In an article on Arc Flash Hazards, this month's NEC Digest recounts the story of an electrician getting burned while attempting to install a commercial meter into the socket.
He verifying the absence of ground faults on the load side he inserted the meter. After the load side stabs engaged, he pushed the line side stabs in. That is when the flash occurred.
What other safety procedures are followed prior to inserting a socket-type meter?
I always (usually) check for voltage between the line & load side socket terminals before plugging in a meter.
Sometimes, 20-30 volts or so is present, between L1 on the line side & L1 on the load side,(same for L2) even though the Main Breaker is open.
Where is this voltage coming from?
I always open the main breaker and any sub panel breakers to prevent current flow/arcing when connecting a meter.Also wear the proper safety stuff(face shield, gloves)in the event there is a fault on the load side ungrounded(hot)conductors that is caused by applied voltage that would not show up with a continuity test from meter output terminals(no meter installed) to ground .
Could the voltage difference of 20-30 volts be caused by a leakage current through your voltmeter to ground?If so it would be interesting to locate the leakage path.
It could be a faulty meter that caused the arc.
Chris
Are you testing with a high-impedance DVM?
If so, then you could easily get a reading due to the capacitance formed between the load-side L1/L2 cables and the neutral or grounded raceway.
I am testing across the line & load terminal of the same pole, with the main open. I have gotten this voltage on several occasions. If there is only 8-10' of load side cable before terminating at the main, can capacitance be an issue?
Disconnect the neutral, then test. It sounds like a low level fault through the neutral. I found this problem several years ago in an older home. The problem was a phase leg rubbing against a tree. The insulation had rubbed barely through and in contact with the neutral, back feeding into the house via the neutral. A new service from the poco fixed the problem.
Capacitance can easily give misleading readings with high-impedance meters.
If you were using a typical DVM with 10 meg input resistance, then by my calcs. a capacitance of only 50pF (0.00005uF) would give a reading of about 22V.
Paul,
Nice deduction/calculation.
Chris
The 20 - 30 or so volts you are reading, with the main open, is probably due to the impedence of your DMM. The easiest way to prove this, is to test voltage with a wiggie in parralel with your DMM meter leads. (The wiggie will provide enough load in the circuit to eliminate false reading.) As for testing the meter socket, most utiities provide their installers with meter socket testers, a unit that plugs into the socket. It contains 30 amp fuses connected line to load, that will blow is more than 30 amps of load is present on any phase or if a short is detected. & two sets of neons light to indicate safe or unsafe conditions exist. These units are invaluable to the metermen, because the only true way of proving a meter socket is safe is to test all phases in the socket at the same time.
And it goes without saying that all Personal Protective Equipment must be worn when working in any energized meter socket or panel (we require 1000v rubber gloves with leather protectors, eye protection, nomex uniforms & boots be worn.)