Hacker News Reader

Comment by lordnacho | Hacker News Reader

How is it possible that the real emission is 40x the allowed limit? Wouldn't you see VW being the only manufacturer under the limit, and everyone else unable to get the car approved? Or alternatively, everyone is cheating? Because you can imagine a world where some firms cheat to get 10-20-30% less than rated, so that they're in line with the industry.

But if VW's normal car is 40x worse than advertised, and most cars were close to the correct standard, wouldn't they just hire a guy who knew how to fix their cars? Did VW's internal testing test competitors' cars?

Is this going to explode across the industry?

Also, this isn't the only kind of test that a car goes through. There's crash tests, MPG tests, and all sorts of things that I wouldn't know about. If you can game an emissions test, you can game the crash test and the MPG test, which are probably both things people care about a fair bit more than emissions.

View on HN

When I was car shopping the VW diesel numbers, for tdi sportwagen in particular, were impressive. Nothing else came close to VW's combination of power, space, and mileage. Now maybe there are people who know more about cars than I do and can dispute that, but that was my perception a couple of years ago.

If I were the other car companies, I'd want to know exactly how VW was pulling that off. They must have looked into it, and surely they're not as easily misled as I apparently was (how would I know if VW was outright lying about the car?).

This suggests to me that the other car companies must have known that VW was doing something wrong. The fact that they didn't rat VW out suggests to me that they were either doing the same thing (maybe not as aggressively as VW) or they hoped to get away with the same thing. The former seems more likely, and if that's true then I expect this to spread beyond VW soon.

It was a researcher from West Virginia who tested it.

Researcher is on the record as "I didn't believe the results" and "I thought I made a mistake".

Also, lol @ efficient market thesis because the results of his research were posted early 2014. EPA took a look at them, and being the slow bureaucracy that it is... took a year and a half to levy charges against Volkswagen.

So this entire "surprise" was over a year in the making. It's been discussed publicly in research groups for a year and no one seemed to care till now.

Slight correction, EPA was not slow to act. When they caught wind of this they approached VW who had one excuse after another:

http://www.washingtonpost.com/news/wonkblog/wp/2015/09/22/an...

The EMH predicts that unless the net probability of effects where totally zero there would be a move. I don't see evidence of there having been one.

> The fact that they didn't rat VW out suggests to me that they were either doing the same thing (maybe not as aggressively as VW) or they hoped to get away with the same thing.

Or (only hypothesizing) they realized they're all in the same industry and are necessarily treading softly wondering what the best course to accuse an 800lb gorilla of cheating.

It will be interesting to see if there are other cheaters, and to what degree, and also if this was an open secret in the industry. Nothing would surprise me.

I'm now searching for a story I'm sure happened, which was Mitsubishi faking crash test data... but damned if I can find it. Best I see is this[0]. Am I imagining a scandal w/ Mitsubishi where they were rigging crash test data ?

[0] https://en.wikipedia.org/wiki/Mitsubishi_Motors#Vehicle_defe...

VW has also built a huge reputation on diesel. They go and win Le Mans with diesels. They have a giant budget and research ability. They've pushed diesel forward in a lot of ways.

Maybe the others knew they were cheating, I can't imagine that it wouldn't leak out some how. I also can't imagine how you wouldn't go buy a hundred VWs and meticulously take them apart and understand them after getting brutalized in the diesel market.

With mid-sized and heavy trucks, there is an entire subculture of guys that mod them for "more power" and such. There is a little industry built on it and nearly a religion surrounding the "better mileage" and "more power." Some of the systems and devices are sophisticated enough that they have integrated on/off switches for passing smog tests and such. If we really really cared about it, that would be illegal, there would be much more stringent emissions testing more frequently.

I'm of the belief that the regulators knew or suspected there was some cheating but it's political to make waves.

I also can't imagine how you wouldn't go buy a hundred VWs and meticulously take them apart and understand them after getting brutalized in the diesel market.

every miniscule detail of the VW car is patented. if you infringe one of those patents accidentally and VW sues you, you'll get slammed with damages proportionate to money you have made or VW has lost (thanks to you). now, you basically cannot infringe a patent incidentally after you've seen it at work. competition buying a bunch of VW cars to take them apart would be an invitation for VW to sue them for willful patent infringement and the accompanying treble damages.

you can read accounts on the internet of programmers working for major corporations (Sun Microsystems among others IIRC) who were prohibited from reading others' patents for exactly this reason.

it's a thouroughly corrupt system.

edit: "the holder" -> VW + some styling

> competition buying a bunch of VW cars to take them apart would be an invitation for VW to sue them for willful patent infringement and the accompanying treble damages.

