By on December 31, 2015

tdiengine

Volkswagen’s emissions cheating program closely follows a set of parameters that are very similar to those defined by the New European Driving Cycle (NEDC), an engineer said this week.

The cheat exists in the ECU’s “main mode,” said Felix Domke, and triggers a normal dosage of urea and other exhaust controls to bring NOx emissions to within acceptable levels.

Domke presented his findings of an unpacked Volkswagen ECU to the 32nd Chaos Communication Congress in Germany.

His findings are mostly in line with what the automaker has already admitted: its 11 million cars worldwide cheated emissions tests by using two different modes for operation, and that its cars could pollute up to 40 times the legal limit of nitrogen oxides when running normally.

But Domke, who said he owns a Volkswagen Sharan equipped with a 2-liter diesel engine, said his own observations showed a severe change in the ECU’s behavior when it exceeded the bounds of what it considered was an emissions test — more than what’s been reported so far.

Domke said he purchased a Volkswagen ECU online and extracted its millions of lines of code. The VW ECU’s behavior is mostly governed by parameters defined outside of its code — not constants that would be hard-coded into the programming — embedded within the computer’s hardware.

He said the car’s main modes for exhaust management were a “normal mode” and an “alternative mode,” the latter of which mostly dominated the car’s behavior. In his own driving, after observing the state of the car, Domke said he found the car ran in a “cold start” cycle for roughly 20 percent of the time and on its “alternative mode” for nearly 80 percent of his driving. Only a small fraction of his normal, everyday driving was in “normal mode.” During “normal mode,” the selective catalytic reduction system used by Volkswagen would consume significantly more diesel emissions fluid than in “alternative” mode. Domke said that finding was supported by additional observations: his van used far less urea than he anticipated.

Domke said he graphed the European emissions testing cycle and overlaid those results with the upper and lower limits of the ECU’s “normal mode” and discovered that the mode aligned perfectly with the limits.

He didn’t test differences in engine performance, nor could he say whether the cheat applied to cars in other countries. But Domke pointed to a parameter in the engine’s code that seemingly always initiated its “alternative” exhaust program: the outside temperature would only need to be suitable for life to exist — above -6,357.9 degrees Fahrenheit (-3,550 degrees Celsius).

Only after a fairly specific set of conditions were met: atmospheric pressure, temperature, speed and distance, which coincide with the NEDC testing parameters, would the car begin its “normal” cycle.

Domke presented his work with Daniel Lange, a former BMW engineer who said the likelihood that the ECU was developed by a “small group” of Volkswagen engineers is hard to believe. Lange said roughly 1,000 hard disks were confiscated from 380 employees, which is hardly a small group.

Also, Lange said the paper trail for documenting changes to an engine’s ECU is prolific — there are too many laws that govern how cars are controlled and automakers fastidiously document those changes for legal reasons.

The presentation is just over 1 hour long. If you want to get to the meaty stuff right away — like how the “defeat device” worked — skip ahead to around 51 minutes in. Otherwise, grab some popcorn.

Get the latest TTAC e-Newsletter!

43 Comments on “This Is How Volkswagen’s Diesel Emissions Cheat Works, According to ECU Hacker (Video)...”


  • avatar
    SCE to AUX

    I don’t understand the notion of -3550C, since everywhere in the universe is above -273C.

  • avatar

    Basically, all of the exhaust gases are forced into a chamber and held there until the emissions test ends.

    Then the car lets out a loud fart that DESTROYS the polar Icecaps.

  • avatar
    Mr.Radar

    For those who don’t want to watch the whole video, here’s a breakdown of the key points:

    Modern diesel vehicles have an emissions control device which uses Selective Catalytic Reduction with urea solution (also called Diesel Exhaust Fluid or AdBlue) to catalyze NOx into (relatively) harmless water vapor and CO2. In order for this process to be effective you need to precisely meter the urea solution. If you use too little you won’t catalyze all of the NOx and if you use too much some of the urea will catalyze into harmful ammonia.

    The Engine Control Unit (ECU) is responsible for determining the correct dose of urea solution at any given time. The Volkswagen ECUs are programmed to perform two different dosage calculations. One is called “normal” mode and the other “alternative” mode. The normal mode incorporates data based on at least 6 different parameters while the alternative mode uses only two and appears to be designed as a “conservative” fallback (to avoid ever overdosing) when one or more of the parameters are outside of the range handled by the normal mode (the presenter gave the example of an overheated engine).

