GM accused of installing devices in diesel trucks to beat emissions tests
cbc.ca
cbc.ca
When the VW story first broke and I learned you could readily control the trade-off between performance and emissions using software in the engine controller of a diesel engine, I knew that everyone would do that. They would rationalize it in what ever way they could "Oh the stomping on the accelerator indicates an emergency so we go with maximum performance rather than emission compliance, it's to save lives don't cha know?"
FWIW, that's common on gas engines so there is some precedent. It's called power enrichment and can be adjusted in GM computers. When it kicks in depends on the car/truck it's installed in. But it uses a combination of manifold pressure and throttle position and rpm to decide whether or not to command extra fuel. You can see the params HP Tuners uses to configure the PE circuit here: https://www.hptuners.com/help/vcm_editor_parameters_gm_eng_f...
Power enrichment is a different function and commands a new afr for peak power (generally around 12.5:1) instead of trying to maintain around stoich.
I've been wondering why the researchers at WVU who found the VW cheating have not gone on to test a larger set of vehicles to see who else is cheating.
You'd think someone would fund broader testing...
But you're right, we don't have proof right now. I think the logical proof, that other companies are selling similar vehicles at similar price points that are "passing" the emissions testing is proof but obviously not of the legal variety.
That is THE tradeoff everyone contends with. The regulations on emissions and fuel economy are directly at odds with each other. The things you do to get more efficient often result in incomplete combustion - strange as that sounds.
The issue is that higher temperature combustion, while more efficient, produces much more NOx. So you get more energy per unit fuel at higher temperatures, but you also get more pollutants. Lower temperatures (less oxygen) result in "cleaner" combustion (essentially, more CO2 and N2, less NO and NO2), but produce much less energy per unit fuel.
https://en.wikipedia.org/wiki/Lean-burn
There really is a lot at play, but I stand by the statement that "clean" and "efficient" are directly at odds with each other.
For example, electronic fuel injection is cleaner and also much more efficient than using a carburetor.
An independent review of the impact of California's CARB emissions regulations found that if you totally ignored the health and climate change benefits, the regulations were still a net benefit just because they encouraged lighter cars, which lead to less death and injury on the roads.
Possibly you're thinking, but if my car was bigger than everyone elses, then I'd kill them, instead of them killing me. But unfortunately, everyone else is thinking the same thing. A classic case for regulation leading to improved outcomes for everyone.
The tradeoff between efficiency and power is ok, but what they did is have software that detected when an emission test was being run and created a fake engine profile that only made fuel efficient tradeoffs, and when the emission test stopped the engine went back to the regular tradeoffs.
That was their crime. They hid what settings they were really picking and specifically wrote software to defeat an emissions test.
> That was their crime. They hid what settings they
> were really picking and specifically wrote software
> to defeat an emissions test.
Except for the tiny detail that the law explicitly forbids using special emission control modes during stationary test bench usage. At least in the VW case, it was not a relaxed interpretation of the rules.
I'm not taking the blame out of carmakers, but it is very easy to end up with regulations that when taken together are not feasible.
Imaginary example: you need a capital/asset ratio of at least 20% (which means capital/liab. of at least that, by definition). But you also require at least 30% of the liabilities to be long-term debt. Another regulation says that 50% of loans must be financed with deposits, and so on. The end effect is that there is little margin for these banks to decide on their liabilities. At the extreme, if we impose unfeasible restrictions, they might just end up skirting some of the rules.
> ... it is very easy to end up with regulations that
> when taken together are not feasible.
It is a classic prisoner's dilemma problem. If everyone is honest, you make cars that are good for the environment but lack the performance, if one can "defect" while looking like they haven't (aka they cheat) they take all the money.
And yet they get caught (and always will) and lose the long game.
Sayonara, VW.....
Big corporations used to have bean-counters that calculated the cost of getting caught and they cheated only when the risk vs. damages had a cost upside.
Corporations now just don't appear to give a flying fuck about what lands in the lap of the next CEO. They'll get their golden parachute, right?
1. Elevated driving platform.
2. Room for extra passengers and cargo.
I agree that the vast majority of pickups (utes) sold in the U.S. are much larger than necessary for their usage and in comparison with pickups sold outside of North America.
