3 ms·
The controlling unit indeed comes from Bosch, but it is a generic unit that can control any kind of engine. The car manufacturer using this unit programs it to
by _ph_ 8y ago
The controlling unit indeed comes from Bosch, but it is a generic unit that can control any kind of engine. The car manufacturer using this unit programs it to perform the actual world, mostly by specifying parameter sets for its operation. The car manufacturer has for example to determine, how exactly the ignition needs to be set up for a given temperature and load etc. There are plenty sensor inputs these profiles can depend on like engine temperature, revs, load, outside temperature, speed of the car. The Bosch software can offer the "test dectection" as one of the inputs. There can be quite reasonable use cases for that information, but that VW used it in the way they did, was VWs choices. The same or a similar control from Bosch unit is probably used by many other car manufacturers, but without abusing this specific "feature".
As the article writes, there are other "strange" engine configurations with other car manufacturers, like using the outside temperature. Reducing or switching off the exhaust cleaning system at outside temperatures below 17C (the tests are guaranteed to be done at 20C) achieves a similar effect to what VW was doing without literally breaking the law by using a defeat device.
- pedrocr 8y ago> There can be quite reasonable use cases for that information What reasonable use case is there for a detection of the test cycle just from ECU information? If you're just testing stuff surely you have access to the ECU to just set a flag manually for anything you need?
- _ph_ 8y agoCould be reasonabe safety features like not performing emergency breaking, or functions aiding the testing process (without interfering with it of course).
- pedrocr 8y agoI can see plenty of reasons to change behavior for the test. But that's not the question. The question is why infer you are being tested through the ECU inputs? Setting a flag manually seems like a much better solution for all those objectives versus a possibly unreliable inference from speed and other data.