* use closer to EPS limits
* raise to upper bound of spread from injection plot
* we've seen ~0.009 at 7 m/s, so allow that
* fix test
* rounding errors cause the 2 to round to the same value
* adjust for latest opendbc change
* we should actually match what safety is doing! fix lack of test coverage near first brake point
fix
* and fix test
3066f93d8a