3. A slanted liftgate that sacrifices cargo space so their crossover doesn't like their parents' station wagon and/or minivan.
It was the emissions restrictions that pushed so many consumers to SUVs in the first place.
After that it was only vehicles with hinged doors and 8 passengers so...SUV or commuter van.
edit: also I live in a northern state in the middle of nowhere. 4 wheel drive and high clearance were not insignificant factors for us.
If you want two bench seats in the 21st century your options are basically all SUVs
Americans for the most part do not like diesel cars. It's really not even worth testing them.
Trucks are another matter, but they have different emissions standards.
The problem isn't that they are changing how they perform depending on circumstance, it's that they've defined a circumstance specifically for "we're being tested" and change how it functions based on that, which means the test no longer has any bearing on reality. That's the difference between a good feature and a defeat device.
The EPA (and other regulators) allow engines to emit more NOx and Particulate under certain circumstances to protect the engine.
Each set of parameters where an engine will do this is called an auxiliary emissions control device [AECD], even where it is just programming, not a separate physical device.
For instance, FCA recently got in trouble for not disclosing all of their AECDs [0] - the AECDs may not have been illegal, but FCA neglected to inform the EPA of all of them.
As noted elsewhere in comments, very very few (if any) vehicle achieve the same emissions or efficiency in the real world as in a test lab.
The disparity between test and real world is not necessarily illegal.
AECDs are not necessarily defeat devices.
0. http://www.detroitnews.com/story/business/autos/chrysler/201...
2017 Shipments of Mercedes Benz Sprinters were actually prevented from being delivered to dealers for a while. The last I'd heard, 6-cylinder models were released, but 4-cylinder models were still being held by the EPA.
While I'm sure at some level the EPA enjoys punishing German auto makers, hopefully the same accusations being leveled towards US car makers will encourage them to adopt a more fair and reasonable stance, assuming these are not actually "defeat devices".
[1] http://autoweek.com/article/diesel/mercedes-parent-company-d...
This is hilarious - this is the firm that made tons of money off the VW case, now they are going after the money and cutting out the middleman (EPA)
VW expects to pay $7.3B for 11 million cheating cars, so less than $700 per vehicle. GM will probably get about the same deal. Considering the profit margin on these vehicles in particular (big trucks), cheating was probably a good call financially.
VW could meet the emissions standards, and went through the actual testing and passed. It had such an effect on either fuel economy or long term reliability however that they cheat on the road.
However it's notable that the only VW vehicles that fail are the ones that don't use diesel exhaust fluid (DEF). This is something that you refill periodically and it's sprayed into the exhaust to reduce a lot of the harmful exhaust. That system costs a couple of thousand dollars, so they eliminated it and claimed that they found a way around it.
The GM trucks, in contrast, have the DEF systems, just like all of the Mercedes and higher end VW diesel cars do.
Too many commentators are taking the word of a legal firm that got some easy prey and now hopes there's blood in the water.
[1] "Dieselgate — a year later" https://youtu.be/SjlHtLux9vc
https://www.honestjohn.co.uk/company-car-tax-advice/news/whi...
VW has since announced a EA288 engine that deprecates the old one, and the new engine uses DEF.
But sure, there is suspicion that a lot of other makers cheat to degrees. Use less DEF and it's more convenient and less costly for your users. Use the EGR less and improve fuel economy and reduce possible maintenance issues. Etc. That was not the primary problem, though.
Of course this truck probably emits more NOx than an emissions-compliant diesel car would, but it also burns about twice as much fuel. The real problem with DEF in passenger cars is those "luxury" manufacturers that require a special tool to fill the DEF tank, as if DEF were some rare or dangerous substance.
Where are you getting this idea from? My recently-relinquished 2014 Passat was one of the affected vehicles, and it most certainly had a DEF system.
The vast bulk of the controversy was the EA189, which did not use DEF in any form. This is literally a seconds search away.
> However it's notable that the only VW vehicles that fail are the ones that don't use diesel exhaust fluid (DEF).
That's not "the bulk", that's all. It is also demonstrably false. Your statement that I quoted previously also did not express or imply "the bulk".
