It looks like VW was the one that got caught first:
Through trickery, the gap between official fuel economy figures and those achieved by an average driver have grown to 40%. For new diesel cars nitrogen oxide emissions are typically five times higher on the road than the allowed limit and just one in 10 cars meets the required level on the road. But for some models the gap is so large T&E suspects that the car is able to detect when it is tested using a “defeat device” and artificially lowers emissions during the test. For example, a diesel Audi A8 tested in Europe produced nitrogen oxide emissions 21.9 times over the legal limit on the road; a BMW X3 diesel was 9.9 times over the limit on the road; an Opal Zafira Tourer, 9.5 times; Citroen C4 Picasso 5.1 times. All these vehicles passed the laboratory test.
In CO tests, on average almost every Mercedes model achieves levels on the road over 50% higher than the laboratory tests; the BMW 5 series and Peugeot 308 achieve just shy of 50% higher than in the lab. For virtually every new model that comes onto the market the gap between test and real-world performance leaps. With the launch of the VW Golf Mark VII the gap between test and real-world CO[SUB]2[/SUB] emissions jumped from 22% to 41%. The gap for the new Mercedes C Class rose from 37% to 53%; for the Renault Clio IV the gap almost doubled from 19% to 34%. These changes are unlikely to be caused solely by the increased use of test flexibilities – the more sinister and illegal defeat devices may also be in use and T&E has initiated a testing programme to demonstrate this as the US authorities have done to expose VW.
From
here.