    This is mostly borne out in the criteria the ECU uses to decide which mode to use. Most of the checks it performs are basic sanity checks on the parameters, however at least one of the checks is being performed “wrong”: if the engine temperature is *above* -3276.8K the alternative mode is selected. The presenter actually had to double-check his reverse-engineering when he discovered that check since it couldn’t possibly be right. The correct sanity check would be if the temperature is *below* that value since that would only trigger for an obviously an incorrect reading while the actual check is triggered for all valid values.

    So if the alternative mode is always used due to a “defective” check how did these vehicles pass emissions tests in the first place? That is because there is a second check after the previous one. If the vehicle detects that the distance it has driven at a given time since engine startup is within certain ranges it will force the “normal” mode to be used even if the first test determined the “alternative” mode should be used. Unsurprisingly one of those ranges matches up *exactly* with the NEDC drive cycle (and presumably the others match up with EPA drive cycles, though that is not mentioned in the video). If you don’t watch anything else from the video, seek to 58:10 and watch the presenter overlay the ranges and NEDC driving distance to see just how precise the match is. That, gentlemen, is the smoking gun.

    • 0 avatar
      Scoutdude

      Further damning evidence is that the program continues to calculate, what appears to be, the correct dosage rate to properly control emissions.

    • 0 avatar
      Pig_Iron

      Thank you for the nutshell explanation. It clarified what I thought I understood from the video.

      Way back when I was in design and development engineering, we would often do work-arounds to get through prototype testing, but those were never meant to be production level test procedures. I wonder if this started as a temporary work-around that got institutionalized.

      Those sorts of engineering deviations usually require cross-functional/departmental and multilevel approval/sign-off; especially for regulatory compliance matters. I find it hard to believe this was just a few people who could orchestrate this across multiple vehicle programs.

  • avatar
    Glenn Mercer

    Excellent summary, thank you. And I do hope that your “smoking gun” usage was a deliberate pun! (grin)

  • avatar
    RideHeight

    Thank you for this video. Lange’s overview is tremendous!

    If anyone else found the TTAC version to have stuttery audio, YouTube has a cleaner English language version available from an early post in the comments section of the German narrated version.

    Find it by entering “Felix Domke”.

  • avatar
    Lorenzo

    I keep hearing “more than 40 times” the legal emissions limit of NOx. Have the limits for oxides of nitrogen been reduced that much for diesel engines? Being a suspicious sort, I’m wondering if Volkswagen was trying an end run around an insidious plot to eliminate diesel engines by reducing emissions to a standard that diesel engines can’t normally meet.

    I don’t want to sound like I’m wearing a tin foil hat (tin foil is expensive, I hope aluminum will do), but the best way to reduce/eliminate something that can’t be banned outright due to politics, is to either tax the daylights out of it, or make its use onerous, or both – like smoking. Tightening the emissions standards looks like a good way to wean Europeans off diesels, and deter their use in passenger cars here.

    I’m not claiming that’s good or bad, just that if Volkswagen was responding to a threat to its passenger diesel business by regulators with an agenda, then the problem is bigger than just Volkswagen.

    • 0 avatar
      RideHeight

      Wouldn’t that insidious plot have to have been launched by the very politicians and government organs that conspired to make diesel the default fuel of European autos in the first place?

      Not that I’m saying government isn’t often schizoid and factionalized to give such results, but the promotion of diesel seems to be the original insidiousness.

    • 0 avatar
      Scoutdude

      The problem with your theory is that at least some other mfgs have been shown to be capable of meeting the standards in normal driving.

      The video shows that if the vehicle isn’t operating within the strict parameters of the emissions drive cycle the DEF dosing is shut off. I’m betting if the guy tracked the EGR system operation it would also cease to operate in the chat mode. So yeah I can see the 40x thing, at or near peak loads, if they are not actively using anything to control NOx.

      Plain and simple VW cheated to gain an advantage because they didn’t think they would get caught.

    • 0 avatar
      wmba

      @ Lorenzo

      The Euro 5 standard, which is the standard the VW EA189 diesels in question in this article had to meet in Europe is about 180mg/km of NOx. It isn’t as if this came out of the blue one day unannounced. It had been proposed for years before coming into effect in 2009. A Particulate Number was introduced as a part of Euro 5 in 2011. and the standard became 5b.