As someone noted above, every auto manufacturer is buying the competitor's cars, tearing them down, figuring out they work, rebuilding them, and benchmarking.

The last point is key - you can run your internal tests against their vehicles and see how you compare in your tests.

You can also get a feel for how many molds/stamps/etc that your competitor has in their factory by looking at the mold IDs. You can figure out how tight their weld tolerances are by X-Ray'ing the welds. There is a wealth of information inside a competitor's product that goes beyond the IP.

> You can also get a feel for how many molds/stamps/etc that your competitor has in their factory by looking at the mold IDs.

Also known as the German Tank Problem[0]

Allied intelligence, trying to estimate the extent of Panzer production, used estimates based on serial numbers of various components in tanks that fell into their hands.

From Wikipedia: "Estimating production was not the only use of this serial number analysis. It was also used to understand German production more generally, including number of factories, relative importance of factories, length of supply chain (based on lag between production and use), changes in production, and use of resources such as rubber."

[0]: https://en.wikipedia.org/wiki/German_tank_problem

The auto industry is so incestuous as it is. If patents came to play at all, which they didn't, then the big money play would be for GM, Ford, etc.. to buy VW motors and or license the technology. Either build as good a diesel or if you somehow couldn't then buy it or the technology. There is tons and tons of precedence for it. (Heck, the Ford "PowerStroke" diesel wasn't built by Ford for the first half of its life) Auto competitors work together on tons of stuff and with something like a motor that has fuel demands, it's mutually beneficial, if Ford and GM sell small diesel cars, it's more likely there will be diesel at your nearest pump. GM and Ford build heavy diesel too, there is very very real demand if they can improve power, improve mileage and meet emissions and VAG doesn't compete in that market.

None of that happened, I think the competitors knew and I also think the EPA knew in some capacity and it wasn't until some "small time" researchers got attention that it came out.

Or: mutually assured destruction. They all have something on each other.

Perhaps it's just the US with the throw weight to go after Toyota and VW in the most widespread, newsworthy issues.

It's also helpful that these are foreign car companies.

I also can't imagine how you wouldn't go buy a hundred VWs and meticulously take them apart and understand them after getting brutalized in the diesel market.

This is exactly how the auto industry works, contrary to the uninformed person(s) who modded you down. People should refrain from moderating posts from users who actually know what they're talking about.

For example, the first thing GM did when they began work on the current-generation Corvette was buy a Porsche 911 (from Volkswagen, no less) and study it in detail. This is an objective fact by GM's own admission (http://www.edmunds.com/porsche/911/2013/comparison-test.html). Competitive analysis is a key engineering strategy, no less important than any other.

It's inconceivable that other manufacturers weren't aware of exactly how VW's seemingly-impossible engineering worked. The only question is why they didn't rat them out to the EPA.

  > The only question is why [other manufacturers] didn't rat 
  > them out to the EPA.

Speculation: Ratting out VW carries some degree of risk that the public will become aware that there are better ways to test the emissions of vehicles. The public will then insist that the government legislate newer/better emissions testing, and the manufacturers will have an expensive new problem to solve. Better to keep quiet and let the existing, implemented and paid for (if flawed) process keep running.

> It's inconceivable that other manufacturers weren't aware of exactly how VW's seemingly-impossible engineering worked. The only question is why they didn't rat them out to the EPA.

Everyone cheats, and preserving status quo is in the best interest of everyone.

What worries me is that if this explodes across the industry, the common refrain will become that EPA guidelines are unrealistic and that's why everyone is cheating. Just what we need as the world is coming around to direly needed environmental regulation.

Isn't it better for any individual/employee to have been a whistleblower under the False Claims Act - https://en.wikipedia.org/wiki/False_Claims_Act ?

According to the Wikipedia page "Persons filing under the Act stand to receive a portion (usually about 15–25 percent) of any recovered damages."

Here is an article with a tour of the GM teardown facility, so automakers unquestionably do it.

http://www.slashgear.com/the-auto-vivisectionist-inside-gms-...

That said, there is pretty much no US market for diesel outside the European imports. So they may not be focused on their diesel competitors enough to have done the teardown/research/testing to discover VW cheating. I find that more plausible than everyone cheats and everyone knows everyone else cheats.

> have integrated on/off switches for passing smog tests and such. If we really really cared about it, that would be illegal, there would be much more stringent emissions testing more frequently.

