Volkswagen has issued a statement regarding the emissions cheating incident:
Discrepancies relate to vehicles with Type EA 189 engines, involving some eleven million vehicles worldwide. A noticeable deviation between bench test results and actual road use was established solely for this type of engine. Volkswagen is working intensely to eliminate these deviations through technical measures. The company is therefore in contact with the relevant authorities and the German Federal Motor Transport Authority (KBA – Kraftfahrtbundesamt).
To cover the necessary service measures and other efforts to win back the trust of our customers, Volkswagen plans to set aside a provision of some 6.5 billion EUR recognized in the profit and loss statement in the third quarter of the current fiscal year. Due to the ongoing investigations the amounts estimated may be subject to revaluation. Earnings targets for the Group for 2015 will be adjusted accordingly.
Volkswagen does not tolerate any kind of violation of laws whatsoever. It is and remains the top priority of the Board of Management to win back lost trust and to avert damage to our customers. The Group will inform the public on the further progress of the investigations constantly and transparently.
From The Register:
To put that in perspective, Volkswagen's profits for the last financial year were €10.85bn (US$12.1bn), so the firm is banking on having to pay out at least half of its profits, and possibly a lot more. The EPA (US Environmental Protection Agency) has already said that the company could be liable for up to $18bn in fine and fix costs, and that was when only half a million cars were thought to be dodgy. As a result, the wheels have fallen off the company's stock price. Shares have nearly halved in value since the firm admitted using the emission-control software, and they are likely to fall further as the scandal unfolds.
Volkswagen's CEO Martin Winterkorn has already issued a public apology for his firm's conduct, and his position is looking increasingly untenable. Rumors of his forced retirement are already circulating, although these are being denied at present.
The case could also have an interesting knock-on effect in the software field. Technically, Volkswagen's software was covered under the US Digital Millennium Copyright Act, meaning tinkerers couldn't have examined and altered the code. The EPA has been lobbying with car companies to make sure the DMCA continues to make engine management software off limits to tinkerers. But based on its experience with Volkswagen, the agency may be changing that stance.
The Electronic Frontier Foundation has been quick to pounce on the DMCA connection.
The BBC reports that this affects 11 million vehicles worldwide, although many of those have passed local emission controls satisfactorily. Neverthless, the same or similar software is believed to be fitted in all those vehicles. The EPA found the "defeat device", the device that allowed VW cars to emit less during tests than they would while driving normally, in diesel cars including the Audi A3 and the VW Jetta, Beetle, Golf and Passat models.
Update: Volkswagen chief executive Martin Winterkorn resigns.
(Score: 1) by pipedwho on Thursday September 24 2015, @07:26AM
They don't need to drive around, they just need to measure all the pertinent variables at the same time. So if the ECU starts pulling power, increasing fuel usage, or overheating the exhaust gas, then the total measurement falls into question.
If all the tests are run separately while ignoring all other variables, then the ECU could exceed the specification on one parameter to allow another to meet the pass threshold. Each test could be gamed in such a way so for example the N2O test could be run with substandard power output or by a big reduction in MPG. Likely the MPG test could be gamed by boosting exhaust temperature to detrimental levels while increasing the output of various pollutants.
But, if you measure everything at the same time, then the car has no option but to remain 'honest'.
This type of thing could easily happen if the test is done while the car is legitimately in a mode where it runs differently that normal (ie. during warmup, or when running in hot ambient weather, or when it is overheating, etc). So a proper test would detect this happening and could be run with all variables properly reported.
For example, if a 30MPG car (specified at a particular load/speed) suddenly runs with 20MPG during the test while passing other tests, then there is a problem.