Non-Contact Meter Lights up on Grounded Metal Surfaces and in the Soil

[Master_HOBrFINCl]

I am commissioning a single axis tracker on a 4.4MVA central inverter and got shocked while performing routine tests on a 24V instrumentation wire which shares a right-of-way with a 1500V DC line and a 480V three-phase AC line. Volt-meter said only 20V AC which doesn't seem like enough to cause a shock. Fluke VoltAlert non-contact meter lights up on the conductors at the rack instrument box side but not at the PLC side back at the central station. It lights up on steel and aluminum grounded surfaces everywhere within the rack too. It even lights up directly in the soil, tens of feet away from any buried conductors. I have never seen a non-contact meter light up in the dirt before.

At the same rack it does not light up at night, even when the inverter runs in VARs at night mode (DC contactors open). There is a 345kV line about a mile away, I don't think it could be capacitively coupled this far away. It's not happening on other racks. The inverter is running and says good insulation resistance on both positive and negative. The 480V transformer is grounded wye - grounded wye though there is not a good zero-sequence source from the MV-AC system (ungrounded wye on the low side). The inverters are negatively grounded, so maybe that could serve ground fault current. All the 480V systems have been functionally tested and verified working 100%.

What's going on here?

 

[tom baker]

You need a low impedance AC voltmeter, like a fluke 115 or a simplson 160

 

[winnie]

When you say it 'lights up directly in the soil', are you holding the thing, sticking the tip into the soil and getting a response, or are you sticking it into the ground, letting go, and still getting a response?

-Jonathan