5 ms·

> There is no suggestion that any of the cars tested broke the law on emissions limits or used any cheat devices. Since all these cars passed the laboratory te

> There is no suggestion that any of the cars tested broke the law on emissions limits or used any cheat devices. Since all these cars passed the laboratory test, that means there must be something very wrong with the test and the way the standards are enforced. Or all car makers do cheat.

stinos 11y ago
I thought it was common knowledge the lab tests are a joke? Has been posted on HN as well IIRC.
- f_allwein 11y ago
  I thought it was more a case of "some car manufacturers cheated" and not "almost all diesel cars emit several times more than they should". How is this possible?
  - Spare_account 11y ago
    The test is accurate and measures emissions very well. The cars emit NOX at levels lower than the legal limits during the test. There is no mistake during the test and no cheating (apart from VW). However, when the cars are used by normal drivers, the usage patterns are so significantly different than the test sequences that the emissions profile changes. In the majority of cases the NOX emissions from cars used by normal people exceed the EU limits. This does not, as I understand it, mean that the cars are illegal. The cars all passed the tests in the test environment. That is the requirement and they met the requirement. No law is being broken but that is only because the law is not fit for purpose.
  - makomk 11y ago
    The tests simply don't match modern real-world driving patterns. For example, apparently real-world drivers tend to accelerate much more rapidly than in the test sequence. It doesn't just affect diesel cars either; they tend to over-estimate the fuel efficiency of all cars and presumably under-estimate their pollution levels too.
  - DanBC 11y ago
    Some manfs were cheating - they increased tire pressure to levels that would be dangerous on roads, for instance. Some manfs were building to the type approval test, but not for real world driving. A car needs to accelerate from 0 kph to 100 kph over a certain time. Normal road drivers would accelerate faster than this test, and acceleration is when the vehicles emit a lot of stuff. The tests are run on engines at their normal operating temperature. So cold engines are allowed to emit more stuff to protect the engine. Many engines in real world use would be at this cold temperature. The US has particularly stringent testing, and it might not be possible to meet those tests. They should probably have done other stuff to reduce diesel vehicles - taxes and duties on the vehicles and the fuel, with exemptions for commercial operators. It's clear now that manfs build to the test, and that type approval testing needs to include better real world driving testing.
  - lb1lf 11y ago
    Also, the number of car manufacturers outnumber the number of car engine manufacturers - so if one company which actually makes engines (cough) optimizes their engines for testing, several car manufacturers may be caught with their trousers down. (Which, I am sure, will lead to all sorts of interesting lawsuits.) For instance, the engine in cars as different as the Land Rover Defender (sadly, no longer among us) and the Volvo, say, S80 is basically the same Ford-engineered unit.
- SideburnsOfDoom 11y ago
  > I thought it was common knowledge the lab tests are a joke it was common knowledge for a long time that the lab tests were not the same as real world driving. But we thought that, even though the lab tests were better, there was some meaningful relationship. As it turns out, they are wildly divergent in a lot of cases. And the repercussions are ongoing. Long-term, this might turn out to a turning point in the history of internal-combustion engines.
Someone 11y ago
A bit/lot of it may be "optimize for the test", not cheat. They know at what temperature the test is done, so they optimize the chemistry for that case. They know it will be done with new cars, so they do not optimize "emission over the car's lifetime". They also have drivers who are above average. That will help with manually shifted cars. And of course, they pick the best car they can produce, and prepare it optimally. Tyre pressure will be optimized, the car will be perfectly polished, if that helps, etc. Cheating starts when they tape shut doors to decrease air resistance, or when they explicitly detect "test is being run" and act upon it. IMO, the real problem is twofold: firstly, the existing tests were designed in such a way that their results do not reflect what happens in the real world, and secondly, some manufacturers went criminally far in optimizing for it.
- jacquesm 11y ago
  "Dynamometer testing is already supplemented by on-road testing using portable emissions equipment, lab testing of cars borrowed from consumers, audits of other industry emissions labs and examination of vehicles chosen randomly from assembly lines, said Valentine, the EPA spokeswoman." from: http://www.autonews.com/article/20150929/OEM11/150929807/epa-should-do-more-road-emissions-tests-critics-say http://www.autonews.com/article/20150929/OEM11/150929807/epa... So the manufacturer apparently does not get to chose the vehicle, which is one reason why the cheating code was present in all VWs with that engine model rather than just in the vehicles used for the testing.
  - cptskippy 11y ago
    So the manufacturer apparently does not get to chose the vehicle, which is one reason why the cheating code was present in all VWs with that engine model rather than just in the vehicles used for the testing. Are you implying that it's a bad thing that a manufacturer is forced to implement cheats on all cars rather than just provide a rigged car for test?
    - jacquesm 11y ago
      No, that's not what I'm implying, no idea where you got that idea.
D_Alex 11y ago
>Since all these cars passed the laboratory test, that means there must be something very wrong with the test and the way the standards are enforced. It depends how the law is written. Case a) - the cars must never emit more pollutants than specified in these rules - then yes, you are correct. The car manufacturers would be rightly concerned about the wide range of situations they would need to account for. Case b) - the cars must not emit more than the specified amount of pollutants under these specific test conditions (implying hopefully "we understand that the real world performance will differ, and are willing to accept this for the sake of simplicity in testing/lawmaking") - in this case the system is working as intended. I am not sure which case we are dealing with, maybe someone can advise us.
TheCondor 11y ago
It's all a bit of a sham, a lot of politics in play with it. If NOx was a real concern, the cars would be required to have realtime sensors, the check engine light would come on when they misbehave. All the manufacturers knew. The testers almost certainly knew
- MadManE 11y ago
  If NOx was really a priority, then the large producers would be subject to increased regulations rather than the tiny fraction that's produced by individual passenger vehicles.