> The vast bulk of the controversy was the EA189, which did not use DEF in any form. This is literally a seconds search away.
My Passat had an EA189 TDI and used DEF. "A seconds search" for EA189 DEF resulted in this link: http://www.autoguide.com/auto-news/2016/02/vw-reportedly-buy...
> Vehicles equipped with updated versions of the EA189 engine that used DEF and SCR systems are considered fixable (this included my car)
So, I ask again, from where did you get the idea that a) only non-DEF-equipped cars were affected and b) that no EA189 engines had a DEF system?
https://youtu.be/7t4paclIwuU?t=13m15s
And for GM, the linked video already shows Opel installed defeat devices, which is a daughter brand. So it's just natural that the same "technology" shows up in other GM cars.
Isn't it less that they couldn't figure out how to meet the standard, but rather that they did not like the consequences? Evidently the cars in question could operate in a lower-emission mode.
This would actually be unfair, since VW does not have to pay a cent in the EU (the prosecutor being an employee of the state of Niedersachsen, which is also a major stockholder in VW and on the advisory board, apparently does not help).
One day a friend brought over special goggles. They were covered in black vinyl to block your vision completely. That day we all found out that every one of us cheated the same way. When playing with the new goggles, suddenly it was: “this game sucks” and: “let’s do something else”.
It was eye-opening (sorry).
Cheating on emissions tests strikes me the same way. “All the other car companies seem to be meeting their targets, how can we keep up?”
Well, it turns out that to be as good as the other players, you have to cheat just like they do.
If you mean doing the same kind of cheating that your competitors are doing then this implies knowledge of the method so you don't need to cheat. You can level the playing field by instead publicizing the fact that your competitors are cheating. You don't even need to name names. Just lobby governments to implement tests that will uncover it. Get tests done on your competitors vehicles by independent laboratories, ponder in public the question that reverse engineering the engine management code might be worthwhile.
On the contrary. When I was cheating at Marco Polo, I thought I was the only one. That I wasn't as good as the other players who were playing honestly, so I cheated to make up for the deficiency.
Your point still stands, though. These aren't kids in a pool. They're engineers who should be testing the competition's product, and discovering the defeat devices.
But if both parties are cheating, what's the incentive to bring light to this problem?
The best case scenario for telling on competitors would be a larger fraction of a smaller market, worse outcomes would involve a similar or smaller fraction of an even smaller market (after symmetric response accusations hitting your own brand). You don't sell Marlboros by pointing out that smoking Lucky Strike might cause cancer.
No, but we could at least hold out on castigation until it's proven they've broken the law. A class action lawsuit isn't proof of anything.
http://money.cnn.com/2016/04/26/news/companies/mitsubishi-ch...:
> The automaker said last week that it had used improper fuel economy tests on hundreds of thousands of vehicles, including some sold to Nissan
This story suggests a possibility I hadn't considered yet, which is that it may already be economically impossible to build a car that complies with the law.
At this point, while you are welcome to castigate the automakers for cheating, one must also consider that the authors of the regulation may bear some responsibility here too.
And I do mean both bear some responsibility; the cheating of the car companies has prevented the feedback from getting to the regulators that they've created an impossible situation. But they may still have created an impossible solution, and it may mean that people just going "Well just comply with the regulations then!" are proposing a non-solution to the problem without realizing it.
It is also worth pointing out that three years ago the idea that regulations were already impossible to comply with would have been met with hoots of derision, and probably someone is going to reply to me to dismiss it even now. We're asking auto companies here to forfeit billions of dollars, but the regulators themselves have no risk if they are creating an impossible situation.
Just because someone wrote a regulation somewhere doesn't mean it's good, or possible, or economical.
I have no hard data, but my suspicion is that it is not economically impossible in any absolute sense. That is, it is possible to create cars that meet the standard, that can be sold to consumers at a price that allows them to be built. They may have lower power than customers desire, but what else are they going to do? Walk?
What is probably not possible is to do that when they have to compete against cars that only pretend to meet the standard. The cars that only pretend will cost less, or have more performance. If the buyer compares a car that meets the regulation and one that only pretends to, the one that only pretends to is a more desirable car, and will get the sale.
