"George" wrote in message news:SMxfe.67538$WI3.26796@attbi_s71...
Sir. If there are 32 VAC in the tank, I assure you that the GFCI will trip.
Obviously he missed something.

Assuming he missed a device - you are right...
But give this man a credit - if he is saying he unplugged EVERYTHING
then I believe him, because I can understand how can high-impedance
voltameter give a false voltage reading. Measuring voltages with digital
meter is not a good idea in this particular case... Better the meter
is, the better chance he has to measure it WRONG, since better
meter will have in most cases, larger input impedance, making it
hard to measure correctly risk in this particular case...

If he said he is measuring 20mA flowing from tank to ground and
nothing is plugged to the wall, then I would be sceptical...
But voltages, without measuring current - that I can believe.

When I first checked out voltages in my tank, I
got a reading in the 30 volt range. One by one I disconnected appliances, and
found that all were contributing small voltages to the tank. When I finally had
all appliances not just turned off, but unplugged, the voltage went to 0.

The only correct measurement of the quality of insulation
is an insulation meter, which applies 600-1000V to the
device and measures the current flowing through the insulator.
If it passes the test at 600V you can be sure nothing wrong
is with the device at 120V. Measuring voltages between water
and copper pipe is not a good way to approach the matter.

It means that he either missed an appliance or that he is picking up stray
voltages. Picking up stray voltages doesn't mean that the tank itself it
charged. If he is using a copper water pipe in the house for a ground
connection, maybe that is a source of the stray readings, since copper pipes
make good AM antannas, for instance. He should use the ground in his wall
outlet, or even better, a filtered ground from a battery backup device for a
computer. My understanding is digital voltmeters are more susceptible to stray
voltage readings, but usually those are in the millivolt, not volts range. I
have radio shack's "best" analogue multimeter, and it seems to be pretty
accurate.

If his copper piping is not grouded he could have the false reading as well.
You are right. But what are the chances his water line is not grounded?
Mosta likely it will be connected with delivery pipe dug under ground.
So usually it is a good source of ground, unless there is a break in
copper lines and you go from metal to plastic (PVC) and back to coper.
Again - blindly assuming anything is wrong... If you want your measurements
be accurate you cannot "assume" anythin, you need to verify everything.
And I would start with checking if what I assume is a good ground (his
copper pipes) is really good ground... Eliminating so trivial mistake
like not unplugging a faulty device giving this 32V reading nothing else
comes to my mind than his tank water, conducting electricity almost as good
as metal acts as antenna, or his pipes are not really grounded...

And I will bet that everything is fine with his devices, George.
His measurement method is wrong! If he had any device leaking,
his high-impedance voltmeter would show voltage close to 120VAC
and only when the suspected device is plugged to the mains.

Well, that has no been my experience, so we'll have to see what his result is.

We will see. But if he trully had unplugged all devices and still had
32VAC than adding GFCI will not change the reading nor cause GFCI to trip.