Actually it is illegal.

And inspections in the US are supposed to check if there are any altered emissions components.

It doesn't help if someone removes and reinstalls it each year, but they do try to check for such things.

According to [1]:

"Volkswagen hasn't explained exactly why it cheated, but outside analysts have a good guess. The NOx emission controls likely degraded the cars' performance when they were switched on — the engines ran hotter, wore out more quickly, and got poorer mileage. Some experts have suggested that the emission controls may have affected the cars' torque and acceleration, making them less fun to drive. (Indeed, some individual car owners have been known to disable their cars' emission controls to boost performance, though this is against the law.)"

So basically, they wanted the cars to have acceptable performance in real-world conditions, which pushes the emissions up.

[1] http://www.vox.com/2015/9/21/9365667/volkswagen-clean-diesel...

When I got my Jetta TDI, it had a nasty stutter when the engine was cold.

It's three blocks from my house to the arterial, and I had a couple of close calls when I tried to pull into traffic and nothing happened.

When I took the car in for 10k miles they had a ROM upgrade that fixed the problem, but I'm wondering now if they cheated to do it.

I had read something about how diesel engines can use Urea to trap emissions, but it must be refilled frequently, so possibly other manufacturers are using that method. This is tested and proved to work well. However, VW was using another type of method to diminish emissions that has not been well proven. Source: a comment I read on Reddit, so take it with a grain of salt.

Edit: found at least one source on how this works: http://www.greencarreports.com/news/1042727_adding-urea-to-c...

It's called "AdBlue": https://en.wikipedia.org/wiki/Diesel_exhaust_fluid

Thus why Merc calls theirs BlueTec

AdBlue is also used in the latest Audi ultra TDI engines. It is supposed to be refilled during normal service (every 30000 km), unfortunately it does not last that long and you have to schedule an appointment much earlier. Refilling it yourself is apparently not possible since a software reset is necessary as well.

The 4 cylinder TDI was for the economy range of vehicles. The more expensive models with bigger engines use urea tanks or BlueTec as Mercedes Benz politely calls it. More expensive to make and service, but those who buy more expensive vehicles expect these kind of added service costs. Those on the lower end do not.

>The 4 cylinder TDI was for the economy range of vehicles

I have a 2014 Jetta TDI (4 cyl) with urea injection. I'm totally confused by the information surrounding what cars are affected.

Model year vs production/purchase year probably doesn't help.

As far as I understood, "model year 2015" cars (made/sold in 2014) are unaffected as of yet, because of the urea injection.

I'm guessing that the cars with urea injection will be fine. It will be tested by the EPA and CARB if you are in California to make sure. Since the urea injection is there specifically to deal with the pollution this issue is causing, the only cheat I would imagine would be to lengthen the amount of time to change the urea tank and if that is the case, VW will be required to cover the added service costs.

Not so sure myself, check-out page 63:

http://www.theicct.org/use-emissions-testing-light-duty-dies...

Car B was a 4cy VW diesel with urea from last year.

I've driven several pickups that require diesel exhaust fluid. It's really not a big deal. If you don't want to pay $3/g for it at Walmart then go to a truck stop where you can get it from a pump for slightly less.

I know someone who recently bought a sportwagen with the urea system. They said it was a $300 service once a year to refill the tank. Though you are probably right that it can be done much cheaper yourself.

It's an ISO-specified solution of CO(NH2)2 and deionized water. If evil automakers want to screw the consumer they could make the fill port difficult to access, but other than that it seems straightforward. The "once a year" thing is also strange, since DEF usage is totally dependent on miles driven and how "hard" those miles were. (High-RPM, high-turbo-pressure driving uses more than low-RPM, low-turbo-pressure driving.)

Mercedes did this on the GLK - the fill port requires a special nozzle. So you buy one of their refill bottles, cut the bottom off, and use it as a funnel with your DEF of choice (since it's a standard ISO solution).

Fun fact: the fill nozzle at the truck stop is also standardized by ISO, so that it will fit all ISO-compliant "commercial vehicles and buses... having a gross vehicle mass of more than 3,5 t". I wonder how how much lobbying it took to convince ISO not to specify anything for smaller vehicles?

Some of the truck stops have DEF nozzles that require a magnet to operate, so the truck drivers don't put it into their diesel tank. There are adapters you can buy that will allow you to fill up your diesel-powered car or RV using the cheaper bulk pump.

That's crazy. I think I probably paid $20 for DEF to cover the ~11,000 miles I drove in my truck last year (and I'd have to imagine it's thirstier than an economy car).

