3 ms·
I'm still waiting for the smoking gun of actual code that detects a testing cycle and changes the operational parameters in a way different from the way paramet
by tbyehl 11y ago
I'm still waiting for the smoking gun of actual code that detects a testing cycle and changes the operational parameters in a way different from the way parameters are adjusted normally for different operating ranges.
Optimizing for test cases is rampant in the car industry. Nobody who has owned a GM car in the OBD II era has ever said "Wow, I get better MPGs than the EPA rating!" Heck, I'd be thrilled if any GM I'd ever owned came close on the highway doing 55 with the cruise control on. One of my GM's had something called Active Fuel Management. Cylinder deactivation. In normal driving it rarely comes on, but when it did there would be an annoying vibration and hum from the exhaust that would make you give the car more throttle in order to make it stop. My real-world MPGs sucked but apparently that anti-feature boosted GM's numbers for MPG and CAFE purposes.
Another GM car I currently own with a manual transmission forces a 1 -> 4 shift if I attempt to shift from 1st below certain throttle position and RPM thresholds. Out on an actual road, even with a giant V8, lugging around in 4th gear at 1,000 RPM is an undesirable condition. Naturally I do my best not to let that happen by applying my foot to the throttle, but somehow, again, this anti-feature improves the government numbers while I average 18MPGs on a mostly highway cruising commute and much less in stop-and-go city driving.
How are these GM anti-features any different from what VW has admitted? In the GM case the cars very obviously are detecting conditions within a test cycle and causing the car to operate in a manner undesirable to the driver, and which will cause the driver to alter their behavior to avoid them.
Is what VW did so blatantly obvious to a driver? What's the legal distinction? Where's the smoking gun code piece of code that makes their actions a "Defeat Device"?
On the premise of this article, the whole automotive testing industry is optimized around testing to the government standards. Just like No Child Left Behind has our teachers "Teaching to the Test", the automotive industry is "Testing to the Test." Nobody has any motivation to care about, or even look at, what happens outside of the test parameters. No grand conspiracy is necessary for this to have gone undetected or ignored. If the government considers this a problem then the solution is to create more tests that better simulate real-world driving conditions.
- jacquesm 11y agoThe legal distinction is that the code in the VW ECU actively detects the testing regime and then switches the engine to a mode of operation it would never use during real world conditions. That's no longer optimizing for a test, that's simply fraud. Other manufacturers performance deviates gradually the further away you get from the test parameters. The VW would never have the performance it has during a test in any real world conditions, not even those matching the test (because the vehicle is moving and it can tell the difference).
- tbyehl 11y agoI see speculation that there exists code that detects the test and activates a special mapping mode. I haven't seen that anyone has actually found the code or map. The map ought to be easy to find, VW ECUs appear to have as many freely-available DIY tuning tools as other platforms. I've also seen it speculated that the code to detecting testing is based on determining that the car is on a dyno, and that all manufacturers perform that detection in order to disable the stability control systems. If it were as straight-forward as VW switching to a different map on a dyno I'd have expected that VW tuners would have blown the lid on this thing long ago. Tuners use dynos.