OT: VW cheating emissions on TDI vehicles

[1centaur]

That last question is the most interesting one. It is somewhat hard to believe this decision was made at the CEO level, though it's possible a cover-up decision was made there. But it also seems unlikely it was made at some minor software designer's level. Somebody in the chain felt some pressure to achieve something that could not be achieved by legitimate means. Was the pressure technical or monetary or both? What sign-off standards does VW have? Did the outside software provider (Bosch) know? And then when the CARB failures started happening, at what point did higher levels of management know? And then what decision(s) did they make?

There is a good reason this is an external investigation.

[oldpotatoe]

Originally Posted by LJohnny

I am confused. So if this has no impact on the performance of the vehicle then why did they turned it off and only to be on while tested for emissions?



Sent from my iPhone using Tapatalk

And how did it 'know' that emissions were being tested? Know to 'turn it on'? And this to just prevent the cost of replacement?

[yngpunk]

Originally Posted by christian

There won't be any performance impact of mileage impact - the NOx catalyzer is effectively a heated element in the exhaust pipe that traps specific exhaust particles. It's all downstream from the combustion chamber. The fix is to put a new program on, or "reflash" the engine management computer so the computer tells the NOx catalyzer to stay on all the time.

The only impact will be you'll have to replace the NOx catalyzer more frequently. Perhaps every 30-40,000 miles rather than never. But given this, consumers will never pay a dime for that.

My understanding is that the effected vehicles use a LNT and not an SCR and in addition to the LNT needing to be replaced more frequently it will also impact performance and fuel economy

[christian]

Originally Posted by yngpunk

My understanding is that the effected vehicles use a LNT and not an SCR and in addition to the LNT needing to be replaced more frequently it will also impact performance and fuel economy

Yes, the operation of the LNT would require something like a max 1.4% penalty in fuel economy during lean engine combustion. This, to me, falls under "no mileage impact." But yes, fuel economy would reduce for some (not all) combustion conditions. If your overall fuel economy goes from 38 to 37.7 mpg, it would be difficult to even detect under normal driving protocols.

[christian]

Originally Posted by oldpotatoe

And how did it 'know' that emissions were being tested? Know to 'turn it on'? And this to just prevent the cost of replacement?

The CARB/EPA test protocol is well known by the industry; the computer was specifically programmed to identify this case, based, presumably, on relative wheel speed, and specific throttle positions and rpm ranges. This would be enough to tell the computer the car is running on a dyno and subject to a specific test. It should be noted that the car isn't programmed to cheat on periodic smog checks (for which most diesels are exempt anyway), but rather to cheat on the specific test the federal government uses to certify the car for sale in the US.

And yes, to prevent the cost of replacement. The part in question is expensive, and VW offered an 8-year warranty on it.

I've programmed engine control units (ECUs) in race cars, including doing something similar (engaging a specific protocol based on a set of difficult-to-replicate conditions) to what VW did. It's not rocket science. It does require a laptop.

[oldpotatoe]

Originally Posted by christian

The CARB/EPA test protocol is well known by the industry; the computer was specifically programmed to identify this case, based, presumably, on relative wheel speed, and specific throttle positions and rpm ranges. This would be enough to tell the computer the car is running on a dyno and subject to a specific test. It should be noted that the car isn't programmed to cheat on periodic smog checks (for which most diesels are exempt anyway), but rather to cheat on the specific test the federal government uses to certify the car for sale in the US.

And yes, to prevent the cost of replacement. The part in question is expensive, and VW offered an 8-year warranty on it.

I've programmed engine control units (ECUs) in race cars, including doing something similar (engaging a specific protocol based on a set of difficult-to-replicate conditions) to what VW did. It's not rocket science. It does require a laptop.

Get it-testing, the CARB/EPA test, not when I get Colorado emissions tested before I get my license plates..10-4. I'm assuming this has nothing to do with this, or does it? Not sure if my TDI is exempt or not.

[goonster]

Originally Posted by 1centaur

But it also seems unlikely it was made at some minor software designer's level. Somebody in the chain felt some pressure to achieve something that could not be achieved by legitimate means.

1. No software engineer has the authority to approve something like this. There would be multiple approvals at the director level, at least.

2. There is more than one chain. All major mfg companies have a separate Quality organization, with separate reporting structure up to at least the VP level, to prevent active or passive collusion. On top of that, there is probably a separate compliance department, again with separated "chain of command" so peers cannot be pressured.

Originally Posted by 1centaur

Did the outside software provider (Bosch) know?

