Especially when testing
security relevant automotive systems it might be
dangerous to apply only a simple acceptance criterion, such as GO / NO GO.
Up to now some pepole are still doing so, in a way like this: a single test in a single static operating mode and a subsequent check for the occurence of errors.
Doing automotive EMC tests that way could be afflicted with some risks: Hidden interferences due to double faults or plausibility checks, overlooked interferences in other operating modes or in dynamic operating modes, EUT susceptibility levels near the test level, etc..
These states could be considered as scare stories, of course. If these faults would not occur ever and anon
after the launch of a product, and if there were not the claims of the customers who detect these faults.
The better way of EMC testing is:
Be aware of what is happening in your automotive ECU!
Notice the behaviour of every single signal under EMC impact!
Only if one is knowing those details about an automotive ECU a responsible manufacturer could estimate the behaviour of a complete system under EMC impact.
Back to "Applications In EMC Testing"