      Euro 6 reduces this NOx limit to 80 mg/km for September 2014. Again, known years in advance.

      Euro 4, btw, was 250 mg/km and came in in 2005. Before that Euro 3 came into effect in 2000 at 500mg/km.

      This was a steady progression downward, and VW as well as everyone else could have complained before each standard came into effect. Instead they chose hubris and kept quiet..

      Now, the latest EPA standard for diesels that recently came into effect is 40mg/km, ONE-HALF Euro 6. The VWs/Audis and Porsche CUVs sitting in the docks obviously do not meet this standard, or the EPA would allow their sale in the US. Again, did VW ever bleat on the federal register when these new standards were proposed? Not a word. They just decided to cheat. It’s simple.

      I really do not understand how people cannot grasp this scenario. It’s not difficult, and 10 minutes googling gets you the answer.

      So far as Euro 5, which this hacker in this video is going on about, on acceleration and high load situations, VW EA189 diesels belched up to 8000 mg/km of NOx. It is invisible, so no, it’s not the black soot you sometimes see, that’s hydrocarbons. NOx causes smog and ozone by photochemical reaction. It is not good stuff.

      And to all the apologists who own these vehicles and couldn’t care less about the pollution they put out, and say so, I have little time. It turns out, if you all had followed the original article here on TTAC about the University of Leeds researcher, you would find out that new buses and trucks following Euro 6 actually meet the standard, because, presumably, they inject enough AdBlue/urea to tame the NOx, rather than going into cheater mode and saving a few cents worth of urea.

      Univ of Leeds also found most other diesel cars were WORSE than VW. So to my obviously linear engineering mind, I’d say diesel is done, finished and washed up. If there is no hope of making the legislated standards without drinking p*ss (urea) by the gallon, diesel isn’t practical for small vehicles, while buses and trucks have space to store the fluid.

      • 0 avatar
        Lorenzo

        Yes, but WHY is the NOx standard being inexorably raised, if not to eventually take passenger car diesels off the road? That’s MY point, the regulators can’t just ban them, they’re using progressively more stringent regulations to achieve a result they couldn’t get from elected officials. That’s lawmaking by unelected bureaucrats.

    • 0 avatar
      Pch101

      The EU hates diesel so much that its emission standards for diesels are less rigorous than they are for gasoline.

      The US hates diesel so much that its requirements for diesel are essentially the same as they are for gas.

      It would seem that Lorenzo wants affirmative action for diesel.

      • 0 avatar
        Lorenzo

        Actually, no, I don’t like diesels and wouldn’t want one. But others do, and if diesel-powered passenger cars are going to be banned, it should be by elected officials following the constitutional procedures required to pass laws.

        It appears the EPA has found a way to gradually outlaw diesel passenger cars, and probably ALL diesel vehicles eventually. The fact that some makers have devised methods of meeting a given standard doesn’t mean the drastic reduction from 500 to 40 can always be met.

        The EPA operates on two irrational principles: 1, if something is deemed “bad” it must ultimately be reduced to zero, and 2. that it doesn’t matter how much it costs to reduce, let alone eliminate that “bad” element. The first is environmentally unnecessary, and the second is financially destructive. Both are the province of our elected government officers, not appointed agencies.

        • 0 avatar
          Pch101

          You obviously didn’t understand my comment. Instead, you just regurgitated your previous comment, which wasn’t any better the second time around.

          Once again: You apparently believe that diesel deserves some kind of pass over gasoline. Neither the US nor Europe regulates diesel pollution emissions more strictly than they do gasoline, and the Europeans actually allow diesel to pollute more. Not less, but more.

          So your comment is nonsensical unless you believe that diesel is some sort of persecuted minority that is worthy of special accommodations. Your views on the subject are otherwise completely detached from reality.

        • 0 avatar
          NickS

          @lorenzo

          It sounds like you are opposed to the aspects of regulating small diesels that have more to do with the political process and democratic principles, i.e., which branch of government decides to ban these. On those grounds someone can sue the regulators and the courts will decide. Mind you, they do legislate from the bench so is that better? Besides, do we want to decide what pollutants to ban based on uninformed popular opinion, politicians who know zero about basic science and are experts at raising campaign money? Would that be better?