You pop a blue cap under the hood, dump in the DEF, and you're done. RAM and Ford (starting this year I think?) make it even easier by putting the fill behind the fuel door right next to the fuel cap.

5 gallons seems to last me about ~5,000 miles.

DEF only became required in US light-trucks ("1-ton" and below) for the 2015 Model-Year I think.

Anyways, the fluid itself is about the same price as gasoline. A $300 service is just plain gouging.

The researchers also tested a diesel BMW X5, which passed emissions tests. [0]

I would guess that its just more expensive to implement, more steps in the filtration process, larger catalytic converters, etc.

[0] http://www.reuters.com/article/2015/09/23/us-usa-volkswagen-...

Smaller, cheaper cars are the problem. The bigger cars have a better catalysator system. All those compact diesel cars from Opel, VW, Renault, Peugeot, Nissan will have the same problem.

What is crazy is that the vast amount of people who buy these small diesels don't need them. You need long runs to get the particle filter hot enough to burn off.

People buying these small diesels are buying them based on the fuel consumption benefits but then only doing short trips like 15 minute school runs. All they are doing is damaging the car.

There have also been no claims that the VW Touareg TDI failed. So 6cyl SUVs pass.

There aren't many 4cyl diesels in the US. Chevy Cruze. New BMW 328d. I imagine the new 328d passes just fine, but supposedly the new-for-2016 VW E288 diesels do as well.. they just never got a chance to sell them!

The BMWs also use a urea based system that separates the NOx into Nitrogen and Oxygen and does it very well. However, those systems are expensive and add a lot of costs to the car. VW wouldn't be able to sell their TDI cars as cheaply if they had to incur this cost. That's why they pass and why they have passed alternative tests.

I just bought a new GTI about a month ago...not sure if their gas cars will take a valuation hit or not due to VWs reputation hit. At least it wasn't a TDI Golf or I'd be extremely pissed.

It is why competitors like diesel Cruze had worse performance and others like Mazda did not quite bring to market. VW knew they could own the US diesel small passenger car market but they had to meet customer demands on fuel efficiency, performance, and price.

it's 40x from "faked condition test" to "road condition". And the "test condition" are totally unrealistic behavior on the road (such as 0-50km/h in 26 seconds)

Most car are above the allowed limit in real situation.

Collision tests can be gamed obviously by designing for the fixtures used... But the variety of fixtures in use and testing protocols imposes a decent all around level. It's hard to cheat when 3000 lbs of steel is crashed at 40mph against a concrete obstacle.

Fuel consumption however, now that is rife for cheating by trading off performance for better fuel consumption in the ECU and automatic transmissions.

LG has cheated Energy Star ratings with its refridgerators

http://www.treehugger.com/clean-technology/lg-under-the-gun-...

Also, weren't the competing car companies curious about how they could also save money on their emissions systems and buy a VW to reverse engineer?

I bet they did and as soon as they saw why just decided to wait to see how long it took for them to get discovered.

That or they decided to implement it too. Only way to know is to test car emissions.

A more interesting plan would have been to short VW while waiting for them to get discovered.

That didn't work out so well for all those people who tried to short VW back in 2008...

The problem affects only diesel cars, and VW is the only company I've heard of selling a significant number of diesel passenger cars in the US.

There are other countries in the world, and in some of those countries, there are significant numbers of diesel passenger cars.

European emission standards are lax, and I would imagine the rest of the world even laxer.

http://www.bloomberg.com/news/articles/2015-09-23/no-cheatin...

Like anything it's a trade off.

Emissions and efficiency are usually conflicting goals. Automakers are under a ton of strain right now to achieve BOTH, which is essentially impossible. The VW TDIs are some of the only vehicles to even get close to the efficiency that will be required in just a few years, regardless of emissions.

Except that poor air quality has a severe impact on people.

Poor is relative. Even a "bad" car by current standards would have been considered extremely clean even 5 years ago.

No. The VWs in question would not have even passed 1981 CARB standards.

For what it's worth, Chevy is selling a diesel version of the Cruze in the US, and it has better performance numbers pretty much all around than the VW. I'm skeptical of the claims that this is "the end of clean diesel" as a result.

Maybe you should be sceptical that Chevy are more honest than VW.

40x is likely a massive overstatement. Nobody's actually released any actual numbers or data yet.

Check-out page 63, looks to be 40x on the graph:

http://www.theicct.org/use-emissions-testing-light-duty-dies...

Reply on news.ycombinator.com