> Just because someone wrote a regulation somewhere doesn't mean it's good, or possible, or economical.
Yup. That's true of regulations on more than just cars, too.
What if the car companies could make a much cleaner diesel engine, but it just wasn't cost efficient?
What if this kind of regulation is the stick to go with the various green regulation carrots (tax credits on fuel efficient vehicles, subsidies for renewables, etc...)?
What if making lawmakers responsible for legislating something out of business has unintended effects? Consider that in some places tobacco companies are suing governments over labeling, imagine the lawsuits if a car company could sue when they didn't like emission laws.
I agree it is easy to look at this with a sympathetic eye to automakers, but I think we cannot condone cheating when other alternatives exists. It may well be impossible to build a clean enough diesel engine. If so we should drop diesel and figure out something with another energy store
Empirically, VW released patches for their cars to bring the non-test characteristics in line with their test-mode characteristics.
The outcome was that the diesel-engine vehicles had reduced fuel economy and acceleration. [1]
Most customers still enjoyed their cars after the patch was applied.
1. http://www.consumerreports.org/cro/cars/guide-to-the-volkswa...
Did the patched cars also meet mileage standards? How close to the limit were they? (Or are there no mileage standards for diesels?)
2. Collision resistance has improved as weight has decreased.
3. Did they meet mileage standards? Maybe. Probably. 46mpg (the resultant fuel economy after the patch) appears to be outstanding even for diesel vehicles. Mileage is imposed on individual models by year targets and also manufacturer-wide. That is why cars in general are sold with much lower margins: to help manufacturers offset the poor fuel economy and emissions of larger vehicles.
We can only dream that we will eventually reach the point where cars cost enough to offset their cost to society.
All the more reason to consider EVs a moral good.
Jokes aside, there is no "offical" source for my claim. But the fact that they use a linux based system yet have no place to download it is all the evidence you need really. You can find a number of people detailing their (unsucessful) GPL requests in forum posts etc. by searching for "tesla gpl" or "tesla gpl violation"
Lets put it in perspective. A lawn mower (Chain Saw, Weed Wacker, Small Engine Device) creates more pollution in 1 hour of use than one of these vehicles creates in 8 hours of driving. Even motor cycles will emit more pollution during use. Can you imagine how much extra pollution is creating daily just because of tire under inflation? or oil changes that are too frequent and not necessary?
Just think of all the equipment out there that doesn't have any pollution controls and these vehicles could be considered green in comparison.
Lets see what the outcome is....
PSA: The best thing you can do for the environment is keep your vehicle in good running condition as long as you can (10-15 years).
That's short. My daily driver is going on 20 and doing just fine.
I think the best thing drivers can do for the environment and society's health is to stop driving and start cycling, walking and using public transport.
Lets not pretend this isn't the first or 10th time they have been caught.
If you own a car, you are involved.
That's over the top in this context. It makes it sound like the trucks are poisoning the owners of those trucks.
I apologize if the above comes across as a rant or a whine. I can at most ask the legislator in my constituency to investigate, but that is yet another game of numbers.
The fact the tests are so reliably circumvented is a failure of regulation. It's evident that until recently regulators were not testing the tests to see if they are predictive of real world performance.
If you're comparing 2 trucks, one with a DEF system and one without, chances are as an average consumer, you're going to take the one without because it's cheaper to run and less effort for you.
Heck, the other day I was looking at some electronics on a very large Chinese shopping site and ran across a device that can emulate DEF systems in 8 different manufacturers vehicles with no actual DEF consumption occurring for ~$15. People are lazy, cheap and just don't give a shit.
http://abcnews.go.com/Business/wireStory/lawsuit-alleges-gm-...
performance vs following rules check
performance directly translates to better perception by public and therefore more money check
small monetary repercussions if discovered in comparison to amount to be gained by cheating check
arms race of detection/regulation no check yet (on the automotive side) but it will be interesting to see how this develops
I understand your sentiment, it feels more bad when a person or group goes to greater lengths to hide wrongdoing.