stinos 11y ago

I thought it was common knowledge the lab tests are a joke? Has been posted on HN as well IIRC.

f_allwein 11y ago
I thought it was more a case of "some car manufacturers cheated" and not "almost all diesel cars emit several times more than they should". How is this possible?
- Spare_account 11y ago
  The test is accurate and measures emissions very well. The cars emit NOX at levels lower than the legal limits during the test. There is no mistake during the test and no cheating (apart from VW). However, when the cars are used by normal drivers, the usage patterns are so significantly different than the test sequences that the emissions profile changes. In the majority of cases the NOX emissions from cars used by normal people exceed the EU limits. This does not, as I understand it, mean that the cars are illegal. The cars all passed the tests in the test environment. That is the requirement and they met the requirement. No law is being broken but that is only because the law is not fit for purpose.
- makomk 11y ago
  The tests simply don't match modern real-world driving patterns. For example, apparently real-world drivers tend to accelerate much more rapidly than in the test sequence. It doesn't just affect diesel cars either; they tend to over-estimate the fuel efficiency of all cars and presumably under-estimate their pollution levels too.
- DanBC 11y ago
  Some manfs were cheating - they increased tire pressure to levels that would be dangerous on roads, for instance. Some manfs were building to the type approval test, but not for real world driving. A car needs to accelerate from 0 kph to 100 kph over a certain time. Normal road drivers would accelerate faster than this test, and acceleration is when the vehicles emit a lot of stuff. The tests are run on engines at their normal operating temperature. So cold engines are allowed to emit more stuff to protect the engine. Many engines in real world use would be at this cold temperature. The US has particularly stringent testing, and it might not be possible to meet those tests. They should probably have done other stuff to reduce diesel vehicles - taxes and duties on the vehicles and the fuel, with exemptions for commercial operators. It's clear now that manfs build to the test, and that type approval testing needs to include better real world driving testing.
- lb1lf 11y ago
  Also, the number of car manufacturers outnumber the number of car engine manufacturers - so if one company which actually makes engines (cough) optimizes their engines for testing, several car manufacturers may be caught with their trousers down. (Which, I am sure, will lead to all sorts of interesting lawsuits.) For instance, the engine in cars as different as the Land Rover Defender (sadly, no longer among us) and the Volvo, say, S80 is basically the same Ford-engineered unit.
SideburnsOfDoom 11y ago
> I thought it was common knowledge the lab tests are a joke it was common knowledge for a long time that the lab tests were not the same as real world driving. But we thought that, even though the lab tests were better, there was some meaningful relationship. As it turns out, they are wildly divergent in a lot of cases. And the repercussions are ongoing. Long-term, this might turn out to a turning point in the history of internal-combustion engines.

Someone 11y ago

A bit/lot of it may be "optimize for the test", not cheat. They know at what temperature the test is done, so they optimize the chemistry for that case. They know it will be done with new cars, so they do not optimize "emission over the car's lifetime". They also have drivers who are above average. That will help with manually shifted cars. And of course, they pick the best car they can produce, and prepare it optimally. Tyre pressure will be optimized, the car will be perfectly polished, if that helps, etc. Cheating starts when they tape shut doors to decrease air resistance, or when they explicitly detect "test is being run" and act upon it. IMO, the real problem is twofold: firstly, the existing tests were designed in such a way that their results do not reflect what happens in the real world, and secondly, some manufacturers went criminally far in optimizing for it.

jacquesm 11y ago
"Dynamometer testing is already supplemented by on-road testing using portable emissions equipment, lab testing of cars borrowed from consumers, audits of other industry emissions labs and examination of vehicles chosen randomly from assembly lines, said Valentine, the EPA spokeswoman." from: http://www.autonews.com/article/20150929/OEM11/150929807/epa-should-do-more-road-emissions-tests-critics-say http://www.autonews.com/article/20150929/OEM11/150929807/epa... So the manufacturer apparently does not get to chose the vehicle, which is one reason why the cheating code was present in all VWs with that engine model rather than just in the vehicles used for the testing.
- cptskippy 11y ago
  So the manufacturer apparently does not get to chose the vehicle, which is one reason why the cheating code was present in all VWs with that engine model rather than just in the vehicles used for the testing. Are you implying that it's a bad thing that a manufacturer is forced to implement cheats on all cars rather than just provide a rigged car for test?
  - jacquesm 11y ago
    No, that's not what I'm implying, no idea where you got that idea.

D_Alex 11y ago

>Since all these cars passed the laboratory test, that means there must be something very wrong with the test and the way the standards are enforced. It depends how the law is written. Case a) - the cars must never emit more pollutants than specified in these rules - then yes, you are correct. The car manufacturers would be rightly concerned about the wide range of situations they would need to account for. Case b) - the cars must not emit more than the specified amount of pollutants under these specific test conditions (implying hopefully "we understand that the real world performance will differ, and are willing to accept this for the sake of simplicity in testing/lawmaking") - in this case the system is working as intended. I am not sure which case we are dealing with, maybe someone can advise us.

TheCondor 11y ago

It's all a bit of a sham, a lot of politics in play with it. If NOx was a real concern, the cars would be required to have realtime sensors, the check engine light would come on when they misbehave. All the manufacturers knew. The testers almost certainly knew

MadManE 11y ago
If NOx was really a priority, then the large producers would be subject to increased regulations rather than the tiny fraction that's produced by individual passenger vehicles.