Like you, I suspect the crank position sensor since it's the only thing down there close enough to arc to the balancer (dampener, actually). However, there shouldn't be much voltage going to it - most likely, the input reference-voltage is only 5 volts. I'll check my Vantage tomorrow and see if it gives a value for that sensor's reference voltage. The crank sensor drives the tach, and is the only input that tells the PCM the engine is running. If weird things are going on with the crank sensor, that explains the tach drop-outs and the alternator light (PCM thinks the ignition is on and engine is off).
I'd pull the sensor and check it for evidence of arcing - replace the sensor if you find said evidence. Maybe the sensor housing is cracked and is allowing an arc.
Do you have any high-voltage sources (like HID-headlight ballasts) that might be near enough to the sensor wiring to induce a higher-than-normal voltage on the sensor? Do you have any way to back-probe the sensor at the connector and capture transitory voltage spikes? To do this, you need a graphing multimeter/lab scope like a Vantage/Modis/Genisys (Actron has a low-cost equivalent) or one of the higher-level automotive multimeters/analyzers that has a rapid scan-rate, high input impedance, and high/low feature or glitch-capture. However, I suspect that if you had a crossover from a high-voltage circuit into the sensor circuit, you would have fried the ECM/PCM from the excessive voltage coming back in from the sensor and you'd be seeing other problems.
Let us know what you find!
on edit: Confirmed via Alldata that the reference voltage for the crankshaft sensor is 5 volts (DC, of course) - so, you should never see more than 5VDC on the reference/input wire.