Just remember that VW is a bit like Nestle; you also have to stay away from Audi, Bentley, Bugatti, Lamborghini, Porsche, SEAT, Škoda, Ducati, MAN and Scania (not that the last two matters to you unless you're buying a bus).
Owners of most other brands don't have to put up with that kind of stuff. Even my friends with pre-bailout low-end GM and Ford cars haven't had as many problems.
They are probably the most honest of any car company at the moment. They cant risk another scandal. Likely the most undervalued by the market as well.
Also, consider a used vehicle. Used vehicle prices are being pushed down due to lots of vehicles coming off lease and subprime auto credit deteriorating simultaneously.
In my case, I've decided my future vehicles will be full-electric - just waiting for the right one (hopefully in a month or two).
What makes you think that?
> VW has one of the biggest R&D divisions in the industry. If they couldn't do it, nobody else could either.
Do what? Make performant cars? Their gasoline (petrol) cars weren't implicated in this scandal. After the patch, the cars were still performing relatively well, losing about a second in the 0-60 time, and losing about 4mpg.
Am I missing something? My understanding is that this was all because they couldn't meet emissions targets without using a certain amount of DEF. They calculated that to do that, the customer would have to refill the DEF tank "too soon" (meaning sooner than every regular service interval at the dealership, which is at least 10,000 miles).
So the simple fix is, in my mind: make the car use a lot more DEF so that it can generate the power and economy advertised, but the DEF tank will have to be refilled far more frequently (perhaps every 1000 miles, guessing). The driver will have to do it themselves obviously. And add some software code so that the car refuses to operate if the tank goes empty. If drivers can figure out that they need to put fuel (gas or diesel) in their car to make it go, and that it's going to die on the side of the road if they run that tank out, then they can do the same with a 2nd tank of fluid too.
The main violation appears to be emitting excess amounts of NO2, a pollutant, and some particulates.
It appears your analysis is right.
What would happen if the DEF was allowed to run empty?
And again, I'm not even sure I'm that well informed, I could very well be mistaken and over-simplifying things. If that's the case, I hope someone will correct me, but so far no one's said anything, but that's why I put that disclaimer in there several times.
1) They couldn't do it.
2) Their margins would be thinner.
3) Their competitors are cheating, and emissions controls reduce performance,
so everyone would buy the cheating competitor.
Also, everyone has dirt on everyone else.
"First they came for the trade unionists,"
"Dunning-Krueger Effect!"
"Wouldn't have happened if they used a secure coding language, like Rust"
"That's a _____ fallacy." [Link to Wikipedia list of logical fallacies]
Sorry for the OT post, I'm getting sick of the oft-repeated HN expressions that don't add to the discussion. We can do better.
And I forgot one, "I know I'll be downvoted for this, but..." ;)
Language is built out of little phrases which come to be associated with an understood meaning and tone, and are treated atomically (sometimes the individual words within the phrases go out of currency and are never used alone). We all use these little atoms as shortcuts all the time, because inventing a completely new phrase to represent every thought would be less efficient. Sometimes the phrases were originally technical terms (for instance, many come from sailing or driving carriages or manual farm labor or ancient styles of combat ...) which have been stripped of their original context; other times they come from literature, like the “shocked, shocked” example, or hundreds of examples from Shakespeare plays.
Douglas Hofstadter has written some nice pieces about this topic, e.g. https://prelectur.stanford.edu/lecturers/hofstadter/analogy....
May 20, 2017 (comments)
An internet accurately describes Google's AMP program for what it is -- a barely-disguised attempt to shove more of the internet into itself, the better to track the living shit out of every human being on earth. Hackernews has attempted to avoid this program by using it heavily. Half the comments are Hackernews insisting that it is not only impossible to compete with Google in any way but also insane to try. Some Googles show up to assure everyone that this is for the greater good, and if you'd all just stop talking and get in line the whole process will be nearly painless. A few Hackernews suggest that AMP's feature set can be replicated with a "stop shoving every fucking possible line of javascript into every single pageload" approach, but they are quickly chloroformed and edited out of past Christmas photos.
How is this anything but genius? Next thing you tell me you do not approve of yosefk's excellent C++ FQA?
Seems like a conflict of interest to me.