          I agree with everything wmba wrote. The way I see it, any diesel application that a) is easily accessible and has the potential for high adoption and b) does not *require* very high torque should be restricted.

          In any event, the EPA may seem like they are picking winners and losers here, but they are the AHJ and as such need to take a stand over which is worse CO2 or NOx. Would you prefer that the people make that decision?

          Edit: never mind that all GDIs should come with a DPF if not more, but you know, we regulate after the horses have left the barn.

  • avatar
    Scoutdude

    I figured it was just a matter of time before someone with the skills and determination dug in and found the two different operating strategies and the conditions to trigger when to use the emissions compliant one and when to use the one that maximizes power and fuel economy instead of worrying about that silly gov’t mandated clean exhaust.

    Meanwhile lest have our marketing dept sing the praises of our wonderful “clean diesel technology”.

    I’m pretty certain someone will crack the US calibration, if they haven’t already and just haven’t made their findings public, yet.

  • avatar
    jthorner

    Impressive work Mr. Domke. Too bad VW hasn’t learned to just be honest and tell the world what they did, who knew, and who approved it.

    The ongoing body part covering attempts out of VW are just more of the same ill behavior that got them into this mess in the first place.

    • 0 avatar
      Kendahl

      You are assuming that VW is a monolithic organization with everyone on the same page. Right now, I’m sure that VW resembles a roach infested room just after the lights are turned on. Everyone is running for cover and cooperating as little as possible. One of the weaknesses of autocratic, top-down organizations, like VW, is that dissent and disaffection are never resolved. They just go underground and fester.

  • avatar
    shaker

    This reminds me of the early days of catalytic converters; introduced to clean up carbureted, and (then poorly-controlled) throttle-body fuel injection systems, when certain conditions would result in the converter being swamped with hydrocarbons. The result of this would be a drastic increase in emissions (that “rotten egg” smell) and the converter actually overheating enough to trigger a warning light.

    Once multi-port FI and better computer control became available, catalytic converters became more effective (and smaller) since they had many fewer excess hydrocarbons to “convert”.

    We’re near the point when the low-end torque requirements of normal driving can be met by battery-hybrid systems (in light vehicles) and pneumo-hydraulic hybrid (for larger trucks with frequent stop-and-go operation) to replace the diesel engine – the continued development of more energy-dense batteries and stronger materials (for hydraulic accumulators) will supplant the need for diesel engines – but the specter of economics is still (somewhat) in favor of cleaning up the crud as it is produced.

    • 0 avatar
      RideHeight

      “the converter actually overheating enough to trigger a warning light”

      Triggered more than that for a lot of rednecks who’d just park any old place off the gravel.

      • 0 avatar
        PrincipalDan

        IIRC someone here or at Curbside Classic talked about their late 70s Nissan Z getting the converter hot enough to smolder the carpet inside the cabin! Now that’s hot…

        • 0 avatar
          golden2husky

          We had an old 81 K car with that nasty Holly feedback carb that never stayed in tune. I started it one morning to let the car warm up. When I came back out the car was full of smoke. The cat was red hot – looked just like an element in an electric oven and the high temperatures were burning off the leaking oil off the bottom of the firewall. The smoke was sucked into the heating system and the fan distributed it into the car…

          • 0 avatar
            RideHeight

            Diesel Simulation Feature!

          • 0 avatar
            Lack Thereof

            A lot of cars from the 70’s (including a Porsche product I owned) had warnings in their owners manuals saying things like “Do not idle the car with a cold engine. Drive immediately upon starting.” This is exactly why. Allowing the car to idle for several minutes with the choke set or the cold-start injector switched on would dump enough extra fuel/soot into the catalytic converters to be a fire hazard. Driving it immediately would warm the engine faster and thus shut off the choke faster.

            My Porsche would actually shut off and refuse to restart for several minutes if you tried to warm it up in the driveway on a cold morning for more than a minute or two.


Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

New Car Research

Get a Free Dealer Quote

Staff

  • Contributors

  • Vojta Dobes, Czech Republic
  • Matthew Guy, Canada
  • Kamil Kaluski, United States
  • Seth Parks, United States
  • Ronnie Schreiber, United States
  • Bozi Tatarevic, United States
  • Chris Tonn, United States
  • Moderators

  • Adam Tonge, United States
  • Kyree Williams, United States