Perhaps. However, it does seem very reasonable to expect the EPA to be up to the times with how computerization has changed the mechanics of cars.

 

[Jimmy Higgins]

I knew he resigned and others probably will as well. What I meant was be held accountable for the criminal fraud.

Hard to tell. This will be a big test for Governments enforcing rules on white collar crime. This could destroy VW.

 

[Jimmy Higgins]

I'm not so sure about this. Most modern cars, not just these VW cars, have settings that detect laboratory testing conditions.

https://www.washingtonpost.com/news...ehind-how-volkswagen-tricked-emissions-tests/

Using a special engine setting for vehicle tests isn’t all that unusual, according to Consumer Reports. Most new vehicles do something similar because otherwise vehicles might interpret some of the testing procedures, like traction issues from being on rollers, as dangerous. But the problem here is that the EPA says the carmaker used its testing mode in an inappropriate attempt to beat the system.

So the software for detecting laboratory settings has been there for some time. The issue is more likely that the EPA functions on the assumption that cars are machines, but the reality is that cars are turning more and more into computers. Is there anybody in the EPA who even considered that computerization might affect testing?

Is it unreasonable to fault the EPA for not finding this cheat? Perhaps. However, it does seem very reasonable to expect the EPA to be up to the times with how computerization has changed the mechanics of cars.

The cheat was a big one. They may start looking at all of the cars harder now.

 

[Bronzeage]

Emission testing for vehicles that are in use by the public is very expensive. Back in the 90's there was a proposal to require real tailpipe emission test for cities with the poorest air quality. Baton Rouge was one of the affected cities.

The test would require a service bay equipped with a dynamometer. This is a machine which holds a car in place and lets it simulate a test drive. At the time, each service bay would cost $77K. The cost for each inspection would be well over $200. The Baton Rouge area's air problem were caused by the combination of refineries, chemical plants, and cars. The real problem was there were not enough cars in the city, so the cost per car was prohibitive.

In the end, a compromise was made. 85% of real life emission failures were due to a worn gas cap, which let gas fumes into the air. A $300 gas cap testing device could eliminate 85% of the violations, so this was chosen over the $77k system.

The reason emission certification testing is done on a dynamometer and not on the road is simple. It's a matter of cost.

What Volkswagen did is a common practice in the world of auto racing. Auto racing rules exist to keep competition fairly even. This is not the same thing as "fair". For the past century, racing teams have exploited the rules by finding things that were violations of the spirit of the rule, but not specifically prohibited. When fuel tank capacity was reduced, which would force a race car to make more pit stops, one team installed a large diameter fuel line and routed it through the longest possible path between the tank and the engine. That's against the rules, today. Racing has hundreds of similar stories.

Volkswagen gamed the rules and got caught. The question becomes one of what the rules actually say. Volkswagen could take the racer's defense and say, "Our cars passed your inspection, so get out of our face."

 

[Derec]

There is talk that the European Commission plans to ban petrol and diesel cars by 2050, at least in capital cities.

Given the rules re congestion charges in places like London that's not that unlikely. Of course, by 2035 or so over half of all the new cars in EU and US will likely be electric (either battery or hydrogen fuel cell) anyway.
It becomes increasingly difficult to extract improvements of ICE engines. You bump up against theoretical cycle Diesel and Otto (i.e. gasoline engine) cycle efficiencies and competing interests of performance, fuel efficiency, pollution and cost. There are also benefits to banishing gas stations from cities with the resultaning vapor pollution among other things.
That said, 35 years is quite a long time.

 

[Jimmy Higgins]

Volkswagen gamed the rules and got caught. The question becomes one of what the rules actually say. Volkswagen could take the racer's defense and say, "Our cars passed your inspection, so get out of our face."

It'll depend heavily on how they marketed their cars. Besides, a BMW was tested and the readings were as expected.

 

[Bronzeage]

Volkswagen gamed the rules and got caught. The question becomes one of what the rules actually say. Volkswagen could take the racer's defense and say, "Our cars passed your inspection, so get out of our face."

It'll depend heavily on how they marketed their cars. Besides, a BMW was tested and the readings were as expected.

That's the court of public opinion, where the rules of evidence are not as strict. Ever since computers became the fuel system drivers, the fuel control program of every manufacturer was tailored to match the actual emission test conditions. The difference between test performance and street performance may have been small, but it's a given that real world emissions will be higher than test results.

The addition of steering wheel position sensors, which are not really part of the fuel control system, creates a data input that indicates the car is not actually on the street. The only time a car's tires are moving and there is not the slightest change in steering wheel position is during and emission test. VW engineers took advantage of this and wrote two different control profiles.

For myself, I don't think a lot of people really care. After several decades of discovering auto companies have ignored defects which actually killed people, this one doesn't seem so bad.

 

[Derec]

Emission testing for vehicles that are in use by the public is very expensive. Back in the 90's there was a proposal to require real tailpipe emission test for cities with the poorest air quality. Baton Rouge was one of the affected cities.

The test would require a service bay equipped with a dynamometer. This is a machine which holds a car in place and lets it simulate a test drive. At the time, each service bay would cost $77K. The cost for each inspection would be well over $200. The Baton Rouge area's air problem were caused by the combination of refineries, chemical plants, and cars. The real problem was there were not enough cars in the city, so the cost per car was prohibitive.

When I first started driving here in Georgia my first two cars were older than 1996 and thus had to get a real dyno test with a tailpipe probe. The testing was $25 everywhere and was obviously subsidized. My later cars were OBD2 equipped and could just be tested at "1996 and newer" stations which only read the code and could thus be cheaper ($15 per test or even lower are common now at "96 and newer" places). Last time I did an emission test the tester actually stuck a tailpipe in and had to rev the engine, but his dyno has long since been removed. He says the state requires (partial, due to no dyno) tailpipe test to be on all cars that come in that particular day (once a month) for verification purposes.

In the end, a compromise was made. 85% of real life emission failures were due to a worn gas cap, which let gas fumes into the air. A $300 gas cap testing device could eliminate 85% of the violations, so this was chosen over the $77k system.

And they miss all the emission failures from failing equipment. Great.

The reason emission certification testing is done on a dynamometer and not on the road is simple. It's a matter of cost.

It is cost prohibitive to do to every car, sure, but not for EPA to certify each model like government does for crash tests (which costs a lot too).

What Volkswagen did is a common practice in the world of auto racing. Auto racing rules exist to keep competition fairly even. This is not the same thing as "fair". For the past century, racing teams have exploited the rules by finding things that were violations of the spirit of the rule, but not specifically prohibited. When fuel tank capacity was reduced, which would force a race car to make more pit stops, one team installed a large diameter fuel line and routed it through the longest possible path between the tank and the engine. That's against the rules, today. Racing has hundreds of similar stories.

Yes, consider the dildo noses of last year's F1 cars due to regulations that said noses had to be low but this not being good for aero performance.

But racing is very different from production. This had a real life effect of additional pollution being released and VW should pay for it dearly and those directly responsible should be criminally charged.

 

[Derec]

The addition of steering wheel position sensors, which are not really part of the fuel control system, creates a data input that indicates the car is not actually on the street. The only time a car's tires are moving and there is not the slightest change in steering wheel position is during and emission test. VW engineers took advantage of this and wrote two different control profiles.

Which is deliberate cheating.

For myself, I don't think a lot of people really care. After several decades of discovering auto companies have ignored defects which actually killed people, this one doesn't seem so bad.

I think they will. And there is big difference between inadvertent defects and deliberate fraud, which is what this is.