They may well have known, but their contribution is compartmentalized, and they should be off the hook. They are responsible for making the engine run the way they are told, not for it being legal in market Z, or what comes out of the tailpipe.

Originally Posted by 1centaur

And then when the CARB failures started happening, at what point did higher levels of management know? And then what decision(s) did they make?

Not CARB failures per se, but regulatory inquiries re. third party findings.

Yup, that's where the coverup could become worse than the crime.

[christian]

I'm not sure we know yet. CARB/EPA isn't saying that that was affected - just the certification test.

But the dumbest (simplest) programming for this would be to use the ABS wheel sensors to tell anytime the car is on a dyno (which would encompass almost all testing):

Get Wheelspeed(Rear-Left) as WRL (Cars have wheelspeed sensors on all four wheels to allow the usage of ABS and traction control systems)
Get Wheelspeed(Rear-Right) as WRR
Get Wheelspeed(Front-Left) as WFL
Get Wheelspeed(Front-Right) as WFR
Get RPM as RPM
WheelspeedRear = avg(WRL, WRR)
WheelspeedFront = avg(WFL, WFR)
If WheelspeedRear < WheelspeedFront
(the car is on a dyno)
And RPM > 0
(the car is running)
Do NOxCatalyzer

That's terrible dummy code, but you get the idea. It's really, really simple. Obviously, the VW code would be more sophisticated, but this would be one way to do it.

[goonster]

With very few exceptions, regulators approach manufacturers with the basic assumption that they are acting in good faith with respect to complying with the letter and spirit of the law. Once that trust breaks down, it is a long, hard battle to get back in the regulators' good graces.

Outright evasion is rare, because the downside is so great.

If a bank takes $5 out of everyone's account illegitimately, the situation is easily remedied, but the big problem for customers and overseers is that they would do it in the first place.

[Climb01742]

I'm the happy owner of an '09 GTI. It's my second VW. My local VW dealer has really upgraded their store and service department/experience the last few years. In adspeak, I was becoming a brand loyalist and was seriously considering the coming Sportswagon diesel w/4-motion or a new Microbus if it ever materialized. But damn, this changes things for me. In a time when corporations' actions become more transparent via social media, how you act is such a part of your 'brand', not just your products. How I'll feel as time passes, I'm not sure, but I've really gone from digging VW to being highly suspicious. I bet I'm not alone. $18B in fines is a lot, but I wonder how much loyalty this might cost VW too. Companies can't be cheating a-holes so easily anymore.

[David Kirk]

How long before the class action lawsuit starts? I expect one of those letters to show up sooner rather than later.

dave

[alancw3]

i wonder if this had anything to do with the ousting of ferdinand piech as chairman of vw many months ago. he had been the ceo and chairman for many years before a power struggle within the company. i would think that a high level executive had to sign off on this decision on the programming.

[grawk]

http://www.bloomberg.com/news/articl...lean-air-group

[rwsaunders]

Here's a copy of the letter that the EPA issued to VW last week. As with anything, it looks like the issue has been on the table since May 2104 following the results of the WVU study. As the issue includes vehicles assembled since 2009, the design certainly predates the assembly date.

It's interesting to note that the "cheat" device isn't a part per se, but an algorithm as indicated in paragraph 2 on page 4. Assemble the Bosch coders and get them in a room...somebody will talk.

http://www3.epa.gov/otaq/cert/docume...a-09-18-15.pdf

In PA we have yearly mandatory safety inspections, and depending on where you live, emissions inspections, if that makes any sense. While the EPA is at work on VW, focusing on ongoing emissions testing for all vehicles would make sense as well.

PS to mjb regarding the PA Turnpike trees...road salt, gypsy moths and the locust leaf miner. Not making excuses for vehicle pollution, but it's deeper than that. You should see the ash borer damage in western PA.

[CNY rider]

Originally Posted by goonster

With very few exceptions, regulators approach manufacturers with the basic assumption that they are acting in good faith with respect to complying with the letter and spirit of the law. Once that trust breaks down, it is a long, hard battle to get back in the regulators' good graces.

Outright evasion is rare, because the downside is so great.

If a bank takes $5 out of everyone's account illegitimately, the situation is easily remedied, but the big problem for customers and overseers is that they would do it in the first place.

Maybe it once worked that way.
SOP in this country now is for politicians (both parties, neither is any better) to accept huge bribes (sorry I meant to say "corporate campaign contributions") and then pick the industry insiders to serve as the regulators after they have been elected.
Then when the corporate folks finish their government work they go become lobbyists for those same industries or go work at law firms that do the companies bidding.
It's a big revolving door.
We live in a synergistic corporate/government kleptocracy.