I like the Hall effect idea (automatically isolated). One thing, however, I believe that same amplification must be integrated at the place of measurement to get the signals large enough to be more than the noise picked up. Also DIFFERENTIAL is the only way to wire industrial sensors.

Erik