ORIGINAL: yl5295
You can use a linear regulator if you can live with the inefficiency of converting battery power to heat, supply enough peak current during CDI charging pulses (measured with a current probe on a scope - nota DVM), and can support 2.5+ amps continous as that is what some ignitions draw. Pretty much everything you said is correct except for the opto isolator. The noise is going to radiate through power and ground and from the wires... Not through the CPU and the signal wire... How does that common mode choke work on differential mode noise?
It works in much the same way as Bumble Bees manage to fly, in spite of some expert opinions that have claimed that itshould not be aerodynamicallypossible. I think that I have provided more than enough technical discussion, and even more in the form of proof of over 2 thousand units that are handling whatever CDI/Engine/Radio combination you throw at it. That's what it's all about.